Internet-Draft | Draft of rfc5661bis | August 2024 |
Noveck | Expires 2 March 2025 | [Page] |
Network File System (NFS) Version 4 Minor Version 1 Protocol
Abstract
This document describes the Network File System (NFS) version 4 minor version 1, including features retained from the base protocol (NFS version 4 minor version 0, which is specified in RFC 7530) and protocol extensions made subsequently. The later minor version has no dependencies on NFS version 4 minor version 0, and was, until recently, documented as a completely separate protocol.¶
This document is part of a set of documents which collectively obsolete RFCs 8881 and 8434. In addition to many corrections and clarifications, it will rely on NFSv4-wide documents to substantially revise the treatment of protocol extension, internationalization, and security, superseding the descriptions of those aspects of the protocol appearing in RFCs 5661 and 8881.¶
Status of This Memo
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.¶
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.¶
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."¶
This Internet-Draft will expire on 2 March 2025.¶
Copyright Notice
Copyright (c) 2024 IETF Trust and the persons identified as the document authors. All rights reserved.¶
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Revised BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Revised BSD License.¶
This document may contain material from IETF Documents or IETF Contributions published or made publicly available before November 10, 2008. The person(s) controlling the copyright in some of this material may not have granted the IETF Trust the right to allow modifications of such material outside the IETF Standards Process. Without obtaining an adequate license from the person(s) controlling the copyright in such materials, this document may not be modified outside the IETF Standards Process, and derivative works of it may not be created outside the IETF Standards Process, except to format it for publication as an RFC or to translate it into languages other than English.¶
1. Introduction to this Update
This document is intended to be the basis for a revised and updated specification of NFSv4.1. Unlike [RFC8881], which provided a limited-function update to [RFC5661], this document has a broader mandate and will do the following:¶
-
Revise any text known to be wrong or otherwise inappropriate. Such text will not be retained as it has been merely because it is outside a limited pre-specified change scope.¶
This includes changes in some errata reports with the status REJECTED, where there is a Working Group Consensus that change is necessary.¶
-
Correct protocol defects, that, by their nature, can be addressed via a limited use of the extension mechanism described in Section 9 of [RFC8178].¶
For more discussion of the handling of existing protocol defects, see Section 1.3. This discussion, which covers protocol defects addressed in NFSv4-wide documents in addition to this document, will focus on how compatibility issues are addressed when defects are corrected. It will also pay attention to the question of when XDR extensions are necessary as part of defect correction.¶
1.1. Requirements Language
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as specified in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.¶
The above differs from the corresponding statement in earlier specification versions ([RFC5661] and [RFC8881]) which only referred to [RFC2119]. For further discussion of this change, see Appendix B.3¶
In some cases, such keywords will appear in all capitals within quotations, direct or indirect, from earlier documents. In such cases, these terms are to be interpreted just as above, except where it is explicitly noted that such an interpretation is not to be inferred. In some cases, it might be that this document's approach to the matter would not use those key words for reasons explained in the text. Such a shift might cause compatibility issues, if the previous keyword were actually relied upon but it also possible that it was not relied upon while the implications of that use were ignored for various reasons.¶
The reader should be aware that, as discussed in Appendix C.1, there are uses of the keywords listed above in RFCs 5661 and 8881 which might not have been appropriate, even though the interpretation specified above was intended when the text was written and submitted for publication. In some cases, the text in this document has been updated to correct the issue but it should be understood that not all such questionable uses have been addressed and that this state of affairs might continue to exist until a later draft of this document is submitted for publication.¶
1.2. The Changed Role of this Specification
Previous specifications for this minor version ([RFC5661], [RFC8881]) have purported to describe the protocol in its entirety, without reference to features common to all minor versions of NFS Version 4. In contrast, this update relies on a set of base documents describing common aspects of the NFSv4 protocol that applies to all minor versions.¶
-
Rules for extensions and creation of new minor versions appear only in [RFC8178], unlike previously in which they appeared in the NFSv4.1 specification. This eliminates the unfortunate situation in which each minor version was allowed to create its own extension rules.¶
-
Handling of internationalization-related matters (for all minor versions) is now discussed in its own document, which is expected to be an RFC derived from [I-D.ietf-nfsv4-internationalization].¶
That document, based in large part on the handling of internationalization for NFSv4 minor version zero outlined in [RFC7530], has been extended to cover all minor versions and enhanced to fully support case-insensitive handling of internationalized file names.¶
This corrects the unfortunate situation in which internationalization for minor version one and subsequent minor versions (in [RFC5661] and [RFC8881]) had never been implemented and could never have been implemented by NFS Version four clients and servers.¶
-
Handling of core security-related matters for all NFSv4 minor versions will be consolidated in a set of documents that are expected to be RFCs derived from [I-D.dnoveck-nfsv4-security] and [I-D.dnoveck-nfsv4-acls].¶
This shift is made necessary by the following issues, many of which are of long standing and make the continuation of previous approaches to these issues insupportable:¶
-
The lack of a substantial threat analysis in the Security Considerations section of any existing minor version specification.¶
-
The unfortunate designation of AUTH_SYS an "OPTIONAL means of authentication" which had the effect of obscuring the severe security problems with its common use. The use of "OPTIONAL" suggested that use of AUTH_SYS had no harmful consequences while the phrase "means of authentication" ignored the fact that no actual authentication took place.¶
-
The assumption that data confidentiality could be satisfactorily addressed as an occasionally-used optional facility.¶
-
The neglect of the need for dependable semantic description of the protocol's authorization semantics.¶
Earlier versions of NFS had avoided the need for such descriptions by relying on POSIX semantics. The addition of non-POSIX semantic elements, including named attributes and ACLs, interfered with that approach although the necessity was not recognized as NFSv4 was being formulated.¶
The availability of new transport-level security features such as those provided by RPC-with-TLS [RFC9289] provides a basis to correct many of the above issues.¶
In providing that sort of correction, we need to be careful not to declare existing implementations non-compliant post facto, while still providing adequate warning of the security consequences of continuing to use the NFS Version 4 protocol insecurely, as described in previous specifications.¶
-
1.3. Addressing Protocol Defects
This section provides an overview of situations in which it was necessary to change the protocol to correct protocol issues that needed to be addressed going forward. Although all these issues can, from today's perspective, be viewed as mistakes, it is not clear whether that description is appropriate for decisions made so long ago, under very different circumstances. In any case, those questions will not be addressed here.¶
The correction of protocol defects often gives rise to compatibility issues and their possible presence will be discussed below. In addition, the question of when it is appropriate to address such issues using the protocol extension mechanism described in [RFC8178] needs to be considered. Section 9 of that document alludes to this possibility but we have to decide when defects are best addressed in that way.¶
These defects can be divided into two groups based on their origin.¶
-
Defects that originated in minor version zero¶
Many of these defects are addressed in the new NFSv4-wide documents ([I-D.dnoveck-nfsv4-security], [I-D.dnoveck-nfsv4-acls], and [I-D.ietf-nfsv4-internationalization]. For these defects , the greater change scope require more attention to compatibility issues. In addition, that greater scope limits the degree to which protocol extension can be used in providing a correction since that extension would need to be propagated to two non-extensible minor versions.¶
-
Defects that originated in minor version one.¶
A major source of defects was the result of the addition of a set of OPTIONAL features in v4.1, that have never been implemented, making it important to eliminate issues in the specification that have led to this situation.¶
The presumptive non-implementation of these features will limit interoperability concerns. However, since we cannot be sure about the possible existence of implementations under development, we will try to provide for the possibility of interoperating with earlier implementations, even if that interoperation is hypothetical. Only in the case of features whose current specification makes implementation impossible can we ignore the possibility of interoperating with such implementations.¶
The following defects were addressed as part of this update effort:¶
- 1:
-
Internationalization is being thoroughly respecified in the NFSv4-wide document [I-D.ietf-nfsv4-internationalization].¶
There were no NFSv4.1 compatibility issues to deal with since the handling of internationalization approach mandated by [RFC8881] had never been implemented.¶
Potential NFSv4.0 compatibility issues were very limited since the approach followed in [RFC7530] was continued and that approach had been used by all implementations.¶
In one particular case, there is a potential compatibility issue arising from the transition of one potential troubling server behavior from being discouraged using SHOULD NOT to being prohibited using MUST NOT. However, in view of the unlikelihood of ongoing use of the discouraged behavior, this has not been considered problematic.¶
The respecification of the fs_charset_cap attribute raises the possibility of within-NFS4.1 compatibility issues. However the very limited use of this attribute by clients combine with the lack of clarity in the previous definition makes it unlikely that the use of protocol extension to support previous uses would be justified.¶
- 2:
-
Because the Persistent reply cache feature could not be implemented as described in [RFC8881], the entire area was respecified in Section 8.¶
The reasons for this respecification are discussed in Section 8.1.¶
Because the existing feature specification was unimplementable there were no compatibility issues to deal with.¶
No protocol extensions were needed since the small set of bits defined for the earlier feature could be coopted.¶
- 3:
-
Security for all minor versions is being thoroughly respecified in the NFSv4-wide document [I-D.dnoveck-nfsv4-security]. In this discussion, issues related to authorization semantics and to ACLs are being dealt with separately.¶
This respecification was made necessary by the lack of threat analyses for all minor versions, the absence of any discussion of the security problems associated with the use of AUTH_SYS, and the half-hearted approach to the security of over-the-wire transmission in which transmission in the clear was the default and the provision of secure transmission was an option requiring per-fs configuration.¶
- 4:
-
As part of the new handling of security, a more serious treatment of authorization semantics was necessary. As part of effecting this, the attributes mode, owner, and owner_group became REQUIRED, as it is impossible to effect security without them.¶
This change was not connected to the shift in terminology in which the attributes incorrectly described as "RECOMMENDED" became "OPTIONAL".¶
No significant compatibility issue are expected, since the existence of servers not supporting these attributes or of clients interacting with such servers, while possible theoretically, has to be considered extremely unlikely and none are known to the working group.¶
- 5:
-
A number of gaps in the description of authorization semantics needed to be addressed. These include the lack of a clear description of authorization for operations on named attribute directories and potential use of the "sticky" bit in controlling authorization of file deletion.¶
These matters are being addressed within [I-D.dnoveck-nfsv4-security] where they are being tracked as Consensus Items #66 and #6 respectively.¶
Working group consideration of the security document will involve resolving those two Consensus Items, as well as others.¶
- 6:
-
The handling of ACLs for all minor versions is being thoroughly respecified in the NFSv4-wide document [I-D.dnoveck-nfsv4-acls].¶
Such a respecification was made necessary by the profound underspecification of the ACL feature that arose from a misguided attempt to support two very different approaches to the provision of ACLs. The problems posed by the different semantics of these two were never clearly addressed since it was erroneously assumed that semantic description could be avoided. As a result, potential interoperability was compromised since there was no way for the client to determined what ACL-based facilities were supported by a particular server, given that the specification treated these differences as if they were quality-of-implementation issues.¶
The development of [I-D.dnoveck-nfsv4-acls] has included a respecification of the area in which support for a subset of draft POSIX ACLs, termed UNIX ACLs, was the core and the various additions to that core were considered additional OPTIONAL features. These included the features that motivated the extensions in the NFSv4 ACL model and further accommodations for the semantics of the draft POSIX ACL model.¶
The development of [I-D.dnoveck-nfsv4-acls] has involved use of protocol extension within NFSv4.1 in addition to necessary structural changes that did not involve XDR changes.¶
The development of [I-D.dnoveck-nfsv4-acls] to support the rfc5661bis effort will most likely be limited to providing interoperability for those using the facilities within the UNIX ACL core or within the draft POSIX acl model. Interoperability for features beyond that set is likely to be delayed to later ACL bis, while the deletion of unneeded proposed features will have to wait for a later minor version, e.g., NFSv4.3.¶
- 7:
-
It has been necessary to define a new read-only per-fs OPTIONAL attribute that will allow clients to determine which of the OPTIONAL extensions to the core UNIX ACL model are supported by the server.¶
While this was essential to make the NFSv4 extensions usable, it also has a critical role in making POSIX ACL support available within NFsv4, albeit with some client mapping/filtering,¶
- 8:
-
A new ACE flag was necessary to address the difference in the handling of partial ALLOW ACE satisfaction in the two ACL models to be supported.¶
No compatibility issues are expected to arise since Aclchoice will indicate to the client whether the server is aware of the new flag.¶
- 9:
-
In order to support the draft POSIX ACL approach to ACL inheritance, it was necessary to provide some protocol extensions despite the fact that, overall, the NFSv4 ACL inheritance model has a wider semantic range.¶
A new ACE flag was used to identify "default" ACEs defined as inherit-only ACEs applying to both file and subdirectories, that are, unlike most ACEs, not normally modified/obliterated by changes to ACL attribute. To control whether such ACEs, considered as part of the "default ACL" within draft POSIX, new flags have been defined within the na41_flag word used by the sacl and dacl attribute.¶
Because this approach is easier to deal with, it will be made available to all ACL users rather than being limited to those using the draft POSIX ACL model, for whom it is pretty much essential.¶
As in the previous case, no compatibility issues are expected to arise since Aclchoice will indicate to the client whether the server is aware of the new flags to support the draft POSIX ACL approach to inheritance.¶
- 10:
-
In order to enable the translation of reverse-slope modes to ACLs in environments which support for DENY ACEs was not available, it was necessary to enable ACLs to contain the special who values, GROUPNOTOWNER@ and OTHERS@.¶
While not, in itself a protocol extension, there were associated protocol extensions to enable Aclchoice to report about the support for these new values.¶
No compatibility issues are expected since the new values are unlikely to be seen by the client and can only be used by the client when Aclchoice indicates that support is present.¶
- 11:
-
The defects described in Appendix C.2.1 needed to be addressed together, in connection with making it clear that the term "Exactly-once Semantics" ignored the fact that there were valid reasons to give up on requests which could leave them unexecuted.¶
This change did not give rise to compatibility issues since the specification was changed to match existing implementations, and these are expected to remain as they are.¶
- 12:
-
The inadvertent prohibition of the use of RoCE in implementing NFSv4.1 using RPC-over-RDMA was removed.¶
This is another case in which compatibility issue are not expected because the spec has been changed to match existing implementations.¶
- 13:
-
The corrections discussed in Appendix C.2.3 had to be made since most of the worries expressed within it were the result of misunderstandings.¶
Although no compatibility issues are expected we will need to review the changes and reach consensus on them.¶
- 14:
-
There were a number of issues in the earlier specification of the directory delegation feature that need to be addressed to enable implementations of this needed feature to be produced. Given the lack of implementation during the long period since they were introduced in [RFC5661] many years ago.¶
While a significant part of the problems could be ascribed to clarity issues, there were also a set of defects, some of which required protocol extensions, as provided for in Section 9 of [RFC8178].¶
The defects which contributed substantially to this long-lasting lack of implementation includes the failure to fully address authorization issues for the use of cached directory data, implementability issues regarding the maintenance of cached attribute data, and the assumption that clients could maintain the cached directory contents only in the same format as used by the server. For more discussion of these defects, see Appendix C.2.2.¶
These issues were addressed by a major rewrite of Section 15.9 in which protocol extension was necessary, including the addition of new values to the enum notify_type4. In addition, there are complementary changes made to Section 23.39 and Section 25.4) and to operations that might result in notifications being sent.¶
Although no client and server implementations of this feature are known to exist, the possibility of them existing cannot be excluded. As a result, the revised specification takes care to deal appropriately with such hypothetical implementations, and to not prohibit their use unless that is necessary to avoid unacceptable system behavior.¶
- 15:
-
Change in the recommendations regarding handling of numeric strings to represent users and groups.¶
Formerly considered troublesome even in the AUTH_SYS case despite the fact that there is no explanation given as to how to effect mapping between numeric ids and strings. Instead, it is assumed that client and server will somehow agree to do this without the specification making it possible or giving a convincing reason that such mapping is needed.¶
Of the above, only the items 7, 8, 9, 10, and 14 required protocol extension to resolve. All will need to be incorporated in the eventual bis document superseding [RFC5662].¶
2. Introduction to this Minor Version Specification
2.1. The NFS Version 4 Minor Version 1 Protocol
The NFS version 4 minor version 1 (NFSv4.1) protocol is the second minor version of the NFS version 4 (NFSv4) protocol. The first minor version, NFSv4.0, is now described in [RFC7530], as modified by [RFC7931] and [RFC8587]. Minor version 1 follows the guidelines for minor versioning presented in [RFC8178].¶
As a minor version, NFSv4.1 is consistent with the overall goals for NFSv4, but extends the protocol so as to better meet those goals, based on experiences with NFSv4.0. In addition, NFSv4.1 has adopted some additional goals, which motivate some of the major extensions in NFSv4.1, such as the use of the sessions model.¶
This minor version adds a considerable number of new operations including some that are not OPTIONAL and makes a number of NFSv4.0 operations MANDATORY to NOT implement. As a result, the vast majority of NFSv4.0 requests are not valid in NFSv4.1 and vice versa. While clients and server that support both minor versions are common, such implementations treat the two versions as distinct protocols sharing a substantial common heritage.¶
2.2. Scope of This Document
This document describes the NFSv4.1 protocol. With respect to NFSv4.0, this document does not:¶
2.3. NFSv4 Goals
The NFSv4 protocol is a further revision of the NFS protocol defined already by NFSv3 [RFC1813]. It retains the essential characteristics of previous versions: easy recovery; independence of transport protocols, operating systems, and file systems; simplicity; and good performance. NFSv4 had the following goals:¶
-
Improved access and good performance on the Internet.¶
The protocol is designed to transit firewalls easily, perform well where latency is high and bandwidth is low, and scale to very large numbers of clients per server.¶
-
Strong security with facilities for negotiation of security handling built into the protocol.¶
The protocol has built on the work of the ONCRPC working group in supporting the RPCSEC_GSS protocol. Additionally, the NFSv4.1 protocol provides a mechanism to allow clients and servers the ability to negotiate security and requires clients and servers to support a minimal set of security schemes.¶
The protocol now takes advantage of the ability of RPC to make confidentiality available by using TLS-based encryption on connections to be used for NFSv4.1, which may limit the need for negotiation regarding facilities such as privacy.¶
-
Good cross-platform interoperability.¶
The protocol embraces a file system model that provides a useful, common set of features that does not unduly favor one file system or operating system over another.¶
-
Designed for protocol extensions via minor versioning.¶
The protocol is designed to accept standard extensions within a framework that enables and encourages backward compatibility.¶
When extensions are OPTIONAL, they can be added to an existing extensible minor version.¶
2.4. NFSv4.1 Goals
NFSv4.1 has the following goals, within the framework established by the overall NFSv4 goals.¶
-
To correct significant structural weaknesses and oversights discovered in the base protocol.¶
-
To add clarity and specificity to areas left unaddressed or not addressed in sufficient detail in the base protocol. However, as stated in Section 2.2, it is not a goal to clarify the NFSv4.0 protocol in the NFSv4.1 specification.¶
-
To add specific features based on experience with the existing protocol and recent industry developments.¶
-
To provide protocol support to take advantage of clustered server deployments including the ability to provide scalable parallel access to sets of files distributed among multiple servers, using a range of data access protocols.¶
This parallel access may involve striping of single files among a set of servers within the cluster.¶
Alternatively, parallel access may be provided by distributing unstriped files within the cluster allowing each client to contact the server holding each particular file directly.¶
2.5. General Definitions
The following definitions provide an appropriate context for the reader.¶
- Byte:
-
In this document, a byte is an octet, i.e., a datum exactly 8 bits in length.¶
- Client:
-
The client is the entity that accesses the NFS server's resources. The client may be an application that contains the logic to access the NFS server directly. The client may also be a traditional operating system client that provides remote file system services for a set of applications.¶
A client is uniquely identified by a client owner.¶
With reference to byte-range locking, the client is also the entity that maintains a set of locks on behalf of one or more applications. This client is responsible for providing crash or failure recovery for those locks it manages, in order to deal with the possibility of server reboot.¶
Note that multiple clients may share the same transport and connection and multiple clients may exist on the same network node.¶
- Client ID:
- The client ID is a 64-bit quantity used as a unique, short-hand reference to a client-supplied client owner, consisting of a verifier and a client owner id. The server is responsible for supplying the client ID.¶
- Client Owner:
- The client owner includes a unique string, the client owner id, opaque to the server, that identifies a client. It also includes a verifier to enable successive instances of the same client to be distinguished. Multiple network connections and source network addresses originating from those connections may share a client owner. The server is expected to treat requests from connections with the same client owner as coming from the same client. See Section 5.5 for more detail.¶
- File System:
- A file system is the collection of objects accessed using a particular server (as identified by the major identifier of a server owner, which is defined later in this section) that share the same value of the fsid attribute (see Section 11.12.1.9).¶
- Lease:
-
A lease is an interval of time defined by the server for which the client is granted locks. At the end of a lease period, unrecallable locks may be revoked if the lease has not been extended. Such a lock must be revoked if a conflicting lock has been granted after the lease interval. Revocation of unrecallable locks within the lease interval is expected to be an unusual event and clients normally expect such revocations to be rare.¶
A server grants a client a single lease for all of its associated locking state. Recallable locks such as layouts and delegations can be revoked within the lease period and are generally not affected by the state of the lease.¶
- Lock:
- The term "lock" is used to refer to byte-range (in UNIX environments, also known as record) locks, share reservations, delegations, or layouts unless specifically stated otherwise.¶
- Secret State Verifier (SSV):
- The SSV is a unique secret key shared between a client and server. The SSV serves as the secret key for an internal (that is, internal to NFSv4.1) Generic Security Services (GSS) mechanism (the SSV GSS mechanism; see Section 7.9). The SSV GSS mechanism uses the SSV to compute message integrity code (MIC) and Wrap tokens. See Section 7.8.3 for more details on how NFSv4.1 uses the SSV and the SSV GSS mechanism.¶
- Server:
- The Server is the entity responsible for coordinating client access to a set of file systems and is identified by a server owner. A server can span multiple network addresses.¶
- Server Owner:
- The server owner identifies the server to the client. The server owner consists of a major identifier and a minor identifier. When the client has two connections each to a peer with the same major identifier, the client assumes that both peers are the same server (the server namespace is the same via each connection) and that lock state is sharable across both connections. When each peer has both the same major and minor identifiers, the client assumes that each connection might be associated with the same session.¶
- Stable Storage:
-
Stable storage is storage from which data stored by an NFSv4.1 server can be recovered without data loss from multiple power failures (including cascading power failures, that is, several power failures in quick succession), operating system failures, and/or hardware failure of components other than the storage medium itself (such as disk, nonvolatile RAM, flash memory, etc.).¶
Some examples of stable storage that are allowable for an NFS server include:¶
- Media commit of data; that is, the modified data has been successfully written to the disk media, for example, the disk platter.¶
- An immediate reply disk drive with battery-backed, on-drive intermediate storage or uninterruptible power system (UPS).¶
- Server commit of data with battery-backed intermediate storage and recovery software.¶
- Cache commit with uninterruptible power system (UPS) and recovery software.¶
- Stateid:
- A stateid is a 128-bit quantity returned by a server that uniquely defines the open and locking states provided by the server for a specific open-owner or lock-owner/open-owner pair for a specific file and type of lock.¶
- Verifier:
-
A verifier is a 64-bit quantity that is changed to indicate that a corresponding change on one of the peers has occurred requiring the other peer to adjust to possibility of change. There are a number of 64-bit quantities identified as verifiers:¶
- In many cases, a verifier is a 64-bit quantity generated by the server that the client can use to determine if the client has restarted and potentially lost writes that had not been reliably committed to stable storage.¶
- Another type of verifier is used to indicate whether the server mapping between directory cookies and directory entries has changed, requiring the client to restart directory interrogations that normally may be continued across multiple READDIR requests.¶
- As described above, client owners include a verifier, allowing the server to determine when a client reboot has occurred¶
The Secret State Verifier is not a verifier in the sense given in this definition.¶
2.6. Overview of NFSv4.1 Features
The major features of the NFSv4.1 protocol will be reviewed in brief. This will be done to provide an appropriate context for both the reader who is familiar with the previous versions of the NFS protocol and the reader who is new to the NFS protocols. For the reader new to the NFS protocols, there is still a set of fundamental knowledge that is expected. The reader should be familiar with the External Data Representation (XDR) and Remote Procedure Call (RPC) protocols as described in [RFC4506] and [RFC5531]. A basic knowledge of file systems and distributed file systems is expected as well.¶
In general, this specification of NFSv4.1 will not distinguish those features added in minor version 1 from those present in the base protocol but will treat NFSv4.1 as a unified whole. See Section 4 for a summary of the differences between NFSv4.0 and NFSv4.1.¶
2.7. RPC and Security
As with previous versions of NFS, the External Data Representation (XDR) and Remote Procedure Call (RPC) mechanisms used for the NFSv4.1 protocol are those defined in [RFC4506] and [RFC5531], as extended by [RFC9289]) to provide TLS-based encryption and client-host authentication. NFSv4.1 security. A description of the basics of NFv4.1 security will appear in an NFSv4-wide security document, to be derived from [I-D.dnoveck-nfsv4-security].¶
NFSv4.1 introduces parallel access (see Section 2.8.2), through the use of pNFS. The security framework described above is significantly modified by the introduction of pNFS (see Section 17.9), because of the addition of additional actors and because data access is sometimes not over RPC. The appropriate handling depends on the data access protocol used (see Section 17.2.5) which depends in turn on the layout type (see Section 17.2.7.) The sections 17.9.1 through 17.9.3 discuss the security implications of using different sorts of data access protocols.¶
2.8. Protocol Structure
2.8.1. Core Protocol
Unlike NFSv3, which relied on a series of ancillary protocols (e.g., NLM, NSM (Network Status Monitor), MOUNT), within all minor versions of NFSv4 a single RPC protocol is available to make requests to the server. Facilities that had been separate protocols, such as locking, are now integrated within a single unified protocol, although, to implement pNFS, different data access protocols may also be used.¶
2.8.2. Parallel Access
Minor version 1 supports high-performance data access to a clustered server implementation by enabling a separation of metadata access and data access, with the latter able to be done to multiple servers in parallel.¶
Such parallel data access is controlled by recallable objects known as "layouts", which are integrated into the protocol locking model. Clients direct requests for data access to a set of data servers specified by the layout via a data storage protocol which may be NFSv4.1 or may be another protocol.¶
Because the protocols used for parallel data access are not necessarily RPC-based, the RPC-based security model (Section 2.7) is impacted (see Section 17.9). The degree of impact varies with the protocol (see Section 17.2.5) used for data access, and can be as low as zero for some RPC-based data access protocols (see Section 18.13).¶
2.9. File System Model
The general file system model used for the NFSv4.1 protocol is the same as for previous minor versions of NFSv4. The server file system is hierarchical with the regular files contained within being treated as opaque byte streams. File names MAY be restricted to UTF-8-encoded strings of Unicode characters or treated as opaque. In addition, for some file systems, name handling MAY reflect the UTF-8 canonical equivalence relation, and in some cases, case-based equivalence relations as well.¶
The NFSv4.1 protocol does not rely on a separate protocol to provide for the initial mapping between path name and filehandle. All file systems exported by a server are presented as a tree so that all file systems are reachable from a special per-server global root filehandle. This allows LOOKUP operations to be used to perform functions previously provided by the MOUNT protocol. The server is responsible for providing any necessary pseudo file systems to bridge any gaps that arise due to unexported portions of the server-local name space that are between exported file systems.¶
2.9.1. Filehandles
As in previous versions of the NFS protocol, opaque filehandles are used to identify individual files and directories. Lookup-type and create operations translate file and directory names to filehandles, which are then used to identify objects in subsequent operations.¶
The NFSv4.1 protocol provides support for persistent filehandles, guaranteed to be valid for the lifetime of the file system object designated and bever to be reused after that. In addition, it provides support to allow servers to provide filehandles with more limited validity guarantees, referred to as volatile filehandles.¶
2.9.2. Numbered File Attributes
The NFSv4.1 protocol has a rich and extensible file object attribute structure in which each attribute is assigned an attribute number. The set of such attributes can be usefully divided in a number of ways, in order to provide helpful context for server implementers choosing to implement or not implement particular attributes which are not REQUIRED and for client implementers deciding how to deal with non-support of particular attributes which are not REQUIRED.¶
-
Attributes differ as to their scope, with only a subset applicable to a particular file object, while others apply to an entire file system.¶
As a practical matter, attributes applicable to a single file object, often require support within the file system proper. While this functionality is most often provided by a file system created initially for local access and only later adapted to remote use through an NFSv4.1 server, there are also file systems purpose-built for remote access.¶
Attributes applicable to an entire file system do not typically require support within the file system proper. One possible exception is when such attributes can be set to indicate the client's desire for some particular feature that inherently require filesystem support.¶
Note that, in all these cases, applications are most likely to be adapted to features that can be accessed using existing file access facilities. As a result, implementers are unlikely to devote efforts to implementation of OPTIONAL features and attributes which require interactions with applications while being more open to attributes usable by the client and server to communicate to optimize data flows without requiring application involvement.¶
-
Attributes differ as to their mutability characteristics, including whether the attribute is question can be modified explicitly by the client and whether the attribute modification happens as a result of performing other operations, such as modifying a file or directory¶
-
In this update of the NFSv4.1 specification, the details for handling authorization-related attributes are the responsibility of the NFSv4-wide security documents expected to be derived from [I-D.dnoveck-nfsv4-security] and [I-D.dnoveck-nfsv4-acls].¶
Although the detailed categorization of such attributes will be the responsibility of the security documents, this document will, in Section 11.3, provide a brief summary and make clear that some of these are more necessary than others, and that they all cannot be reasonably treated as having the being of the same class regarding the need for server support.¶
Attributes are divided into a number of classes based on the protocol's requirements/recommendations for server implementation and the client's expected response to a server's non-support of those attributes. This categorization differs from that appearing previously for a number of reasons with the specific differences explained in Section 11.2.¶
-
A significant number of attributes are described as REQUIRED so that servers MUST provide support for them.¶
These include the same set of attributes described in this way in RFCs [RFC7530] and [RFC8881]. In addition, there are a set of authorization-related attributes that need to be included in this group for reasons explained in [I-D.dnoveck-nfsv4-security]. The inclusion of all these attributes is discussed in more detail in Section 11.4.¶
-
Many attributes are truly OPTIONAL, even though such attributes have been erroneously categorized as "RECOMMENDED" in the past. These attributes are discussed in more detail in Section 11.5.¶
For a more detailed explanation of these shifts in terminology, see Section 11.2.¶
-
It appears necessary to designate certain authorization-related attributes as Experimental.¶
These attributes are discussed in more detail in Sections 11.3 and 11.6.¶
Descriptions of each specific attribute appears in the following places:¶
-
Those which are authorization-related are described in [I-D.dnoveck-nfsv4-security] and [I-D.dnoveck-nfsv4-acls].¶
These descriptions can be found using the three sources described below:¶
- The list of Authorization-related attributes appears in Section 11.18¶
- The list of REQUIRED Attributes appears in Table 4 within Section 11.10.¶
- The list of attributes which are not REQUIRED) appears in Table 5 within Section 11.11.¶
"Named attributes", despite this designation, which will be retained, differ substantially from file attributes per se and are explained in Section 2.9.3. All such attributes are OPTIONAL as is the entire named attribute feature and such attributes are not part of the categorization above.¶
2.9.3. Named Attributes
A named attribute is an opaque byte stream that is associated with a directory or file and referred to by a string name. Named attributes were intended to be used by client applications as a method to associate application-specific data with a regular file or directory. Servers providing support for the named attribute feature, which is OPTIONAL, allow a number of opaque byte stream to be associated with a directory or file. This feature allows applications to define extended attributes which could be opened, read and written just as files are.¶
For further information about the use of named attributes, see Section 11.7¶
2.9.4. Multi-Server Namespace
NFSv4.1 contains a number of features to allow implementation of namespaces that cross server boundaries and that allow and facilitate a nondisruptive transfer of support for individual file systems between servers. They are all based upon attributes that allow one file system to specify alternate, additional, and new location information that specifies how the client may access that file system.¶
These attributes can be used to provide for individual active file systems:¶
- Alternate network addresses to access the current file system instance.¶
- The locations of alternate file system instances or replicas to be used in the event that the current file system instance becomes unavailable.¶
These file system location attributes may be used together with the concept of absent file systems, in which a position in the server namespace is associated with locations on other servers without there being any corresponding file system instance on the current server. For example,¶
- These attributes may be used with absent file systems to implement referrals whereby one server may direct the client to a file system provided by another server. This allows extensive multi-server namespaces to be constructed.¶
- These attributes may be provided when a previously present file system becomes absent. This allows nondisruptive migration of file systems to alternate servers.¶
3. Locking Facilities
As mentioned previously, NFSv4.1 is a single protocol that includes locking facilities. These locking facilities include support for many types of locks including a number of sorts of recallable locks. Recallable locks such as delegations allow the client to be assured that certain events will not occur so long as that lock is held. When circumstances change, the lock is recalled via a callback request. The assurances provided by delegations allow more extensive caching to be done safely when circumstances allow it.¶
The types of locks are:¶
- Share reservations as established by OPEN operations.¶
- Byte-range locks.¶
- File delegations, which are recallable locks that assure the holder that inconsistent opens and file changes cannot occur so long as the delegation is held.¶
- Directory delegations, which are recallable locks that assure the holder that inconsistent directory modifications cannot occur so long as the delegation is held.¶
- Layouts, which are recallable objects that assure the holder that direct access to the file data may be performed directly by the client and that no change to the data's location that is inconsistent with that access may be made so long as the layout is held.¶
All non-recallable locks for a given client are tied together under a single client-wide lease. All requests made on sessions associated with the client renew that lease. When the client's lease is not promptly renewed, the client's locks are subject to revocation. In the event of server restart, clients have the opportunity to safely reclaim their locks within a special grace period.¶
Recallable locks are subject to revocation irrespective of lease state. Servers often need to revoke such locks when recalling them does not result in their prompt return.¶
4. Differences from NFSv4.0
The following summarizes the major differences between minor version 1 and the base protocol:¶
- Implementation of the sessions model (Section 7).¶
- Parallel access to data (Section 17).¶
- Addition of the RECLAIM_COMPLETE operation to better structure the lock reclamation process (Section 23.51).¶
-
Enhanced delegation support as follows.¶
- Delegations on directories and other file types in addition to regular files (Section 23.39, Section 23.49).¶
- Operations to optimize acquisition of recalled or denied delegations (Section 23.49, Section 25.5, Section 25.7).¶
- Notifications of changes to files and directories (Section 23.39, Section 25.4).¶
- A method to allow a server to indicate that it is recalling one or more delegations for resource management reasons, and thus a method to allow the client to pick which delegations to return (Section 25.6).¶
- Attributes can be set atomically during exclusive file create via the OPEN operation (see the new EXCLUSIVE4_1 creation method in Section 23.16).¶
- Open files can be preserved if removed and the hard link count ("hard link" is defined in an Open Group [hardlink] standard) goes to zero, thus obviating the need for clients to rename deleted files to partially hidden names, a technique colloquially called "silly rename" (see the new OPEN4_RESULT_PRESERVE_UNLINKED reply flag in Section 23.16).¶
- Improved compatibility with Microsoft Windows for Access Control Lists (sacl and dacl attributes, acl inheritance).¶
- Data retention (Section 11.17).¶
- Identification of the implementation of the NFS client and server (Section 23.35).¶
- Support for notification of the availability of byte-range locks (see the new OPEN4_RESULT_MAY_NOTIFY_LOCK reply flag in Section 23.16 and see Section 25.11).¶
- In NFSv4.1, LIPKEY and SPKM-3 are not required security mechanisms [RFC2847].¶
5. Core Infrastructure
5.1. Introduction
NFSv4.1 relies on core infrastructure common to nearly every operation. This core infrastructure is described in the remainder of this section.¶
5.2. RPC and XDR
The NFSv4.1 protocol is a Remote Procedure Call (RPC) application that uses RPC version 2 and the corresponding eXternal Data Representation (XDR) as defined in [RFC5531] and [RFC4506]. The transport-level encryption and client-host authentication facilities described in [RFC9289] can also be used.¶
5.3. RPC-Based Security
In addition to the above, as discussed in Section 5.3.1, some security flavors provide additional security services.¶
NFSv4.1 clients and servers MUST implement RPCSEC_GSS. (This requirement to implement is not a requirement to use.) Other flavors, such as AUTH_NONE and AUTH_SYS, can be implemented as well, although the security implications of doing so need to be carefully considered, particularly when the client host is not itself authenticated. In particular, it is RECOMMENDED by rpc-tls [RFC9289] that AUTH_SYS not be used when client host authentication is not in effect.¶
5.3.1. RPCSEC_GSS and Security Services
RPCSEC_GSS [RFC2203] uses the functionality of GSS-API [RFC2743]. This allows for the use of various security mechanisms by the RPC layer without the additional implementation overhead of adding RPC security flavors.¶
5.3.1.1. GSS Server Principal
Regardless of what security mechanism under RPCSEC_GSS is being used, the NFS server MUST identify itself in GSS-API via a GSS_C_NT_HOSTBASED_SERVICE name type. GSS_C_NT_HOSTBASED_SERVICE names are of the form:¶
service@hostname¶
For NFS, the "service" element is¶
nfs¶
Implementations of security mechanisms will convert nfs@hostname to various different forms. For Kerberos V5, the following form is RECOMMENDED:¶
nfs/hostname¶
5.4. COMPOUND and CB_COMPOUND
A significant departure from the versions of the NFS protocol before NFSv4 is the introduction of the COMPOUND procedure. For the NFSv4 protocol, in all minor versions, there are exactly two RPC procedures, NULL and COMPOUND. The COMPOUND procedure is defined as a series of individual operations and these operations perform the sorts of functions performed by traditional NFS procedures.¶
The operations combined within a COMPOUND request are evaluated in order by the server, without any atomicity guarantees. A limited set of facilities exist to pass results from one operation to another. Once an operation returns a failing result, the evaluation ends and the results of all evaluated operations are returned to the client.¶
With the use of the COMPOUND procedure, the client is able to build simple or complex requests. These COMPOUND requests allow for a reduction in the number of RPCs needed for logical file system operations. For example, multi-component look up requests can be constructed by combining multiple LOOKUP operations. Those can be further combined with operations such as GETATTR, READDIR, or OPEN plus READ to do more complicated sets of operation without incurring additional latency.¶
NFSv4.1 also contains a considerable set of callback operations in which the server makes an RPC directed at the client. Callback RPCs have a similar structure to that of the normal server requests. In all minor versions of the NFSv4 protocol, there are two callback RPC procedures: CB_NULL and CB_COMPOUND. The CB_COMPOUND procedure is defined in an analogous fashion to that of COMPOUND with its own set of callback operations.¶
The addition of new server and callback operations within the COMPOUND and CB_COMPOUND request framework provides a means of extending the protocol in subsequent minor versions.¶
Except for a small number of operations needed for session creation, server requests and callback requests are performed within the context of a session. Sessions provide a client context for every request and support robust replay protection for non-idempotent requests.¶
5.5. Client Identifiers and Client Owners
For each operation that obtains or depends on locking state, the specific client needs to be identifiable by the server.¶
Each distinct client instance is represented by a client ID. A client ID is a 64-bit identifier representing a specific client at a given time. The client ID is changed whenever the client re-initializes, and may change when the server re-initializes. Client IDs are used to support lock identification and crash recovery.¶
During steady state operation, the client ID associated with each operation is derived from the session (see Section 7) on which the operation is sent. A session is associated with a client ID when the session is created.¶
Unlike NFSv4.0, the only NFSv4.1 operations possible before a client ID is established are those needed to establish the client ID.¶
A sequence of an EXCHANGE_ID operation followed by a CREATE_SESSION operation using that client ID (eir_clientid as returned from EXCHANGE_ID) is required to establish and confirm the client ID on the server. Establishment of identification by a new incarnation of the client also has the effect of immediately releasing any locking state that a previous incarnation of that same client might have had on the server. Such released state would include all byte-range lock, share reservation, layout state. Also, where the server supports neither the CLAIM_DELEGATE_PREV nor the CLAIM_DELEG_PREV_FH claim types, all delegation state associated with the same client is released as well. For discussion of delegation state recovery, see Section 15.2.1. For discussion of layout state recovery, see Section 17.7.1.¶
Releasing such state requires that the server be able to determine that one client instance is the successor of another. Where this cannot be done, for any of a number of reasons, the locking state will remain for a time subject to lease expiration (see Section 13.3) and the new client will need to wait for such state to be removed, if it makes conflicting lock requests.¶
Client identification is encapsulated in the following client owner data type:¶
struct client_owner4 { verifier4 co_verifier; opaque co_ownerid<NFS4_OPAQUE_LIMIT>; };¶
The first field, co_verifier, is a client incarnation verifier, allowing the server to distinguish successive incarnations (e.g., reboots) of the same client. The server will start the process of canceling the client's leased state if co_verifier is different than what the server has previously recorded for the identified client (as specified in the co_ownerid field).¶
The second field, co_ownerid, contains the client owner id. This is a variable-length string that uniquely defines the client so that subsequent instances of the same client bear the same co_ownerid with a different verifier.¶
There are several considerations for how the client generates the co_ownerid string:¶
- The string should be unique so that multiple clients do not present the same string. The consequences of two clients presenting the same string range from one client getting an error to one client having its leased state abruptly and unexpectedly cancelled.¶
- The string should be selected so that subsequent incarnations (e.g., restarts) of the same client cause the client to present the same string. The implementer is cautioned from an approach that requires the string to be recorded in a local file because this precludes the use of the implementation in an environment where there is no local disk and all file access is from an NFSv4.1 server.¶
- The string should be the same for each server network address that the client accesses. This way, if a server has multiple interfaces, the client can trunk traffic over multiple network paths as described in Section 7.5. (Note: the precise opposite was advised in the NFSv4.0 specification [RFC3530].)¶
- The algorithm for generating the string should not assume that the client's network address will not change, unless the client implementation knows it is using statically assigned network addresses. This includes changes between client incarnations and even changes while the client is still running in its current incarnation. Thus, with dynamic address assignment, if the client includes just the client's network address in the co_ownerid string, there is a real risk that after the client gives up the network address, another client, using a similar algorithm for generating the co_ownerid string, would generate a conflicting co_ownerid string.¶
Given the above considerations, an example of a well-generated co_ownerid string is one that includes:¶
- If applicable, the client's statically assigned network address.¶
-
Additional information that tends to be unique, such as one or more of:¶
- The client machine's serial number (for privacy reasons, it is best to perform some one-way function on the serial number).¶
- A Media Access Control (MAC) address (again, a one-way function should be performed).¶
- The timestamp of when the NFSv4.1 software was first installed on the client (though this is subject to the previously mentioned caution about using information that is stored in a file, because the file might only be accessible over NFSv4.1).¶
- A true random number. However, since this number ought to be the same between client incarnations, this shares the same problem as that of using the timestamp of the software installation.¶
- For a user-level NFSv4.1 client, it should contain additional information to distinguish the client from other user-level clients running on the same host, such as a process identifier or other unique sequence.¶
The client ID is assigned by the server (the eir_clientid result from EXCHANGE_ID) and should be chosen so that it will not conflict with a client ID previously assigned by the server. This applies across server restarts.¶
In the event of a server restart, a client may find out that its current client ID is no longer valid when it receives an NFS4ERR_STALE_CLIENTID error. The precise circumstances depend on the characteristics of the sessions involved, specifically whether the session is persistent (see Section 8), but in each case the client will receive this error when it attempts to establish a new session with the existing client ID and receives the error NFS4ERR_STALE_CLIENTID, indicating that a new client ID needs to be obtained via EXCHANGE_ID and the new session established with that client ID.¶
When a session is not persistent, the client will find out that it needs to create a new session as a result of getting an NFS4ERR_BADSESSION, since the session in question was lost as part of a server restart. When the existing client ID is presented to a server as part of creating a session and that client ID is not recognized, as would happen after a server restart, the server will reject the request with the error NFS4ERR_STALE_CLIENTID.¶
In the case of the session being persistent, the client will re-establish communication using the existing session after the restart. This session will be associated with the existing client ID but may only be used to retransmit operations that the client previously transmitted and did not see replies to. Replies to operations that the server previously performed will come from the reply cache; otherwise, NFS4ERR_DEADSESSION will be returned. Hence, such a session is referred to as "dead". In this situation, in order to perform new operations, the client needs to establish a new session. If an attempt is made to establish this new session with the existing client ID, the server will reject the request with NFS4ERR_STALE_CLIENTID.¶
When NFS4ERR_STALE_CLIENTID is received in either of these situations, the client needs to obtain a new client ID by use of the EXCHANGE_ID operation, then use that client ID as the basis of a new session, and then proceed to any other necessary recovery for the server restart case (see Section 13.4.2).¶
See the descriptions of EXCHANGE_ID (Section 23.35) and CREATE_SESSION (Section 23.36) for a complete specification of these operations.¶
5.5.1. Upgrade from NFSv4.0 to NFSv4.1
To facilitate upgrade from NFSv4.0 to NFSv4.1, a server may compare a value of data type client_owner4 in an EXCHANGE_ID with a value of data type nfs_client_id4 that was established using the SETCLIENTID operation of NFSv4.0. A server that does so will allow an upgraded client to avoid waiting until the lease (i.e., the lease established by the NFSv4.0 instance client) expires. This requires that the value of data type client_owner4 be constructed the same way as the value of data type nfs_client_id4. If the latter's contents included the server's network address (per the recommendations of the NFSv4.0 specification [RFC3530]), and the NFSv4.1 client does not wish to use a client ID that prevents trunking, it should send two EXCHANGE_ID operations. The first EXCHANGE_ID will have a client_owner4 equal to the nfs_client_id4. This will clear the state created by the NFSv4.0 client. The second EXCHANGE_ID will not have the server's network address. The state created for the second EXCHANGE_ID will not have to wait for lease expiration, because there will be no state to expire.¶
5.5.2. Server Release of Client ID
NFSv4.1 introduces a new operation called DESTROY_CLIENTID (Section 23.50), which the client uses to destroy a client ID it no longer needs. This permits graceful, bilateral release of a client ID. The operation cannot be used if there are sessions associated with the client ID, or state with an unexpired lease.¶
If the server determines that the client holds no associated state for its client ID (associated state includes unrevoked sessions, opens, locks, delegations, layouts, and wants), the server MAY choose to unilaterally release the client ID in order to conserve resources. If the client contacts the server after this release, the server MUST ensure that the client receives the appropriate error so that it will use the EXCHANGE_ID/CREATE_SESSION sequence to establish a new client ID. The server ought to be very hesitant to release a client ID since the resulting work on the client to recover from such an event will be the same burden as if the server had failed and restarted. Typically, a server would not release a client ID unless there had been no activity from that client for many minutes. As long as there are sessions, opens, locks, delegations, layouts, or wants, the server MUST NOT release the client ID. See Section 7.13.1.4 for discussion on releasing inactive sessions.¶
5.5.3. Resolving Client Owner Conflicts
When the server gets an EXCHANGE_ID for a client owner that currently has no state, or that has state but the lease has expired, the server MUST allow the EXCHANGE_ID and confirm the new client ID if followed by the appropriate CREATE_SESSION.¶
When the server gets an EXCHANGE_ID for a new incarnation of a client owner that currently has an old incarnation with state and an unexpired lease, the server is allowed to dispose of the state of the previous incarnation of the client owner if one of the following is true:¶
- The client ID was created without explicit state protection (i.e. SP4_NONE was used) and without client host authentication while the current EXCHANGE_ID shares those characteristics and the principals used for client Id creation and the current EXCHANGE_ID match as well.¶
- The client ID was created without explicit state protection (i.e. SP4_NONE was used) and with client host authentication while the current EXCHANGE_ID shares those characteristics with the EXCHANGE_ID used to create the client ID while the authenticated client hosts match as well.¶
- The principal that created the client ID for the client owner is the same as the principal that is sending the EXCHANGE_ID operation. Note that if the client ID was created with SP4_MACH_CRED state protection (Section 23.35), the principal MUST be based on RPCSEC_GSS authentication, the RPCSEC_GSS service used MUST be integrity or privacy, and the same GSS mechanism and principal MUST be used as that used when the client ID was created.¶
- The client ID was established with SP4_SSV protection (Section 23.35, Section 7.8.3) and the client sends the EXCHANGE_ID with the security flavor set to RPCSEC_GSS using the GSS SSV mechanism (Section 7.9).¶
- The client ID was established with SP4_SSV protection, and under the conditions described herein, the EXCHANGE_ID was sent with SP4_MACH_CRED state protection. Because the SSV might not persist across client and server restart, and because the first time a client sends EXCHANGE_ID to a server it does not have an SSV, the client MAY send the subsequent EXCHANGE_ID without an SSV RPCSEC_GSS handle. Instead, as with SP4_MACH_CRED protection, the principal MUST be based on RPCSEC_GSS authentication, the RPCSEC_GSS service used MUST be integrity or privacy, and the same GSS mechanism and principal MUST be used as that used when the client ID was created.¶
If none of the above situations apply, the server MUST return NFS4ERR_CLID_INUSE.¶
If the server accepts the principal and co_ownerid as matching that which created the client ID, and the co_verifier in the EXCHANGE_ID differs from the co_verifier used when the client ID was created, then after the server receives a CREATE_SESSION that confirms the client ID, the server deletes state. If the co_verifier values are the same (e.g., the client either is updating properties of the client ID (Section 23.35) or is attempting trunking (Section 7.5), the server MUST NOT delete state.¶
5.6. Server Owners
The server owner is similar to a client owner (Section 5.5), but unlike the client owner, there is no shorthand server ID. The server owner is defined in the following data type:¶
struct server_owner4 { uint64_t so_minor_id; opaque so_major_id<NFS4_OPAQUE_LIMIT>; };¶
The server owner is returned from EXCHANGE_ID. When the so_major_id fields are the same in two EXCHANGE_ID results, the connections that each EXCHANGE_ID were sent over can be assumed to address the same server (as defined in Section 2.5). If the so_minor_id fields are also the same, then not only do both connections connect to the same server, but the session can be shared across both connections. The reader is cautioned that multiple servers may deliberately or accidentally claim to have the same so_major_id or so_major_id/so_minor_id; the reader should examine Sections 7.5 and 23.35 in order to avoid acting on falsely matching server owner values.¶
The considerations for generating an so_major_id are similar to that for generating a co_ownerid string (see Section 5.5). The consequences of two servers generating conflicting so_major_id values are less dire than they are for co_ownerid conflicts because the client can use RPCSEC_GSS to compare the authenticity of each server (see Section 7.5).¶
5.7. Transport Layers
5.7.1. REQUIRED and RECOMMENDED Properties of Transports
NFSv4.1 works over Remote Direct Memory Access (RDMA) and non-RDMA-based transports with the following attributes:¶
- The transport supports reliable delivery of data, which NFSv4.1 requires. However the possibility of connections breaking is addressed in NFSv4.1 by a session-based replay cache to prevent the spurious re-execution of non-idempotent requests or modifying idempotent requests.¶
- The transport delivers data in the order it was sent. Ordered delivery simplifies detection of transmit errors, and simplifies the sending of arbitrary sized requests and responses via the record marking protocol [RFC5531].¶
Because efficient handling is required when sending large amounts of data, congestion control facilities are a significant concern.¶
- When NFSv4.1 is used over an IP-based network protocol, it is REQUIRED that the transport provide congestion control.¶
- When NFSv4.1 is used over a non-IP network protocol, it is RECOMMENDED that the transport provide congestion control.¶
To enhance the possibilities for interoperability, it is strongly recommended that NFSv4.1 client and server implementations support operation over the TCP transport protocol.¶
It is permissible for a connectionless transport to be used under NFSv4.1; however, reliable and in-order delivery of data combined with congestion control by the connectionless transport is REQUIRED. As a consequence, UDP by itself MUST NOT be used as an NFSv4.1 transport, although transports to be used for NFSv4.1 may be layered on UDP. NFSv4.1 assumes that a client transport address and server transport address used to send data over a transport together constitute a connection, even if the underlying transport eschews the concept of a connection.¶
5.7.2. Client and Server Transport Behavior
[Author Aside]: Section substantially revised to address unjustified use of RFC2119-defined keywords regarding retries and replace that with appropriate implementation advice.¶
When a connection-oriented transport (e.g., TCP) is used, the client and server are normally expected to maintain the use of connections already established for a considerable length of time. This is for a number reasons:¶
- This will prevent the weakening of the transport's congestion control mechanisms by the confusion resulting from dispersing a single burst of network traffic into multiple connections.¶
- This will improve performance for the WAN environment by eliminating the need for connection setup handshakes. This is particularly so given that each new connection requires the re-establishment of a connection id, and setting up new connections.¶
- The NFSv4.1 callback model requires the client and server to maintain a client-created backchannel (see Section 7.3.1) for the server to use so that any dropping of the connection will interfere with the use of the backchannel.¶
Although it is not fatal for a requester to retry without a disconnect between the request and retry, there are good reasons to avoid this practice. The retry does consume resources, especially with RDMA, where each request, retry or not, consumes a credit. Retries for no reason, especially retries sent shortly after the previous attempt, are a poor use of network bandwidth and defeat the purpose of a transport's inherent congestion control system.¶
There is no situation in which a replier is allowed to silently drop a request, whether the request is a retry or not. (The silent drop behavior of RPCSEC_GSS [RFC2203] is not relevant here since this behavior happens at the RPCSEC_GSS layer, which is at a lower layer in the request processing.) While the replier MAY disconnect the connection, if it does not do so, it is obligated to execute the request or return an appropriate error based on the contents of the reply cache (see Section 7.6.1).¶
When sending a reply, the replier MUST send the reply to the same full network address (e.g., if using an IP-based transport, the source port of the requester is part of the full network address) from which the requester sent the request. If using a connection-oriented transport, replies MUST be sent on the connection from which the request was received.¶
If a connection is dropped after the replier receives the request but before the replier sends the reply, the replier might have a pending reply. If a connection is established with the same source and destination full network address as the dropped connection, then the replier MUST NOT send the reply until the requester retries the request. The reason for this prohibition is that the requester MAY retry a request over a different connection (provided that connection is associated with the original request's session).¶
When using RDMA transports, there are other reasons for avoiding retries over the same connection:¶
- RDMA transports use "credits" to enforce flow control, where a credit is a right to a peer to transmit a message. If one peer were to retransmit a request (or reply), it would consume an additional credit. If the replier retransmitted a reply, it would certainly result in an RDMA connection loss, since the requester would typically only post a single receive buffer for each request. If the requester retransmitted a request, the additional credit consumed on the server might lead to RDMA connection failure unless the client accounted for it and decreased its available credit, leading to wasted resources.¶
- RDMA credits present a new issue to the reply cache in NFSv4.1. The reply cache may be used when a connection within a session is lost, such as after the client reconnects. Credit information is a dynamic property of the RDMA connection, and stale values must not be replayed from the cache. This implies that the reply cache contents must not be blindly used when replies are sent from it, and credit information appropriate to the channel must be refreshed by the RPC layer.¶
In addition, as described in Section 7.6.2, while a session is active, an NFSv4.1 requester that ceases to wait for an outstanding reply MUST take appropriate care to avoid that situation vitiating guarantees needed to maintain the exactly-once semantics needed for the successful operation of the session-based reply cache.¶
5.7.3. Ports
Historically, NFSv3 servers have listened over TCP port 2049. The registered port 2049 [RFC3232] for the NFS protocol should be the default configuration for NFSv4.1, although the port 20049 is used for NFSv4.1 layered on RPC-over-RDMA.¶
The use of a reserved port has been common for NFS implementations and it is expected that this will apply to NFSv4.1 as well. While the use of RPC binding protocols as described in [RFC1833] is a possibility, there is no requirement that servers provide support for such use.¶
In light of this, a client should avoid this sort of use unless it has good reason to expect such support to be present on the server, while accessing NFS services at the appropriate well-known port depending on the transport to be used.¶
6. Security-related Infrastructure
NFSv4.1 relies on the Infrastructure described by the NFSv4-wide security-related documents, currently [I-D.dnoveck-nfsv4-security] and [I-D.dnoveck-nfsv4-acls]. This infrastructure includes:¶
-
The RPC-based facilities to provide authentication, privacy and integrity, including facilities provided by the various authentication flavors and those provided at the transport layer.¶
-
The security negotiation facilities described in section 16 of the security document, as they have been enhanced to support selection of transport-layer facilities, as well as authentication flavors and associated services.¶
-
The authorization facilities described in Sections 5.3, 5.4, and 7 of the security document.¶
-
The audit and alarm facilities described in Section 13 of the security document.¶
There are, however, a number of places where the NFSv4-wide treatment needs to be supplemented to deal with NFSv4.1-specific features, requirements, and recommendations as discussed below:¶
-
The parallel NFS feature is described in Section 17, with security for it dealt with in Section 17.9. The case of the file layout type is described in Section 17 with security for it dealt with in Section 18.13.¶
The security for parallel NFS is dealt with in this specification even though it relies on the security infrastructure describe in the NFSv4-wide security document. As a result, it will be discussed in a restructured Section 26.¶
-
The handling of SECINFO and SECINFO_NO_NAME is complicated because both share the extensions to the negotiation process described in Section 12 of the NFSv4-wide security document, currently [I-D.dnoveck-nfsv4-security], while the former operation in present in all minor versions while the latter is specific to NFSv4.1.¶
-
The requirements and recommendations regarding associated security services are discussed in Section 6.1. The discussion had been modified to include the possibility that encryption at the RPC transport layer might obviate the need for these services, although the existing requirements and recommendations still stand.¶
[Author Aside]: Further work in this area is likely and there should be working group discussion of possible changes. Of particular concern is the use of "SHOULD" in connection with support for privacy, as it is not clear what might be valid reasons not to support this. The provision of confidentiality using transport-based encryption further complicates the matter, although it needs to be clear that the need that confidentiality be available in some form is strongly recommended.¶
6.1. NFSv4.1-specific Recommendations and Requirements Regarding Security Services
[Author Aside]: Significant revisions have been made to address the hole created by the fact that the discussion of client support of data privacy uses the word "SHOULD".¶
Via the GSS-API, RPCSEC_GSS can be used to identify and authenticate users on clients to servers, and servers to users. Authentication of the client itself is not provided but can be provided by RPC independently of the use of RPCSEC_GSS.¶
GSS-API can also perform integrity checking on the entire RPC message, including the RPC header, and on the arguments or results. Finally, privacy/confidentiality, usually via encryption, is a service available with RPCSEC_GSS. Privacy is provided for the arguments and results. Note that if privacy is selected, integrity, authentication, and identification are enabled. If privacy is not selected, but integrity is selected, authentication and identification are enabled. If integrity and privacy are not selected, but authentication is enabled, identification is enabled. RPCSEC_GSS does not provide identification as a separate service.¶
Although GSS-API has an authentication service distinct from its privacy and integrity services, GSS-API's authentication service is not used for RPCSEC_GSS's authentication service. Instead, each RPC request and response header is integrity protected with the GSS-API integrity service, and this allows RPCSEC_GSS to offer per-RPC authentication and identity. See [RFC2203] for more information.¶
NFSv4.1 client and servers MUST support RPCSEC_GSS's integrity and authentication service. NFSv4.1 servers MUST support RPCSEC_GSS's privacy service.¶
NFSv4.1 clients SHOULD support RPCSEC_GSS's privacy service. Given that it is has never been made clear, as required by the definition of "SHOULD in [RFC2119], it has to be assumed that this statement, appearing in previous specifications has been treated as providing permission for clients not to support RPCSEC_GSS privacy. In light of this situation, it needs to be understood that, with regard to the use of "SHOULD" above, valid reasons to bypass this recommendation are limited to the reliance of implementors on those previous specifications and the difficulty of changing them now.¶
The following consequences need to be kept in mind by those not providing such support:¶
- Any data accessed on connections for which rpc-tls support is not provided will be available, in the clear, to those with the ability to monitor network traffic on a network segment used to effect the access.¶
- The existence of clients without privacy support would make it difficult or impossible to enforce privacy constraints that would otherwise be straightforward. The effect is to undercut the process of security negotiation, which is the only possible way to provide confidentiality when rpc-tls encryption is not in effect.¶
The reader is directed to Section 18.3.1 of [I-D.dnoveck-nfsv4-security] for a more complete discussion of security issues regarding data in flight.¶
6.2. NFSv4.1-specific Details of Security Negotiation
Unlike NFSv4.0, which only has the SECINFO operation, NFSv4.1 has the SECINFO_NO_NAME operation as well. As a result, many of the details of performing security negotiation will different from those in other minor versions and need to be discussed in this document, in the sections below.¶
6.2.1. Put Filehandle Operations
The term "put filehandle operation" refers to PUTROOTFH, PUTPUBFH, PUTFH, and RESTOREFH. Each of the subsections herein describes how the server handles a subseries of operations that starts with a put filehandle operation.¶
6.2.1.1. Put Filehandle Operation + SAVEFH
The client is saving a filehandle for a future RESTOREFH, LINK, or RENAME. SAVEFH MUST NOT return NFS4ERR_WRONGSEC. To determine whether or not the put filehandle operation returns NFS4ERR_WRONGSEC, the server implementation pretends SAVEFH is not in the series of operations and examines which of the situations described in the other subsections of Section 6.2.1 apply.¶
6.2.1.2. Two or More Put Filehandle Operations
For a series of N put filehandle operations, the server MUST NOT return NFS4ERR_WRONGSEC to the first N-1 put filehandle operations. The Nth put filehandle operation is handled as if it is the first in a subseries of operations. For example, if the server received a COMPOUND request with this series of operations -- PUTFH, PUTROOTFH, LOOKUP -- then the PUTFH operation is ignored for NFS4ERR_WRONGSEC purposes, and the PUTROOTFH, LOOKUP subseries is processed as according to Section 6.2.1.3.¶
6.2.1.3. Put Filehandle Operation + LOOKUP (or OPEN of an Existing Name)
This situation also applies to a put filehandle operation followed by a LOOKUP or an OPEN operation that specifies an existing component name.¶
In this situation, the client is potentially crossing a security policy boundary, and the set of security tuples the parent directory supports may differ from those of the child. The server implementation may decide whether to impose any restrictions on security policy administration. There are at least three approaches (sec_policy_child is the tuple set of the child export, sec_policy_parent is that of the parent).¶
- (a)
- sec_policy_child <= sec_policy_parent (<= for subset). This means that the set of security tuples specified on the security policy of a child directory is always a subset of its parent directory.¶
- (b)
- sec_policy_child ^ sec_policy_parent != {} (^ for intersection, {} for the empty set). This means that the set of security tuples specified on the security policy of a child directory always has a non-empty intersection with that of the parent.¶
- (c)
- sec_policy_child ^ sec_policy_parent == {}. This means that the set of security tuples specified on the security policy of a child directory may not intersect with that of the parent. In other words, there are no restrictions on how the system administrator may set up these tuples.¶
In order for a server to support approaches (b) (for the case when a client chooses a flavor that is not a member of sec_policy_parent) and (c), the put filehandle operation cannot return NFS4ERR_WRONGSEC when there is a security tuple mismatch. Instead, it should be returned from the LOOKUP (or OPEN by existing component name) that follows.¶
Since the above guideline does not contradict approach (a), it should be followed in general. Even if approach (a) is implemented, it is possible for the security tuple used to be acceptable for the target of LOOKUP but not for the filehandles used in the put filehandle operation. The put filehandle operation could be a PUTROOTFH or PUTPUBFH, where the client cannot know the security tuples for the root or public filehandle. Or the security policy for the filehandle used by the put filehandle operation could have changed since the time the filehandle was obtained.¶
Therefore, an NFSv4.1 server MUST NOT return NFS4ERR_WRONGSEC in response to the put filehandle operation if the operation is immediately followed by a LOOKUP or an OPEN by component name.¶
6.2.1.4. Put Filehandle Operation + LOOKUPP
Since SECINFO only works its way down, there is no way LOOKUPP can return NFS4ERR_WRONGSEC without SECINFO_NO_NAME. SECINFO_NO_NAME solves this issue via style SECINFO_STYLE4_PARENT, which works in the opposite direction as SECINFO. As with Section 6.2.1.3, a put filehandle operation that is followed by a LOOKUPP MUST NOT return NFS4ERR_WRONGSEC. If the server does not support SECINFO_NO_NAME, the client's only recourse is to send the put filehandle operation, LOOKUPP, GETFH sequence of operations with every security tuple it supports.¶
Regardless of whether SECINFO_NO_NAME is supported, an NFSv4.1 server MUST NOT return NFS4ERR_WRONGSEC in response to a put filehandle operation if the operation is immediately followed by a LOOKUPP.¶
6.2.1.5. Put Filehandle Operation + SECINFO/SECINFO_NO_NAME
A security-sensitive client is allowed to choose a strong security tuple when querying a server to determine a file object's permitted security tuples. The security tuple chosen by the client does not have to be included in the tuple list of the security policy of either the parent directory indicated in the put filehandle operation or the child file object indicated in SECINFO (or any parent directory indicated in SECINFO_NO_NAME). Of course, the server has to be configured for whatever security tuple the client selects; otherwise, the request will fail at the RPC layer with an appropriate authentication error.¶
In theory, there is no connection between the security flavor used by SECINFO or SECINFO_NO_NAME and those supported by the security policy. But in practice, the client may start looking for strong flavors from those supported by the security policy, followed by those in the REQUIRED set.¶
The NFSv4.1 server MUST NOT return NFS4ERR_WRONGSEC to a put filehandle operation that is immediately followed by SECINFO or SECINFO_NO_NAME. The NFSv4.1 server MUST NOT return NFS4ERR_WRONGSEC from SECINFO or SECINFO_NO_NAME.¶
6.2.1.6. Put Filehandle Operation + Nothing
The NFSv4.1 server MUST NOT return NFS4ERR_WRONGSEC.¶
6.2.1.7. Put Filehandle Operation + Anything Else
"Anything Else" includes OPEN by filehandle.¶
The security policy enforcement applies to the filehandle specified in the put filehandle operation. Therefore, the put filehandle operation MUST return NFS4ERR_WRONGSEC when there is a security tuple mismatch. This avoids the complexity of adding NFS4ERR_WRONGSEC as an allowable error to every other operation.¶
A COMPOUND containing the series put filehandle operation + SECINFO_NO_NAME (style SECINFO_STYLE4_CURRENT_FH) is an efficient way for the client to recover from NFS4ERR_WRONGSEC.¶
The NFSv4.1 server MUST NOT return NFS4ERR_WRONGSEC to any operation other than a put filehandle operation, LOOKUP, LOOKUPP, and OPEN (by component name).¶
6.2.1.8. Operations after SECINFO and SECINFO_NO_NAME
Suppose a client sends a COMPOUND procedure containing the series SEQUENCE, PUTFH, SECINFO_NO_NAME, READ, and suppose the security tuple used does not match that required for the target file. By rule (see Section 6.2.1.5), neither PUTFH nor SECINFO_NO_NAME can return NFS4ERR_WRONGSEC. By rule (see Section 6.2.1.7), READ cannot return NFS4ERR_WRONGSEC. The issue is resolved by the fact that SECINFO and SECINFO_NO_NAME consume the current filehandle (note that this is a change from NFSv4.0). This leaves no current filehandle for READ to use, and READ returns NFS4ERR_NOFILEHANDLE.¶
6.2.2. LINK and RENAME
The LINK and RENAME operations use both the current and saved filehandles. If the security policy of the saved filehandle rejects the security flavor used in the COMPOUND request's credentials, the server MAY return NFS4ERR_WRONGSEC from LINK or RENAME. When the server does so, if there is no intersection between the security policies of saved and current filehandles, this implies that it will be impossible for the client to perform the intended LINK or RENAME operation.¶
For example, suppose the client sends this COMPOUND request: SEQUENCE, PUTFH bFH, SAVEFH, PUTFH aFH, RENAME "c" "d", where filehandles bFH and aFH refer to different directories. Suppose no common security tuple exists between the security policies of aFH and bFH. If the client sends the request using credentials acceptable to bFH's security policy but not aFH's policy, then the PUTFH aFH operation will fail with NFS4ERR_WRONGSEC. After a SECINFO_NO_NAME request, the client sends SEQUENCE, PUTFH bFH, SAVEFH, PUTFH aFH, RENAME "c" "d", using credentials acceptable to aFH's security policy but not bFH's policy. The server returns NFS4ERR_WRONGSEC on the RENAME operation.¶
To prevent a client from an endless sequence of a request containing LINK or RENAME, followed by a request containing SECINFO_NO_NAME or SECINFO, the server MUST detect when the security policies of the current and saved filehandles have no mutually acceptable security tuple, and MUST NOT return NFS4ERR_WRONGSEC from LINK or RENAME in that situation. Instead the server MUST do one of two things:¶
7. Session
NFSv4.1 clients and servers MUST support and MUST use the session feature as described in this section.¶
7.1. Motivation and Overview
Previous versions and minor versions of NFS have suffered from the following:¶
- Lack of support for Exactly Once Semantics (EOS). This includes lack of support for EOS through server failure and recovery.¶
- Limited callback support, including no support for sending callbacks through firewalls, and races between replies to normal requests and callbacks.¶
- Limited trunking over multiple network paths.¶
- Requiring machine credentials for fully secure operation.¶
Through the introduction of a session, NFSv4.1 addresses the above shortfalls with practical solutions:¶
- EOS is enabled by a reply cache with a bounded size, making it feasible to keep the cache in persistent storage and enable EOS through server failure and recovery. One reason than previous revisions of NFS did not support EOS was because some EOS approaches often limited parallelism. As will be explained in Section 7.6, NFSv4.1 supports EOS without unduly limiting parallelism.¶
- The NFSv4.1 client (defined in Section 2.5) creates transport connections and provides them to the server to use for sending callback requests, thus solving the firewall issue (Section 23.34). Races between responses from client requests and callbacks caused by the requests are detected via the session's sequencing properties that are a consequence of EOS (Section 7.6.3).¶
- The NFSv4.1 client can associate an arbitrary number of connections with the session, and thus provide trunking (Section 7.5).¶
- The NFSv4.1 client and server produce a session key independent of client and server machine credentials which can be used to compute a digest for protecting critical session management operations (Section 7.8.3).¶
- The NFSv4.1 client can also create secure RPCSEC_GSS contexts for use by the session's backchannel that do not require the server to authenticate to a client machine principal (Section 7.8.2).¶
A session is a dynamically created, long-lived server object created by a client and used over time from one or more transport connections. Its function is to maintain the server's state relative to the connection(s) belonging to a client instance. This state is entirely independent of the connection itself, and indeed the state exists whether or not the connection exists. A client may have one or more sessions associated with it so that client-associated state may be accessed using any of the sessions associated with that client's client ID, when connections are associated with those sessions. When no connections are associated with any of a client ID's sessions for an extended time, such objects as locks, opens, delegations, layouts, etc. are subject to expiration. The session serves as an object representing a means of access by a client to the associated client state on the server, independent of the physical means of access to that state.¶
A single client may create multiple sessions. A single session MUST NOT serve multiple clients.¶
7.2. NFSv4 Integration
Sessions are part of NFSv4.1 and not NFSv4.0. Normally, a major infrastructure change such as sessions would require a new major version number to an Open Network Computing (ONC) RPC program like NFS. However, because NFSv4 encapsulates its functionality in a single procedure, COMPOUND, and because COMPOUND can support an arbitrary number of operations, sessions have been added to NFSv4.1 with little difficulty. COMPOUND includes a minor version number field, and for NFSv4.1 this minor version is set to 1. When the NFSv4 server processes a COMPOUND with the minor version set to 1, it expects a different set of operations than it does for NFSv4.0. NFSv4.1 defines the SEQUENCE operation, which is required for every COMPOUND that operates over an established session, with the exception of some session administration operations, such as DESTROY_SESSION (Section 23.37).¶
7.2.1. SEQUENCE and CB_SEQUENCE
In NFSv4.1, when the SEQUENCE operation is present, it MUST be the first operation in the COMPOUND procedure. The primary purpose of SEQUENCE is to carry the session identifier. The session identifier associates all other operations in the COMPOUND procedure with a particular session. SEQUENCE also contains required information for maintaining EOS (see Section 7.6). Session-enabled NFSv4.1 COMPOUND requests thus have the form:¶
+-----+--------------+-----------+------------+-----------+---- | tag | minorversion | numops |SEQUENCE op | op + args | ... | | (== 1) | (limited) | + args | | +-----+--------------+-----------+------------+-----------+----¶
and the replies have the form:¶
+------------+-----+--------+-------------------------------+--// |last status | tag | numres |status + SEQUENCE op + results | // +------------+-----+--------+-------------------------------+--// //-----------------------+---- // status + op + results | ... //-----------------------+----¶
A CB_COMPOUND procedure request and reply has a similar form to COMPOUND, but instead of a SEQUENCE operation, there is a CB_SEQUENCE operation. CB_COMPOUND also has an additional field called "callback_ident", which is superfluous in NFSv4.1 and MUST be ignored by the client. CB_SEQUENCE has the same information as SEQUENCE, and also includes other information needed to resolve callback races (Section 7.6.3).¶
7.2.2. Client ID and Session Association
Each client ID (Section 5.5) can have zero or more active sessions. A client ID and associated session are required to perform file access in NFSv4.1. Each time a session is used (whether by a client sending a request to the server or the client replying to a callback request from the server), the state leased to its associated client ID is automatically renewed.¶
State (which can consist of share reservations, locks, delegations, and layouts (Section 3)) is tied to the client ID. Client state is not tied to any individual session. Successive state changing operations from a given state owner MAY go over different sessions, provided the session is associated with the same client ID. A callback MAY arrive over a different session than that of the request that originally acquired the state pertaining to the callback. For example, if session A is used to acquire a delegation, a request to recall the delegation MAY arrive over session B if both sessions are associated with the same client ID. Sections 7.8.1 and 7.8.2 discuss the security considerations around callbacks.¶
7.3. Channels
A channel is not a connection. A channel represents a single direction in which ONC RPC requests are sent as part of a session..¶
Each session has one or two channels: the fore channel and the backchannel. Because there are at most two channels per session, and because each channel has a distinct purpose, channels are not assigned identifiers.¶
The fore channel is used for ordinary requests from the client to the server, and carries COMPOUND requests and responses. A session always has a fore channel.¶
The backchannel is used for callback requests from server to client, and carries CB_COMPOUND requests and responses. Whether or not there is a backchannel is decided by the client; however, many features of NFSv4.1 require a backchannel. NFSv4.1 servers MUST support backchannels.¶
Each session has resources for each channel, including separate reply caches (see Section 7.6.1). Note that even the backchannel requires a reply cache (or, at least, a slot table in order to detect retries) because some callback operations are non-idempotent.¶
7.3.1. Association of Connections, Channels, and Sessions
Each channel is associated with zero or more transport connections (whether of the same transport protocol or different transport protocols). A connection can be associated with one channel or both channels of a session; the client and server negotiate whether a connection will carry traffic for one channel or both channels via the CREATE_SESSION (Section 23.36) and the BIND_CONN_TO_SESSION (Section 23.34) operations. When a session is created via CREATE_SESSION, the connection that transported the CREATE_SESSION request is automatically associated with the fore channel, and optionally the backchannel. If the client specifies no state protection (Section 23.35) when the session is created, then when SEQUENCE is transmitted on a different connection, the connection is automatically associated with the fore channel of the session specified in the SEQUENCE operation.¶
A connection's association with a session is not exclusive. A connection associated with the channel(s) of one session may be simultaneously associated with the channel(s) of other sessions including sessions associated with other client IDs.¶
It is permissible for connections of multiple transport types to be associated with the same channel. For example, both TCP and RDMA connections can be associated with the fore channel. In the event an RDMA and non-RDMA connection are associated with the same channel, it is desirable for the maximum number slots to be at least one more than the total number of RDMA credits (Section 7.6.1). This way, if all RDMA credits are used, the non-RDMA connection can have at least one outstanding request. If a server supports multiple transport types, it MUST allow a client to associate connections from each transport to a channel.¶
It is permissible for a connection of one type of transport to be associated with the fore channel, and a connection of a different type to be associated with the backchannel.¶
7.4. Server Scope
Servers each specify a server scope value in the form of an opaque string eir_server_scope returned as part of the results of an EXCHANGE_ID operation. The purpose of the server scope is to allow a group of servers to indicate to clients that a set of servers sharing the same server scope value has arranged to use distinct values of opaque identifiers so that the two servers never assign the same value to two distinct objects. Thus, the identifiers generated by two servers within that set can be assumed compatible so that, in certain important cases, identifiers generated by one server in that set may be presented to another server of the same scope.¶
The use of such compatible values does not imply that a value generated by one server will always be accepted by another. In most cases, it will not. However, a server will not inadvertently accept a value generated by another server. When it does accept it, it will be because it is recognized as valid and carrying the same meaning as on another server of the same scope.¶
When servers are of the same server scope, this compatibility of values applies to the following identifiers:¶
- Filehandle values. A filehandle value accepted by two servers of the same server scope denotes the same object. A WRITE operation sent to one server is reflected immediately in a READ sent to the other.¶
- Server owner values. When the server scope values are the same, server owner value may be validly compared. In cases where the server scope values are different, server owner values are treated as different even if they contain identical strings of bytes.¶
The coordination among servers required to provide such compatibility can be quite minimal, and limited to a simple partition of the ID space. The recognition of common values requires additional implementation, but this can be tailored to the specific situations in which that recognition is desired.¶
Clients will have occasion to compare the server scope values of multiple servers under a number of circumstances, each of which will be discussed under the appropriate functional section:¶
- When server owner values received in response to EXCHANGE_ID operations sent to multiple network addresses are compared for the purpose of determining the validity of various forms of trunking, as described in Section 16.5.2.¶
- When network or server reconfiguration causes the same network address to possibly be directed to different servers, with the necessity for the client to determine when lock reclaim should be attempted, as described in Section 13.4.2.1.¶
When two replies from EXCHANGE_ID, each from two different server network addresses, have the same server scope, there are a number of ways a client can validate that the common server scope is due to two servers cooperating in a group.¶
- If both EXCHANGE_ID requests were sent with RPCSEC_GSS ([RFC2203], [RFC5403], [RFC7861]) authentication and the server principal is the same for both targets, the equality of server scope is validated. It is RECOMMENDED that two servers intending to share the same server scope and server_owner major_id also share the same principal name. In some cases, this simplifies the client's task of validating server scope.¶
- The client may accept the appearance of the second server in the fs_locations or fs_locations_info attribute for a relevant file system. For example, if there is a migration event for a particular file system or there are locks to be reclaimed on a particular file system, the attributes for that particular file system may be used. The client sends the GETATTR request to the first server for the fs_locations or fs_locations_info attribute with RPCSEC_GSS authentication. It may need to do this in advance of the need to verify the common server scope. If the client successfully authenticates the reply to GETATTR, and the GETATTR request and reply containing the fs_locations or fs_locations_info attribute refers to the second server, then the equality of server scope is supported. A client may choose to limit the use of this form of support to information relevant to the specific file system involved (e.g. a file system being migrated).¶
7.5. Trunking
Trunking is the use of multiple connections between a client and server in order to increase the speed of data transfer. NFSv4.1 supports two types of trunking: session trunking and client ID trunking.¶
In the context of a single server network address, it can be assumed that all connections are accessing the same server, and NFSv4.1 servers MUST support both forms of trunking. When multiple connections use a set of network addresses to access the same server, the server MUST support both forms of trunking. NFSv4.1 servers in a clustered configuration MAY allow network addresses for different servers to use client ID trunking.¶
Clients may use either form of trunking as long as they do not, when trunking between different server network addresses, violate the servers' mandates as to the kinds of trunking to be allowed (see below). With regard to callback channels, the client MUST allow the server to choose among all callback channels valid for a given client ID and MUST support trunking when the connections supporting the backchannel allow session or client ID trunking to be used for callbacks.¶
Session trunking is essentially the association of multiple connections, each with potentially different target and/or source network addresses, to the same session. When the target network addresses (server addresses) of the two connections are the same, the server MUST support such session trunking. When the target network addresses are different, the server MAY indicate such support using the data returned by the EXCHANGE_ID operation (see below).¶
Client ID trunking is the association of multiple sessions to the same client ID. Servers MUST support client ID trunking for two target network addresses whenever they allow session trunking for those same two network addresses. In addition, a server MAY, by presenting the same major server owner ID (Section 5.6) and server scope (Section 7.4), allow an additional case of client ID trunking. When two servers return the same major server owner and server scope, it means that the two servers are cooperating on locking state management, which is a prerequisite for client ID trunking.¶
Distinguishing when the client is allowed to use session and client ID trunking requires understanding how the results of the EXCHANGE_ID (Section 23.35) operation identify a server. Suppose a client sends EXCHANGE_IDs over two different connections, each with a possibly different target network address, but each EXCHANGE_ID operation has the same value in the eia_clientowner field. If the same NFSv4.1 server is listening over each connection, then each EXCHANGE_ID result MUST return the same values of eir_clientid, eir_server_owner.so_major_id, and eir_server_scope. The client can then treat each connection as referring to the same server (subject to verification; see Section 7.5.1 below), and it can use each connection to trunk requests and replies. The client's choice is whether session trunking or client ID trunking applies.¶
- Session Trunking.
-
If the eia_clientowner argument is the same in two different EXCHANGE_ID requests, and the eir_clientid, eir_server_owner.so_major_id, eir_server_owner.so_minor_id, and eir_server_scope results match in both EXCHANGE_ID results, then the client is permitted to perform session trunking. If the client has no session mapping to the tuple of eir_clientid, eir_server_owner.so_major_id, eir_server_scope, and eir_server_owner.so_minor_id, then it creates the session via a CREATE_SESSION operation over one of the connections, which associates the connection to the session. If there is a session for the tuple, the client can send BIND_CONN_TO_SESSION to associate the connection to the session.¶
Of course, if the client does not desire to use session trunking, it is not required to do so. It can invoke CREATE_SESSION on the connection. This will result in client ID trunking as described below. It can also decide to drop the connection if it does not choose to use trunking.¶
- Client ID Trunking.
-
If the eia_clientowner argument is the same in two different EXCHANGE_ID requests, and the eir_clientid, eir_server_owner.so_major_id, and eir_server_scope results match in both EXCHANGE_ID results, then the client is permitted to perform client ID trunking (regardless of whether the eir_server_owner.so_minor_id results match). The client can associate each connection with different sessions, where each session is associated with the same server.¶
The client completes the act of client ID trunking by invoking CREATE_SESSION on each connection, using the same client ID that was returned in eir_clientid. These invocations create two sessions and also associate each connection with its respective session. The client is free to decline to use client ID trunking by simply dropping the connection at this point.¶
When doing client ID trunking, locking state is shared across sessions associated with that same client ID. This requires the server to coordinate state across sessions and the client to be able to associate the same locking state with multiple sessions.¶
It is always possible that, as a result of various sorts of reconfiguration events, eir_server_scope and eir_server_owner values may be different on subsequent EXCHANGE_ID requests made to the same network address.¶
In most cases, such reconfiguration events will be disruptive and indicate that an IP address formerly connected to one server is now connected to an entirely different one.¶
Some guidelines on client handling of such situations follow:¶
- When eir_server_scope changes, the client has no assurance that any IDs that it obtained previously (e.g., filehandles) can be validly used on the new server, and, even if the new server accepts them, there is no assurance that this is not due to accident. Thus, it is best to treat all such state as lost or stale, although a client may assume that the probability of inadvertent acceptance is low and treat this situation as within the next case.¶
- When eir_server_scope remains the same and eir_server_owner.so_major_id changes, the client can use the filehandles it has, consider its locking state lost, and attempt to reclaim or otherwise re-obtain its locks. It might find that its filehandle is now stale. However, if NFS4ERR_STALE is not returned, it can proceed to reclaim or otherwise re-obtain its open locking state.¶
- When eir_server_scope and eir_server_owner.so_major_id remain the same, the client has to use the now-current values of eir_server_owner.so_minor_id in deciding on appropriate forms of trunking. This may result in connections being dropped or new sessions being created.¶
7.5.1. Verifying Claims of Matching Server Identity
When the server responds using two different connections that claim matching or partially matching eir_server_owner, eir_server_scope, and eir_clientid values, the client does not have to trust the servers' claims. The client may verify these claims before trunking traffic in the following ways:¶
-
For session trunking, clients SHOULD reliably verify if connections between different network paths are in fact associated with the same NFSv4.1 server and usable on the same session, and servers MUST allow clients to perform reliable verification. When a client ID is created, the client SHOULD, unless client host authentication is in effect, specify that BIND_CONN_TO_SESSION is to be verified according to the SP4_SSV or SP4_MACH_CRED (Section 23.35) state protection options. For SP4_SSV, reliable verification depends on a shared secret (the SSV) that is established via the SET_SSV (see Section 23.47) operation.¶
When a new connection is associated with the session (via the BIND_CONN_TO_SESSION operation, see Section 23.34), if the client specified SP4_SSV state protection for the BIND_CONN_TO_SESSION operation, the client MUST send the BIND_CONN_TO_SESSION with RPCSEC_GSS protection, using integrity or privacy, and an RPCSEC_GSS handle created with the GSS SSV mechanism (see Section 7.9).¶
If the client mistakenly tries to associate a connection to a session of a wrong server, the server will either reject the attempt because it is not aware of the session identifier of the BIND_CONN_TO_SESSION arguments, or it will reject the attempt because the RPCSEC_GSS authentication fails. Even if the server mistakenly or maliciously accepts the connection association attempt, the RPCSEC_GSS verifier it computes in the response will not be verified by the client, so the client will know it cannot use the connection for trunking the specified session.¶
If the client specified SP4_MACH_CRED state protection, the BIND_CONN_TO_SESSION operation will use RPCSEC_GSS integrity or privacy, using the same credential that was used when the client ID was created. Mutual authentication via RPCSEC_GSS assures the client that the connection is associated with the correct session of the correct server.¶
-
For client ID trunking, the client has at least two options for verifying that the same client ID obtained from two different EXCHANGE_ID operations came from the same server. The first option is to use RPCSEC_GSS authentication when sending each EXCHANGE_ID operation. Each time an EXCHANGE_ID is sent with RPCSEC_GSS authentication, the client notes the principal name of the GSS target. If the EXCHANGE_ID results indicate that client ID trunking is possible, and the GSS targets' principal names are the same, the servers are the same and client ID trunking is allowed.¶
The second option for verification is to use SP4_SSV protection. When the client sends EXCHANGE_ID, it specifies SP4_SSV protection. The first EXCHANGE_ID the client sends always has to be confirmed by a CREATE_SESSION call. The client then sends SET_SSV. Later, the client sends EXCHANGE_ID to a second destination network address different from the one the first EXCHANGE_ID was sent to. The client checks that each EXCHANGE_ID reply has the same eir_clientid, eir_server_owner.so_major_id, and eir_server_scope. If so, the client verifies the claim by sending a CREATE_SESSION operation to the second destination address, protected with RPCSEC_GSS integrity using an RPCSEC_GSS handle returned by the second EXCHANGE_ID. If the server accepts the CREATE_SESSION request, and if the client verifies the RPCSEC_GSS verifier and integrity codes, then the client has proof the second server knows the SSV, and thus the two servers are cooperating for the purposes of specifying server scope and client ID trunking.¶
7.6. Exactly Once Semantics
[Author Aside]: This section, including some subsections, has been substantially modified from the corresponding section appearing in previous specifications [RFC5661] [RFC8881] and earlier drafts of this document. Change has been driven primarily by the incorrect use of RFC2119-defined keywords, most importantly in the case in which RPC requests need to be aborted, leading to some related changes to clarify the appropriate level of checking for the possibility of false retry. As part of this revised description, it is explained that, given the possibility of requests being aborted, the term "Exactly-once semantics" describes an aspiration and that what is really provided would better be called "at-most-once semantics. Also, the description of retry has been revised to properly use RFC2119 keywords. For more detailed information regarding changes which have been made, see Appendix C.2.1.¶
Via the session, NFSv4.1 offers what is termed "exactly once semantics" (EOS) for requests sent over a channel. EOS is supported on both the fore channel and backchannel.¶
Although this term is well-established and will not be changed, it should be noted that what is actually provided is at-most-once semantics to accommodate the possibility that the client will need to abort RPC requests, remaining unsure about whether the requested actions have been performed one time or not at all.¶
Each COMPOUND or CB_COMPOUND request that is sent with a leading SEQUENCE or CB_SEQUENCE operation needs to be executed by the receiver at most once. This requirement holds regardless of whether the request is sent with reply caching specified (see Section 7.6.1.3). The requirement also holds in the case in which NFSv4.1 is a pNFS data access protocol and the requester is sending the request over a session created between a pNFS data client and pNFS data server. To help understand the need for this requirement, we divide the requests sent to be executed into three categories:¶
An example of a non-idempotent request is RENAME. Obviously, if a replier executes the same RENAME request twice, and the first execution succeeds, the re-execution will fail. If the replier returns the result from the re-execution, this result is incorrect. Therefore, EOS is required for non-idempotent requests.¶
An example of an idempotent modifying request is a COMPOUND request containing a WRITE operation. Repeated execution of the same WRITE has the same effect as execution of that WRITE a single time. Nevertheless, enforcing EOS for WRITEs and other idempotent modifying requests is necessary to avoid data corruption, which could result from executing the same write request multiple times including some executions that occur after the completion of the first is noted by the requester.¶
Suppose a client sends WRITE A to a noncompliant server that does not enforce EOS, and receives no response, perhaps due to a network partition. The client reconnects to the server and re-sends WRITE A. Now, the server has outstanding two instances of A. The server can be in a situation in which it executes and replies to the retry of A, while the first A is still waiting in the server's internal I/O system for some resource. Upon receiving the reply to the second attempt of WRITE A, the client believes its WRITE is done so it is free to send WRITE B, which overlaps the byte-range of A. When the original A is dispatched from the server's I/O system and executed (thus the second time A will have been written), then what has been written by B can be overwritten and thus corrupted.¶
An example of an idempotent non-modifying request is a COMPOUND containing SEQUENCE, PUTFH, READLINK, and nothing else. The re-execution of such a request will not cause data corruption or produce an incorrect result. Nonetheless, to keep the implementation simple, the replier MUST enforce EOS for all requests, whether or not they are idempotent or modifying.¶
Note that fully complete EOS is not possible unless the server persists the reply cache in stable storage, and unless the server is somehow implemented to never require a restart (indeed, if such a server exists, the distinction between a reply cache kept in stable storage versus one that is not is one without meaning). See Section 8 for a discussion of persistence in the reply cache. Regardless, even if the server does not persist the reply cache, EOS improves robustness and correctness relative to previous versions of NFS because the earlier duplicate request/reply caches were based on the ONC RPC transaction identifier (XID). Section 7.6.1 explains the shortcomings of the XID as a basis for a reply cache and describes how NFSv4.1 sessions improve upon the XID.¶
7.6.1. Slot Identifiers and Reply Cache
The RPC layer provides a transaction ID (XID), which, while required to be unique, is not convenient for tracking requests for two reasons. First, the XID is only meaningful to the requester; it cannot be interpreted by the replier except to test for equality with previously sent requests. When consulting an RPC-based duplicate request cache, the opaqueness of the XID requires a computationally expensive look up (often via a hash that includes XID and source address). NFSv4.1 requests include a non-opaque slot ID, which can be used as an index into a slot table, which is far more efficient. Second, because RPC requests can be executed by the replier in any order, there is no bound on the number of requests that may be outstanding at any time. To achieve perfect EOS, using ONC RPC would require storing all replies in the reply cache. XIDs are 32 bits; storing over four billion (232) replies in the reply cache is not practical. In practice, previous versions of NFS have chosen to store a fixed number of replies in the cache, and to use a least recently used (LRU) approach to replacing cache entries with new entries when the cache is full. In NFSv4.1, the number of outstanding requests is bounded by the size of the slot table, and a sequence ID per slot is used to tell the replier when it is safe to delete a cached reply.¶
In the NFSv4.1 reply cache, when the requester sends a new request, it selects a slot ID in the range 0..N, where N is the replier's current maximum slot ID granted to the requester on the session over which the request is to be sent. The value of N starts out as equal to ca_maxrequests - 1 (Section 23.36), but can be adjusted by the response to SEQUENCE or CB_SEQUENCE as described later in this section. The slot ID must be unused by any of the requests that the requester has already active on the session. "Unused" here means the requester has no outstanding request for that slot ID.¶
A slot contains a sequence ID and the cached reply corresponding to the request sent with that sequence ID. The sequence ID is a 32-bit unsigned value, and is therefore in the range 0..0xFFFFFFFF (232 - 1). The first time a slot is used, the requester MUST specify a sequence ID of one (Section 23.36). Each time a slot is reused, the request MUST specify a sequence ID that is one greater than that of the previous request on the slot. If the previous sequence ID was 0xFFFFFFFF, then the next request for the slot MUST have the sequence ID set to zero (i.e., (232 - 1) + 1 mod 232).¶
The sequence ID accompanies the slot ID in each request. It is for the critical check at the replier: it used to efficiently determine whether a request using a certain slot ID is a retransmit or a new, never-before-seen request. It is not feasible for the requester to assert that it is retransmitting to implement this, because for any given request the requester cannot know whether the replier has seen it unless the replier actually replies. Of course, if the requester has seen the reply, the requester would not retransmit.¶
The replier compares each received request's sequence ID with the last one previously received for that slot ID, to see if the new request is:¶
- A new request, in which the sequence ID is one greater than that previously seen in the slot (accounting for sequence wraparound). The replier proceeds to execute the new request, and the replier MUST increase the slot's sequence ID by one.¶
- A retransmitted request, in which the sequence ID is equal to that currently recorded in the slot. If the original request has executed to completion, the replier returns the cached reply. See Section 7.6.2 for direction on how the replier deals with retries of requests that are still in progress.¶
- A misordered retry, in which the sequence ID is less than (accounting for sequence wraparound) that previously seen in the slot. The replier MUST return NFS4ERR_SEQ_MISORDERED (as the result from SEQUENCE or CB_SEQUENCE).¶
- A misordered new request, in which the sequence ID is two or more than (accounting for sequence wraparound) that previously seen in the slot. Note that because the sequence ID MUST wrap around to zero once it reaches 0xFFFFFFFF, a misordered new request and a misordered retry cannot be distinguished. Thus, the replier MUST return NFS4ERR_SEQ_MISORDERED (as the result from SEQUENCE or CB_SEQUENCE).¶
Unlike the XID, the slot ID is always within a specific range; this has two implications. The first implication is that for a given session, the replier need only cache the results of a limited number of COMPOUND requests. The second implication derives from the first, which is that unlike XID-indexed reply caches (also known as duplicate request caches - DRCs), the slot ID-based reply cache cannot be overflowed. Through use of the sequence ID to identify retransmitted requests, the replier does not need to actually cache the request itself, reducing the storage requirements of the reply cache further. These facilities make it practical to maintain all the required entries for an effective reply cache.¶
As a result, the slot ID, sequence ID, and session ID take over the traditional role of the XID and source network address in the replier's reply cache implementation. This approach is considerably more portable and completely robust -- it is not subject to the reassignment of ports as clients reconnect over IP networks. In addition, the RPC XID is not used in the reply cache, enhancing robustness of the cache in the face of any rapid reuse of XIDs by the requester. While the replier does not care about the XID for the purposes of reply cache management (but the replier MUST return the same XID that was in the request), nonetheless there are considerations for the XID in NFSv4.1 that are the same as all other previous versions of NFS. The RPC XID remains in each message and needs to be formulated in NFSv4.1 requests as in any other ONC RPC request. The reasons include:¶
- The RPC layer retains its existing semantics and implementation.¶
- The requester and replier must be able to interoperate at the RPC layer, prior to the NFSv4.1 decoding of the SEQUENCE or CB_SEQUENCE operation.¶
- If an operation is being used that does not start with SEQUENCE or CB_SEQUENCE (e.g., BIND_CONN_TO_SESSION), then the RPC XID is needed for correct operation to match the reply to the request.¶
- The SEQUENCE or CB_SEQUENCE operation may generate an error. If so, the embedded slot ID, sequence ID, and session ID (if present) in the request will not be in the reply, and the requester has only the XID to match the reply to the request.¶
Given that well-formulated XIDs continue to be required, this raises the question: why do SEQUENCE and CB_SEQUENCE replies have a session ID, slot ID, and sequence ID? Having the session ID in the reply means that the requester does not have to use the XID to look up the session ID, which would be necessary if the connection were associated with multiple sessions. Having the slot ID and sequence ID in the reply means that the requester does not have to use the XID to look up the slot ID and sequence ID. Furthermore, since the XID is only 32 bits, it is too small to guarantee the re-association of a reply with its request (See [rpc_xid_issues]); having session ID, slot ID, and sequence ID in the reply allows the client to validate that the reply in fact belongs to the matched request.¶
The SEQUENCE (and CB_SEQUENCE) operation also carries a "highest_slotid" value, which carries additional requester slot usage information. The requester MUST always indicate the slot ID representing the outstanding request with the highest-numbered slot value. The requester should in all cases provide the most conservative value possible, although it can be increased somewhat above the actual instantaneous usage to maintain some minimum or optimal level. This provides a way for the requester to yield unused request slots back to the replier, which in turn can use the information to reallocate resources.¶
The replier responds with both a new target highest_slotid and an enforced highest_slotid, described as follows:¶
-
The target highest_slotid is an indication to the requester of the highest_slotid the replier wishes the requester to be using. This permits the replier to withdraw (or add) resources from a requester that has been found to not be using them, in order to more fairly share resources among a varying level of demand from other requesters. The requester must always comply with the replier's value updates, since they indicate newly established hard limits on the requester's access to session resources. However, because of request pipelining, the requester might have active requests in flight reflecting prior values; therefore, the replier cannot immediately require the requester to comply.¶
-
The enforced highest_slotid indicates the highest slot ID the requester is permitted to use on a subsequent SEQUENCE or CB_SEQUENCE operation. The replier's enforced highest_slotid SHOULD be no less than the highest_slotid the requester indicated in the SEQUENCE or CB_SEQUENCE arguments.¶
A requester can be intransigent with respect to lowering its highest_slotid argument to a Sequence operation, i.e. the requester continues to ignore the target highest_slotid in the response to a Sequence operation, and continues to set its highest_slotid argument to be higher than the target highest_slotid. This can be considered particularly egregious behavior when the replier knows there are no outstanding requests with slot IDs higher than its target highest_slotid. When faced with such intransigence, the replier is free to take more forceful action, and MAY reply with a new enforced highest_slotid that is less than its previous enforced highest_slotid. Thereafter, if the requester continues to send requests with a highest_slotid that is greater than the replier's new enforced highest_slotid, the server MAY return NFS4ERR_BAD_HIGH_SLOT, unless the slot ID in the request is greater than the new enforced highest_slotid and the request is a retry.¶
The replier should retain the slots it wants to retire until the requester sends a request with a highest_slotid less than or equal to the replier's new enforced highest_slotid.¶
The requester can also be intransigent with respect to sending non-retry requests that have a slot ID that exceeds the replier's highest_slotid. Once the replier has forcibly lowered the enforced highest_slotid, the requester is only allowed to send retries on slots that exceed the replier's highest_slotid. If a request is received with a slot ID that is higher than the new enforced highest_slotid, and the sequence ID is one higher than what is in the slot's reply cache, then the server can both retire the slot and return NFS4ERR_BADSLOT (however, the server MUST NOT do one and not the other). The reason it is safe to retire the slot is because by using the next sequence ID, the requester is indicating it has received the previous reply for the slot.¶
- The requester is better off using the lowest available slot when sending a new request. This way, the replier may be able to retire slot entries faster. However, where the replier is actively adjusting its granted highest_slotid, it will not be able to use only the receipt of the slot ID and highest_slotid in the request. Neither the slot ID nor the highest_slotid used in a request may reflect the replier's current idea of the requester's session limit, because the request may have been sent from the requester before the update was received. Therefore, in the downward adjustment case, the replier may have to retain a number of reply cache entries at least as large as the old value of maximum requests outstanding, until it can infer that the requester has seen a reply containing the new granted highest_slotid. The replier can infer that the requester has seen such a reply when it receives a new request with the same slot ID as the request replied to and the next higher sequence ID.¶
7.6.1.1. Caching of SEQUENCE and CB_SEQUENCE Replies
When a SEQUENCE or CB_SEQUENCE operation is successfully executed, its reply MUST always be cached. Specifically, session ID, sequence ID, and slot ID MUST be cached in the reply cache. The reply from SEQUENCE also includes the highest slot ID, target highest slot ID, and status flags. Instead of caching these values, the server MAY re-compute the values from the current state of the fore channel, session, and/or client ID as appropriate. Similarly, the reply from CB_SEQUENCE includes a highest slot ID and target highest slot ID. The client MAY re-compute the values from the current state of the session as appropriate.¶
Regardless of whether or not a replier is re-computing highest slot ID, target slot ID, and status on replies to retries, the requester cannot assume that the values are being re-computed whenever it receives a reply after a retry is sent, since it has no way of knowing whether the reply it has received was sent by the replier in response to the retry or is a delayed response to the original request. Therefore, it may be the case that highest slot ID, target slot ID, or status bits may reflect the state of affairs when the request was first executed. Although acting based on such delayed information is valid, it may cause the receiver of the reply to do unneeded work. Requesters MAY choose to send additional requests to get the current state of affairs or use the state of affairs reported by subsequent requests, in preference to acting immediately on data that might be out of date.¶
7.6.1.2. Errors from SEQUENCE and CB_SEQUENCE
Any time SEQUENCE or CB_SEQUENCE returns an error, the sequence ID of the slot MUST NOT change. The replier MUST NOT modify the reply cache entry for the slot whenever an error is returned from SEQUENCE or CB_SEQUENCE.¶
7.6.1.3. Optional Reply Caching
On a per-request basis, the requester can choose to direct the replier to cache the reply to all operations after the first operation (SEQUENCE or CB_SEQUENCE) via the sa_cachethis or csa_cachethis fields of the arguments to SEQUENCE or CB_SEQUENCE. The reason it would not direct the replier to cache the entire reply is that the request is composed of all idempotent operations [Chet]. Caching the reply may offer little benefit. If the reply is too large (see Section 7.6.4), it may not be cacheable anyway. Even if the reply to an idempotent request is small enough to cache, unnecessarily caching the reply slows down the server and increases RPC latency.¶
Whether or not the requester requests the reply to be cached has no effect on the slot processing. If the result of SEQUENCE or CB_SEQUENCE is NFS4_OK, then the slot's sequence ID MUST be incremented by one. If a requester does not direct the replier to cache the reply, the replier MUST do one of following:¶
- The replier can cache the entire original reply. Even though sa_cachethis or csa_cachethis is FALSE, the replier is always free to cache. It may choose this approach in order to simplify implementation.¶
-
The replier enters into its reply cache a reply consisting of the original results to the SEQUENCE or CB_SEQUENCE operation, and with the next operation in COMPOUND or CB_COMPOUND having the error NFS4ERR_RETRY_UNCACHED_REP. Thus, if the requester later retries the request, it will get NFS4ERR_RETRY_UNCACHED_REP. If a replier receives a retried Sequence operation where the reply to the COMPOUND or CB_COMPOUND was not cached, then the replier,¶
- MAY return NFS4ERR_RETRY_UNCACHED_REP in reply to a Sequence operation if the Sequence operation is not the first operation (granted, a requester that does so is in violation of the NFSv4.1 protocol).¶
- MUST NOT return NFS4ERR_RETRY_UNCACHED_REP in reply to a Sequence operation if the Sequence operation is the first operation.¶
- If the second operation is an illegal operation, or an operation that was legal in a previous minor version of NFSv4 and MUST NOT be supported in the current minor version (e.g., SETCLIENTID), the replier MUST NOT ever return NFS4ERR_RETRY_UNCACHED_REP. Instead the replier MUST return NFS4ERR_OP_ILLEGAL or NFS4ERR_BADXDR or NFS4ERR_NOTSUPP as appropriate.¶
- If the second operation can result in another error status, the replier MAY return a status other than NFS4ERR_RETRY_UNCACHED_REP, provided the operation is not executed in such a way that the state of the replier is changed. Examples of such error statuses include NFS4ERR_SEQUENCE_POS,NFS4ERR_REQ_TOO_BIG, and NFS4ERR_NOTSUPP returned for an operation that is legal but not REQUIRED in the current minor version and but is not supported by the replier or its file system.¶
The discussion above assumes that the retried request matches the original one. Section 7.6.1.3.1 discusses what the replier might do, and MUST do when it is aware that original and retried requests do not match. Since the replier might only cache a small amount of the information that would be required to determine whether this is a case of a false retry, the replier may send to the client any of the following responses:¶
-
The cached reply to the original request. This done if users of the original request and retry match, and there is no evidence that there is in fact a mismatch between the original request and retry.¶
This can occur if the server caches the entire request and compares it to the retry but also in situations in which only a limited comparison or no comparison is possible. For details see Section 7.6.1.3.1¶
-
A reply that consists only of the Sequence operation with the error NFS4ERR_SEQ_FALSE_RETRY.¶
This is done if the users of the original request and putative retry do not match, or if there is the server has sufficient data to indicate that that supposed retry does not match the original request.¶
- A reply consisting of the response to Sequence with the status NFS4_OK, together with the second operation as it appeared in the retried request with an error of NFS4ERR_RETRY_UNCACHED_REP or other error as described above.¶
- A reply that consists of the response to Sequence with the status NFS4_OK, together with the second operation as it appeared in the original request with an error of NFS4ERR_RETRY_UNCACHED_REP or other error as described above.¶
7.6.1.3.1. False Retry
[Author Aside]: Section substantially revised to explain why false retries can occur, even though EOS is designed to avoid them. This is used as a basis for explaining the potential need for false retry detection while avoiding a level of checking that would be a performance issue.¶
The mechanisms described in Section 7.6 are designed to ensure that if a Sequence operation is sent and matches a request in the reply cache with the same slot ID and sequence ID then, it is a retry of that original request. However, it is possible, although quite unlikely, that servers will encounter requests where this is not the case, in which case the request is considered a "false retry".¶
-
False reties can occur if the client does not implement request sequencing as described in Section 7.6.¶
-
They can also occur as a result of situations in which large number of requests are aborted and considered complete, even though no response has been received by the requester. However, for this situation to result in a false retry there would have to be a sequence of over four billion such requests being processed using the same slot ID with that sequence followed by a long-delayed transmission of an abandoned request.¶
If a requester sent a Sequence operation with a slot ID and sequence ID that are in the reply cache but the replier detects that the retried request is not the same as the original request, including a retry that has different operations or different arguments in the operations from the original and a retry that uses a different principal in the RPC request's credential field that translates to a different user, then this is a false retry.¶
Given the low expected frequency of such false retries, the replier is not obligated to check for their existence although it is prudent to do so with requesters whose implementation of EOD is any way suspect or where the requests are transmitted over a network capable of delivering a request a very long time after it was sent. When the replier does detect a false retry, it is permitted (but not always obligated) to return NFS4ERR_SEQ_FALSE_RETRY in response to the Sequence operation when it detects a false retry.¶
Translations of particularly privileged user values to other users due to the lack of appropriately secure credentials, as configured on the replier, should be applied before determining whether the users are the same or different. If the replier determines the users are different between the original request and a retry, then the replier MUST return NFS4ERR_SEQ_FALSE_RETRY.¶
Regardless of whether such user mismatches do occur, the occurrence of false retries is an indication that the EOS logic is faulty, has not been implemented correctly, or that there is an extraordinary frequency of aborted requests. In light of this fact, there are practical limits to the information that might be saved in order to determine whether a particular request is a false retry. In the case of large requests recording the entire request might not be practical while a recording a compact form in the form of a checksum might unacceptably limit performance.¶
In the case of requests for which the reply is cached, comparing the operations in the cached response to those in the putative retry can serve to detect interactions with clients not properly implementing EOS or aborting requests inappropriately. In other cases, recording the operation count and the identity of the first non-SEQUENCE operation can make a simple check for false retry feasible.¶
If an operation of the retry is an illegal operation, or an operation that was legal in a previous minor version of NFSv4 and MUST NOT be supported in the current minor version (e.g., SETCLIENTID), the replier MAY return NFS4ERR_SEQ_FALSE_RETRY (and MUST do so if the users of the original request and retry differ). Otherwise, the replier MAY return NFS4ERR_OP_ILLEGAL or NFS4ERR_BADXDR or NFS4ERR_NOTSUPP as appropriate. Note that the handling is in contrast for how the replier deals with retries requests with no cached reply. The difference is due to NFS4ERR_SEQ_FALSE_RETRY being a valid error for only Sequence operations, whereas NFS4ERR_RETRY_UNCACHED_REP is a valid error for all operations except illegal operations and operations that MUST NOT be supported in the current minor version of NFSv4.¶
7.6.2. Retry and Replay of Reply
Because NFSv4.1 is used on transports providing reliable delivery, retrying requests within an existing connection is unlikely to be helpful. Requesters will not normally retry a request, unless the connection it used to send the request disconnects. The requester can then reconnect and re-send the request, or it can re-send the request over a different connection that is associated with the same session, to deal with the possibility that the original connection is no longer functioning appropriately.¶
If the requester is a server wanting to re-send a callback operation over the backchannel of a session, the requester of course cannot reconnect because only the client can associate connections with the backchannel. The server can re-send the request over another connection that is bound to the same session's backchannel. If there is no such connection, the server is forced to indicate that the session has no backchannel by setting the SEQ4_STATUS_CB_PATH_DOWN_SESSION flag bit in the response to the next SEQUENCE operation from the client. The client then has no option but to associate a new connection with the session (or destroy the session).¶
Note that it is not, in general, fatal for a requester to retry without a disconnect between the request and retry. However, in order to prevent false retries (see Section 7.6.1.3.1), the requester MUST NOT retry a request once the slot used to send that request has been used to send a new request.¶
Nevertheless, the retry does consume resources, especially with RDMA, where each request, retry or not, consumes a credit. Retries for no reason, especially retries sent shortly after the previous attempt, are a poor use of network bandwidth and defeat the purpose of a transport's inherent congestion control system.¶
A requester will normally wait for a reply to a request before using the slot for another request and MUST do so unless events such as termination of the issuing process makes it impossible to do so. If no such situation were to arise, then the protocol design would ensure no false retry situation could occur (see Section 7.6.1.3.1 for details. When it does not wait for a reply, the requester cannot be sure that using the next sequence ID for the slot chosen, as it normally does, will always be accepted. For example, suppose a requester sends a request with sequence ID 1, and does not wait for the response. The next time it uses the slot, it sends the new request with sequence ID 2. If the replier has not seen the request with sequence ID 1, then the replier is not expecting sequence ID 2, and rejects the requester's new request with NFS4ERR_SEQ_MISORDERED (as the result from SEQUENCE or CB_SEQUENCE).¶
In light of the above, clients that do not wait for a reply before reusing the slot need to be aware of the possibility of receiving NFS4ERRR_SEQ_MISORDERED as a result and infer the probable existence of a request not received by the server. The client will then adjust the current sequence id sent, using successful execution as an indication that seqids on that slot are again correctly aligned.¶
RDMA fabrics do not guarantee that the memory handles (Steering Tags) within each RPC/RDMA "chunk" [RFC8166] are valid on a scope outside that of a single connection. Therefore, handles used by the direct operations become invalid after connection loss. The server must ensure that any RDMA operations that must be replayed from the reply cache use the newly provided handle(s) from the most recent request.¶
A retry might be sent while the original request is still in progress on the replier. In this case, the replier SHOULD deal with the issue by returning NFS4ERR_DELAY as the reply to SEQUENCE or CB_SEQUENCE operation, but implementations MAY return NFS4ERR_MISORDERED. Since errors from SEQUENCE and CB_SEQUENCE are never recorded in the reply cache, this approach allows the results of the execution of the original request to be properly recorded in the reply cache (assuming that the requester specified the reply to be cached).¶
7.6.3. Resolving Server Callback Races
It is possible for server callbacks to arrive at the client before the reply from related fore channel operations. For example, a client may have been granted a delegation to a file it has opened, but the reply to the OPEN (informing the client of the granting of the delegation) may be delayed in the network. If a conflicting operation arrives at the server, it will recall the delegation using the backchannel, which may be on a different transport connection, perhaps even a different network, or even a different session associated with the same client ID.¶
The presence of a session between the client and server alleviates this issue. When a session is in place, each client request is uniquely identified by its { session ID, slot ID, sequence ID } triple. By the rules under which slot entries (reply cache entries) are retired, the server has knowledge whether the client has "seen" each of the server's replies. The server can therefore provide sufficient information to the client to allow it to disambiguate between an erroneous or conflicting callback race condition.¶
For each client operation that might result in some sort of server callback, the server SHOULD keep track of the { session ID, slot ID, sequence ID } triple of the client request until the slot ID retirement rules allow the server to determine that the client has, in fact, seen the server's reply. Until the time the { session ID, slot ID, sequence ID } request triple can be retired, any recalls of the associated object MUST carry an array of these referring identifiers (in the CB_SEQUENCE operation's arguments), for the benefit of the client. After this time, it is not necessary for the server to provide this information in related callbacks, since it is certain that a race condition can no longer occur.¶
The CB_SEQUENCE operation that begins each server callback carries a list of "referring" { session ID, slot ID, sequence ID } triples. If the client finds the request corresponding to the referring session ID, slot ID, and sequence ID to be currently outstanding (i.e., the server's reply has not been seen by the client), it can determine that the callback has raced the reply, and act accordingly. If the client does not find the request corresponding to the referring triple to be outstanding (including the case of a session ID referring to a destroyed session), then there is no race with respect to this triple. The server SHOULD limit the referring triples to requests that refer to just those that apply to the objects referred to in the CB_COMPOUND procedure.¶
The client must not simply wait forever for the expected server reply to arrive before responding to the CB_COMPOUND that won the race, because it is possible that it will be delayed indefinitely. The client should assume the likely case that the reply will arrive within the average round-trip time for COMPOUND requests to the server, and wait that period of time. If that period of time expires, it can respond to the CB_COMPOUND with NFS4ERR_DELAY. There are other scenarios under which callbacks may race replies. Among them are pNFS layout recalls as described in Section 17.5.5.2.¶
7.6.4. COMPOUND and CB_COMPOUND Construction Issues
Very large requests and replies may pose both buffer management issues (especially with RDMA) and reply cache issues. When the session is created (Section 23.36), for each channel (fore and back), the client and server negotiate the maximum-sized request they will send or process (ca_maxrequestsize), the maximum-sized reply they will return or process (ca_maxresponsesize), and the maximum-sized reply they will store in the reply cache (ca_maxresponsesize_cached).¶
If a request exceeds ca_maxrequestsize, the reply will have the status NFS4ERR_REQ_TOO_BIG. A replier MAY return NFS4ERR_REQ_TOO_BIG as the status for the first operation (SEQUENCE or CB_SEQUENCE) in the request (which means that no operations in the request executed and that the state of the slot in the reply cache is unchanged), or it MAY opt to return it on a subsequent operation in the same COMPOUND or CB_COMPOUND request (which means that at least one operation did execute and that the state of the slot in the reply cache does change). The replier SHOULD set NFS4ERR_REQ_TOO_BIG on the operation that exceeds ca_maxrequestsize.¶
If a reply exceeds ca_maxresponsesize, the reply will have the status NFS4ERR_REP_TOO_BIG. A replier MAY return NFS4ERR_REP_TOO_BIG as the status for the first operation (SEQUENCE or CB_SEQUENCE) in the request, or it MAY opt to return it on a subsequent operation (in the same COMPOUND or CB_COMPOUND reply). A replier MAY return NFS4ERR_REP_TOO_BIG in the reply to SEQUENCE or CB_SEQUENCE, even if the response would still exceed ca_maxresponsesize.¶
If sa_cachethis or csa_cachethis is TRUE, then the replier MUST cache a reply except if an error is returned by the SEQUENCE or CB_SEQUENCE operation (see Section 7.6.1.2). If the reply exceeds ca_maxresponsesize_cached (and sa_cachethis or csa_cachethis is TRUE), then the server MUST return NFS4ERR_REP_TOO_BIG_TO_CACHE. Even if NFS4ERR_REP_TOO_BIG_TO_CACHE (or any other error for that matter) is returned on an operation other than the first operation (SEQUENCE or CB_SEQUENCE), then the reply MUST be cached if sa_cachethis or csa_cachethis is TRUE. For example, if a COMPOUND has eleven operations, including SEQUENCE, the fifth operation is a RENAME, and the tenth operation is a READ for one million bytes, the server may return NFS4ERR_REP_TOO_BIG_TO_CACHE on the tenth operation. Since the server executed several operations, especially the non-idempotent RENAME, the client's request to cache the reply needs to be honored in order for the correct operation of exactly once semantics. If the client retries the request, the server will have cached a reply that contains results for ten of the eleven requested operations, with the tenth operation having a status of NFS4ERR_REP_TOO_BIG_TO_CACHE.¶
A client needs to take care that, when sending operations that change the current filehandle (except for PUTFH, PUTPUBFH, PUTROOTFH, and RESTOREFH), it does not exceed the maximum reply buffer before the GETFH operation. Otherwise, the client will have to retry the operation that changed the current filehandle, in order to obtain the desired filehandle. For the OPEN operation (see Section 23.16), retry is not always available as an option. The following guidelines for the handling of filehandle-changing operations are advised:¶
- Within the same COMPOUND procedure, a client SHOULD send GETFH immediately after a current filehandle-changing operation. A client MUST send GETFH after a current filehandle-changing operation that is also non-idempotent (e.g., the OPEN operation), unless the operation is RESTOREFH. RESTOREFH is an exception, because even though it is non-idempotent, the filehandle RESTOREFH produced originated from an operation that is either idempotent (e.g., PUTFH, LOOKUP), or non-idempotent (e.g., OPEN, CREATE). If the origin is non-idempotent, then because the client MUST send GETFH after the origin operation, the client can recover if RESTOREFH returns an error.¶
- A server MAY return NFS4ERR_REP_TOO_BIG or NFS4ERR_REP_TOO_BIG_TO_CACHE (if sa_cachethis is TRUE) on a filehandle-changing operation if the reply would be too large on the next operation.¶
- A server SHOULD return NFS4ERR_REP_TOO_BIG or NFS4ERR_REP_TOO_BIG_TO_CACHE (if sa_cachethis is TRUE) on a filehandle-changing, non-idempotent operation if the reply would be too large on the next operation, especially if the operation is OPEN.¶
- A server MAY return NFS4ERR_UNSAFE_COMPOUND to a non-idempotent current filehandle-changing operation, if it looks at the next operation (in the same COMPOUND procedure) and finds it is not GETFH. The server SHOULD do this if it is unable to determine in advance whether the total response size would exceed ca_maxresponsesize_cached or ca_maxresponsesize.¶
7.7. RDMA Considerations
A complete discussion of the operation of RPC-based protocols over RDMA transports is in [RFC8166]. A discussion of the operation of NFSv4, including NFSv4.1, over RDMA is in [RFC8267]. Where RDMA is considered, this specification assumes the use of such a layering; it addresses only the upper-layer issues relevant to making best use of RPC/RDMA.¶
7.7.1. RDMA Connection Resources
RDMA requires its consumers to register memory and post buffers of a specific size and number for receive operations.¶
Registration of memory can be a relatively high-overhead operation, since it requires pinning of buffers, assignment of attributes (e.g., readable/writable), and initialization of hardware translation. Preregistration is desirable to reduce overhead. These registrations are specific to hardware interfaces and even to RDMA connection endpoints; therefore, negotiation of their limits is desirable to manage resources effectively.¶
Following basic registration, these buffers must be posted by the RPC layer to handle receives. These buffers remain in use by the RPC/NFSv4.1 implementation; the size and number of them must be known to the remote peer in order to avoid RDMA errors that would cause a fatal error on the RDMA connection.¶
NFSv4.1 manages slots as resources on a per-session basis (see Section 7), while RDMA connections manage credits on a per-connection basis. This means that in order for a peer to send data over RDMA to a remote buffer, it has to have both an NFSv4.1 slot and an RDMA credit. If multiple RDMA connections are associated with a session, then if the total number of credits across all RDMA connections associated with the session is X, and the number of slots in the session is Y, then the maximum number of outstanding requests is the lesser of X and Y.¶
7.7.2. Flow Control
Previous versions of NFS do not provide flow control; instead, they rely on the windowing provided by transports like TCP to throttle requests. This does not work with RDMA, which provides no operation flow control and will terminate a connection in error when limits are exceeded. Limits such as maximum number of requests outstanding are therefore negotiated when a session is created (see the ca_maxrequests field in Section 23.36). These limits then provide the maxima within which each connection associated with the session's channel(s) must remain. RDMA connections are managed within these limits as described in Section 3.3 of [RFC8166]; if there are multiple RDMA connections, then the maximum number of requests for a channel will be divided among the RDMA connections. Put a different way, the onus is on the replier to ensure that the total number of RDMA credits across all connections associated with the replier's channel does exceed the channel's maximum number of outstanding requests.¶
The limits may also be modified dynamically at the replier's choosing by manipulating certain parameters present in each NFSv4.1 reply. In addition, the CB_RECALL_SLOT callback operation (see Section 25.8) can be sent by a server to a client to return RDMA credits to the server, thereby lowering the maximum number of requests a client can have outstanding to the server.¶
7.7.3. Padding
Header padding is requested by each peer at session initiation (see the ca_headerpadsize argument to CREATE_SESSION in Section 23.36), and subsequently used by the RPC RDMA layer, as described in [RFC8166]. Zero padding is permitted.¶
Padding leverages the useful property that RDMA preserve alignment of data, even when they are placed into anonymous (untagged) buffers. If requested, client inline writes will insert appropriate pad bytes within the request header to align the data payload on the specified boundary. The client is encouraged to add sufficient padding (up to the negotiated size) so that the "data" field of the WRITE operation is aligned. Most servers can make good use of such padding, which allows them to chain receive buffers in such a way that any data carried by client requests will be placed into appropriate buffers at the server, ready for file system processing. The receiver's RPC layer encounters no overhead from skipping over pad bytes, and the RDMA layer's high performance makes the insertion and transmission of padding on the sender a significant optimization. In this way, the need for servers to perform RDMA Read to satisfy all but the largest client writes is obviated. An added benefit is the reduction of message round trips on the network -- a potentially good trade, where latency is present.¶
The value to choose for padding is subject to a number of criteria. A primary source of variable-length data in the RPC header is the authentication information, the form of which is client-determined, possibly in response to server specification. The contents of COMPOUNDs, sizes of strings such as those passed to RENAME, etc. all go into the determination of a maximal NFSv4.1 request size and therefore minimal buffer size. The client must select its offered value carefully, so as to avoid overburdening the server, and vice versa. The benefit of an appropriate padding value is higher performance.¶
Sender gather: |RPC Request|Pad bytes|Length| -> |User data...| \------+----------------------/ \ \ \ \ Receiver scatter: \-----------+- ... /-----+----------------\ \ \ |RPC Request|Pad|Length| -> |FS buffer|->|FS buffer|->...¶
In the above case, the server may recycle unused buffers to the next posted receive if unused by the actual received request, or may pass the now-complete buffers by reference for normal write processing. For a server that can make use of it, this removes any need for data copies of incoming data, without resorting to complicated end-to-end buffer advertisement and management. This includes most kernel-based and integrated server designs, among many others. The client may perform similar optimizations, if desired.¶
7.7.4. Dual RDMA and Non-RDMA Transports
Some RDMA transports (e.g. [RFC5040]) permit a "streaming" (non-RDMA) phase, where ordinary traffic might flow before "stepping up" to RDMA mode, commencing RDMA traffic. Some RDMA transports start connections always in RDMA mode. NFSv4.1 allows, but does not assume, a streaming phase before RDMA mode. When a connection is associated with a session, the client and server negotiate whether the connection is used in RDMA or non-RDMA mode (see Sections 23.36 and 23.34).¶
7.8. Session Security
7.8.1. Session Callback Security
Via session/connection association, NFSv4.1 improves security over that provided by NFSv4.0 for the backchannel. The connection is client-initiated (see Section 23.34) and subject to the same firewall and routing checks as the fore channel. At the client's option (see Section 23.35), connection association is fully authenticated before being activated (see Section 23.34). Traffic from the server over the backchannel is authenticated exactly as the client specifies (see Section 7.8.2).¶
7.8.2. Backchannel RPC Security
When the NFSv4.1 client establishes the backchannel, it informs the server of the security flavors and principals to use when sending requests. If the security flavor is RPCSEC_GSS, the client expresses the principal in the form of an established RPCSEC_GSS context. The server is free to use any of the flavor/principal combinations the client offers, but it MUST NOT use combinations not offered. This way, the client need not provide a target GSS principal for the backchannel as it did with NFSv4.0, nor does the server have to implement an RPCSEC_GSS initiator as it did with NFSv4.0 [RFC3530].¶
The CREATE_SESSION (Section 23.36) and BACKCHANNEL_CTL (Section 23.33) operations allow the client to specify flavor/principal combinations.¶
Also note that the SP4_SSV state protection mode (see Sections 23.35 and 7.8.3) has the side benefit of providing SSV-derived RPCSEC_GSS contexts (Section 7.9).¶
7.8.3. Protection from Unauthorized State Changes
As described to this point in the specification, the state model of NFSv4.1 is vulnerable to an attacker that sends a SEQUENCE operation with a forged session ID and with a slot ID that it expects the legitimate client to use next. When the legitimate client uses the slot ID with the same sequence number, the server returns the attacker's result from the reply cache, which disrupts the legitimate client and thus denies service to it. Similarly, an attacker could send a CREATE_SESSION with a forged client ID to create a new session associated with the client ID. The attacker could send requests using the new session that change locking state, such as LOCKU operations to release locks the legitimate client has acquired. Setting a security policy on the file that requires RPCSEC_GSS credentials when manipulating the file's state is one potential work around, but has the disadvantage of preventing a legitimate client from releasing state when RPCSEC_GSS is required to do so, but a GSS context cannot be obtained (possibly because the user has logged off the client).¶
NFSv4.1 provides three options to a client for state protection, which are specified when a client creates a client ID via EXCHANGE_ID (Section 23.35).¶
The first (SP4_NONE) is to simply waive state protection, except for that provided by client host authentication.¶
The other two options (SP4_MACH_CRED and SP4_SSV) share several traits:¶
- An RPCSEC_GSS-based credential is used to authenticate client ID and session maintenance operations, including creating and destroying a session, associating a connection with the session, and destroying the client ID.¶
- Because RPCSEC_GSS is used to authenticate client ID and session maintenance, the attacker cannot associate a rogue connection with a legitimate session, or associate a rogue session with a legitimate client ID in order to maliciously alter the client ID's lock state via CLOSE, LOCKU, DELEGRETURN, LAYOUTRETURN, etc.¶
- In cases where the server's security policies on a portion of its namespace require RPCSEC_GSS authentication, a client may have to use an RPCSEC_GSS credential to remove per-file state (e.g., LOCKU, CLOSE, etc.). The server may require that the principal that removes the state match certain criteria (e.g., the principal might have to be the same as the one that acquired the state). However, the client might not have an RPCSEC_GSS context for such a principal, and might not be able to create such a context (perhaps because the user has logged off). When the client establishes SP4_MACH_CRED or SP4_SSV protection, it can specify a list of operations that the server MUST allow using the machine credential (if SP4_MACH_CRED is used) or the SSV credential (if SP4_SSV is used).¶
The SP4_MACH_CRED state protection option uses a machine credential where the principal that creates the client ID MUST also be the principal that performs client ID and session maintenance operations. The security of the machine credential state protection approach depends entirely on safeguarding the per-machine credential. Assuming a proper safeguard using the per-machine credential for operations like CREATE_SESSION, BIND_CONN_TO_SESSION, DESTROY_SESSION, and DESTROY_CLIENTID will prevent an attacker from associating a rogue connection with a session, or associating a rogue session with a client ID.¶
There are at least three scenarios for the SP4_MACH_CRED option:¶
- The system administrator configures a unique, permanent per-machine credential for one of the mandated GSS mechanisms (e.g., if Kerberos V5 is used, a "keytab" containing a principal derived from a client host name could be used).¶
- The client is used by a single user, and so the client ID and its sessions are used by just that user. If the user's credential expires, then session and client ID maintenance cannot occur, but since the client has a single user, only that user is inconvenienced.¶
- The physical client has multiple users, but the client implementation has a unique client ID for each user. This is effectively the same as the second scenario, but a disadvantage is that each user needs to be allocated at least one session each, so the approach suffers from lack of economy.¶
The SP4_SSV protection option uses the SSV (Section 2.5), via RPCSEC_GSS and the SSV GSS mechanism (Section 7.9), to protect state from attack. The SP4_SSV protection option is intended for the situation comprised of a client that has multiple active users and a system administrator who wants to avoid the burden of installing a permanent machine credential on each client. The SSV is established and updated on the server via SET_SSV (see Section 23.47). To prevent eavesdropping, a client SHOULD send SET_SSV via RPCSEC_GSS with the privacy service or use tls encryption on the connection making the request. Several aspects of the SSV make it intractable for an attacker to guess the SSV, and thus associate rogue connections with a session, and rogue sessions with a client ID:¶
- The arguments to and results of SET_SSV include digests of the old and new SSV, respectively.¶
- Because the initial value of the SSV is zero, therefore known, the client that opts for SP4_SSV protection and opts to apply SP4_SSV protection to BIND_CONN_TO_SESSION and CREATE_SESSION MUST send at least one SET_SSV operation before the first BIND_CONN_TO_SESSION operation or before the second CREATE_SESSION operation on a client ID. If it does not, the SSV mechanism will not generate tokens (Section 7.9). A client SHOULD send SET_SSV as soon as a session is created.¶
- A SET_SSV request does not replace the SSV with the argument to SET_SSV. Instead, the current SSV on the server is logically exclusive ORed (XORed) with the argument to SET_SSV. Each time a new principal uses a client ID for the first time, the client SHOULD send a SET_SSV with that principal's RPCSEC_GSS credentials, with RPCSEC_GSS service set to RPC_GSS_SVC_PRIVACY.¶
Here are the types of attacks that can be attempted by an attacker named Eve on a victim named Bob, and how SP4_SSV protection foils each attack:¶
-
Suppose Eve is the first user to log into a legitimate client. Eve's use of an NFSv4.1 file system will cause the legitimate client to create a client ID with SP4_SSV protection, specifying that the BIND_CONN_TO_SESSION operation MUST use the SSV credential. Eve's use of the file system also causes an SSV to be created. The SET_SSV operation that creates the SSV will be protected by the RPCSEC_GSS context created by the legitimate client, which uses Eve's GSS principal and credentials. Eve can eavesdrop on the network while her RPCSEC_GSS context is created and the SET_SSV using her context is sent. Even if the legitimate client sends the SET_SSV with RPC_GSS_SVC_PRIVACY, because Eve knows her own credentials, she can decrypt the SSV. Eve can compute an RPCSEC_GSS credential that BIND_CONN_TO_SESSION will accept, and so associate a new connection with the legitimate session. Eve can change the slot ID and sequence state of a legitimate session, and/or the SSV state, in such a way that when Bob accesses the server via the same legitimate client, the legitimate client will be unable to use the session.¶
The client's only recourse is to create a new client ID for Bob to use, and establish a new SSV for the client ID. The client will be unable to delete the old client ID, and will let the lease on the old client ID expire.¶
Once the legitimate client establishes an SSV over the new session using Bob's RPCSEC_GSS context, Eve can use the new session via the legitimate client, but she cannot disrupt Bob. Moreover, because the client SHOULD have modified the SSV due to Eve using the new session, Bob cannot get revenge on Eve by associating a rogue connection with the session.¶
The question is how did the legitimate client detect that Eve has hijacked the old session? When the client detects that a new principal, Bob, wants to use the session, it SHOULD have sent a SET_SSV, which leads to the following sub-scenarios:¶
-
Let us suppose that from the rogue connection, Eve sent a SET_SSV with the same slot ID and sequence ID that the legitimate client later uses. The server will assume the SET_SSV sent with Bob's credentials is a retry, and return to the legitimate client the reply it sent Eve. However, unless Eve can correctly guess the SSV the legitimate client will use, the digest verification checks in the SET_SSV response will fail. That is an indication to the client that the session has apparently been hijacked.¶
-
Alternatively, Eve sent a SET_SSV with a different slot ID than the legitimate client uses for its SET_SSV. Then the digest verification of the SET_SSV sent with Bob's credentials fails on the server, and the error returned to the client makes it apparent that the session has been hijacked.¶
-
Alternatively, Eve sent an operation other than SET_SSV, but with the same slot ID and sequence that the legitimate client uses for its SET_SSV. The server returns to the legitimate client the response it sent Eve. The client sees that the response is not at all what it expects. The client assumes either session hijacking or a server bug, and either way destroys the old session.¶
-
-
Eve associates a rogue connection with the session as above, and then destroys the session. Again, Bob goes to use the server from the legitimate client, which sends a SET_SSV using Bob's credentials. The client receives an error that indicates that the session does not exist. When the client tries to create a new session, this will fail because the SSV it has does not match that which the server has, and now the client knows the session was hijacked. The legitimate client establishes a new client ID.¶
-
If Eve creates a connection before the legitimate client establishes an SSV, because the initial value of the SSV is zero and therefore known, Eve can send a SET_SSV that will pass the digest verification check. However, because the new connection has not been associated with the session, the SET_SSV is rejected for that reason.¶
In summary, an attacker's disruption of state when SP4_SSV protection is in use is limited to the formative period of a client ID, its first session, and the establishment of the SSV. Once a non-malicious user uses the client ID, the client quickly detects any hijack and rectifies the situation. Once a non-malicious user successfully modifies the SSV, the attacker cannot use NFSv4.1 operations to disrupt the non-malicious user.¶
Note that neither the SP4_MACH_CRED nor SP4_SSV protection approaches prevent hijacking of a transport connection that has previously been associated with a session. If the goal of a counter-threat strategy is to prevent connection hijacking, the use of IPsec or TLS is RECOMMENDED.¶
If a connection hijack occurs, the hijacker could in theory change locking state and negatively impact the service to legitimate clients. However, if the server is configured to require the use of RPCSEC_GSS with integrity or privacy on the affected file objects, and if EXCHGID4_FLAG_BIND_PRINC_STATEID capability (Section 23.35) is in force, this will thwart unauthorized attempts to change locking state.¶
7.9. The Secret State Verifier (SSV) GSS Mechanism
The SSV provides the secret key for a GSS mechanism internal to NFSv4.1 that NFSv4.1 uses for state protection. Contexts for this mechanism are not established via the RPCSEC_GSS protocol. Instead, the contexts are automatically created when EXCHANGE_ID specifies SP4_SSV protection. The only tokens defined are the PerMsgToken (emitted by GSS_GetMIC) and the SealedMessage token (emitted by GSS_Wrap).¶
The mechanism OID for the SSV mechanism is iso.org.dod.internet.private.enterprise.Michael Eisler.nfs.ssv_mech (1.3.6.1.4.1.28882.1.1). While the SSV mechanism does not define any initial context tokens, the OID can be used to let servers indicate that the SSV mechanism is acceptable whenever the client sends a SECINFO or SECINFO_NO_NAME operation (see Section 6.2).¶
The SSV mechanism defines four subkeys derived from the SSV value. Each time SET_SSV is invoked, the subkeys are recalculated by the client and server. The calculation of each of the four subkeys depends on each of the four respective ssv_subkey4 enumerated values. The calculation uses the HMAC [RFC2104] algorithm, using the current SSV as the key, the one-way hash algorithm as negotiated by EXCHANGE_ID, and the input text as represented by the XDR encoded enumeration value for that subkey of data type ssv_subkey4. If the length of the output of the HMAC algorithm exceeds the length of key of the encryption algorithm (which is also negotiated by EXCHANGE_ID), then the subkey MUST be truncated from the HMAC output, i.e., if the subkey is of N bytes long, then the first N bytes of the HMAC output MUST be used for the subkey. The specification of EXCHANGE_ID states that the length of the output of the HMAC algorithm MUST NOT be less than the length of subkey needed for the encryption algorithm (see Section 23.35).¶
/* Input for computing subkeys */ enum ssv_subkey4 { SSV4_SUBKEY_MIC_I2T = 1, SSV4_SUBKEY_MIC_T2I = 2, SSV4_SUBKEY_SEAL_I2T = 3, SSV4_SUBKEY_SEAL_T2I = 4 };¶
The subkey derived from SSV4_SUBKEY_MIC_I2T is used for calculating message integrity codes (MICs) that originate from the NFSv4.1 client, whether as part of a request over the fore channel or a response over the backchannel. The subkey derived from SSV4_SUBKEY_MIC_T2I is used for MICs originating from the NFSv4.1 server. The subkey derived from SSV4_SUBKEY_SEAL_I2T is used for encryption text originating from the NFSv4.1 client, and the subkey derived from SSV4_SUBKEY_SEAL_T2I is used for encryption text originating from the NFSv4.1 server.¶
The PerMsgToken description is based on an XDR definition:¶
/* Input for computing smt_hmac */ struct ssv_mic_plain_tkn4 { uint32_t smpt_ssv_seq; opaque smpt_orig_plain<>; };¶
/* SSV GSS PerMsgToken token */ struct ssv_mic_tkn4 { uint32_t smt_ssv_seq; opaque smt_hmac<>; };¶
The field smt_hmac is an HMAC calculated by using the subkey derived from SSV4_SUBKEY_MIC_I2T or SSV4_SUBKEY_MIC_T2I as the key, the one-way hash algorithm as negotiated by EXCHANGE_ID, and the input text as represented by data of type ssv_mic_plain_tkn4. The field smpt_ssv_seq is the same as smt_ssv_seq. The field smpt_orig_plain is the "message" input passed to GSS_GetMIC() (see Section 2.3.1 of [RFC2743]). The caller of GSS_GetMIC() provides a pointer to a buffer containing the plain text. The SSV mechanism's entry point for GSS_GetMIC() encodes this into an opaque array, and the encoding will include an initial four-byte length, plus any necessary padding. Prepended to this will be the XDR encoded value of smpt_ssv_seq, thus making up an XDR encoding of a value of data type ssv_mic_plain_tkn4, which in turn is the input into the HMAC.¶
The token emitted by GSS_GetMIC() is XDR encoded and of XDR data type ssv_mic_tkn4. The field smt_ssv_seq comes from the SSV sequence number, which is equal to one after SET_SSV (Section 23.47) is called the first time on a client ID. Thereafter, the SSV sequence number is incremented on each SET_SSV. Thus, smt_ssv_seq represents the version of the SSV at the time GSS_GetMIC() was called. As noted in Section 23.35, the client and server can maintain multiple concurrent versions of the SSV. This allows the SSV to be changed without serializing all RPC calls that use the SSV mechanism with SET_SSV operations. Once the HMAC is calculated, it is XDR encoded into smt_hmac, which will include an initial four-byte length, and any necessary padding. Prepended to this will be the XDR encoded value of smt_ssv_seq.¶
The SealedMessage description is based on an XDR definition:¶
/* Input for computing ssct_encr_data and ssct_hmac */ struct ssv_seal_plain_tkn4 { opaque sspt_confounder<>; uint32_t sspt_ssv_seq; opaque sspt_orig_plain<>; opaque sspt_pad<>; };¶
/* SSV GSS SealedMessage token */ struct ssv_seal_cipher_tkn4 { uint32_t ssct_ssv_seq; opaque ssct_iv<>; opaque ssct_encr_data<>; opaque ssct_hmac<>; };¶
The token emitted by GSS_Wrap() is XDR encoded and of XDR data type ssv_seal_cipher_tkn4.¶
The ssct_ssv_seq field has the same meaning as smt_ssv_seq.¶
The ssct_encr_data field is the result of encrypting a value of the XDR encoded data type ssv_seal_plain_tkn4. The encryption key is the subkey derived from SSV4_SUBKEY_SEAL_I2T or SSV4_SUBKEY_SEAL_T2I, and the encryption algorithm is that negotiated by EXCHANGE_ID.¶
The ssct_iv field is the initialization vector (IV) for the encryption algorithm (if applicable) and is sent in clear text. The content and size of the IV MUST comply with the specification of the encryption algorithm. For example, the id-aes256-CBC algorithm MUST use a 16-byte initialization vector (IV), which MUST be unpredictable for each instance of a value of data type ssv_seal_plain_tkn4 that is encrypted with a particular SSV key.¶
The ssct_hmac field is the result of computing an HMAC using the value of the XDR encoded data type ssv_seal_plain_tkn4 as the input text. The key is the subkey derived from SSV4_SUBKEY_MIC_I2T or SSV4_SUBKEY_MIC_T2I, and the one-way hash algorithm is that negotiated by EXCHANGE_ID.¶
The sspt_confounder field is a random value.¶
The sspt_ssv_seq field is the same as ssvt_ssv_seq.¶
The field sspt_orig_plain field is the original plaintext and is the "input_message" input passed to GSS_Wrap() (see Section 2.3.3 of [RFC2743]). As with the handling of the plaintext by the SSV mechanism's GSS_GetMIC() entry point, the entry point for GSS_Wrap() expects a pointer to the plaintext, and will XDR encode an opaque array into sspt_orig_plain representing the plain text, along with the other fields of an instance of data type ssv_seal_plain_tkn4.¶
The sspt_pad field is present to support encryption algorithms that require inputs to be in fixed-sized blocks. The content of sspt_pad is zero filled except for the length. Beware that the XDR encoding of ssv_seal_plain_tkn4 contains three variable-length arrays, and so each array consumes four bytes for an array length, and each array that follows the length is always padded to a multiple of four bytes per the XDR standard.¶
For example, suppose the encryption algorithm uses 16-byte blocks, and the sspt_confounder is three bytes long, and the sspt_orig_plain field is 15 bytes long. The XDR encoding of sspt_confounder uses eight bytes (4 + 3 + 1-byte pad), the XDR encoding of sspt_ssv_seq uses four bytes, the XDR encoding of sspt_orig_plain uses 20 bytes (4 + 15 + 1-byte pad), and the smallest XDR encoding of the sspt_pad field is four bytes. This totals 36 bytes. The next multiple of 16 is 48; thus, the length field of sspt_pad needs to be set to 12 bytes, or a total encoding of 16 bytes. The total number of XDR encoded bytes is thus 8 + 4 + 20 + 16 = 48.¶
GSS_Wrap() emits a token that is an XDR encoding of a value of data type ssv_seal_cipher_tkn4. Note that regardless of whether or not the caller of GSS_Wrap() requests confidentiality, the token always has confidentiality. This is because the SSV mechanism is for RPCSEC_GSS, and RPCSEC_GSS never produces GSS_wrap() tokens without confidentiality.¶
There is one SSV per client ID. There is a single GSS context for a client ID / SSV pair. All SSV mechanism RPCSEC_GSS handles of a client ID / SSV pair share the same GSS context. SSV GSS contexts do not expire except when the SSV is destroyed (causes would include the client ID being destroyed or a server restart). Since one purpose of context expiration is to replace keys that have been in use for "too long", hence vulnerable to compromise by brute force or accident, the client can replace the SSV key by sending periodic SET_SSV operations, which is done by cycling through different users' RPCSEC_GSS credentials. This way, the SSV is replaced without destroying the SSV's GSS contexts.¶
SSV RPCSEC_GSS handles can be expired or deleted by the server at any time, and the EXCHANGE_ID operation can be used to create more SSV RPCSEC_GSS handles. Expiration of SSV RPCSEC_GSS handles does not imply that the SSV or its GSS context has expired.¶
The client MUST establish an SSV via SET_SSV before the SSV GSS context can be used to emit tokens from GSS_Wrap() and GSS_GetMIC(). If SET_SSV has not been successfully called, attempts to emit tokens MUST fail.¶
The SSV mechanism does not support replay detection and sequencing in its tokens because RPCSEC_GSS does not use those features (see "Context Creation Requests", Section 5.2.2 of [RFC2203]). However, Section 7.10 discusses special considerations for the SSV mechanism when used with RPCSEC_GSS.¶
7.10. Security Considerations for RPCSEC_GSS When Using the SSV Mechanism
When a client ID is created with SP4_SSV state protection (see Section 23.35), the client is permitted to associate multiple RPCSEC_GSS handles with the single SSV GSS context (see Section 7.9). Because of the way RPCSEC_GSS (both version 1 and version 2, see [RFC2203] and [RFC5403]) calculate the verifier of the reply, special care must be taken by the implementation of the NFSv4.1 client to prevent attacks by a man-in-the-middle. The verifier of an RPCSEC_GSS reply is the output of GSS_GetMIC() applied to the input value of the seq_num field of the RPCSEC_GSS credential (data type rpc_gss_cred_ver_1_t) (see Section 5.3.3.2 of [RFC2203]). If multiple RPCSEC_GSS handles share the same GSS context, then if one handle is used to send a request with the same seq_num value as another handle, an attacker could block the reply, and replace it with the verifier used for the other handle.¶
There are multiple ways to prevent the attack on the SSV RPCSEC_GSS verifier in the reply. The simplest is believed to be as follows.¶
- Each time one or more new SSV RPCSEC_GSS handles are created via EXCHANGE_ID, the client SHOULD send a SET_SSV operation to modify the SSV. By changing the SSV, the new handles will not result in the re-use of an SSV RPCSEC_GSS verifier in a reply.¶
- When a requester decides to use N SSV RPCSEC_GSS handles, it SHOULD assign a unique and non-overlapping range of seq_nums to each SSV RPCSEC_GSS handle. The size of each range SHOULD be equal to MAXSEQ / N (see Section 5 of [RFC2203] for the definition of MAXSEQ). When an SSV RPCSEC_GSS handle reaches its maximum, it SHOULD force the replier to destroy the handle by sending a NULL RPC request with seq_num set to MAXSEQ + 1 (see Section 5.3.3.3 of [RFC2203]).¶
- When the requester wants to increase or decrease N, it SHOULD force the replier to destroy all N handles by sending a NULL RPC request on each handle with seq_num set to MAXSEQ + 1. If the requester is the client, it SHOULD send a SET_SSV operation before using new handles. If the requester is the server, then the client SHOULD send a SET_SSV operation when it detects that the server has forced it to destroy a backchannel's SSV RPCSEC_GSS handle. By sending a SET_SSV operation, the SSV will change, and so the attacker will be unavailable to successfully replay a previous verifier in a reply to the requester.¶
Note that if the replier carefully creates the SSV RPCSEC_GSS handles, the related risk of a man-in-the-middle splicing a forged SSV RPCSEC_GSS credential with a verifier for another handle does not exist. This is because the verifier in an RPCSEC_GSS request is computed from input that includes both the RPCSEC_GSS handle and seq_num (see Section 5.3.1 of [RFC2203]). Provided the replier takes care to avoid re-using the value of an RPCSEC_GSS handle that it creates, such as by including a generation number in the handle, the man-in-the-middle will not be able to successfully replay a previous verifier in the request to a replier.¶
7.11. Session Mechanics - Steady State
7.11.1. Obligations of the Server
The server has the primary obligation to monitor the state of backchannel resources that the client has created for the server (RPCSEC_GSS contexts and backchannel connections). If these resources vanish, the server takes action as specified in Section 7.13.2.¶
7.11.2. Obligations of the Client
The client SHOULD honor the following obligations in order to utilize the session:¶
- Keep a necessary session from going idle on the server. A client that requires a session but nonetheless is not sending operations risks having the session be destroyed by the server. This is because sessions consume resources, and resource limitations may force the server to cull an inactive session. A server MAY consider a session to be inactive if the client has not used the session before the session inactivity timer (Section 7.12) has expired.¶
- Destroy the session when not needed. If a client has multiple sessions, one of which has no requests waiting for replies, and has been idle for some period of time, it SHOULD destroy the session.¶
- Maintain GSS contexts and RPCSEC_GSS handles for the backchannel. If the client requires the server to use the RPCSEC_GSS security flavor for callbacks, then it needs to be sure the RPCSEC_GSS handles and/or their GSS contexts that are handed to the server via BACKCHANNEL_CTL or CREATE_SESSION are unexpired.¶
- Preserve a connection for a backchannel. The server requires a backchannel in order to gracefully recall recallable state or notify the client of certain events. Note that if the connection is not being used for the fore channel, there is no way for the client to tell if the connection is still alive (e.g., the server restarted without sending a disconnect). The onus is on the server, not the client, to determine if the backchannel's connection is alive, and to indicate in the response to a SEQUENCE operation when the last connection associated with a session's backchannel has disconnected.¶
7.11.3. Steps the Client Takes to Establish a Session
If the client does not have a client ID, the client sends EXCHANGE_ID to establish a client ID. If it opts for SP4_MACH_CRED or SP4_SSV protection, in the spo_must_enforce list of operations, it SHOULD at minimum specify CREATE_SESSION, DESTROY_SESSION, BIND_CONN_TO_SESSION, BACKCHANNEL_CTL, and DESTROY_CLIENTID. If it opts for SP4_SSV protection, the client needs to ask for SSV-based RPCSEC_GSS handles.¶
The client uses the client ID to send a CREATE_SESSION on a connection to the server. The results of CREATE_SESSION indicate whether or not the server undertakes to persist the session reply cache in which a server restarts, and the client notes this for future reference.¶
If the client specified SP4_SSV state protection when the client ID was created, then it SHOULD send SET_SSV in the first COMPOUND after the session is created. Each time a new principal goes to use the client ID, it SHOULD send a SET_SSV again.¶
If the client wants to use delegations, layouts, directory notifications, or any other state that requires a backchannel, then it needs to add a connection to the backchannel if CREATE_SESSION did not already do so. The client creates a connection, and calls BIND_CONN_TO_SESSION to associate the connection with the session and the session's backchannel. If CREATE_SESSION did not already do so, the client MUST tell the server what security is required in order for the client to accept callbacks. The client does this via BACKCHANNEL_CTL. If the client selected SP4_MACH_CRED or SP4_SSV protection when it called EXCHANGE_ID, then the client SHOULD specify that the backchannel use RPCSEC_GSS contexts for security.¶
If the client wants to use additional connections for the backchannel, then it needs to call BIND_CONN_TO_SESSION on each connection it wants to use with the session. If the client wants to use additional connections for the fore channel, then it needs to call BIND_CONN_TO_SESSION if it specified SP4_SSV or SP4_MACH_CRED state protection when the client ID was created.¶
At this point, the session has reached steady state.¶
7.12. Session Inactivity Timer
The server MAY maintain a session inactivity timer for each session. If the session inactivity timer expires, then the server MAY destroy the session. To avoid losing a session due to inactivity, the client MUST renew the session inactivity timer. The length of session inactivity timer MUST NOT be less than the lease_time attribute (Section 11.12.1.11). As with lease renewal (Section 13.3), when the server receives a SEQUENCE operation, it resets the session inactivity timer, and MUST NOT allow the timer to expire while the rest of the operations in the COMPOUND procedure's request are still executing. Once the last operation has finished, the server MUST set the session inactivity timer to expire no sooner than the sum of the current time and the value of the lease_time attribute.¶
7.13. Session Mechanics - Recovery
7.13.1. Events Requiring Client Action
The following events require client action to recover.¶
7.13.1.1. RPCSEC_GSS Context Loss by Callback Path
If all RPCSEC_GSS handles granted by the client to the server for callback use have expired, the client MUST establish a new handle via BACKCHANNEL_CTL. The sr_status_flags field of the SEQUENCE results indicates when callback handles are nearly expired, or fully expired (see Section 23.46.3).¶
7.13.1.2. Connection Loss
If the client loses the last connection of the session and wants to retain the session, then it needs to create a new connection, and if, when the client ID was created, BIND_CONN_TO_SESSION was specified in the spo_must_enforce list, the client MUST use BIND_CONN_TO_SESSION to associate the connection with the session.¶
If there was a request outstanding at the time of connection loss, then if the client wants to continue to use the session, it MUST retry the request, as described in Section 7.6.2. Note that it is not necessary to retry requests over a connection with the same source network address or the same destination network address as the lost connection. As long as the session ID, slot ID, and sequence ID in the retry match that of the original request, the server will recognize the request as a retry if it executed the request prior to disconnect.¶
If the connection that was lost was the last one associated with the backchannel, and the client wants to retain the backchannel and/or prevent revocation of recallable state, the client needs to reconnect, and if it does, it MUST associate the connection to the session and backchannel via BIND_CONN_TO_SESSION. The server SHOULD indicate when it has no callback connection via the sr_status_flags result from SEQUENCE.¶
7.13.1.3. Backchannel GSS Context Loss
Via the sr_status_flags result of the SEQUENCE operation or other means, the client will learn if some or all of the RPCSEC_GSS contexts it assigned to the backchannel have been lost. If the client wants to retain the backchannel and/or not put recallable state subject to revocation, the client needs to use BACKCHANNEL_CTL to assign new contexts.¶
7.13.1.4. Loss of Session
The replier might lose a record of the session. Causes include:¶
- Replier failure and restart.¶
- A catastrophe that causes the reply cache to be corrupted or lost on the media on which it was stored. This applies even if the replier indicated in the CREATE_SESSION results that it would persist the cache.¶
- The server purges the session of a client that has been inactive for a very extended period of time.¶
- As a result of configuration changes among a set of clustered servers, a network address previously connected to one server becomes connected to a different server that has no knowledge of the session in question. Such a configuration change will generally only happen when the original server ceases to function for a time.¶
Loss of reply cache often leads to loss of session. The replier indicates loss of session to the requester by returning NFS4ERR_BADSESSION on the next operation that uses the session ID that refers to the lost session.¶
Although loss of session is often associated with loss of the associated clientid and corresponding locking state, this is not always the case. A session can be lost without loss of the corresponding clientid-based locking state in the event of clientid trunking, or when locking state is stored persistently but the reply cache is not. See Section 8 for details.¶
In the event of server restart, in the absence of clientid trunking, the following situations can arise: can arise:¶
-
If neither the reply cache nor locking state is being stored persistently both the session and clientid are lost and new ones need to be established to continue operation.¶
-
If the reply cache is persistent, it is possible that existing locking state is available so the existing session id and clientid can be tried going forward to determine if operation can be continued with existing locking state or a new clientid needs to be established and locks reclaimed.¶
-
If the reply cache is not persistent, and the locking state is available in persistent storage the session is lost and a new session can be created for the existing clientid.¶
After an event like a server restart, the client may have lost its connections. The client assumes for the moment that the session has not been lost. It reconnects, and if it specified connection association enforcement when the session was created, it invokes BIND_CONN_TO_SESSION using the session ID. Otherwise, it invokes SEQUENCE. If BIND_CONN_TO_SESSION or SEQUENCE returns NFS4ERR_BADSESSION, the client knows the session is not available to it when communicating with that network address. If the connection survives session loss, then the next SEQUENCE operation the client sends over the connection will get back NFS4ERR_BADSESSION. The client again knows the session was lost.¶
Here is one suggested algorithm for the client when it gets NFS4ERR_BADSESSION. It is not obligatory in that, if a client does not want to take advantage of such features as trunking, it may omit parts of it. However, it is a useful example that draws attention to various possible recovery issues:¶
- If the client has other connections to other server network addresses associated with the same session, attempt a COMPOUND with a single operation, SEQUENCE, on each of the other connections.¶
- If the attempts succeed, the session is still alive, and this is a strong indicator that the server's network address has moved. The client might send an EXCHANGE_ID on the connection that returned NFS4ERR_BADSESSION to see if there are opportunities for client ID trunking (i.e., the same client ID and so_major_id value are returned). The client might use DNS to see if the moved network address was replaced with another, so that the performance and availability benefits of session trunking can continue.¶
- If the SEQUENCE requests fail with NFS4ERR_BADSESSION, then the session no longer exists on any of the server network addresses for which the client has connections associated with that session ID. It is possible the session is still alive and available on other network addresses. The client sends an EXCHANGE_ID on all the connections to see if the server owner is still listening on those network addresses. If the same server owner is returned but a new client ID is returned, this is a strong indicator of a server restart. If both the same server owner and same client ID are returned, then this is a strong indication that the server did delete the session, and the client will need to send a CREATE_SESSION if it has no other sessions for that client ID. If a different server owner is returned, the client can use DNS to find other network addresses. If it does not, or if DNS does not find any other addresses for the server, then the client will be unable to provide NFSv4.1 service, and fatal errors should be returned to processes that were using the server. If the client is using a "mount" paradigm, unmounting the server is advised.¶
- If the client knows of no other connections associated with the session ID and server network addresses that are, or have been, associated with the session ID, then the client can use DNS to find other network addresses. If it does not, or if DNS does not find any other addresses for the server, then the client will be unable to provide NFSv4.1 service, and fatal errors should be returned to processes that were using the server. If the client is using a "mount" paradigm, unmounting the server is advised.¶
If there is a reconfiguration event that results in the same network address being assigned to servers where the eir_server_scope value is different, it cannot be guaranteed that a session ID generated by the first will be recognized as invalid by the first. Therefore, in managing server reconfigurations among servers with different server scope values, it is necessary to make sure that all clients have disconnected from the first server before effecting the reconfiguration. Nonetheless, clients should not assume that servers will always adhere to this requirement; clients MUST be prepared to deal with unexpected effects of server reconfigurations. Even where a session ID is inappropriately recognized as valid, it is likely either that the connection will not be recognized as valid or that a sequence value for a slot will not be correct. Therefore, when a client receives results indicating such unexpected errors, the use of EXCHANGE_ID to determine the current server configuration is RECOMMENDED.¶
A variation on the above is that after a server's network address moves, there is no NFSv4.1 server listening, e.g., no listener on port 2049. In this example, one of the following occur: the NFSv4 server returns NFS4ERR_MINOR_VERS_MISMATCH, the NFS server returns a PROG_MISMATCH error, the RPC listener on 2049 returns PROG_UNVAIL, or attempts to reconnect to the network address timeout. These SHOULD be treated as equivalent to SEQUENCE returning NFS4ERR_BADSESSION for these purposes.¶
When the client detects session loss, it needs to call CREATE_SESSION to recover. Any non-idempotent operations that were in progress might have been performed on the server at the time of session loss. The client has no general way to recover from this.¶
Note that loss of session does not imply loss of byte-range lock, open, delegation, or layout state because locks, opens, delegations, and layouts are tied to the client ID and depend on the client ID, not the session. Nor does loss of byte-range lock, open, delegation, or layout state imply loss of session state, because the session depends on the client ID; loss of client ID however does imply loss of session, byte-range lock, open, delegation, and layout state. See Section 13.4.2. A session can survive a server restart, but lock recovery may still be needed.¶
It is possible that CREATE_SESSION will fail with NFS4ERR_STALE_CLIENTID (e.g., the server restarts and does not preserve client ID state). If so, the client needs to call EXCHANGE_ID, followed by CREATE_SESSION.¶
7.13.2. Events Requiring Server Action
The following events require server action to recover.¶
7.13.2.1. Client Crash and Restart
As described in Section 23.35, a restarted client sends EXCHANGE_ID in such a way that it causes the server to delete any sessions it had.¶
7.13.2.2. Client Crash with No Restart
If a client crashes and never comes back, it will never send EXCHANGE_ID with its old client owner. Thus, the server has session state that will never be used again. After an extended period of time, and if the server has resource constraints, it MAY destroy the old session as well as locking state.¶
7.13.2.3. Extended Network Partition
To the server, the extended network partition may be no different from a client crash with no restart (see Section 7.13.2.2). Unless the server can discern that there is a network partition, it is free to treat the situation as if the client has crashed permanently.¶
7.13.2.4. Backchannel Connection Loss
If there were callback requests outstanding at the time of a connection loss, then the server MUST retry the requests, as described in Section 7.6.2. Note that it is not necessary to retry requests over a connection with the same source network address or the same destination network address as the lost connection. As long as the session ID, slot ID, and sequence ID in the retry match that of the original request, the callback target will recognize the request as a retry even if it did see the request prior to disconnect.¶
If the connection lost is the last one associated with the backchannel, then the server MUST indicate that in the sr_status_flags field of every SEQUENCE reply until the backchannel is re-established. There are two situations, each of which uses different status flags: no connectivity for the session's backchannel and no connectivity for any session backchannel of the client. See Section 23.46 for a description of the appropriate flags in sr_status_flags.¶
7.13.2.5. GSS Context Loss
The server SHOULD monitor when the number of RPCSEC_GSS handles assigned to the backchannel reaches one, and when that one handle is near expiry (i.e., between one and two periods of lease time), and indicate so in the sr_status_flags field of all SEQUENCE replies. The server MUST indicate when all of the backchannel's assigned RPCSEC_GSS handles have expired via the sr_status_flags field of all SEQUENCE replies.¶
7.14. Parallel NFS and Sessions
A client and server can potentially be a non-pNFS implementation, a metadata server implementation, a data server implementation, or two or three types of implementations. The EXCHGID4_FLAG_USE_NON_PNFS, EXCHGID4_FLAG_USE_PNFS_MDS, and EXCHGID4_FLAG_USE_PNFS_DS flags (not mutually exclusive) are passed in the EXCHANGE_ID arguments and results to allow the client to indicate how it wants to use sessions created under the client ID, and to allow the server to indicate how it will allow the sessions to be used. See Section 18.1 for pNFS sessions considerations.¶
8. Persistence
[Author Aside]: This is a new top-level section which is based on the Persistence section previously within the discussion of Exactly-once Semantics. Essentially, it deletes the feature described in [RFC8881] which could never be implemented in that form and addresses the need with a new feature having the same goals.¶
While file data and metadata are typically stored persistently and are not affected by server restart, with the exception of certain optimizations for writing data, there are two sorts of data not normally stored persistently, that often are affected by server restart. Since [RFC8881] did not address either of these in a way that could be implemented, the entire area has been respecified for reasons discussed in Section 8.1.¶
For each of these type of data, the protocol provides an OPTIONAL feature whereby the server can provide persistent storage to eliminate functional problems when the data is lost or to simplify the process of reconstructing the data based on the client's knowledge.¶
-
Reply caches may be stored persistently, as described in Section 8.2, allowing the same at-most-once semantics (often called "EOS") provided by the session-based reply cache to be maintained across server restarts.¶
As discussed below, the server may provide a persistent reply cache allowing EOS across server restarts or fully persistent sessions that allow the use of existing sessions to be continued across server restart.¶
-
Per-client locking state may be stored persistently, as described in Section 8.3, allowing clients to continue after server restart without the delay caused by interposing a grace period during which all new lock requests are to be rejected.¶
If per-client locking state is not stored persistently, a grace period is provided to allow clients time to reclaim their locks. When this period is needed, requests to obtain new locks (e.g. when opening a file) are delayed until all clients have had a chance to reclaim their locks.¶
Although the incremental cost of supporting lock persistence is generally low enough that servers providing persistent sessions would provide persistent locking state as well, these two features are independent and the client cannot always assume lock persistence is available when an associated session is persistent and successfully recovered. For a discussion of how the client would be able to determine what state has been stored persistently and continue operation without unnecessary disruption, see Section 8.4¶
8.1. Need for Feature Respecification
The original material has been modified substantially and extended in order address the three items listed below. As a result, the focus of the section has shifted to include all elements relevant to persistence across server failure, rather than dealing only with reply cache issues.¶
-
Eliminate elements of the description that made the feature essentially unimplementable. These include overbroad requirements for atomicity and the assumption that all requests needed to be continued across server restart.¶
-
Appropriately discuss lock persistence and its relation to reply cache persistence and session persistence.¶
-
Provide new material describing the process by which the client finds out about the presence of persistence- related features in the event of server restart.¶
8.2. Persistence of Reply Cache
Since the reply cache is bounded, it is possible for the reply cache to be maintained in persistent storage so that it can be made available across server restarts. When the server undertakes to provide this support when the session is created (see Section 23.36 for details), it is uncertain whether what will provided is either:¶
-
Persistence of the reply cache only.¶
-
Persistent of the session including its membership within the clientid of which it is a part.¶
The replier needs to persist the following information if it agreed to provide persistence for the session (when the session was created;¶
- The session ID.¶
-
The slot table¶
This need to include the sequence ID and cached reply for each slot.¶
-
Information about the connection(s) used by the server with is sufficient to determine whether a client attempting to connect after a server Restart.¶
This sort of information can be used to provide either of the two distinct sorts of session-based persistence. The server provides no specific commitment to provide either of these, although, as described in Section 8.4, the client will be able to determine which form, if any, has actually been provided, and respond appropriately¶
In describing persistence-related semantics it will be helpful to define the following two terms:¶
-
An operation is said "reply-caching relevant" if it is either non-idempotent, modifying, or is the final operation (including the case of request termination because of an error) of a request that is specifically requested to be cached (i.e., has a SEQUENCE operation with sa_cachethis set to true).¶
-
A request is said "reply-caching relevant" if it contains one or more operations which are non-idempotent or modifying or it is specifically requested to be cached (i.e., has a SEQUENCE operation with sa_cachethis set to true).¶
Whichever form of session-based persistence is provided by the server, any requests the client retries after the server restarts will return the results that are cached in the reply cache, However, these two forms differ with regard to the handling of new requests and the possible use of clientid-based persistence facilities:¶
-
If only reply cache persistence is provided, any new requests will fail with NFS4ERR_DEADSESSION being returned as the result of the initial SEQUENCE operation.¶
Because there is no need to use the sequence id to order future request the server does not need to update persistent storage, if two successive requests using the same slot are both not reply-caching relevant, although it does if one or both of the request is reply-cache relevant.¶
-
If session persistence is provided, the existing session can be used after connection re-establishment to support the execution of new requests so that the client will be able to continue just as it would have if no session restart had occurred.¶
A persistent reply cache places certain demands on the server. Although it is not it is not necessary to execute successive operations within a COMPOUND atomically, the transfer of the results of a set of operations and their installation in the persistent cache must be immediate following the execution of any reply-cache relevant operation so that it is impossible for operations to be executed or have other visible effects while not appearing in persistent reply cache.¶
If a client were to retry a sequence of operations that was issued to the server, the only acceptable outcomes are:¶
-
an indication that the request is still being processed.¶
-
a cached reply reflecting the completion of the request,¶
-
a cached reply reflecting the interruption of the request due to server failure.¶
-
an indication that the client ID or session has been lost (indicating a catastrophic loss of the reply cache or a session that has been deleted because the client failed to use the session for an extended period of time).¶
The possibility exists of situations in which a server could fail and restart in the middle of a COMPOUND procedure that contains one or more non-idempotent or idempotent-but-modifying operations. If the server allows COMPOUND procedures to be continued after server failure, it creates significantly greater challenges for the execution of such requests and the atomic placement of results in the reply cache.¶
When a server providing a persistent reply cache does not continue a COMPOUND procedure that was interrupted by a server failure, the error NFS4ERR_DEADSESSION is returned on the last operation which was executed.¶
8.3. Persistence of Locking State
Servers may make locking state available across a server restart in a number of ways including the following:¶
-
Data related to the existence of locks and their corresponding characteristics can be stored in persistent RAM and then used after restart if the address of that storage can be reliably obtained after restart.¶
-
The storage of locking-related state can be integrated with the file system by treating locking state in the same fashion used for other metadata.¶
-
Locking state information may be periodically logged to block-based low-latency persistent storage with logging of individual updates.¶
Although the details will vary with the means of providing persistence that is adopted, it is important that locking state made available across the server restart be consistent with locking state reflected in the results of requests made by clients.¶
The simplest part of this is to ensure that all locking state changes are effectively made available persistently before returning to the requester. In addition, when lock state additions or deletions are reflected in the processing of other operations, the state changes must be available persistently before allowing or denying some operation done by another client. For example, when opens denying write prevent file removal, granting such opens or doing corresponding closes need to be reflected persistently before denying or allowing corresponding file removal. Similar consideration apply to doing IO when mandatory byte-range locks are supported¶
The following items need to be kept in mind:¶
-
There is no commitment by the server to provide this persistence and it may be dropped if for a particular client if unusual situations make it advisable.¶
This decision is made separately for each client so that it is possible there will be server restarts where some, but not all, clients have persistent locking state available.¶
-
While the fact that a reclaim on a reclaimable lock is part of the locking state which is to be persistent, the client's state of awareness of that need not be.¶
There is thus no need for the reclaiming client to inform the server that it has completed specific individual reclaims after receiving the response.¶
8.4. Client Handling of Server Failure When Persistence Can be Used
When server failure occurs, the connection to the client will be disconnected and the client can then find out, as described below, whether server failure has occurred and what steps are necessary to continue use of the client with minimal disruption to those using the client.¶
This process includes the potential use of a persistent reply cache, as described in Section 8.5. The same process is followed depending on whether the server provided only a persistent reply cache or full session persistence.¶
If the server did not promise any session persistence, the client instead immediately does an EXCHANGE_ID followed by a CREATE_SESSION. On the other hand, if there was a possible use of a persistent reply cache, the use of EXCHANGE_ID/CREATE_SESSION is conditional and only happens if a new request has been completed with the error NFS4ERR_STALECLIENTID.¶
In either case, the next step depends on whether the clientid is the same as the one before the disconnection. If it is, then recovery is complete and new requests can be issued. This could happen if there were no server restart but also could if a combination of session-based and clientid-based persistence allowed the server failure to be dealt with essentially transparently.¶
In the case in which the clientid is different, the client need to reclaim its locks, as described in Section 8.6.¶
Even in the case in which lock persistence is available for a client, it is still possible that attempts to obtain new locks will fail with NFS4ERR_GRACE if other clients do not have their locks made available persistently.¶
8.5. Client Use of Session-based Persistence
After the connection to the server is re-established, the server will try to re-establish the connection, as the connection breakage occurred at a lower layer, without server restart. Although it is theoretically possible for an intermediary to hide such a disconnection, it would cause problems if it were to do so and the client had no knowledge of the server failure The discussion here assumes that no such disconnection-hiding implementation is in effect¶
After re-establishing the connection to the server, the client would initially attempt to continue use of the session, since it has no knowledge of whether the disconnection was the result of a server restart. If persistence not was requested when creating the session or the server indicated it was not present, then the client can legitimately conclude that EOS semantics was not available across server restart and needs to operate in that environment.¶
The continued use of the existing session could include both retries of requests issued before the disconnection and issuing new requests. As a result, the discussion below will deal with both type of requests. Given that context, one needs to note the following:¶
-
Whether a given request is a retry or a new one may be judged differently by the client and the server.¶
While it is virtually certain that a new request issued by the client will be perceived as such by the server, the reverse is not the case. Retries issued by the client might be perceived as new requests, if the original requests was lost before it was executed or its existence was noted in persistent storage.¶
-
Although it might be desirable for a client to obtain information about existing requests before issuing new one, the discussion will not assume that clients take steps to prevent new requests from being issued.¶
Since retries, as perceived by the client, may be considered as new requests by the server, the prevention of new requests by the client does not ensure that the server will not see and respond to such requests.¶
After re-establishing the connection, the client will be able to issue requests, including retries of requests already issued before the disconnection occurred. These retries need to be issued since there is no way the results of these requests could be communicated back to the client in the absence of a retry since the connection on which it was received no longer exists.¶
When responses to these requests are received, what is to be done depends primarily about the error, if any, associated with the response:¶
-
In all the cases except the two special error codes noted in the bulleted items below, including receiving no error, the client can conclude that the request was executed to completion as reflected in the response. By design, the client is not aware of whether the execution occurred before or after the serve restart, or whether a server restart, in fact, occurred. However, if persistence was requested when the session was created and the server indicated it was present, the client can assume that the request was executed exactly once with the result reflected in the response.¶
When this is the result that is returned for new requests, it can be because the server has provided full session persistence or because no server restart has occurred. In the former case, it must be true that the server has provided persistent storge of locking state for the d associated clientid since, if it had not, the error NFS4ERR_STALECLIENTID would have been returned.¶
-
In the case that NFS4ERR_DEADSESSION is returned on the SEQUENCE operation, the most likely cause is that the request was, from the server's point of view, a new request and that session persistence was not provided by the server. In this case, the current request should be deferred until the results of all retried requests known to the client have been resolved. Others that are considered new by the server also need to be deferred until are reply cache information is obtained.¶
In the case that NFS4ERR_DEADSESSION is returned on another operation, the request is one that was discontinued as a result of server restart. It is most likely that the request was one that contained more than one non-idempotent or modifying operations, with the server failing after one had been completed but before later operations were started. In this case the client has been informed of a partially complete request and needs to issue a new request to include the operations that were not performed as part of the initial request.¶
-
In the case that the error NFS4ERR_STALECLIENTID is returned, the server has recognized a new request but was unable to continue its execution because the locking information it would use has been destroyed as part of the server restart. This can occur if no persistence was provided for the session, if the persistence was limited to the reply cache or if there was session persistence and client locking state was not maintained persistently.¶
In this case lock recovery will be required but it will need to be delayed until all requests that were issued before the disconnection have been marked completed using the persisted reply cache.¶
Once the existing pending requests are disposed of, the client can proceed to doing new requests, although it might have to do lock recovery first. This can occur after a persistent reply cache is used to provide EOS or after it is found that there is no session persistence provided by the server.¶
8.6. Client Use of Clientid-based Persistence
At this point, lock recovery needs to begin if a new request is processed and completes returning the error NFS4ERR_STALECLIENTID. If no new requests have been issued at this point, the client can issue a request consisting only of a SEQUENCE operation to provide a test. If NFS4ERR_STALECLIENTID is not returned then the client will assume either that there has been no server restart or thar server restart as been accompanied with the recovery of locking state for the current clientid. Otherwise, lock recovery can be done as part of a server-provided grace period. The following three steps need to be taken:¶
When lock recovery is necessary, the client need to inform the new server of the existence of its locks before using stateids it obtained before the server restart. This process is referred to as reclaiming the client's locks, which is accomplished using the method listed below, depending on the type of lock to be reclaimed.¶
-
Opens can generally be reclaimed by doing an OPEN with the claim type CLAIM_PREVIOUS.¶
This includes the case of opens associated with delegation. For details, see Section 15.2.1,¶
There is no specific way to reclaim delegations that have no associated open. In such cases, the client can open the file asking for an associated delegation, and return it immediately¶
-
To reclaim byte-range locks, a LOCK operation with the reclaim parameter set to true is used.¶
The associated open will need to be reclaimed first.¶
-
There is no provision regarding reclaiming of layouts and thus no way to obtain them during a grace period.¶
As a result, in case in which locking state is not made available by the server across a server failure, use of the data server is not immediately available and the client is best off doing IO through the MDS until obtaining needed layouts once the rest of lock reclamation is complete.¶
Once all reclaimable locks have been reclaimed, the client need to do a global RECLAIM_COMPLETE to indicate that process is complete. The is necessary to allow new locks to be obtained. However, such request might still be rejected with NFS4ERR_GRACE if other client have not completed their lock reclamations.¶
9. Protocol Constants and Data Types
The syntax and semantics to describe the data types of the NFSv4.1 protocol are defined in the XDR ([RFC4506]) and RPC ([RFC5531]) documents. The next sections build upon the XDR data types to define constants, types, and structures specific to this protocol. The full list of XDR data types is in [RFC5662].¶
9.1. Basic Constants
const NFS4_FHSIZE = 128; const NFS4_VERIFIER_SIZE = 8; const NFS4_OPAQUE_LIMIT = 1024; const NFS4_SESSIONID_SIZE = 16; const NFS4_INT64_MAX = 0x7fffffffffffffff; const NFS4_UINT64_MAX = 0xffffffffffffffff; const NFS4_INT32_MAX = 0x7fffffff; const NFS4_UINT32_MAX = 0xffffffff; const NFS4_MAXFILELEN = 0xffffffffffffffff; const NFS4_MAXFILEOFF = 0xfffffffffffffffe;¶
Except where noted, all these constants are defined in bytes.¶
- NFS4_FHSIZE is the maximum size of a filehandle.¶
- NFS4_VERIFIER_SIZE is the fixed size of a verifier.¶
- NFS4_OPAQUE_LIMIT is the maximum size of certain opaque information.¶
- NFS4_SESSIONID_SIZE is the fixed size of a session identifier.¶
- NFS4_INT64_MAX is the maximum value of a signed 64-bit integer.¶
- NFS4_UINT64_MAX is the maximum value of an unsigned 64-bit integer.¶
- NFS4_INT32_MAX is the maximum value of a signed 32-bit integer.¶
- NFS4_UINT32_MAX is the maximum value of an unsigned 32-bit integer.¶
- NFS4_MAXFILELEN is the maximum length of a regular file.¶
- NFS4_MAXFILEOFF is the maximum offset into a regular file.¶
9.2. Basic Data Types
These are the base NFSv4.1 data types.¶
Data Type | Definition |
---|---|
int32_t | typedef int int32_t; |
uint32_t | typedef unsigned int uint32_t; |
int64_t | typedef hyper int64_t; |
uint64_t | typedef unsigned hyper uint64_t; |
attrlist4 |
typedef opaque attrlist4<>;¶ Used for file/directory attributes.¶ |
bitmap4 |
typedef uint32_t bitmap4<>;¶ Used in attribute array encoding.¶ |
changeid4 |
typedef uint64_t changeid4;¶ Used in the definition of change_info4.¶ |
clientid4 |
typedef uint64_t clientid4;¶ Shorthand reference to client identification.¶ |
count4 |
typedef uint32_t count4;¶ Various count parameters (READ, WRITE, COMMIT).¶ |
length4 |
typedef uint64_t length4;¶ The length of a byte-range within a file.¶ |
mode4 |
typedef uint32_t mode4;¶ Mode attribute data type.¶ |
nfs_cookie4 |
typedef uint64_t nfs_cookie4;¶ Opaque cookie value for READDIR.¶ |
nfs_fh4 |
typedef opaque nfs_fh4<NFS4_FHSIZE>;¶ Filehandle definition.¶ |
nfs_ftype4 |
enum nfs_ftype4;¶ Various defined file types.¶ |
nfsstat4 |
enum nfsstat4;¶ Return value for operations.¶ |
offset4 |
typedef uint64_t offset4;¶ Various offset designations (READ, WRITE, LOCK, COMMIT).¶ |
qop4 |
typedef uint32_t qop4;¶ Quality of protection designation in SECINFO.¶ |
sec_oid4 |
typedef opaque sec_oid4<>;¶ Security Object Identifier. The sec_oid4 data type is not really opaque. Instead, it contains an ASN.1 OBJECT IDENTIFIER as used by GSS-API in the mech_type argument to GSS_Init_sec_context. See [RFC2743] for details.¶ |
sequenceid4 |
typedef uint32_t sequenceid4;¶ Sequence number used for various session operations (EXCHANGE_ID, CREATE_SESSION, SEQUENCE, CB_SEQUENCE).¶ |
seqid4 |
typedef uint32_t seqid4;¶ Sequence identifier used for locking.¶ |
sessionid4 |
typedef opaque sessionid4[NFS4_SESSIONID_SIZE];¶ Session identifier.¶ |
slotid4 |
typedef uint32_t slotid4;¶ Sequencing artifact for various session operations (SEQUENCE, CB_SEQUENCE).¶ |
utf8string |
typedef opaque utf8string<>;¶ UTF-8 encoding for strings.¶ |
utf8str_cis |
typedef utf8string utf8str_cis;¶ Case-insensitive UTF-8 string.¶ |
utf8str_cs |
typedef utf8string utf8str_cs;¶ Case-sensitive UTF-8 string.¶ |
utf8str_mixed |
typedef utf8string utf8str_mixed;¶ UTF-8 strings with a domain or host prefix and an server or file name suffix. Domains can be internationalized as described in [I-D.ietf-nfsv4-internationalization].¶ |
utf8pref |
typedef opaque utf8pref<>;¶ String for which UTF-8 encoding is preferred, although other encodings can be used,¶ |
component4 |
typedef utf8pref component4;¶ Represents pathname components, which may be either encoded using UTF-8 or nor, with use of UTF-8 needed to support normalization and case-insensitivity.¶ |
linktext4 |
typedef opaque linktext4<>¶ Symbolic link contents ("symbolic link" is defined in an Open Group [symlink] standard).¶ |
pathname4 |
typedef component4 pathname4<>;¶ Represents pathname for fs_locations.¶ |
verifier4 |
typedef opaque verifier4[NFS4_VERIFIER_SIZE];¶ Verifier used for various operations (COMMIT, CREATE, EXCHANGE_ID, OPEN, READDIR, WRITE) NFS4_VERIFIER_SIZE is defined as 8.¶ |
End of Base Data Types¶
9.3. Structured Data Types
9.3.1. nfstime4
struct nfstime4 { int64_t seconds; uint32_t nseconds; };¶
The nfstime4 data type gives the number of seconds and nanoseconds since midnight or zero hour January 1, 1970 Coordinated Universal Time (UTC). Values greater than zero for the seconds field denote dates after the zero hour January 1, 1970. Values less than zero for the seconds field denote dates before the zero hour January 1, 1970. In both cases, the nseconds field is to be added to the seconds field for the final time representation. For example, if the time to be represented is one-half second before zero hour January 1, 1970, the seconds field would have a value of negative one (-1) and the nseconds field would have a value of one-half second (500000000). Values greater than 999,999,999 for nseconds are invalid.¶
This data type is used to pass time and date information. A server converts to and from its local representation of time when processing time values, preserving as much accuracy as possible. If the precision of timestamps stored for a file system object is less than defined, loss of precision can occur. An adjunct time maintenance protocol is RECOMMENDED to reduce skew between client and server times.¶
9.3.3. settime4
union settime4 switch (time_how4 set_it) { case SET_TO_CLIENT_TIME4: nfstime4 time; default: void; };¶
The time_how4 and settime4 data types are used for setting timestamps in file object attributes. If set_it is SET_TO_SERVER_TIME4, then the server uses its local representation of time for the time value.¶
9.3.4. specdata4
struct specdata4 { uint32_t specdata1; /* major device number */ uint32_t specdata2; /* minor device number */ };¶
This data type represents the device numbers for the device file types NF4CHR and NF4BLK.¶
9.3.6. change_policy4
struct change_policy4 { uint64_t cp_major; uint64_t cp_minor; };¶
The change_policy4 data type is used for the change_policy OPTIONAL attribute. It provides change sequencing indication analogous to the change attribute. To enable the server to present a value valid across server re-initialization without requiring persistent storage, two 64-bit quantities are used, allowing one to be a server instance ID and the second to be incremented non-persistently, within a given server instance.¶
9.3.7. fattr4
struct fattr4 { bitmap4 attrmask; attrlist4 attr_vals; };¶
The fattr4 data type is used to represent sets of protocol-defined attributes.¶
The bitmap is a counted array of 32-bit integers used to contain bit values. The position of the integer in the array that contains bit n can be computed from the expression (n / 32), and its bit within that integer is (n mod 32).¶
0 1 +-----------+-----------+-----------+-- | count | 31 .. 0 | 63 .. 32 | +-----------+-----------+-----------+--¶
9.3.8. change_info4
struct change_info4 { bool atomic; changeid4 before; changeid4 after; };¶
This data type is used with the CREATE, LINK, OPEN, REMOVE, and RENAME operations to let the client know the value of the change attribute for the directory in which the target file system object resides.¶
9.3.9. netaddr4
struct netaddr4 { /* see struct rpcb in RFC 1833 */ string na_r_netid<>; /* network id */ string na_r_addr<>; /* universal address */ };¶
The netaddr4 data type is used to identify network transport endpoints. The na_r_netid and na_r_addr fields respectively contain a netid and uaddr. The netid and uaddr concepts are defined in [RFC5665]. The netid and uaddr formats for TCP over IPv4 and TCP over IPv6 are defined in [RFC5665], specifically Tables 2 and 3 and in Sections 5.2.3.3 and 5.2.3.4.¶
9.3.10. state_owner4
struct state_owner4 { clientid4 clientid; opaque owner<NFS4_OPAQUE_LIMIT>; }; typedef state_owner4 open_owner4; typedef state_owner4 lock_owner4;¶
The state_owner4 data type is the base type for the open_owner4 (Section 9.3.10.1) and lock_owner4 (Section 9.3.10.2).¶
9.3.10.1. open_owner4
This data type is used to identify the owner of OPEN state.¶
9.3.10.2. lock_owner4
This structure is used to identify the owner of byte-range locking state.¶
9.3.11. open_to_lock_owner4
struct open_to_lock_owner4 { seqid4 open_seqid; stateid4 open_stateid; seqid4 lock_seqid; lock_owner4 lock_owner; };¶
This data type is used for the first LOCK operation done for an open_owner4. It provides both the open_stateid and lock_owner, such that the transition is made from a valid open_stateid sequence to that of the new lock_stateid sequence. Using this mechanism avoids the confirmation of the lock_owner/lock_seqid pair since it is tied to established state in the form of the open_stateid/open_seqid.¶
9.3.12. stateid4
struct stateid4 { uint32_t seqid; opaque other[12]; };¶
This data type is used for the various state sharing mechanisms between the client and server. The client never modifies a value of data type stateid. The starting value of the "seqid" field is undefined. The server is required to increment the "seqid" field by one at each transition of the stateid. This is important since the client will inspect the seqid in OPEN stateids to determine the order of OPEN processing done by the server.¶
9.3.13. layouttype4
enum layouttype4 { LAYOUT4_NFSV4_1_FILES = 0x1, LAYOUT4_OSD2_OBJECTS = 0x2, LAYOUT4_BLOCK_VOLUME = 0x3 };¶
This data type indicates what type of layout is being used. The file server advertises the layout types it supports through the fs_layout_type file system attribute (Section 11.16.1). A client asks for layouts of a particular type in LAYOUTGET, and processes those layouts using layout-type-specific logic.¶
The layouttype4 data type is 32 bits in length. The range represented by the layout type is split into three parts. Type 0x0 is reserved. Types within the range 0x00000001-0x7FFFFFFF are globally unique and are assigned according to the description in Section 27.5; they are maintained by IANA. Types within the range 0x80000000-0xFFFFFFFF are site specific and for private use only.¶
The LAYOUT4_NFSV4_1_FILES enumeration specifies that the NFSv4.1 file layout type, as defined in Section 18, is to be used. The LAYOUT4_OSD2_OBJECTS enumeration specifies that the object layout, as defined in [RFC5664], is to be used. Similarly, the LAYOUT4_BLOCK_VOLUME enumeration specifies that the block/volume layout, as defined in [RFC5663], is to be used.¶
9.3.14. deviceid4
const NFS4_DEVICEID4_SIZE = 16; typedef opaque deviceid4[NFS4_DEVICEID4_SIZE];¶
Layout information includes device IDs that specify a storage device through a compact handle. Addressing and type information is obtained with the GETDEVICEINFO operation. Device IDs are not guaranteed to be valid across metadata server restarts. A device ID is unique per client ID and layout type. See Section 17.2.10 for more details.¶
9.3.15. device_addr4
struct device_addr4 { layouttype4 da_layout_type; opaque da_addr_body<>; };¶
The device address is used to set up a communication channel with the storage device. Different layout types will require different data types to define how they communicate with storage devices. The opaque da_addr_body field is interpreted based on the specified da_layout_type field.¶
This document defines the device address for the NFSv4.1 file layout (see Section 18.3), which identifies a storage device by network IP address and port number. This is sufficient for the clients to communicate with the NFSv4.1 storage devices, and may be sufficient for other layout types as well. Device types for object-based storage devices and block storage devices (e.g., Small Computer System Interface (SCSI) volume labels) are defined by their respective layout specifications.¶
9.3.16. layout_content4
struct layout_content4 { layouttype4 loc_type; opaque loc_body<>; };¶
The loc_body field is interpreted based on the layout type (loc_type). This document defines the loc_body for the NFSv4.1 file layout type; see Section 18.3 for its definition.¶
9.3.17. layout4
struct layout4 { offset4 lo_offset; length4 lo_length; layoutiomode4 lo_iomode; layout_content4 lo_content; };¶
The layout4 data type defines a layout for a file. The layout type specific data is opaque within lo_content. Since layouts are sub-dividable, the offset and length together with the file's filehandle, the client ID, iomode, and layout type identify the layout.¶
9.3.18. layoutupdate4
struct layoutupdate4 { layouttype4 lou_type; opaque lou_body<>; };¶
The layoutupdate4 data type is used by the client to return updated layout information to the metadata server via the LAYOUTCOMMIT (Section 23.42) operation. This data type provides a channel to pass layout type specific information (in field lou_body) back to the metadata server. For example, for the block/volume layout type, this could include the list of reserved blocks that were written. The contents of the opaque lou_body argument are determined by the layout type. The NFSv4.1 file-based layout does not use this data type; if lou_type is LAYOUT4_NFSV4_1_FILES, the lou_body field MUST have a zero length.¶
9.3.19. layouthint4
struct layouthint4 { layouttype4 loh_type; opaque loh_body<>; };¶
The layouthint4 data type is used by the client to pass in a hint about the type of layout it would like created for a particular file. It is the data type specified by the layout_hint attribute described in Section 11.16.4. The metadata server may ignore the hint or may selectively ignore fields within the hint. This hint should be provided at create time as part of the initial attributes within OPEN. The loh_body field is specific to the type of layout (loh_type). The NFSv4.1 file-based layout uses the nfsv4_1_file_layouthint4 data type as defined in Section 18.3.¶
9.3.20. layoutiomode4
enum layoutiomode4 { LAYOUTIOMODE4_READ = 1, LAYOUTIOMODE4_RW = 2, LAYOUTIOMODE4_ANY = 3 };¶
The iomode specifies whether the client intends to just read or both read and write the data represented by the layout. While the LAYOUTIOMODE4_ANY iomode MUST NOT be used in the arguments to the LAYOUTGET operation, it MAY be used in the arguments to the LAYOUTRETURN and CB_LAYOUTRECALL operations. The LAYOUTIOMODE4_ANY iomode specifies that layouts pertaining to both LAYOUTIOMODE4_READ and LAYOUTIOMODE4_RW iomodes are being returned or recalled, respectively. The metadata server's use of the iomode may depend on the layout type being used. The storage devices MAY validate I/O accesses against the iomode and reject invalid accesses.¶
9.3.21. nfs_impl_id4
struct nfs_impl_id4 { utf8str_cis nii_domain; utf8str_cs nii_name; nfstime4 nii_date; };¶
This data type is used to identify client and server implementation details. The nii_domain field is the DNS domain name with which the implementer is associated. The nii_name field is the product name of the implementation and is completely free form. It is RECOMMENDED that the nii_name be used to distinguish machine architecture, machine platforms, revisions, versions, and patch levels. The nii_date field is the timestamp of when the software instance was published or built.¶
9.3.22. threshold_item4
struct threshold_item4 { layouttype4 thi_layout_type; bitmap4 thi_hintset; opaque thi_hintlist<>; };¶
This data type contains a list of hints specific to a layout type for helping the client determine when it should send I/O directly through the metadata server versus the storage devices. The data type consists of the layout type (thi_layout_type), a bitmap (thi_hintset) describing the set of hints supported by the server (they may differ based on the layout type), and a list of hints (thi_hintlist) whose content is determined by the hintset bitmap. See the mdsthreshold attribute for more details.¶
The thi_hintset field is a bitmap of the following values:¶
name | # | Data Type | Description |
---|---|---|---|
threshold4_read_size | 0 | length4 | If a file's length is less than the value of threshold4_read_size, then it is RECOMMENDED that the client read from the file via the MDS and not a storage device. |
threshold4_write_size | 1 | length4 | If a file's length is less than the value of threshold4_write_size, then it is RECOMMENDED that the client write to the file via the MDS and not a storage device. |
threshold4_read_iosize | 2 | length4 | For read I/O sizes below this threshold, it is RECOMMENDED that the client read data using the MDS. |
threshold4_write_iosize | 3 | length4 | For write I/O sizes below this threshold, it is RECOMMENDED that the client write data using the MDS. |
9.3.23. mdsthreshold4
struct mdsthreshold4 { threshold_item4 mth_hints<>; };¶
This data type holds an array of elements of data type threshold_item4, each of which is valid for a particular layout type. An array is necessary because a server can support multiple layout types for a single file.¶
10. Filehandles
The filehandle in the NFS protocol is a per-server unique identifier for a file system object. The contents of the filehandle are opaque to the client. Therefore, the server is responsible for translating the filehandle to an internal representation of the file system object.¶
10.1. Obtaining the First Filehandle
The operations of the NFS protocol are defined in terms of one or more filehandles. Therefore, the client needs a filehandle to initiate communication with the server. With the NFSv3 protocol ([RFC1813]), there exists an ancillary protocol to obtain this first filehandle. The MOUNT protocol, RPC program number 100005, provides the mechanism of translating a string-based file system pathname to a filehandle, which can then be used by the NFS protocols.¶
The MOUNT protocol has deficiencies in the area of security and use via firewalls. This is one reason that the use of the public filehandle was introduced in [RFC2054] and [RFC2055]. With the use of the public filehandle in combination with the LOOKUP operation in the NFSv3 protocol, it has been demonstrated that the MOUNT protocol is unnecessary for viable interaction between NFS client and server.¶
Therefore, the NFSv4.1 protocol will not use an ancillary protocol for translation from string-based pathnames to a filehandle. Two special filehandles will be used as starting points for the NFS client.¶
10.1.1. Root Filehandle
The first of the special filehandles is the ROOT filehandle. The ROOT filehandle is the "conceptual" root of the file system namespace at the NFS server. The client uses or starts with the ROOT filehandle by employing the PUTROOTFH operation. The PUTROOTFH operation instructs the server to set the "current" filehandle to the ROOT of the server's file tree. Once this PUTROOTFH operation is used, the client can then traverse the entirety of the server's file tree with the LOOKUP operation. A complete discussion of the server namespace is in Section 12.¶
10.1.2. Public Filehandle
The second special filehandle is the PUBLIC filehandle. Unlike the ROOT filehandle, the PUBLIC filehandle may be bound or represent an arbitrary file system object at the server. The server is responsible for this binding. It may be that the PUBLIC filehandle and the ROOT filehandle refer to the same file system object. However, it is up to the administrative software at the server and the policies of the server administrator to define the binding of the PUBLIC filehandle and server file system object. The client may not make any assumptions about this binding. The client uses the PUBLIC filehandle via the PUTPUBFH operation.¶
10.2. Filehandle Types
In the NFSv3 protocol, there was one type of filehandle with a single set of semantics. This type of filehandle is termed "persistent" in NFSv4.1. The semantics of a persistent filehandle remain the same as before. A new type of filehandle introduced in NFSv4.1 is the "volatile" filehandle, which attempts to accommodate certain server environments.¶
The volatile filehandle type was introduced to address server functionality or implementation issues that make correct implementation of a persistent filehandle infeasible. Some server environments do not provide a file-system-level invariant that can be used to construct a persistent filehandle. The underlying server file system may not provide the invariant or the server's file system programming interfaces may not provide access to the needed invariant. Volatile filehandles may ease the implementation of server functionality such as hierarchical storage management or file system reorganization or migration. However, the volatile filehandle increases the implementation burden for the client.¶
Since the client will need to handle persistent and volatile filehandles differently, a file attribute is defined that may be used by the client to determine the filehandle types being returned by the server.¶
10.2.1. General Properties of a Filehandle
The filehandle contains all the information the server needs to distinguish an individual file. To the client, the filehandle is opaque. The client stores filehandles for use in a later request and can compare two filehandles from the same server for equality by doing a byte-by-byte comparison. However, the client MUST NOT otherwise interpret the contents of filehandles. If two filehandles from the same server are equal, they MUST refer to the same file. Servers SHOULD try to maintain a one-to-one correspondence between filehandles and files, but this is not required. Clients MUST use filehandle comparisons only to improve performance, not for correct behavior. All clients need to be prepared for situations in which it cannot be determined whether two filehandles denote the same object and in such cases, avoid making invalid assumptions that might cause incorrect behavior. Further discussion of filehandle and attribute comparison in the context of data caching is presented in Section 15.3.4.¶
As an example, in the case that two different pathnames when traversed at the server terminate at the same file system object, the server SHOULD return the same filehandle for each path. This can occur if a hard link (see [hardlink]) is used to create two file names that refer to the same underlying file object and associated data. For example, if paths /a/b/c and /a/d/c refer to the same file, the server SHOULD return the same filehandle for both pathnames' traversals.¶
10.2.2. Persistent Filehandle
A persistent filehandle is defined as having a fixed value for the lifetime of the file system object to which it refers. Once the server creates the filehandle for a file system object, the server MUST accept the same filehandle for the object for the lifetime of the object. If the server restarts, the NFS server MUST honor the same filehandle value as it did in the server's previous instantiation. Similarly, if the file system is migrated, the new NFS server MUST honor the same filehandle as the old NFS server.¶
The persistent filehandle will be become stale or invalid when the file system object is removed. When the server is presented with a persistent filehandle that refers to a deleted object, it MUST return an error of NFS4ERR_STALE. A filehandle may become stale when the file system containing the object is no longer available. The file system may become unavailable if it exists on removable media and the media is no longer available at the server or the file system in whole has been destroyed or the file system has simply been removed from the server's namespace (i.e., unmounted in a UNIX environment).¶
10.2.3. Volatile Filehandle
A volatile filehandle does not share the same longevity characteristics of a persistent filehandle. The server may determine that a volatile filehandle is no longer valid at many different points in time. If the server can definitively determine that a volatile filehandle refers to an object that has been removed, the server should return NFS4ERR_STALE to the client (as is the case for persistent filehandles). In all other cases where the server determines that a volatile filehandle can no longer be used, it should return an error of NFS4ERR_FHEXPIRED.¶
The REQUIRED attribute "fh_expire_type" is used by the client to determine what type of filehandle the server is providing for a particular file system. This attribute is a bitmask with the following values:¶
- FH4_PERSISTENT
- The value of FH4_PERSISTENT is used to indicate a persistent filehandle, which is valid until the object is removed from the file system. The server will not return NFS4ERR_FHEXPIRED for this filehandle. FH4_PERSISTENT is defined as a value in which none of the bits specified below are set.¶
- FH4_VOLATILE_ANY
- The filehandle may expire at any time, except as specifically excluded (i.e., FH4_NO_EXPIRE_WITH_OPEN).¶
- FH4_NOEXPIRE_WITH_OPEN
- May only be set when FH4_VOLATILE_ANY is set. If this bit is set, then the meaning of FH4_VOLATILE_ANY is qualified to exclude any expiration of the filehandle when it is open.¶
- FH4_VOL_MIGRATION
- The filehandle will expire as a result of a file system transition (migration or replication), in those cases in which the continuity of filehandle use is not specified by handle class information within the fs_locations_info attribute. When this bit is set, clients without access to fs_locations_info information should assume that filehandles will expire on file system transitions.¶
- FH4_VOL_RENAME
- The filehandle will expire during rename. This includes a rename by the requesting client or a rename by any other client. If FH4_VOL_ANY is set, FH4_VOL_RENAME is redundant.¶
Servers that provide volatile filehandles that can expire while open require special care as regards handling of RENAMEs and REMOVEs. This situation can arise if FH4_VOL_MIGRATION or FH4_VOL_RENAME is set, if FH4_VOLATILE_ANY is set and FH4_NOEXPIRE_WITH_OPEN is not set, or if a non-read-only file system has a transition target in a different handle class. In these cases, the server should deny a RENAME or REMOVE that would affect an OPEN file of any of the components leading to the OPEN file. In addition, the server should deny all RENAME or REMOVE requests during the grace period, in order to make sure that reclaims of files where filehandles may have expired do not do a reclaim for the wrong file.¶
Volatile filehandles are especially suitable for implementation of the pseudo file systems used to bridge exports. See Section 12.5 for a discussion of this.¶
10.3. One Method of Constructing a Volatile Filehandle
A volatile filehandle, while opaque to the client, could contain:¶
[volatile bit = 1 | server boot time | slot | generation number]¶
- slot is an index in the server volatile filehandle table¶
- generation number is the generation number for the table entry/slot¶
When the client presents a volatile filehandle, the server makes the following checks, which assume that the check for the volatile bit has passed. If the server boot time is less than the current server boot time, return NFS4ERR_FHEXPIRED. If slot is out of range, return NFS4ERR_BADHANDLE. If the generation number does not match, return NFS4ERR_FHEXPIRED.¶
When the server restarts, the table is gone (it is volatile).¶
If the volatile bit is 0, then it is a persistent filehandle with a different structure following it.¶
10.4. Client Recovery from Filehandle Expiration
If possible, the client SHOULD recover from the receipt of an NFS4ERR_FHEXPIRED error. The client must take on additional responsibility so that it may prepare itself to recover from the expiration of a volatile filehandle. If the server returns persistent filehandles, the client does not need these additional steps.¶
For volatile filehandles, most commonly the client will need to store the component names leading up to and including the file system object in question. With these names, the client should be able to recover by finding a filehandle in the namespace that is still available or by starting at the root of the server's file system namespace.¶
If the expired filehandle refers to an object that has been removed from the file system, obviously the client will not be able to recover from the expired filehandle.¶
It is also possible that the expired filehandle refers to a file that has been renamed. If the file was renamed by another client, again it is possible that the original client will not be able to recover. However, in the case that the client itself is renaming the file and the file is open, it is possible that the client may be able to recover. The client can determine the new pathname based on the processing of the rename request. The client can then regenerate the new filehandle based on the new pathname. The client could also use the COMPOUND procedure to construct a series of operations like:¶
RENAME A B LOOKUP B GETFH¶
Note that the COMPOUND procedure does not provide atomicity. This example only reduces the overhead of recovering from an expired filehandle.¶
11. File Attributes
To meet the requirements of extensibility and increased interoperability with non-UNIX platforms, attributes are being handled in a more flexible manner than NFSv3. The NFSv3 fattr3 structure consists of a fixed list of attributes some of which that might not all be supported by some potential servers and includes some attributes that not all clients have an interest in. The fattr3 structure and similar fixed structures cannot be extended as new needs arise and provide no way to indicate non-support of particular attributes. Within the NFSv4.1 protocol, the client is able to query what attributes the server supports and construct requests that deal only with those supported attributes (or a subset thereof). This raises the issues, discussed in Sections 11.1 through 11.3 and 11.5 through 11.6, of determining how the non-support of particular attributes is to be dealt with.¶
11.1. Categorization of File Attributes
In order to clarify the requirements for server support of particular attributes, and to provide guidance for clients dealing with non-support of particular attributes, all NFSv4.1 attributes are divided into the groups listed below:¶
All of these attributes are accommodated in the NFSv4.1 protocol by a specific, well-defined encoding and are identified by a number. They are interrogated by setting a bit in the bit vector sent in a GETATTR, request. The server response includes a bit vector to indicate which attributes were returned in the response.¶
The following attribute categories are defined:¶
- REQUIRED attributes, as discussed in Section 11.4.¶
- OPTIONAL attributes, as discussed in Section 11.5.¶
- Experimental attributes, as discussed in Section 11.6.¶
New attributes of any of these categories may be added to the NFSv4 protocol as part of a new minor version by publishing a Standards Track RFC that allocates a new attribute number value and defines the encoding for the attribute. In addition, new minor versions can move attributes between categories or make formerly OPTIONAL and Experimental attributes MANDATORY to NOT implement. Similarly, OPTIONAL attributes may be added to an existing extensible version by publishing a Standards Track RFC that allocates a new attribute number value and defines the encoding for the attribute. See [RFC8178] for further details¶
11.2. Changes in the Categorization of File Attributes
The categorization of file attributes appearing in this specification differs from that previously published for a number of reasons:¶
-
The description of the attributes for which support is not REQUIRED no longer uses the RFC2119 keyword "RECOMMENDED" as this is not in accord with the definition of that term in [RFC2119].¶
We now describe such attributes as OPTIONAL, leaving it to server to decide which are worthy of support and to clients to decide whether they wish to use servers on which they are not supported.¶
-
The categorization of requirements/recommendation as to support for authorization-related attributes is now the responsibility of the NFSv4-wide security documents, to be derived from [I-D.dnoveck-nfsv4-security] and [I-D.dnoveck-nfsv4-acls].¶
Currently, given the likely lack of agreement on the semantics of ACLs, it is likely that acl would best be described as an Experimental attribute. See Section 11.3 for further discussion.¶
As one illustration of the new approach to these matters, and its differences from older approaches, let us consider the following statement from Section 5.1 of [RFC8881]. Referring to the REQUIRED attributes, it states:¶
- The client is expected to be able to function with an attribute set limited to these attributes. With just the REQUIRED attributes some client functionality may be impaired or limited in some ways. In the case of servers not supporting the owner, mode, or acl-related attributes, there would be no ability to provide substantial security-related functionality.¶
This expectation was not a reasonable one when first formulated and as the NFSv4 protocols have been developed, there have never been any cases of it being realized. There is no reason to implement a server without the minimal authorization-related attributes derived from NFSv3 and no point in working to develop clients capable of interoperating with it. There is no motivation for the working group to devote any time to defining how such a combination is to operate or for implementers to experiment to try to implement remote file access without any meaningful authorization process.¶
Further, the above also seems to conflict with the following, appearing in Section 5.2 of [RFC8881]:¶
- It is expected that servers will support all attributes they comfortably can and only fail to support attributes that are difficult to support in their operating environments.¶
Together, these imply that there are operating environments in which it difficult to support all of mode, owner, group, and acl attributes. It is hard to believe that any such environments exist or that there would be any point in implementing an NFSv4.1 server using then, if they did exist.¶
11.3. Categorization of Authorization-related Attributes
This section provides an overview of the issues in involved in appropriately categorizing the authorization-related attributes, although the final categorization of these will appear in NFSv4-wide security documents, expected to be based on [I-D.dnoveck-nfsv4-security] and [I-D.dnoveck-nfsv4-acls].¶
Authorization-related attributes that are part of NFSv4.1 can be divided into those connected to he POSIX-based authorization model used in NFSv3 and those related to the use of ACLs to provide a more flexible authorization model. Within the context of NFSv4.1, the following should be noted:¶
-
The attributes mode, owner, and owner_group need to be considered REQUIRED, as they are in [I-D.dnoveck-nfsv4-security].¶
This is despite the fact that previous specifications have considered these attributes as OPTIONAL, although the word "RECOMMENDED" was sometime used. In any case, the new categorization in [I-D.dnoveck-nfsv4-acls] has to be considered dispositive both with regard to NFSv4.1 an other minor versions.¶
-
The attributes acl, sacl, and dacl, although designated as OPTIONAL, have never been documented in a manner allowing effective client-server interoperability, suggesting that they would more appropriately be designated as "Experimental".¶
While it is possible that tightening of the specifications being done in [I-D.dnoveck-nfsv4-security] and [I-D.dnoveck-nfsv4-acls] as part of the rfc5661bis effort might allow this to change, that is not yet assured¶
In any case, efforts to provide a path to interoperability will continue and might affect this categorization in later minor versions, even if NFSv4.1 is not affected. See [I-D.dnoveck-nfsv4-acls] for details,¶
-
The attribute aclsupport is appropriately designated as OPTIONAL, as it in [I-D.dnoveck-nfsv4-security].¶
11.4. REQUIRED Attributes
These MUST be supported by every NFSv4.1 client and server in order to ensure a minimum level of interoperability. The server MUST store and return these attributes when requested. A client may ask for the value of any of these attributes to be returned by setting a bit in the GETATTR request, and the server MUST return their value.¶
The client is expected to be able to function with an attribute set limited to these attributes. With just the REQUIRED attributes some client functionality may be unavailable or functionally limited .¶
11.5. OPTIONAL Attributes
These attributes are understood well enough to warrant support in the NFSv4.1 protocol. However, they might not be supported on all servers or used by all clients. A client may ask for any of these attributes to be returned by setting a bit in the GETATTR request but need to handle the case where the server does not return them. A client MAY ask for the set of attributes the server supports within a given file system and has no reason to request attributes the server does not support. A server is REQUIRED to be deal with requests for unsupported attributes by not returning values for them rather than by considering the request an error.¶
Previous versions of the NFSv4.1 specification [RFC5661] [RFC8881] have described these attributes as "RECOMMENDED" even though that description is not accord with [RFC2119]. The NFSv4.0 specification [RFC7530] still uses "RECOMMENDED" although explicitly disclaiming the assumption that the RFC2119 definition applies in this case. The description of these attribute as OPTIONAL connects them appropriately to provisions for protocol extension and minor versioning in which attributes are to be treated as OPTIONAL.¶
11.6. Experimental Attributes
While the vast majority of attributes are, as described in Section 11.5, "understood well enough to warrant support in the NFSv4.1 protocol", it appears to be the case that, for several attributes, that understanding was never properly recorded in existing NFSv4.1 specification documents. While it might be possibly to rectify that issue before eventual publication of this document, the likely existence of multiple incompatible implementations of such attributes make that unlikely¶
Although the existence of such attributes has never been acknowledged before as part of the categorization of NFSv4 attributes. Nevertheless, such attributes have existed in all NFSv4 minor versions and the necessary clarification, if it occurs, is not likely to be complete for some time.¶
While the intention has always been that attribute not be included in Proposed Standards unless they are described adequately to allow interoperable implementation to be developed. Despite that intention, such attributes have been included in multiple minor versions. Given the need to correct that situation, we need to be clear about the issues that have led to these unfortunate situations, so that we can, over time, address them.¶
11.7. Named Attributes
These attributes are not supported by direct encoding in the NFSv4 protocol but are accessed using string names rather than numbers and each corresponds to an uninterpreted stream of bytes that is stored in its own file system object. The namespace for these attributes may be accessed by using the OPENATTR operation as described below. The OPENATTR operation returns a filehandle for a "named attribute directory", and further perusal and modification of the namespace may be done using operations that work on more typical directories, subject to restrictions discussed below. In particular, READDIR may be used to get a list of such named attributes, and LOOKUP and OPEN may select a particular attribute. Creation of a new named attribute can be accomplished using an OPEN specifying file creation.¶
OPENATTR takes a filehandle for the object and returns the filehandle for the attribute directory. The filehandle for the named attributes designates a directory object accessible by LOOKUP or READDIR and contains files whose names identify the named attributes and whose data bytes are the value of those attributes. For example:¶
LOOKUP | "foo" | ; look up file |
GETATTR | attrbits | |
OPENATTR | ; access foo's named attributes | |
LOOKUP | "x11icon" | ; look up specific attribute |
READ | 0,4096 | ; read stream of bytes |
Named attributes are intended for data needed by applications rather than by NFS client implementations. NFS implementers who wish to define new attributes need to specify them as OPTIONAL attributes using the protocol extension facilities specified in [RFC8178].¶
Once an OPEN is done, named attributes may be examined and changed using READ and WRITE operations referencing the filehandles and stateids returned by OPEN.¶
Named attributes may have their own (non-named) attributes. Each of these objects MUST have all of the REQUIRED attributes and may have additional attributes which are not REQUIRED. However, the sets of supported attributes for named attributes need not be, and typically will not be, as large as that for other objects in that file system. Nevertheless, the value of the supported_attrs attribute should reflect the supported attributes for the file system and will not reflect the restricted attribute sets for these special objects.¶
Named attributes and the named attribute directories can be the target of delegations (in the case of the named attribute directory, these will be directory delegations). However, since granting of delegations is at the server's discretion, a server need not support delegations on named attributes or on named attribute directories.¶
Support for named attributes is OPTIONAL and clients need to be prepared to deal with servers that do not support them. However, clients are entitled to assume that if OPENATTR is supported, there will be support for arbitrarily named attributes, rather than support for a few specific names known to the server. If a server does support named attributes, a client that is also able to handle them should be able to copy a file's data and metadata with complete transparency from one location to another since names allowed for regular directory entries are expected to be valid for named attribute names as well.¶
In NFSv4.1, the structure of named attribute directories is restricted in a number of ways, in order to prevent the development of non-interoperable implementations in which some servers support a fully general hierarchical directory structure for named attributes while others support a limited, non-hierarchal structure for named attributes. In such a mixed environment, clients or applications might come to depend on non-portable extensions. The restrictions are:¶
- CREATE is not allowed in a named attribute directory. Thus, such objects as symbolic links and special files are not allowed to be named attributes. Further, directories may not be created in a named attribute directory, so a hierarchical structure of named attributes for a single object is not allowed.¶
- If OPENATTR is done on a named attribute directory or on a named attribute, the server MUST return NFS4ERR_WRONG_TYPE.¶
- Doing a RENAME of a named attribute to a different named attribute directory or to an ordinary (i.e., non-named-attribute) directory is not allowed.¶
- Creating hard links between named attribute directories or between named attribute directories and ordinary directories is not allowed.¶
Names of attributes will not be controlled by this document or other IETF Standards Track documents, beyond what is necessary to regulate the names of files within directories to handle internationalization and case-insensitivity. See Section 27.2 for further discussion.¶
11.8. Classification of Attributes
Each of the protocol-defined attributes can be classified in one of three categories: per server (i.e., the value of the attribute will be the same for all file objects that share the same server owner; see Section 5.6 for a definition of server owner), per file system (i.e., the value of the attribute will be the same for some or all file objects that share the same fsid attribute (Section 11.12.1.9) and server owner), or per file system object. Note that it is possible that some per file system attributes may vary within the file system, depending on the value of the "homogeneous" (Section 11.12.2.16) attribute. Note that the attributes time_access_set and time_modify_set are not listed in this section because they are write-only attributes corresponding to time_access and time_modify, and are used in a special instance of SETATTR.¶
-
The per-server attribute is:¶
- lease_time¶
-
The per-file system attributes are:¶
- supported_attrs, suppattr_exclcreat, fh_expire_type, link_support, symlink_support, unique_handles, aclsupport, cansettime, case_insensitive, case_preserving, chown_restricted, files_avail, files_free, files_total, fs_locations, homogeneous, maxfilesize, maxname, maxread, maxwrite, no_trunc, space_avail, space_free, space_total, time_delta, change_policy, fs_status, fs_layout_type, fs_locations_info, fs_charset_cap¶
-
The per-file system object attributes are:¶
- type, change, size, named_attr, fsid, rdattr_error, filehandle, acl, archive, fileid, hidden, maxlink, mimetype, mode, numlinks, owner, owner_group, rawdev, space_used, system, time_access, time_backup, time_create, time_metadata, time_modify, mounted_on_fileid, dir_notif_delay, dirent_notif_delay, dacl, sacl, layout_type, layout_hint, layout_blksize, layout_alignment, mdsthreshold, retention_get, retention_set, retentevt_get, retentevt_set, retention_hold, mode_set_masked¶
For quota_avail_hard, quota_avail_soft, and quota_used, see their definitions below for the appropriate classification.¶
11.9. Set-Only and Get-Only Attributes
Some of the protocol-defined attributes are set-only; i.e., they can be set via SETATTR but not retrieved via GETATTR. Similarly, some protocol-defined attributes are get-only; i.e., they can be retrieved via GETATTR but not set via SETATTR. If a client attempts to set a get-only attribute or get a set-only attributes, the server MUST return NFS4ERR_INVAL.¶
11.10. REQUIRED Attributes - List and Definition References
The list of REQUIRED attributes appears in Table 4. The meaning of the columns of the table are:¶
- Name:
- The name of the attribute.¶
- Id:
- The number assigned to the attribute. In the event of conflicts between the assigned number and [RFC5662], the latter is likely authoritative, but should be resolved with Errata to this document and/or [RFC5662]. See [errata] for the Errata process.¶
- Data Type:
- The XDR data type of the attribute.¶
- Acc:
- Access allowed to the attribute. R means read-only (GETATTR may retrieve, SETATTR may not set). W means write-only (SETATTR may set, GETATTR may not retrieve). R W means read/write (GETATTR may retrieve, SETATTR may set).¶
- Defined in:
- The section of this specification that describes the attribute.¶
Name | Id | Data Type | Acc | Defined in: |
---|---|---|---|---|
supported_attrs | 0 | bitmap4 | R | Section 11.12.1.1 |
type | 1 | nfs_ftype4 | R | Section 11.12.1.2 |
fh_expire_type | 2 | uint32_t | R | Section 11.12.1.3 |
change | 3 | uint64_t | R | Section 11.12.1.4 |
size | 4 | uint64_t | R W | Section 11.12.1.5 |
link_support | 5 | bool | R | Section 11.12.1.6 |
symlink_support | 6 | bool | R | Section 11.12.1.7 |
named_attr | 7 | bool | R | Section 11.12.1.8 |
fsid | 8 | fsid4 | R | Section 11.12.1.9 |
unique_handles | 9 | bool | R | Section 11.12.1.10 |
lease_time | 10 | nfs_lease4 | R | Section 11.12.1.11 |
rdattr_error | 11 | enum | R | Section 11.12.1.12 |
filehandle | 19 | nfs_fh4 | R | Section 11.12.1.13 |
mode | 33 | mode4 | R W | Section 11.18 |
owner | 36 | utf8str_mixed | R W | Section 11.18 |
owner_group | 37 | utf8str_mixed | R W | Section 11.18 |
suppattr_exclcreat | 75 | bitmap4 | R | Section 11.12.1.14 |
11.11. OPTIONAL Attributes - List and Definition References
The OPTIONAL attributes are defined in Table 5. The meanings of the column headers are the same as Table 4; see Section 11.10 for the meanings.¶
Name | Id | Data Type | Acc | Defined in: |
---|---|---|---|---|
acl | 12 | nfsace4<> | R W | Section 11.18 |
aclsupport | 13 | uint32_t | R | Section 11.18 |
archive | 14 | bool | R W | Section 11.12.2.1 |
cansettime | 15 | bool | R | Section 11.12.2.2 |
case_insensitive | 16 | bool | R | Section 11.12.2.3 |
case_preserving | 17 | bool | R | Section 11.12.2.4 |
change_policy | 60 | chg_policy4 | R | Section 11.12.2.5 |
chown_restricted | 18 | bool | R | Section 11.12.2.6 |
dacl | 58 | nfsacl41 | R W | Section 11.18 |
dir_notif_delay | 56 | nfstime4 | R | Section 11.15.1 |
dirent_notif_delay | 57 | nfstime4 | R | Section 11.15.2 |
fileid | 20 | uint64_t | R | Section 11.12.2.7 |
files_avail | 21 | uint64_t | R | Section 11.12.2.8 |
files_free | 22 | uint64_t | R | Section 11.12.2.9 |
files_total | 23 | uint64_t | R | Section 11.12.2.10 |
fs_charset_cap | 76 | uint32_t | R | Section 11.12.2.11 |
fs_layout_type | 62 | layouttype4<> | R | Section 11.16.1 |
fs_locations | 24 | fs_locations | R | Section 11.12.2.12 |
fs_locations_info | 67 | fs_locations_info4 | R | Section 11.12.2.13 |
fs_status | 61 | fs4_status | R | Section 11.12.2.14 |
hidden | 25 | bool | R W | Section 11.12.2.15 |
homogeneous | 26 | bool | R | Section 11.12.2.16 |
layout_alignment | 66 | uint32_t | R | Section 11.16.2 |
layout_blksize | 65 | uint32_t | R | Section 11.16.3 |
layout_hint | 63 | layouthint4 | W | Section 11.16.4 |
layout_type | 64 | layouttype4<> | R | Section 11.16.5 |
maxfilesize | 27 | uint64_t | R | Section 11.12.2.17 |
maxlink | 28 | uint32_t | R | Section 11.12.2.18 |
maxname | 29 | uint32_t | R | Section 11.12.2.19 |
maxread | 30 | uint64_t | R | Section 11.12.2.20 |
maxwrite | 31 | uint64_t | R | Section 11.12.2.21 |
mdsthreshold | 68 | mdsthreshold4 | R | Section 11.16.6 |
mimetype | 32 | utf8str_cs | R W | Section 11.12.2.22 |
mode | 33 | mode4 | R W | Section 11.18 |
mode_set_masked | 74 | mode_masked4 | W | Section 11.18 |
mounted_on_fileid | 55 | uint64_t | R | Section 11.12.2.23 |
no_trunc | 34 | bool | R | Section 11.12.2.24 |
numlinks | 35 | uint32_t | R | Section 11.12.2.25 |
quota_avail_hard | 38 | uint64_t | R | Section 11.12.2.26 |
quota_avail_soft | 39 | uint64_t | R | Section 11.12.2.27 |
quota_used | 40 | uint64_t | R | Section 11.12.2.28 |
rawdev | 41 | specdata4 | R | Section 11.12.2.29 |
retentevt_get | 71 | retention_get4 | R | Section 11.17.3 |
retentevt_set | 72 | retention_set4 | W | Section 11.17.4 |
retention_get | 69 | retention_get4 | R | Section 11.17.1 |
retention_hold | 73 | uint64_t | R W | Section 11.17.5 |
retention_set | 70 | retention_set4 | W | Section 11.17.2 |
sacl | 59 | nfsacl41 | R W | Section 11.18 |
space_avail | 42 | uint64_t | R | Section 11.12.2.30 |
space_free | 43 | uint64_t | R | Section 11.12.2.31 |
space_total | 44 | uint64_t | R | Section 11.12.2.32 |
space_used | 45 | uint64_t | R | Section 11.12.2.33 |
system | 46 | bool | R W | Section 11.12.2.34 |
time_access | 47 | nfstime4 | R | Section 11.12.2.35 |
time_access_set | 48 | settime4 | W | Section 11.12.2.36 |
time_backup | 49 | nfstime4 | R W | Section 11.12.2.37 |
time_create | 50 | nfstime4 | R W | Section 11.12.2.38 |
time_delta | 51 | nfstime4 | R | Section 11.12.2.39 |
time_metadata | 52 | nfstime4 | R | Section 11.12.2.40 |
time_modify | 53 | nfstime4 | R | Section 11.12.2.41 |
time_modify_set | 54 | settime4 | W | Section 11.12.2.42 |
11.12. Attribute Definitions
11.12.1. Definitions of REQUIRED Attributes
11.12.1.1. Attribute 0: supported_attrs
The bit vector that would retrieve all protocol-defined attributes that are supported for this object. The scope of this attribute applies to all objects with a matching fsid.¶
11.12.1.2. Attribute 1: type
Designates the type of an object in terms of one of a number of special constants:¶
- NF4REG designates a regular file.¶
- NF4DIR designates a directory.¶
- NF4BLK designates a block device special file.¶
- NF4CHR designates a character device special file.¶
- NF4LNK designates a symbolic link.¶
- NF4SOCK designates a named socket special file.¶
- NF4FIFO designates a fifo special file.¶
- NF4ATTRDIR designates a named attribute directory.¶
- NF4NAMEDATTR designates a named attribute.¶
Within the explanatory text and operation descriptions, the following phrases will be used with the meanings given below:¶
- The phrase "is a directory" means that the object's type attribute is NF4DIR or NF4ATTRDIR.¶
- The phrase "is a special file" means that the object's type attribute is NF4BLK, NF4CHR, NF4SOCK, or NF4FIFO.¶
- The phrases "is an ordinary file" and "is a regular file" mean that the object's type attribute is NF4REG or NF4NAMEDATTR.¶
11.12.1.3. Attribute 2: fh_expire_type
Server uses this to specify filehandle expiration behavior to the client. See Section 10 for additional description.¶
11.12.1.4. Attribute 3: change
A value created by the server that the client can use to determine if file data, directory contents, or attributes of the object have been modified. The server may return the object's time_metadata attribute for this attribute's value, but only if the file system object cannot be updated more frequently than the resolution of time_metadata.¶
11.12.1.5. Attribute 4: size
The size of the object in bytes.¶
11.12.1.6. Attribute 5: link_support
TRUE, if the object's file system supports hard links.¶
11.12.1.7. Attribute 6: symlink_support
TRUE, if the object's file system supports symbolic links.¶
11.12.1.8. Attribute 7: named_attr
TRUE, if this object has named attributes. In other words, object has a non-empty named attribute directory.¶
11.12.1.9. Attribute 8: fsid
Unique file system identifier for the file system holding this object. The fsid attribute has major and minor components, each of which are of data type uint64_t.¶
11.12.1.10. Attribute 9: unique_handles
TRUE, if two distinct filehandles are guaranteed to refer to two different file system objects.¶
11.12.1.11. Attribute 10: lease_time
Duration of the lease at server in seconds.¶
11.12.1.12. Attribute 11: rdattr_error
Error returned from an attempt to retrieve attributes during a READDIR operation.¶
11.12.1.13. Attribute 19: filehandle
The filehandle of this object (primarily for READDIR requests).¶
11.12.1.14. Attribute 75: suppattr_exclcreat
The bit vector that would set all protocol-defined attributes that are supported by the EXCLUSIVE4_1 method of file creation via the OPEN operation. The scope of this attribute applies to all objects with a matching fsid.¶
11.12.2. Definitions of Uncategorized OPTIONAL Attributes
The definitions of most of the OPTIONAL attributes follow. Collections that share a common category are defined in other sections.¶
11.12.2.1. Attribute 14: archive
TRUE, if this file has been archived since the time of last modification (deprecated in favor of time_backup).¶
11.12.2.2. Attribute 15: cansettime
TRUE, if the server is able to change the times for a file system object as specified in a SETATTR operation.¶
11.12.2.3. Attribute 16: case_insensitive
TRUE, if file name comparisons on this file system are case insensitive.¶
11.12.2.4. Attribute 17: case_preserving
TRUE, if file name case on this file system is preserved.¶
11.12.2.5. Attribute 60: change_policy
A value created by the server that the client can use to determine if some server policy related to the current file system has been subject to change. If the value remains the same, then the client can be sure that the values of the attributes related to fs location and the fss_type field of the fs_status attribute have not changed. On the other hand, a change in this value does necessarily imply a change in policy. It is up to the client to interrogate the server to determine if some policy relevant to it has changed. See Section 9.3.6 for details.¶
This attribute MUST change when the value returned by the fs_locations or fs_locations_info attribute changes, when a file system goes from read-only to writable or vice versa, or when the allowable set of security flavors for the file system or any part thereof is changed.¶
11.12.2.6. Attribute 18: chown_restricted
If TRUE, the server will reject any request to change either the owner or the group associated with a file if the caller is not a privileged user (for example, "root" in UNIX operating environments or, in Windows 2000, the "Take Ownership" privilege).¶
11.12.2.7. Attribute 20: fileid
A number uniquely identifying the file within the file system.¶
11.12.2.8. Attribute 21: files_avail
File slots available to this user on the file system containing this object -- this should be the smallest relevant limit.¶
11.12.2.9. Attribute 22: files_free
Free file slots on the file system containing this object -- this should be the smallest relevant limit.¶
11.12.2.10. Attribute 23: files_total
Total file slots on the file system containing this object.¶
11.12.2.11. Attribute 76: fs_charset_cap
Character set capabilities for this file system. See Section 19.1.¶
11.12.2.12. Attribute 24: fs_locations
Locations where this file system may be found. If the server returns NFS4ERR_MOVED as an error, this attribute MUST be supported. See Section 16.16 for more details.¶
11.12.2.13. Attribute 67: fs_locations_info
Full function file system location. See Section 16.17.2 for more details.¶
11.12.2.14. Attribute 61: fs_status
Generic file system type information. See Section 16.18 for more details.¶
11.12.2.16. Attribute 26: homogeneous
TRUE, if this object's file system is homogeneous; i.e., all objects in the file system (all objects on the server with the same fsid) have common values for all per-file-system attributes.¶
11.12.2.17. Attribute 27: maxfilesize
Maximum supported file size for the file system of this object.¶
11.12.2.18. Attribute 28: maxlink
Maximum number of links for this object.¶
11.12.2.19. Attribute 29: maxname
Maximum file name size supported for this object.¶
11.12.2.20. Attribute 30: maxread
Maximum amount of data the READ operation will return for this object.¶
11.12.2.21. Attribute 31: maxwrite
Maximum amount of data the WRITE operation will accept for this object. This attribute SHOULD be supported if the file is writable. Lack of this attribute can lead to the client either wasting bandwidth or not receiving the best performance.¶
11.12.2.22. Attribute 32: mimetype
MIME body type/subtype of this object.¶
11.12.2.23. Attribute 55: mounted_on_fileid
Like fileid, but if the target filehandle is the root of a file system, this attribute represents the fileid of the underlying directory.¶
UNIX-based operating environments connect a file system into the namespace by connecting (mounting) the file system onto the existing file object (the mount point, usually a directory) of an existing file system. When the mount point's parent directory is read via an API like readdir(), the return results are directory entries, each with a component name and a fileid. The fileid of the mount point's directory entry will be different from the fileid that the stat() system call returns. The stat() system call is returning the fileid of the root of the mounted file system, whereas readdir() is returning the fileid that stat() would have returned before any file systems were mounted on the mount point.¶
Unlike NFSv3, NFSv4.1 allows a client's LOOKUP request to cross other file systems. The client detects the file system crossing whenever the filehandle argument of LOOKUP has an fsid attribute different from that of the filehandle returned by LOOKUP. A UNIX-based client will consider this a "mount point crossing". UNIX has a legacy scheme for allowing a process to determine its current working directory. This relies on readdir() of a mount point's parent and stat() of the mount point returning fileids as previously described. The mounted_on_fileid attribute corresponds to the fileid that readdir() would have returned as described previously.¶
While the NFSv4.1 client could simply fabricate a fileid corresponding to what mounted_on_fileid provides (and if the server does not support mounted_on_fileid, the client has no choice), there is a risk that the client will generate a fileid that conflicts with one that is already assigned to another object in the file system. Instead, if the server can provide the mounted_on_fileid, the potential for client operational problems in this area is eliminated.¶
If the server detects that there is no mounted point at the target file object, then the value for mounted_on_fileid that it returns is the same as that of the fileid attribute.¶
The mounted_on_fileid attribute is OPTIONAL, and the server should provide it if possible. For a UNIX-based server, this is straightforward. Usually, mounted_on_fileid will be requested during a READDIR operation, in which case it is trivial (at least for UNIX-based servers) to return mounted_on_fileid since it is equal to the fileid of a directory entry returned by readdir(). If mounted_on_fileid is requested in a GETATTR operation, the server should obey an invariant that has it returning a value that is equal to the file object's entry in the object's parent directory, i.e., what readdir() would have returned. Some operating environments allow a series of two or more file systems to be mounted onto a single mount point. In this case, for the server to obey the aforementioned invariant, it will need to find the base mount point, and not the intermediate mount points.¶
11.12.2.24. Attribute 34: no_trunc
If this attribute is TRUE, then if the client uses a file name longer than name_max, an error will be returned instead of the name being truncated.¶
11.12.2.25. Attribute 35: numlinks
Number of hard links to this object.¶
11.12.2.26. Attribute 38: quota_avail_hard
The value in bytes that represents the amount of additional disk space beyond the current allocation that can be allocated to this file or directory before further allocations will be refused. It is understood that this space may be consumed by allocations to other files or directories.¶
11.12.2.27. Attribute 39: quota_avail_soft
The value in bytes that represents the amount of additional disk space that can be allocated to this file or directory before the user may reasonably be warned. It is understood that this space may be consumed by allocations to other files or directories though there is a rule as to which other files or directories.¶
11.12.2.28. Attribute 40: quota_used
The value in bytes that represents the amount of disk space used by this file or directory and possibly a number of other similar files or directories, where the set of "similar" meets at least the criterion that allocating space to any file or directory in the set will reduce the "quota_avail_hard" of every other file or directory in the set.¶
Note that there may be a number of distinct but overlapping sets of files or directories for which a quota_used value is maintained, e.g., "all files with a given owner", "all files with a given group owner", etc. The server is at liberty to choose any of those sets when providing the content of the quota_used attribute, but should do so in a repeatable way. The rule may be configured per file system or may be "choose the set with the smallest quota".¶
11.12.2.29. Attribute 41: rawdev
Raw device number of file of type NF4BLK or NF4CHR. The device number is split into major and minor numbers. If the file's type attribute is not NF4BLK or NF4CHR, the value returned SHOULD NOT be considered useful.¶
11.12.2.30. Attribute 42: space_avail
Disk space in bytes available to this user on the file system containing this object -- this should be the smallest relevant limit.¶
11.12.2.31. Attribute 43: space_free
Free disk space in bytes on the file system containing this object -- this should be the smallest relevant limit.¶
11.12.2.32. Attribute 44: space_total
Total disk space in bytes on the file system containing this object.¶
11.12.2.33. Attribute 45: space_used
Number of file system bytes allocated to this object.¶
11.12.2.34. Attribute 46: system
This attribute is TRUE if this file is a "system" file with respect to the Windows operating environment.¶
11.12.2.35. Attribute 47: time_access
The time_access attribute represents the time of last access to the object by a READ operation sent to the server. The notion of what is an "access" depends on the server's operating environment and/or the server's file system semantics. For example, for servers obeying Portable Operating System Interface (POSIX) semantics, time_access would be updated only by the READ and READDIR operations and not any of the operations that modify the content of the object [read_atime], [readdir_atime], [write_atime]. Of course, setting the corresponding time_access_set attribute is another way to modify the time_access attribute.¶
Whenever the file object resides on a writable file system, the server should make its best efforts to record time_access into stable storage. However, to mitigate the performance effects of doing so, and most especially whenever the server is satisfying the read of the object's content from its cache, the server MAY cache access time updates and lazily write them to stable storage. It is also acceptable to give administrators of the server the option to disable time_access updates.¶
11.12.2.36. Attribute 48: time_access_set
Sets the time of last access to the object. SETATTR use only.¶
11.12.2.37. Attribute 49: time_backup
The time of last backup of the object.¶
11.12.2.38. Attribute 50: time_create
The time of creation of the object. This attribute does not have any relation to the traditional UNIX file attribute "ctime" or "change time".¶
11.12.2.39. Attribute 51: time_delta
Smallest useful server time granularity.¶
11.12.2.40. Attribute 52: time_metadata
The time of last metadata modification of the object.¶
11.12.2.41. Attribute 53: time_modify
The time of last modification to the object.¶
11.12.2.42. Attribute 54: time_modify_set
Sets the time of last modification to the object. SETATTR use only.¶
11.13. Interpreting owner and owner_group
The attributes "owner" and "owner_group" (and also users and groups within the "acl" attribute) are transferred in the form of a UTF-8 string. This string can be used to identify users and groups in several ways:¶
-
A string of the form "name@domain" can be used to give a user or group name together with a domain in which those are defined.¶
This form provides greater degree of extensibility than was possible in NFSv3 which limited these identifiers to 32-bit unsigned integers whose values are all centrally administered as members within a common domain.¶
-
Numeric ids converted to string form.¶
Using this format maintains the strengths and weaknesses of the NFSv3 approach.¶
The following issues are relevant in selected the form to use.¶
-
The use of the form "name@domain" provides greater flexibility, both with regard to the number of users that can be accommodated and to the management of multiple sets of users in separate domains.¶
Taking advantage of this flexibility often requires extensive work because of limitations of the API's used to reference users and groups.¶
-
The use of the form "name@domain" allows clients and servers to work together even if they have different internal formats for user and groups.¶
In many cases, there is no need for such mapping.¶
Providing this mapping requires extra implementation and raises potential security issues.¶
For detailed discussions regarding which of the forms clients and server are to use for these values, see Section 5.1 of [I-D.dnoveck-nfsv4-security].¶
11.14. Character Case Attributes
With respect to the case_insensitive and case_preserving attributes, each UCS-4 character (which UTF-8 encodes) can be mapped to an equivalent character of different case or compared in a case-insensitive manner. The details vary based on the Unicode version implemented by the server for the current file system. Details of the process and how the client can best deal with uncertainty about the process will be discussed in the NFSv4-wide internationalization document (See [I-D.ietf-nfsv4-internationalization] for the latest version)¶
11.15. Directory Notification Attributes
As described in Section 23.39, the client can request a minimum delay for notifications of changes to attributes, but the server is free to ignore what the client requests. The client can determine in advance what notification delays the server will accept by sending a GETATTR operation for either or both of two directory notification attributes. When the client calls the GET_DIR_DELEGATION operation and asks for attribute change notifications, it should request notification delays that are no less than the values in the server-provided attributes.¶
11.15.1. Attribute 56: dir_notif_delay
The dir_notif_delay attribute is the minimum number of seconds the server will delay before notifying the client of a change to the directory's attributes.¶
11.15.2. Attribute 57: dirent_notif_delay
The dirent_notif_delay attribute is the minimum number of seconds the server will delay before notifying the client of a change to a file object that has an entry in the directory.¶
11.16. pNFS Attribute Definitions
11.16.1. Attribute 62: fs_layout_type
The fs_layout_type attribute (see Section 9.3.13) applies to a file system and indicates what layout types are supported by the file system. When the client encounters a new fsid, the client SHOULD obtain the value for the fs_layout_type attribute associated with the new file system. This attribute is used by the client to determine if the layout types supported by the server match any of the client's supported layout types.¶
11.16.2. Attribute 66: layout_alignment
When a client holds layouts on files of a file system, the layout_alignment attribute indicates the preferred alignment for I/O to files on that file system. Where possible, the client should send READ and WRITE operations with offsets that are whole multiples of the layout_alignment attribute.¶
11.16.3. Attribute 65: layout_blksize
When a client holds layouts on files of a file system, the layout_blksize attribute indicates the preferred block size for I/O to files on that file system. Where possible, the client should send READ operations with a count argument that is a whole multiple of layout_blksize, and WRITE operations with a data argument of size that is a whole multiple of layout_blksize.¶
11.16.4. Attribute 63: layout_hint
The layout_hint attribute (see Section 9.3.19) may be set on newly created files to influence the metadata server's choice for the file's layout. If possible, this attribute is one of those set in the initial attributes within the OPEN operation. The metadata server may choose to ignore this attribute. The layout_hint attribute is a subset of the layout structure returned by LAYOUTGET. For example, instead of specifying particular devices, this would be used to suggest the stripe width of a file. The server implementation determines which fields within the layout will be used.¶
11.16.5. Attribute 64: layout_type
This attribute lists the layout type(s) available for a file. The value returned by the server is for informational purposes only. The client will use the LAYOUTGET operation to obtain the information needed in order to perform I/O, for example, the specific device information for the file and its layout.¶
11.16.6. Attribute 68: mdsthreshold
This attribute is a server-provided hint used to communicate to the client when it is more efficient to send READ and WRITE operations to the metadata server or the data server. The two types of thresholds described are file size thresholds and I/O size thresholds. If a file's size is smaller than the file size threshold, data accesses SHOULD be sent to the metadata server. If an I/O request has a length that is below the I/O size threshold, the I/O SHOULD be sent to the metadata server. Each threshold type is specified separately for read and write.¶
The server MAY provide both types of thresholds for a file. If both file size and I/O size are provided, the client SHOULD reach or exceed both thresholds before sending its read or write requests to the data server. Alternatively, if only one of the specified thresholds is reached or exceeded, the I/O requests are sent to the metadata server.¶
For each threshold type, a value of zero indicates no READ or WRITE should be sent to the metadata server, while a value of all ones indicates that all READs or WRITEs should be sent to the metadata server.¶
The attribute is available on a per-filehandle basis. If the current filehandle refers to a non-pNFS file or directory, the metadata server should return an attribute that is representative of the filehandle's file system. It is suggested that this attribute is queried as part of the OPEN operation. Due to dynamic system changes, the client should not assume that the attribute will remain constant for any specific time period; thus, it should be periodically refreshed.¶
11.17. Retention Attributes
Retention is a concept whereby a file object can be placed in an immutable, undeletable, unrenamable state for a fixed or infinite duration of time. Once in this "retained" state, the file cannot be moved out of the state until the duration of retention has been reached.¶
When retention is enabled, retention MUST extend to the data of the file, and the name of file. The server MAY extend retention to any other property of the file, including any subset of REQUIRED, OPTIONAL, and named attributes, with the exceptions noted in this section.¶
Servers MAY support or not support retention on any file object type.¶
The five retention attributes are explained in the next subsections.¶
11.17.1. Attribute 69: retention_get
If retention is enabled for the associated file, this attribute's value represents the retention begin time of the file object. This attribute's value is only readable with the GETATTR operation and MUST NOT be modified by the SETATTR operation (Section 11.9). The value of the attribute consists of:¶
const RET4_DURATION_INFINITE = 0xffffffffffffffff; struct retention_get4 { uint64_t rg_duration; nfstime4 rg_begin_time<1>; };¶
The field rg_duration is the duration in seconds indicating how long the file will be retained once retention is enabled. The field rg_begin_time is an array of up to one absolute time value. If the array is zero length, no beginning retention time has been established, and retention is not enabled. If rg_duration is equal to RET4_DURATION_INFINITE, the file, once retention is enabled, will be retained for an infinite duration.¶
If (as soon as) rg_duration is zero, then rg_begin_time will be of zero length, and again, retention is not (no longer) enabled.¶
11.17.2. Attribute 70: retention_set
This attribute is used to set the retention duration and optionally enable retention for the associated file object. This attribute is only modifiable via the SETATTR operation and MUST NOT be retrieved by the GETATTR operation (Section 11.9). This attribute corresponds to retention_get. The value of the attribute consists of:¶
struct retention_set4 { bool rs_enable; uint64_t rs_duration<1>; };¶
If the client sets rs_enable to TRUE, then it is enabling retention on the file object with the begin time of retention starting from the server's current time and date. The duration of the retention can also be provided if the rs_duration array is of length one. The duration is the time in seconds from the begin time of retention, and if set to RET4_DURATION_INFINITE, the file is to be retained forever. If retention is enabled, with no duration specified in either this SETATTR or a previous SETATTR, the duration defaults to zero seconds. The server MAY restrict the enabling of retention or the duration of retention on the basis of the ACE4_WRITE_RETENTION ACL permission. The enabling of retention MUST NOT prevent the enabling of event-based retention or the modification of the retention_hold attribute.¶
The following rules apply to both the retention_set and retentevt_set attributes.¶
- As long as retention is not enabled, the client is permitted to decrease the duration.¶
- The duration can always be set to an equal or higher value, even if retention is enabled. Note that once retention is enabled, the actual duration (as returned by the retention_get or retentevt_get attributes; see Section 11.17.1 or Section 11.17.3) is constantly counting down to zero (one unit per second), unless the duration was set to RET4_DURATION_INFINITE. Thus, it will not be possible for the client to precisely extend the duration on a file that has retention enabled.¶
- While retention is enabled, attempts to disable retention or decrease the retention's duration MUST fail with the error NFS4ERR_INVAL.¶
- If the principal attempting to change retention_set or retentevt_set does not have ACE4_WRITE_RETENTION permissions, the attempt MUST fail with NFS4ERR_ACCESS.¶
11.17.3. Attribute 71: retentevt_get
Gets the event-based retention duration, and if enabled, the event-based retention begin time of the file object. This attribute is like retention_get, but refers to event-based retention. The event that triggers event-based retention is not defined by the NFSv4.1 specification.¶
11.17.4. Attribute 72: retentevt_set
Sets the event-based retention duration, and optionally enables event-based retention on the file object. This attribute corresponds to retentevt_get and is like retention_set, but refers to event-based retention. When event-based retention is set, the file MUST be retained even if non-event-based retention has been set, and the duration of non-event-based retention has been reached. Conversely, when non-event-based retention has been set, the file MUST be retained even if event-based retention has been set, and the duration of event-based retention has been reached. The server MAY restrict the enabling of event-based retention or the duration of event-based retention on the basis of the ACE4_WRITE_RETENTION ACL permission. The enabling of event-based retention MUST NOT prevent the enabling of non-event-based retention or the modification of the retention_hold attribute.¶
11.17.5. Attribute 73: retention_hold
Gets or sets administrative retention holds, one hold per bit position.¶
This attribute allows one to 64 administrative holds, one hold per bit on the attribute. If retention_hold is not zero, then the file MUST NOT be deleted, renamed, or modified, even if the duration on enabled event or non-event-based retention has been reached. The server MAY restrict the modification of retention_hold on the basis of the ACE4_WRITE_RETENTION_HOLD ACL permission. The enabling of administration retention holds does not prevent the enabling of event-based or non-event-based retention.¶
If the principal attempting to change retention_hold does not have ACE4_WRITE_RETENTION_HOLD permissions, the attempt MUST fail with NFS4ERR_ACCESS.¶
11.18. Access Control Attributes
The use of the access control attributes are fully described in various sections of the NFSv4-wide security documents [I-D.dnoveck-nfsv4-security] [I-D.dnoveck-nfsv4-acls].¶
-
The mode, mode_set_masked, owner, and owner_group attributes are described in Sections 5.3.1 though 5.3.4 of [I-D.dnoveck-nfsv4-security].¶
-
The acl, aclsupport, sacl, and dacl attributes are described in Sections 3.4, 3.5, 3.6, and 3.8 of [I-D.dnoveck-nfsv4-acls].¶
12. Single-Server Namespace
This section describes the NFSv4 single-server namespace. Single-server namespaces may be presented directly to clients, or they may be used as a basis to form larger multi-server namespaces (e.g., site-wide or organization-wide) to be presented to clients, as described in Section 16.¶
12.1. Server Exports
On a UNIX server, the namespace describes all the files reachable by pathnames under the root directory or "/". On a Windows server, the namespace constitutes all the files on disks named by mapped disk letters. NFS server administrators rarely make the entire server's file system namespace available to NFS clients. More often, portions of the namespace are made available via an "export" feature. In previous versions of the NFS protocol, the root filehandle for each export is obtained through the MOUNT protocol; the client sent a string that identified the export name within the namespace and the server returned the root filehandle for that export. The MOUNT protocol also provided an EXPORTS procedure that enumerated the server's exports.¶
12.2. Browsing Exports
The NFSv4.1 protocol provides a root filehandle that clients can use to obtain filehandles for the exports of a particular server, via a series of LOOKUP operations within a COMPOUND, to traverse a path. A common user experience is to use a graphical user interface (perhaps a file "Open" dialog window) to find a file via progressive browsing through a directory tree. The client must be able to move from one export to another export via single-component, progressive LOOKUP operations.¶
This style of browsing is not well supported by the NFSv3 protocol. In NFSv3, the client expects all LOOKUP operations to remain within a single server file system. For example, the device attribute will not change. This prevents a client from taking namespace paths that span exports.¶
In the case of NFSv3, an automounter on the client can obtain a snapshot of the server's namespace using the EXPORTS procedure of the MOUNT protocol. If it understands the server's pathname syntax, it can create an image of the server's namespace on the client. The parts of the namespace that are not exported by the server are filled in with directories that might be constructed similarly to an NFSv4.1 "pseudo file system" (see Section 12.3) that allows the user to browse from one mounted file system to another. There is a drawback to this representation of the server's namespace on the client: it is static. If the server administrator adds a new export, the client will be unaware of it.¶
12.3. Server Pseudo File System
NFSv4.1 servers avoid this namespace inconsistency by presenting all the exports for a given server within the framework of a single namespace for that server. An NFSv4.1 client uses LOOKUP and READDIR operations to browse seamlessly from one export to another.¶
Where there are portions of the server namespace that are not exported, clients require some way of traversing those portions to reach actual exported file systems. A technique that servers may use to provide for this is to bridge the unexported portion of the namespace via a "pseudo file system" that provides a view of exported directories only. A pseudo file system has a unique fsid and behaves like a normal, read-only file system.¶
Based on the construction of the server's namespace, it is possible that multiple pseudo file systems may exist. For example,¶
/a pseudo file system /a/b real file system /a/b/c pseudo file system /a/b/c/d real file system¶
Each of the pseudo file systems is considered a separate entity and therefore MUST have its own fsid, unique among all the fsids for that server.¶
12.4. Multiple Roots
Certain operating environments are sometimes described as having "multiple roots". In such environments, individual file systems are commonly represented by disk or volume names. NFSv4 servers for these platforms can construct a pseudo file system above these root names so that disk letters or volume names are simply directory names in the pseudo root.¶
12.5. Filehandle Volatility
The nature of the server's pseudo file system is that it is a logical representation of file system(s) available from the server. Therefore, the pseudo file system is most likely constructed dynamically when the server is first instantiated. It is expected that the pseudo file system may not have an on-disk counterpart from which persistent filehandles could be constructed. Even though it is preferable that the server provide persistent filehandles for the pseudo file system, the NFS client should expect that pseudo file system filehandles are volatile. This can be confirmed by checking the associated "fh_expire_type" attribute for those filehandles in question. If the filehandles are volatile, the NFS client must be prepared to recover a filehandle value (e.g., with a series of LOOKUP operations) when receiving an error of NFS4ERR_FHEXPIRED.¶
Because it is quite likely that servers will implement pseudo file systems using volatile filehandles, clients need to be prepared for them, rather than assuming that all filehandles will be persistent.¶
12.6. Exported Root
If the server's root file system is exported, one might conclude that a pseudo file system is unneeded. This is not necessarily so. Assume the following file systems on a server:¶
/ fs1 (exported) /a fs2 (not exported) /a/b fs3 (exported)¶
Because fs2 is not exported, fs3 cannot be reached with simple LOOKUPs. The server must bridge the gap with a pseudo file system.¶
12.7. Mount Point Crossing
The server file system environment may be constructed in such a way that one file system contains a directory that is 'covered' or mounted upon by a second file system. For example:¶
/a/b (file system 1) /a/b/c/d (file system 2)¶
The pseudo file system for this server may be constructed to look like:¶
/ (place holder/not exported) /a/b (file system 1) /a/b/c/d (file system 2)¶
It is the server's responsibility to present the pseudo file system that is complete to the client. If the client sends a LOOKUP request for the path /a/b/c/d, the server's response is the filehandle of the root of the file system /a/b/c/d. In previous versions of the NFS protocol, the server would respond with the filehandle of directory /a/b/c/d within the file system /a/b.¶
The NFS client will be able to determine if it crosses a server mount point by a change in the value of the "fsid" attribute.¶
12.8. Security Policy and Namespace Presentation
Because NFSv4 clients possess the ability to change the security mechanisms used, after determining what is allowed, by using SECINFO and SECINFO_NO_NAME, the server SHOULD NOT present a different view of the namespace based on the security mechanism being used by a client. Instead, it should present a consistent view and return NFS4ERR_WRONGSEC if an attempt is made to access data with an inappropriate security mechanism.¶
If security considerations make it necessary to hide the existence of a particular file system, as opposed to all of the data within it, the server can apply the security policy of a shared resource in the server's namespace to components of the resource's ancestors. For example:¶
/ (place holder/not exported) /a/b (file system 1) /a/b/MySecretProject (file system 2)¶
The /a/b/MySecretProject directory is a real file system and is the shared resource. Suppose the security policy for /a/b/MySecretProject is Kerberos with integrity and it is desired to limit knowledge of the existence of this file system. In this case, the server should apply the same security policy to /a/b. This allows for knowledge of the existence of a file system to be secured when desirable.¶
For the case of the use of multiple, disjoint security mechanisms in the server's resources, applying that sort of policy would result in the higher-level file system not being accessible using any security flavor. Therefore, that sort of configuration is not compatible with hiding the existence (as opposed to the contents) from clients using multiple disjoint sets of security flavors.¶
In other circumstances, a desirable policy is for the security of a particular object in the server's namespace to include the union of all security mechanisms of all direct descendants. A common and convenient practice, unless strong security requirements dictate otherwise, is to make the entire the pseudo file system accessible by all of the valid security mechanisms.¶
Where there is concern about the security of data on the network, clients should use strong security mechanisms to access the pseudo file system in order to prevent man-in-the-middle attacks.¶
13. State Management
Integrating locking into the NFS protocol necessarily causes it to be stateful. With the inclusion of such features as share reservations, file and directory delegations, recallable layouts, and support for mandatory byte-range locking, the protocol becomes substantially more dependent on proper management of state than the traditional combination of NFS and NLM (Network Lock Manager) [xnfs]. These features include expanded locking facilities, which provide some measure of inter-client exclusion, but the state also offers features not readily providable using a stateless model. There are three components to making this state manageable:¶
- clear division between client and server¶
- ability to reliably detect inconsistency in state between client and server¶
- simple and robust recovery mechanisms¶
In this model, the server owns the state information. The client requests changes in locks and the server responds with the changes made. Non-client-initiated changes in locking state are infrequent. The client receives prompt notification of such changes and can adjust its view of the locking state to reflect the server's changes.¶
Individual pieces of state created by the server and passed to the client at its request are represented by 128-bit stateids. These stateids may represent a particular open file, a set of byte-range locks held by a particular owner, or a recallable delegation of privileges to access a file in particular ways or at a particular location.¶
In all cases, there is a transition from the most general information that represents a client as a whole to the eventual lightweight stateid used for most client and server locking interactions. The details of this transition will vary with the type of object but it always starts with a client ID.¶
13.1. Client and Session ID
A client must establish a client ID (see Section 5.5) and then one or more sessionids (see Section 7) before performing any operations to open, byte-range lock, delegate, or obtain a layout for a file object. Each session ID is associated with a specific client ID, and thus serves as a shorthand reference to an NFSv4.1 client.¶
For some types of locking interactions, the client will represent some number of internal locking entities called "owners", which normally correspond to processes internal to the client. For other types of locking-related objects, such as delegations and layouts, no such intermediate entities are provided for, and the locking-related objects are considered to be transferred directly between the server and a unitary client.¶
13.2. Stateid Definition
When the server grants a lock of any type (including opens, byte-range locks, delegations, and layouts), it responds with a unique stateid that represents a set of locks (often a single lock) for the same file, of the same type, and sharing the same ownership characteristics. Thus, opens of the same file by different open-owners each have an identifying stateid. Similarly, each set of byte-range locks on a file owned by a specific lock-owner has its own identifying stateid. Delegations and layouts also have associated stateids by which they may be referenced. The stateid is used as a shorthand reference to a lock or set of locks, and given a stateid, the server can determine the associated state-owner or state-owners (in the case of an open-owner/lock-owner pair) and the associated filehandle. When stateids are used, the current filehandle must be the one associated with that stateid.¶
All stateids associated with a given client ID are associated with a common lease that represents the claim of those stateids and the objects they represent to be maintained by the server. See Section 13.3 for a discussion of the lease.¶
The server may assign stateids independently for different clients. A stateid with the same bit pattern for one client may designate an entirely different set of locks for a different client. The stateid is always interpreted with respect to the client ID associated with the current session. Stateids apply to all sessions associated with the given client ID, and the client may use a stateid obtained from one session on another session associated with the same client ID.¶
13.2.1. Stateid Types
With the exception of special stateids (see Section 13.2.3), each stateid represents locking objects of one of a set of types defined by the NFSv4.1 protocol. Note that in all these cases, where we speak of guarantee, it is understood there are situations such as a client restart, or lock revocation, that allow the guarantee to be voided.¶
-
Stateids may represent opens of files.¶
Each stateid in this case represents the OPEN state for a given client ID/open-owner/filehandle triple. Such stateids are subject to change (with consequent incrementing of the stateid's seqid) in response to OPENs that result in upgrade and OPEN_DOWNGRADE operations.¶
-
Stateids may represent sets of byte-range locks.¶
All locks held on a particular file by a particular owner and gotten under the aegis of a particular open file are associated with a single stateid with the seqid being incremented whenever LOCK and LOCKU operations affect that set of locks.¶
-
Stateids may represent file delegations, which are recallable guarantees by the server to the client that other clients will not reference or modify a particular file, until the delegation is returned. In NFSv4.1, file delegations may be obtained on both regular and non-regular files.¶
A stateid represents a single delegation held by a client for a particular filehandle.¶
-
Stateids may represent directory delegations, which are recallable guarantees by the server to the client that other clients will not modify the directory, until the delegation is returned.¶
A stateid represents a single delegation held by a client for a particular directory filehandle.¶
-
Stateids may represent layouts, which are recallable guarantees by the server to the client that particular files may be accessed via an alternate data access protocol at specific locations. Such access is limited to particular sets of byte-ranges and may proceed until those byte-ranges are reduced or the layout is returned.¶
A stateid represents the set of all layouts held by a particular client for a particular filehandle with a given layout type. The seqid is updated as the layouts of that set of byte-ranges change, via layout stateid changing operations such as LAYOUTGET and LAYOUTRETURN.¶
13.2.2. Stateid Structure
Stateids are divided into two fields, a 96-bit "other" field identifying the specific set of locks and a 32-bit "seqid" sequence value. Except in the case of special stateids (see Section 13.2.3), a particular value of the "other" field denotes a set of locks of the same type (for example, byte-range locks, opens, delegations, or layouts), for a specific file or directory, and sharing the same ownership characteristics. The seqid designates a specific instance of such a set of locks, and is incremented to indicate changes in such a set of locks, either by the addition or deletion of locks from the set, a change in the byte-range they apply to, or an upgrade or downgrade in the type of one or more locks.¶
When such a set of locks is first created, the server returns a stateid with seqid value of one. On subsequent operations that modify the set of locks, the server is required to increment the "seqid" field by one whenever it returns a stateid for the same state-owner/file/type combination and there is some change in the set of locks actually designated. In this case, the server will return a stateid with an "other" field the same as previously used for that state-owner/file/type combination, with an incremented "seqid" field. This pattern continues until the seqid is incremented past NFS4_UINT32_MAX, and one (not zero) is the next seqid value.¶
The purpose of the incrementing of the seqid is to allow the server to communicate to the client the order in which operations that modified locking state associated with a stateid have been processed and to make it possible for the client to send requests that are conditional on the set of locks not having changed since the stateid in question was returned.¶
Except for layout stateids (Section 17.5.3), when a client sends a stateid to the server, it has two choices with regard to the seqid sent. It may set the seqid to zero to indicate to the server that it wishes the most up-to-date seqid for that stateid's "other" field to be used. This would be the common choice in the case of a stateid sent with a READ or WRITE operation. It also may set a non-zero value, in which case the server checks if that seqid is the correct one. In that case, the server is required to return NFS4ERR_OLD_STATEID if the seqid is lower than the most current value and NFS4ERR_BAD_STATEID if the seqid is greater than the most current value. This would be the common choice in the case of stateids sent with a CLOSE or OPEN_DOWNGRADE. Because OPENs may be sent in parallel for the same owner, a client might close a file without knowing that an OPEN upgrade had been done by the server, changing the lock in question. If CLOSE were sent with a zero seqid, the OPEN upgrade would be cancelled before the client even received an indication that an upgrade had happened.¶
When a stateid is sent by the server to the client as part of a callback operation, it is not subject to checking for a current seqid and returning NFS4ERR_OLD_STATEID. This is because the client is not in a position to know the most up-to-date seqid and thus cannot verify it. Unless specially noted, the seqid value for a stateid sent by the server to the client as part of a callback is required to be zero with NFS4ERR_BAD_STATEID returned if it is not.¶
In making comparisons between seqids, both by the client in determining the order of operations and by the server in determining whether the NFS4ERR_OLD_STATEID is to be returned, the possibility of the seqid being swapped around past the NFS4_UINT32_MAX value needs to be taken into account. When two seqid values are being compared, the total count of slots for all sessions associated with the current client is used to do this. When one seqid value is less than this total slot count and another seqid value is greater than NFS4_UINT32_MAX minus the total slot count, the former is to be treated as lower than the latter, despite the fact that it is numerically greater.¶
13.2.3. Special Stateids
Stateid values whose "other" field is either all zeros or all ones are reserved. They may not be assigned by the server but have special meanings defined by the protocol. The particular meaning depends on whether the "other" field is all zeros or all ones and the specific value of the "seqid" field.¶
The following combinations of "other" and "seqid" are defined in NFSv4.1:¶
- When "other" and "seqid" are both zero, the stateid is treated as a special anonymous stateid, which can be used in READ, WRITE, and SETATTR requests to indicate the absence of any OPEN state associated with the request. When an anonymous stateid value is used and an existing open denies the form of access requested, then access will be denied to the request. This stateid MUST NOT be used on operations to data servers (Section 18.6).¶
- When "other" and "seqid" are both all ones, the stateid is a special READ bypass stateid. When this value is used in WRITE or SETATTR, it is treated like the anonymous value. When used in READ, the server MAY grant access, even if access would normally be denied to READ operations. This stateid MUST NOT be used on operations to data servers.¶
- When "other" is zero and "seqid" is one, the stateid represents the current stateid, which is whatever value is the last stateid returned by an operation within the COMPOUND. In the case of an OPEN, the stateid returned for the open file and not the delegation is used. The stateid passed to the operation in place of the special value has its "seqid" value set to zero, except when the current stateid is used by the operation CLOSE or OPEN_DOWNGRADE. If there is no operation in the COMPOUND that has returned a stateid value, the server MUST return the error NFS4ERR_BAD_STATEID. As illustrated in Figure 6, if the value of a current stateid is a special stateid and the stateid of an operation's arguments has "other" set to zero and "seqid" set to one, then the server MUST return the error NFS4ERR_BAD_STATEID.¶
- When "other" is zero and "seqid" is NFS4_UINT32_MAX, the stateid represents a reserved stateid value defined to be invalid. When this stateid is used, the server MUST return the error NFS4ERR_BAD_STATEID.¶
If a stateid value is used that has all zeros or all ones in the "other" field but does not match one of the cases above, the server MUST return the error NFS4ERR_BAD_STATEID.¶
Special stateids, unlike other stateids, are not associated with individual client IDs or filehandles and can be used with all valid client IDs and filehandles. In the case of a special stateid designating the current stateid, the current stateid value substituted for the special stateid is associated with a particular client ID and filehandle, and so, if it is used where the current filehandle does not match that associated with the current stateid, the operation to which the stateid is passed will return NFS4ERR_BAD_STATEID.¶
13.2.4. Stateid Lifetime and Validation
Stateids must remain valid until either a client restart or a server restart or until the client returns all of the locks associated with the stateid by means of an operation such as CLOSE or DELEGRETURN. If the locks are lost due to revocation, as long as the client ID is valid, the stateid remains a valid designation of that revoked state until the client frees it by using FREE_STATEID. Stateids associated with byte-range locks are an exception. They remain valid even if a LOCKU frees all remaining locks, so long as the open file with which they are associated remains open, unless the client frees the stateids via the FREE_STATEID operation.¶
It should be noted that there are situations in which the client's locks become invalid, without the client requesting they be returned. These include lease expiration and a number of forms of lock revocation within the lease period. It is important to note that in these situations, the stateid remains valid and the client can use it to determine the disposition of the associated lost locks.¶
An "other" value must never be reused for a different purpose (i.e., different filehandle, owner, or type of locks) within the context of a single client ID. A server may retain the "other" value for the same purpose beyond the point where it may otherwise be freed, but if it does so, it must maintain "seqid" continuity with previous values.¶
One mechanism that may be used to satisfy the requirement that the server recognize invalid and out-of-date stateids is for the server to divide the "other" field of the stateid into two fields.¶
- an index into a table of locking-state structures.¶
- a generation number that is incremented on each allocation of a table entry for a particular use.¶
And then store in each table entry,¶
- the client ID with which the stateid is associated.¶
- the current generation number for the (at most one) valid stateid sharing this index value.¶
- the filehandle of the file on which the locks are taken.¶
- an indication of the type of stateid (open, byte-range lock, file delegation, directory delegation, layout).¶
- the last "seqid" value returned corresponding to the current "other" value.¶
- an indication of the current status of the locks associated with this stateid, in particular, whether these have been revoked and if so, for what reason.¶
With this information, an incoming stateid can be validated and the appropriate error returned when necessary. Special and non-special stateids are handled separately. (See Section 13.2.3 for a discussion of special stateids.)¶
Note that stateids are implicitly qualified by the current client ID, as derived from the client ID associated with the current session. Note, however, that the semantics of the session will prevent stateids associated with a previous client or server instance from being analyzed by this procedure.¶
If server restart has resulted in an invalid client ID or a session ID that is invalid, SEQUENCE will return an error and the operation that takes a stateid as an argument will never be processed.¶
If there has been a server restart where there is a persistent session and all leased state has been lost, then the session in question will, although valid, be marked as dead, and any operation not satisfied by means of the reply cache will receive the error NFS4ERR_DEADSESSION, and thus not be processed as indicated below.¶
When a stateid is being tested and the "other" field is all zeros or all ones, a check that the "other" and "seqid" fields match a defined combination for a special stateid is done and the results determined as follows:¶
- If the "other" and "seqid" fields do not match a defined combination associated with a special stateid, the error NFS4ERR_BAD_STATEID is returned.¶
- If the special stateid is one designating the current stateid and there is a current stateid, then the current stateid is substituted for the special stateid and the checks appropriate to non-special stateids are performed.¶
- If the combination is valid in general but is not appropriate to the context in which the stateid is used (e.g., an all-zero stateid is used when an OPEN stateid is required in a LOCK operation), the error NFS4ERR_BAD_STATEID is also returned.¶
- Otherwise, the check is completed and the special stateid is accepted as valid.¶
When a stateid is being tested, and the "other" field is neither all zeros nor all ones, the following procedure could be used to validate an incoming stateid and return an appropriate error, when necessary, assuming that the "other" field would be divided into a table index and an entry generation.¶
- If the table index field is outside the range of the associated table, return NFS4ERR_BAD_STATEID.¶
- If the selected table entry is of a different generation than that specified in the incoming stateid, return NFS4ERR_BAD_STATEID.¶
- If the selected table entry does not match the current filehandle, return NFS4ERR_BAD_STATEID.¶
- If the client ID in the table entry does not match the client ID associated with the current session, return NFS4ERR_BAD_STATEID.¶
- If the stateid represents revoked state, then return NFS4ERR_EXPIRED, NFS4ERR_ADMIN_REVOKED, or NFS4ERR_DELEG_REVOKED, as appropriate.¶
- If the stateid type is not valid for the context in which the stateid appears, return NFS4ERR_BAD_STATEID. Note that a stateid may be valid in general, as would be reported by the TEST_STATEID operation, but be invalid for a particular operation, as, for example, when a stateid that doesn't represent byte-range locks is passed to the non-from_open case of LOCK or to LOCKU, or when a stateid that does not represent an open is passed to CLOSE or OPEN_DOWNGRADE. In such cases, the server MUST return NFS4ERR_BAD_STATEID.¶
- If the "seqid" field is not zero and it is greater than the current sequence value corresponding to the current "other" field, return NFS4ERR_BAD_STATEID.¶
- If the "seqid" field is not zero and it is less than the current sequence value corresponding to the current "other" field, return NFS4ERR_OLD_STATEID.¶
- Otherwise, the stateid is valid and the table entry should contain any additional information about the type of stateid and information associated with that particular type of stateid, such as the associated set of locks, e.g., open-owner and lock-owner information, as well as information on the specific locks, e.g., open modes and byte-ranges.¶
13.2.5. Stateid Use for I/O Operations
Clients performing I/O operations need to select an appropriate stateid based on the locks (including opens and delegations) held by the client and the various types of state-owners sending the I/O requests. SETATTR operations that change the file size are treated like I/O operations in this regard.¶
The following rules, applied in order of decreasing priority, govern the selection of the appropriate stateid. In following these rules, the client will only consider locks of which it has actually received notification by an appropriate operation response or callback. Note that the rules are slightly different in the case of I/O to data servers when file layouts are being used (see Section 18.10.1).¶
- If the client holds a delegation for the file in question, the delegation stateid SHOULD be used.¶
- Otherwise, if the entity corresponding to the lock-owner (e.g., a process) sending the I/O has a byte-range lock stateid for the associated open file, then the byte-range lock stateid for that lock-owner and open file SHOULD be used.¶
- If there is no byte-range lock stateid, then the OPEN stateid for the open file in question SHOULD be used.¶
- Finally, if none of the above apply, then a special stateid SHOULD be used.¶
Ignoring these rules may result in situations in which the server does not have information necessary to properly process the request. For example, when mandatory byte-range locks are in effect, if the stateid does not indicate the proper lock-owner, via a lock stateid, a request might be avoidably rejected.¶
The server however should not try to enforce these ordering rules and should use whatever information is available to properly process I/O requests. In particular, when a client has a delegation for a given file, it SHOULD take note of this fact in processing a request, even if it is sent with a special stateid.¶
13.2.6. Stateid Use for SETATTR Operations
Because each operation is associated with a session ID and from that the clientid can be determined, operations do not need to include a stateid for the server to be able to determine whether they should cause a delegation to be recalled or are to be treated as done within the scope of the delegation.¶
In the case of SETATTR operations, a stateid is present. In cases other than those that set the file size, the client may send either a special stateid or, when a delegation is held for the file in question, a delegation stateid. While the server SHOULD validate the stateid and may use the stateid to optimize the determination as to whether a delegation is held, it SHOULD note the presence of a delegation even when a special stateid is sent, and MUST accept a valid delegation stateid when sent.¶
13.3. Lease Renewal
Each client/server pair, as represented by a client ID, has a single lease. The purpose of the lease is to allow the client to indicate to the server, in a low-overhead way, that it is active, and thus that the server is to retain the client's locks. This arrangement allows the server to remove stale locking-related objects that are held by a client that has crashed or is otherwise unreachable, once the relevant lease expires. This in turn allows other clients to obtain conflicting locks without being delayed indefinitely by inactive or unreachable clients. It is not a mechanism for cache consistency and lease renewals may not be denied if the lease interval has not expired.¶
Since each session is associated with a specific client (identified by the client's client ID), any operation sent on that session is an indication that the associated client is reachable. When a request is sent for a given session, successful execution of a SEQUENCE operation (or successful retrieval of the result of SEQUENCE from the reply cache) on an unexpired lease will result in the lease being implicitly renewed, for the standard renewal period (equal to the lease_time attribute).¶
If the client ID's lease has not expired when the server receives a SEQUENCE operation, then the server MUST renew the lease. If the client ID's lease has expired when the server receives a SEQUENCE operation, the server MAY renew the lease; this depends on whether any state was revoked as a result of the client's failure to renew the lease before expiration.¶
Absent other activity that would renew the lease, a COMPOUND consisting of a single SEQUENCE operation will suffice. The client should also take communication-related delays into account and take steps to ensure that the renewal messages actually reach the server in good time. For example:¶
- When trunking is in effect, the client should consider sending multiple requests on different connections, in order to ensure that renewal occurs, even in the event of blockage in the path used for one of those connections.¶
-
Transport retransmission delays might become so large as to approach or exceed the length of the lease period. This may be particularly likely when the server is unresponsive due to a restart; see Section 13.4.2.1. If the client implementation is not careful, transport retransmission delays can result in the client failing to detect a server restart before the grace period ends. The scenario is that the client is using a transport with exponential backoff, such that the maximum retransmission timeout exceeds both the grace period and the lease_time attribute. A network partition causes the client's connection's retransmission interval to back off, and even after the partition heals, the next transport-level retransmission is sent after the server has restarted and its grace period ends.¶
The client MUST either recover from the ensuing NFS4ERR_NO_GRACE errors or it MUST ensure that, despite transport-level retransmission intervals that exceed the lease_time, a SEQUENCE operation is sent that renews the lease before expiration. The client can achieve this by associating a new connection with the session, and sending a SEQUENCE operation on it. However, if the attempt to establish a new connection is delayed for some reason (e.g., exponential backoff of the connection establishment packets), the client will have to abort the connection establishment attempt before the lease expires, and attempt to reconnect.¶
If the server renews the lease upon receiving a SEQUENCE operation, the server MUST NOT allow the lease to expire while the rest of the operations in the COMPOUND procedure's request are still executing. Once the last operation has finished, and the response to COMPOUND has been sent, the server MUST set the lease to expire no sooner than the sum of current time and the value of the lease_time attribute.¶
A client ID's lease can expire when it has been at least the lease interval (lease_time) since the last lease-renewing SEQUENCE operation was sent on any of the client ID's sessions and there are no active COMPOUND operations on any such sessions.¶
Because the SEQUENCE operation is the basic mechanism to renew a lease, and because it must be done at least once for each lease period, it is the natural mechanism whereby the server will inform the client of changes in the lease status that the client needs to be informed of. The client should inspect the status flags (sr_status_flags) returned by sequence and take the appropriate action (see Section 23.46.3 for details).¶
- The status bits SEQ4_STATUS_CB_PATH_DOWN and SEQ4_STATUS_CB_PATH_DOWN_SESSION indicate problems with the backchannel that the client may need to address in order to receive callback requests.¶
- The status bits SEQ4_STATUS_CB_GSS_CONTEXTS_EXPIRING and SEQ4_STATUS_CB_GSS_CONTEXTS_EXPIRED indicate problems with GSS contexts or RPCSEC_GSS handles for the backchannel that the client might have to address in order to allow callback requests to be sent.¶
- The status bits SEQ4_STATUS_EXPIRED_ALL_STATE_REVOKED, SEQ4_STATUS_EXPIRED_SOME_STATE_REVOKED, SEQ4_STATUS_ADMIN_STATE_REVOKED, and SEQ4_STATUS_RECALLABLE_STATE_REVOKED notify the client of lock revocation events. When these bits are set, the client should use TEST_STATEID to find what stateids have been revoked and use FREE_STATEID to acknowledge loss of the associated state.¶
- The status bit SEQ4_STATUS_LEASE_MOVE indicates that responsibility for lease renewal has been transferred to one or more new servers.¶
- The status bit SEQ4_STATUS_RESTART_RECLAIM_NEEDED indicates that due to server restart the client must reclaim locking state.¶
- The status bit SEQ4_STATUS_BACKCHANNEL_FAULT indicates that the server has encountered an unrecoverable fault with the backchannel (e.g., it has lost track of a sequence ID for a slot in the backchannel).¶
13.4. Crash Recovery
A critical requirement in crash recovery is that both the client and the server know when the other has failed. Additionally, it is required that a client sees a consistent view of data across server restarts. All READ and WRITE operations that may have been queued within the client or network buffers must wait until the client has successfully recovered the locks protecting the READ and WRITE operations. Any that reach the server before the server can safely determine that the client has recovered enough locking state to be sure that such operations can be safely processed must be rejected. This will happen because either:¶
- The state presented is no longer valid since it is associated with a now invalid client ID. In this case, the client will receive either an NFS4ERR_BADSESSION or NFS4ERR_DEADSESSION error, and any attempt to attach a new session to that invalid client ID will result in an NFS4ERR_STALE_CLIENTID error.¶
- Subsequent recovery of locks may make execution of the operation inappropriate (NFS4ERR_GRACE).¶
13.4.1. Client Failure and Recovery
In the event that a client fails, the server may release the client's locks when the associated lease has expired. Conflicting locks from another client may only be granted after this lease expiration. As discussed in Section 13.3, when a client has not failed and re-establishes its lease before expiration occurs, requests for conflicting locks will not be granted.¶
To minimize client delay upon restart, lock requests are associated with an instance of the client by a client-supplied verifier. This verifier is part of the client_owner4 sent in the initial EXCHANGE_ID call made by the client. The server returns a client ID as a result of the EXCHANGE_ID operation. The client then confirms the use of the client ID by establishing a session associated with that client ID (see Section 23.36.3 for a description of how this is done). All locks, including opens, byte-range locks, delegations, and layouts obtained by sessions using that client ID, are associated with that client ID.¶
Since the verifier will be changed by the client upon each initialization, the server can compare a new verifier to the verifier associated with currently held locks and determine that they do not match. This signifies the client's new instantiation and subsequent loss (upon confirmation of the new client ID) of locking state. As a result, the server is free to release all locks held that are associated with the old client ID that was derived from the old verifier. At this point, conflicting locks from other clients, kept waiting while the lease had not yet expired, can be granted. In addition, all stateids associated with the old client ID can also be freed, as they are no longer reference-able.¶
Note that the verifier must have the same uniqueness properties as the verifier for the COMMIT operation.¶
13.4.2. Server Failure and Recovery
If the server loses locking state (usually as a result of a restart), it must allow clients time to discover this fact and re-establish the lost locking state. The client must be able to re-establish the locking state without having the server deny valid requests because the server has granted conflicting access to another client. Likewise, if there is a possibility that clients have not yet re-established their locking state for a file and that such locking state might make it invalid to perform READ or WRITE operations. For example, if mandatory locks are a possibility, the server must disallow READ and WRITE operations for that file.¶
A client can determine that loss of locking state has occurred via several methods.¶
- When a SEQUENCE (most common) or other operation returns NFS4ERR_BADSESSION, this may mean that the session has been destroyed but the client ID is still valid. The client sends a CREATE_SESSION request with the client ID to re-establish the session. If CREATE_SESSION fails with NFS4ERR_STALE_CLIENTID, the client must establish a new client ID (see Section 13.1) and re-establish its lock state with the new client ID, after the CREATE_SESSION operation succeeds (see Section 13.4.2.1).¶
- When a SEQUENCE (most common) or other operation on a persistent session returns NFS4ERR_DEADSESSION, this indicates that a session is no longer usable for new, i.e., not satisfied from the reply cache, operations. Once all pending operations are determined to be either performed before the retry or not performed, the client sends a CREATE_SESSION request with the client ID to re-establish the session. If CREATE_SESSION fails with NFS4ERR_STALE_CLIENTID, the client must establish a new client ID (see Section 13.1) and re-establish its lock state after the CREATE_SESSION, with the new client ID, succeeds (Section 13.4.2.1).¶
- When an operation, neither SEQUENCE nor preceded by SEQUENCE (for example, CREATE_SESSION, DESTROY_SESSION), returns NFS4ERR_STALE_CLIENTID, the client MUST establish a new client ID (Section 13.1) and re-establish its lock state (Section 13.4.2.1).¶
13.4.2.1. State Reclaim
When state information and the associated locks are lost as a result of a server restart, the protocol must provide a way to cause that state to be re-established. The approach used is to define, for most types of locking state (layouts are an exception), a request whose function is to allow the client to re-establish on the server a lock first obtained from a previous instance. Generally, these requests are variants of the requests normally used to create locks of that type and are referred to as "reclaim-type" requests, and the process of re-establishing such locks is referred to as "reclaiming" them.¶
Because each client must have an opportunity to reclaim all of the locks that it has without the possibility that some other client will be granted a conflicting lock, a "grace period" is devoted to the reclaim process. During this period, requests creating client IDs and sessions are handled normally, but locking requests are subject to special restrictions. Only reclaim-type locking requests are allowed, unless the server can reliably determine (through state persistently maintained across restart instances) that granting any such lock cannot possibly conflict with a subsequent reclaim. When a request is made to obtain a new lock (i.e., not a reclaim-type request) during the grace period and such a determination cannot be made, the server must return the error NFS4ERR_GRACE.¶
Once a session is established using the new client ID, the client will use reclaim-type locking requests (e.g., LOCK operations with reclaim set to TRUE and OPEN operations with a claim type of CLAIM_PREVIOUS; see Section 14.11) to re-establish its locking state. Once this is done, or if there is no such locking state to reclaim, the client sends a global RECLAIM_COMPLETE operation, i.e., one with the rca_one_fs argument set to FALSE, to indicate that it has reclaimed all of the locking state that it will reclaim. Once a client sends such a RECLAIM_COMPLETE operation, it may attempt non-reclaim locking operations, although it might get an NFS4ERR_GRACE status result from each such operation until the period of special handling is over. See Section 16.11.9 for a discussion of the analogous handling lock reclamation in the case of file systems transitioning from server to server.¶
During the grace period, the server must reject READ and WRITE operations and non-reclaim locking requests (i.e., other LOCK and OPEN operations) with an error of NFS4ERR_GRACE, unless it can guarantee that these may be done safely, as described below.¶
The grace period may last until all clients that are known to possibly have had locks have done a global RECLAIM_COMPLETE operation, indicating that they have finished reclaiming the locks they held before the server restart. This means that a client that has done a RECLAIM_COMPLETE must be prepared to receive an NFS4ERR_GRACE when attempting to acquire new locks. In order for the server to know that all clients with possible prior lock state have done a RECLAIM_COMPLETE, the server must maintain in stable storage a list clients that may have such locks. The server may also terminate the grace period before all clients have done a global RECLAIM_COMPLETE. The server SHOULD NOT terminate the grace period before a time equal to the lease period in order to give clients an opportunity to find out about the server restart, as a result of sending requests on associated sessions with a frequency governed by the lease time. Note that when a client does not send such requests (or they are sent by the client but not received by the server), it is possible for the grace period to expire before the client finds out that the server restart has occurred.¶
Some additional time in order to allow a client to establish a new client ID and session and to effect lock reclaims may be added to the lease time. Note that analogous rules apply to file system-specific grace periods discussed in Section 16.11.9.¶
If the server can reliably determine that granting a non-reclaim request will not conflict with reclamation of locks by other clients, the NFS4ERR_GRACE error does not have to be returned even within the grace period, although NFS4ERR_GRACE must always be returned to clients attempting a non-reclaim lock request before doing their own global RECLAIM_COMPLETE. For the server to be able to service READ and WRITE operations during the grace period, it must again be able to guarantee that no possible conflict could arise between a potential reclaim locking request and the READ or WRITE operation. If the server is unable to offer that guarantee, the NFS4ERR_GRACE error must be returned to the client.¶
For a server to provide simple, valid handling during the grace period, the easiest method is to simply reject all non-reclaim locking requests and READ and WRITE operations by returning the NFS4ERR_GRACE error. However, a server may keep information about granted locks in stable storage. With this information, the server could determine if a locking, READ or WRITE operation can be safely processed.¶
For example, if the server maintained on stable storage summary information on whether mandatory locks exist, either mandatory byte-range locks, or share reservations specifying deny modes, many requests could be allowed during the grace period. If it is known that no such share reservations exist, OPEN request that do not specify deny modes may be safely granted. If, in addition, it is known that no mandatory byte-range locks exist, either through information stored on stable storage or simply because the server does not support such locks, READ and WRITE operations may be safely processed during the grace period. Another important case is where it is known that no mandatory byte-range locks exist, either because the server does not provide support for them or because their absence is known from persistently recorded data. In this case, READ and WRITE operations specifying stateids derived from reclaim-type operations may be validly processed during the grace period because of the fact that the valid reclaim ensures that no lock subsequently granted can prevent the I/O.¶
To reiterate, for a server that allows non-reclaim lock and I/O requests to be processed during the grace period, it MUST determine that no lock subsequently reclaimed will be rejected and that no lock subsequently reclaimed would have prevented any I/O operation processed during the grace period.¶
Clients should be prepared for the return of NFS4ERR_GRACE errors for non-reclaim lock and I/O requests. In this case, the client should employ a retry mechanism for the request. A delay (on the order of several seconds) between retries should be used to avoid overwhelming the server. Further discussion of the general issue is included in [Floyd]. The client must account for the server that can perform I/O and non-reclaim locking requests within the grace period as well as those that cannot do so.¶
A reclaim-type locking request outside the server's grace period can only succeed if the server can guarantee that no conflicting lock or I/O request has been granted since restart.¶
A server may, upon restart, establish a new value for the lease period. Therefore, clients should, once a new client ID is established, refetch the lease_time attribute and use it as the basis for lease renewal for the lease associated with that server. However, the server must establish, for this restart event, a grace period at least as long as the lease period for the previous server instantiation. This allows the client state obtained during the previous server instance to be reliably re-established.¶
The possibility exists that, because of server configuration events, the client will be communicating with a server different than the one on which the locks were obtained, as shown by the combination of eir_server_scope and eir_server_owner. This leads to the issue of if and when the client should attempt to reclaim locks previously obtained on what is being reported as a different server. The rules to resolve this question are as follows:¶
- If the server scope is different, the client should not attempt to reclaim locks. In this situation, no lock reclaim is possible. Any attempt to re-obtain the locks with non-reclaim operations is problematic since there is no guarantee that the existing filehandles will be recognized by the new server, or that if recognized, they denote the same objects. It is best to treat the locks as having been revoked by the reconfiguration event.¶
- If the server scope is the same, the client should attempt to reclaim locks, even if the eir_server_owner value is different. In this situation, it is the responsibility of the server to return NFS4ERR_NO_GRACE if it cannot provide correct support for lock reclaim operations, including the prevention of edge conditions.¶
The eir_server_owner field is not used in making this determination. Its function is to specify trunking possibilities for the client (see Section 7.5) and not to control lock reclaim.¶
13.4.2.1.1. Security Issues for State Reclaim
During the grace period, a client can reclaim state that it believes or asserts it had before the server restarted. Unless the server has maintained a complete record of all the state the client had, the server has little choice but to trust the client's requests. (Of course, if the server maintained a complete record, then there would be no need to force the client to reclaim state after server restart.) While the server has to trust the client to tell the truth, the negative consequences for security are limited to enabling denial-of-service attacks in situations in which AUTH_SYS, particularly AUTH_SYS in the clear, is supported. The fundamental rule for the server when processing reclaim requests is that it MUST NOT grant the reclaim if an equivalent non-reclaim request would not be granted during steady state due to access control or access conflict issues. For example, an OPEN request during a reclaim will be refused with NFS4ERR_ACCESS if the principal making the request does not have sufficient access to open the file according to the acl, dacl, or mode attributes of the file.¶
Nonetheless, it is possible that a client operating in error or maliciously could, during reclaim, prevent another client from reclaiming access to state. For example, an attacker could send an OPEN reclaim operation with a deny mode that prevents another client from reclaiming the OPEN state it had before the server restarted. The attacker could perform the same denial of service during steady state prior to server restart, as long as the attacker had permissions. Given that the attack vectors are equivalent, the grace period does not offer any additional opportunity for denial of service, and any concerns about this attack vector, whether during grace or steady state, are addressed in the same way, by using RPCSEC_GSS for authentication and limiting access to the file only to principals that the owner of the file trusts.¶
Note that if prior to restart the server had client IDs with the EXCHGID4_FLAG_BIND_PRINC_STATEID (Section 23.35) capability set, then the server SHOULD record in stable storage the client owner id and the principal that established the client ID via EXCHANGE_ID. If the server does not do so, then there is a risk a client will be unable to reclaim state if it does not have a credential for a principal that was originally authorized to establish the state.¶
13.4.3. Network Partitions and Recovery
If the duration of a network partition is greater than the lease period provided by the server, the server will not have received a lease renewal from the client. If this occurs, the server may free all locks held for the client or it may allow the lock state to remain for a considerable period, subject to the constraint that if a request for a conflicting lock is made, locks associated with an expired lease do not prevent such a conflicting lock from being granted but MUST be revoked as necessary so as to avoid interfering with such conflicting requests.¶
If the server chooses to delay freeing of lock state until there is a conflict, it may either free all of the client's locks once there is a conflict or it may only revoke the minimum set of locks necessary to allow conflicting requests. When it adopts the finer-grained approach, it must revoke all locks associated with a given stateid, even if the conflict is with only a subset of locks.¶
When the server chooses to free all of a client's lock state, either immediately upon lease expiration or as a result of the first attempt to obtain a conflicting a lock, the server may report the loss of lock state in a number of ways.¶
The server may choose to invalidate the session and the associated client ID. In this case, once the client can communicate with the server, it will receive an NFS4ERR_BADSESSION error. Upon attempting to create a new session, it would get an NFS4ERR_STALE_CLIENTID. Upon creating the new client ID and new session, the client will attempt to reclaim locks. Normally, the server will not allow the client to reclaim locks, because the server will not be in its recovery grace period.¶
Another possibility is for the server to maintain the session and client ID but for all stateids held by the client to become invalid or stale. Once the client can reach the server after such a network partition, the status returned by the SEQUENCE operation will indicate a loss of locking state; i.e., the flag SEQ4_STATUS_EXPIRED_ALL_STATE_REVOKED will be set in sr_status_flags. In addition, all I/O submitted by the client with the now invalid stateids will fail with the server returning the error NFS4ERR_EXPIRED. Once the client learns of the loss of locking state, it will suitably notify the applications that held the invalidated locks. The client should then take action to free invalidated stateids, either by establishing a new client ID using a new verifier or by doing a FREE_STATEID operation to release each of the invalidated stateids.¶
When the server adopts a finer-grained approach to revocation of locks when a client's lease has expired, only a subset of stateids will normally become invalid during a network partition. When the client can communicate with the server after such a network partition heals, the status returned by the SEQUENCE operation will indicate a partial loss of locking state (SEQ4_STATUS_EXPIRED_SOME_STATE_REVOKED). In addition, operations, including I/O submitted by the client, with the now invalid stateids will fail with the server returning the error NFS4ERR_EXPIRED. Once the client learns of the loss of locking state, it will use the TEST_STATEID operation on all of its stateids to determine which locks have been lost and then suitably notify the applications that held the invalidated locks. The client can then release the invalidated locking state and acknowledge the revocation of the associated locks by doing a FREE_STATEID operation on each of the invalidated stateids.¶
When a network partition is combined with a server restart, there are edge conditions that place requirements on the server in order to avoid silent data corruption following the server restart. Two of these edge conditions are known, and are discussed below.¶
The first edge condition arises as a result of the scenarios such as the following:¶
- Client A acquires a lock.¶
- Client A and server experience mutual network partition, such that client A is unable to renew its lease.¶
- Client A's lease expires, and the server releases the lock.¶
- Client B acquires a lock that would have conflicted with that of client A.¶
- Client B releases its lock.¶
- Server restarts.¶
- Network partition between client A and server heals.¶
- Client A connects to a new server instance and finds out about server restart.¶
- Client A reclaims its lock within the server's grace period.¶
Thus, at the final step, the server has erroneously granted client A's lock reclaim. If client B modified the object the lock was protecting, client A will experience object corruption.¶
The second known edge condition arises in situations such as the following:¶
- Client A acquires one or more locks.¶
- Server restarts.¶
- Client A and server experience mutual network partition, such that client A is unable to reclaim all of its locks within the grace period.¶
- Server's reclaim grace period ends. Client A has either no locks or an incomplete set of locks known to the server.¶
- Client B acquires a lock that would have conflicted with a lock of client A that was not reclaimed.¶
- Client B releases the lock.¶
- Server restarts a second time.¶
- Network partition between client A and server heals.¶
- Client A connects to new server instance and finds out about server restart.¶
- Client A reclaims its lock within the server's grace period.¶
As with the first edge condition, the final step of the scenario of the second edge condition has the server erroneously granting client A's lock reclaim.¶
Solving the first and second edge conditions requires either that the server always assumes after it restarts that some edge condition occurs, and thus returns NFS4ERR_NO_GRACE for all reclaim attempts, or that the server record some information in stable storage. The amount of information the server records in stable storage is in inverse proportion to how harsh the server intends to be whenever edge conditions arise. The server that is completely tolerant of all edge conditions will record in stable storage every lock that is acquired, removing the lock record from stable storage only when the lock is released. For the two edge conditions discussed above, the harshest a server can be, and still support a grace period for reclaims, requires that the server record in stable storage some minimal information. For example, a server implementation could, for each client, save in stable storage a record containing:¶
- the co_ownerid field from the client_owner4 presented in the EXCHANGE_ID operation.¶
- a boolean that indicates if the client's lease expired or if there was administrative intervention (see Section 13.5) to revoke a byte-range lock, share reservation, or delegation and there has been no acknowledgment, via FREE_STATEID, of such revocation.¶
- a boolean that indicates whether the client may have locks that it believes to be reclaimable in situations in which the grace period was terminated, making the server's view of lock reclaimability suspect. The server will set this for any client record in stable storage where the client has not done a suitable RECLAIM_COMPLETE (global or file system-specific depending on the target of the lock request) before it grants any new (i.e., not reclaimed) lock to any client.¶
Assuming the above record keeping, for the first edge condition, after the server restarts, the record that client A's lease expired means that another client could have acquired a conflicting byte-range lock, share reservation, or delegation. Hence, the server must reject a reclaim from client A with the error NFS4ERR_NO_GRACE.¶
For the second edge condition, after the server restarts for a second time, the indication that the client had not completed its reclaims at the time at which the grace period ended means that the server must reject a reclaim from client A with the error NFS4ERR_NO_GRACE.¶
When either edge condition occurs, the client's attempt to reclaim locks will result in the error NFS4ERR_NO_GRACE. When this is received, or after the client restarts with no lock state, the client will send a global RECLAIM_COMPLETE. When the RECLAIM_COMPLETE is received, the server and client are again in agreement regarding reclaimable locks and both booleans in persistent storage can be reset, to be set again only when there is a subsequent event that causes lock reclaim operations to be questionable.¶
Regardless of the level and approach to record keeping, the server MUST implement one of the following strategies (which apply to reclaims of share reservations, byte-range locks, and delegations):¶
- Reject all reclaims with NFS4ERR_NO_GRACE. This is extremely unforgiving, but necessary if the server does not record lock state in stable storage.¶
-
Record sufficient state in stable storage such that all known edge conditions involving server restart, including the two noted in this section, are detected. It is acceptable to erroneously recognize an edge condition and not allow a reclaim, when, with sufficient knowledge, it would be allowed. The error the server would return in this case is NFS4ERR_NO_GRACE. Note that it is not known if there are other edge conditions.¶
In the event that, after a server restart, the server determines there is unrecoverable damage or corruption to the information in stable storage, then for all clients and/or locks that may be affected, the server MUST return NFS4ERR_NO_GRACE.¶
A mandate for the client's handling of the NFS4ERR_NO_GRACE error is outside the scope of this specification, since the strategies for such handling are very dependent on the client's operating environment. However, one potential approach is described below.¶
When the client receives NFS4ERR_NO_GRACE, it could examine the change attribute of the objects for which the client is trying to reclaim state, and use that to determine whether to re-establish the state via normal OPEN or LOCK operations. This is acceptable provided that the client's operating environment allows it. In other words, the client implementer is advised to document for his users the behavior. The client could also inform the application that its byte-range lock or share reservations (whether or not they were delegated) have been lost, such as via a UNIX signal, a Graphical User Interface (GUI) pop-up window, etc. See Section 15.5 for a discussion of what the client should do for dealing with unreclaimed delegations on client state.¶
For further discussion of revocation of locks, see Section 13.5.¶
13.5. Server Revocation of Locks
At any point, the server can revoke locks held by a client, and the client must be prepared for this event. When the client detects that its locks have been or may have been revoked, the client is responsible for validating the state information between itself and the server. Validating locking state for the client means that it must verify or reclaim state for each lock currently held.¶
The first occasion of lock revocation is upon server restart. Note that this includes situations in which sessions are persistent and locking state is lost. In this class of instances, the client will receive an error (NFS4ERR_STALE_CLIENTID) on an operation that takes client ID, usually as part of recovery in response to a problem with the current session), and the client will proceed with normal crash recovery as described in the Section 13.4.2.1.¶
The second occasion of lock revocation is the inability to renew the lease before expiration, as discussed in Section 13.4.3. While this is considered a rare or unusual event, the client must be prepared to recover. The server is responsible for determining the precise consequences of the lease expiration, informing the client of the scope of the lock revocation decided upon. The client then uses the status information provided by the server in the SEQUENCE results (field sr_status_flags, see Section 23.46.3) to synchronize its locking state with that of the server, in order to recover.¶
The third occasion of lock revocation can occur as a result of revocation of locks within the lease period, either because of administrative intervention or because a recallable lock (a delegation or layout) was not returned within the lease period after having been recalled. While these are considered rare events, they are possible, and the client must be prepared to deal with them. When either of these events occurs, the client finds out about the situation through the status returned by the SEQUENCE operation. Any use of stateids associated with locks revoked during the lease period will receive the error NFS4ERR_ADMIN_REVOKED or NFS4ERR_DELEG_REVOKED, as appropriate.¶
In all situations in which a subset of locking state may have been revoked, which include all cases in which locking state is revoked within the lease period, it is up to the client to determine which locks have been revoked and which have not. It does this by using the TEST_STATEID operation on the appropriate set of stateids. Once the set of revoked locks has been determined, the applications can be notified, and the invalidated stateids can be freed and lock revocation acknowledged by using FREE_STATEID.¶
13.6. Short and Long Leases
When determining the time period for the server lease, the usual lease trade-offs apply. A short lease is good for fast server recovery at a cost of increased operations to effect lease renewal (when there are no other operations during the period to effect lease renewal as a side effect). A long lease is certainly kinder and gentler to servers trying to handle very large numbers of clients. The number of extra requests to effect lock renewal drops in inverse proportion to the lease time. The disadvantages of a long lease include the possibility of slower recovery after certain failures. After server failure, a longer grace period may be required when some clients do not promptly reclaim their locks and do a global RECLAIM_COMPLETE. In the event of client failure, the longer period for a lease to expire will force conflicting requests to wait longer.¶
A long lease is practical if the server can store lease state in stable storage. Upon recovery, the server can reconstruct the lease state from its stable storage and continue operation with its clients.¶
13.7. Clocks, Propagation Delay, and Calculating Lease Expiration
To avoid the need for synchronized clocks, lease times are granted by the server as a time delta. However, there is a requirement that the client and server clocks do not drift excessively over the duration of the lease. There is also the issue of propagation delay across the network, which could easily be several hundred milliseconds, as well as the possibility that requests will be lost and need to be retransmitted.¶
To take propagation delay into account, the client should subtract it from lease times (e.g., if the client estimates the one-way propagation delay as 200 milliseconds, then it can assume that the lease is already 200 milliseconds old when it gets it). In addition, it will take another 200 milliseconds to get a response back to the server. So the client must send a lease renewal or write data back to the server at least 400 milliseconds before the lease would expire. If the propagation delay varies over the life of the lease (e.g., the client is on a mobile host), the client will need to continuously subtract the increase in propagation delay from the lease times.¶
The server's lease period configuration should take into account the network distance of the clients that will be accessing the server's resources. It is expected that the lease period will take into account the network propagation delays and other network delay factors for the client population. Since the protocol does not allow for an automatic method to determine an appropriate lease period, the server's administrator may have to tune the lease period.¶
13.8. Obsolete Locking Infrastructure from NFSv4.0
There are a number of operations and fields within existing operations that no longer have a function in NFSv4.1. In one way or another, these changes are all due to the implementation of sessions that provide client context and exactly once semantics as a base feature of the protocol, separate from locking itself.¶
The following NFSv4.0 operations MUST NOT be implemented in NFSv4.1. The server MUST return NFS4ERR_NOTSUPP if these operations are found in an NFSv4.1 COMPOUND.¶
- SETCLIENTID since its function has been replaced by EXCHANGE_ID.¶
- SETCLIENTID_CONFIRM since client ID confirmation now happens by means of CREATE_SESSION.¶
- OPEN_CONFIRM because state-owner-based seqids have been replaced by the sequence ID in the SEQUENCE operation.¶
- RELEASE_LOCKOWNER because lock-owners with no associated locks do not have any sequence-related state and so can be deleted by the server at will.¶
- RENEW because every SEQUENCE operation for a session causes lease renewal, making a separate operation superfluous.¶
Also, there are a number of fields, present in existing operations, related to locking that have no use in minor version 1. They were used in minor version 0 to perform functions now provided in a different fashion.¶
- Sequence ids used to sequence requests for a given state-owner and to provide retry protection, now provided via sessions.¶
- Client IDs used to identify the client associated with a given request. Client identification is now available using the client ID associated with the current session, without needing an explicit client ID field.¶
Such vestigial fields in existing operations have no function in NFSv4.1 and are ignored by the server. Note that client IDs in operations new to NFSv4.1 (such as CREATE_SESSION and DESTROY_CLIENTID) are not ignored.¶
14. File Locking and Share Reservations
To support Win32 share reservations, it is necessary to provide operations that atomically open or create files. Having a separate share/unshare operation would not allow correct implementation of the Win32 OpenFile API. In order to correctly implement share semantics, the previous NFS protocol mechanisms used when a file is opened or created (LOOKUP, CREATE, ACCESS) need to be replaced. The NFSv4.1 protocol defines an OPEN operation that is capable of atomically looking up, creating, and locking a file on the server.¶
14.1. Opens and Byte-Range Locks
It is assumed that manipulating a byte-range lock is rare when compared to READ and WRITE operations. It is also assumed that server restarts and network partitions are relatively rare. Therefore, it is important that the READ and WRITE operations have a lightweight mechanism to indicate if they possess a held lock. A LOCK operation contains the heavyweight information required to establish a byte-range lock and uniquely define the owner of the lock.¶
14.1.1. State-Owner Definition
When opening a file or requesting a byte-range lock, the client must specify an identifier that represents the owner of the requested lock. This identifier is in the form of a state-owner, represented in the protocol by a state_owner4, a variable-length opaque array that, when concatenated with the current client ID, uniquely defines the owner of a lock managed by the client. This may be a thread ID, process ID, or other unique value.¶
Owners of opens and owners of byte-range locks are separate entities and remain separate even if the same opaque arrays are used to designate owners of each. The protocol distinguishes between open-owners (represented by open_owner4 structures) and lock-owners (represented by lock_owner4 structures).¶
Each open is associated with a specific open-owner while each byte-range lock is associated with a lock-owner and an open-owner, the latter being the open-owner associated with the open file under which the LOCK operation was done. Delegations and layouts, on the other hand, are not associated with a specific owner but are associated with the client as a whole (identified by a client ID).¶
14.1.2. Use of the Stateid and Locking
All READ, WRITE, and SETATTR operations contain a stateid. For the purposes of this section, SETATTR operations that change the size attribute of a file are treated as if they are writing the area between the old and new sizes (i.e., the byte-range truncated or added to the file by means of the SETATTR), even where SETATTR is not explicitly mentioned in the text. The stateid passed to one of these operations must be one that represents an open, a set of byte-range locks, or a delegation, or it may be a special stateid representing anonymous access or the special bypass stateid.¶
If the state-owner performs a READ or WRITE operation in a situation in which it has established a byte-range lock or share reservation on the server (any OPEN constitutes a share reservation), the stateid (previously returned by the server) must be used to indicate what locks, including both byte-range locks and share reservations, are held by the state-owner. If no state is established by the client, either a byte-range lock or a share reservation, a special stateid for anonymous state (zero as the value for "other" and "seqid") is used. (See Section 13.2.3 for a description of 'special' stateids in general.) Regardless of whether a stateid for anonymous state or a stateid returned by the server is used, if there is a conflicting share reservation or mandatory byte-range lock held on the file, the server MUST refuse to service the READ or WRITE operation.¶
Share reservations are established by OPEN operations and by their nature are mandatory in that when the OPEN denies READ or WRITE operations, that denial results in such operations being rejected with error NFS4ERR_LOCKED. Byte-range locks may be implemented by the server as either mandatory or advisory, or the choice of mandatory or advisory behavior may be determined by the server on the basis of the file being accessed (for example, some UNIX-based servers support a "mandatory lock bit" on the mode attribute such that if set, byte-range locks are required on the file before I/O is possible). When byte-range locks are advisory, they only prevent the granting of conflicting lock requests and have no effect on READs or WRITEs. Mandatory byte-range locks, however, prevent conflicting I/O operations. When they are attempted, they are rejected with NFS4ERR_LOCKED. When the client gets NFS4ERR_LOCKED on a file for which it knows it has the proper share reservation, it will need to send a LOCK operation on the byte-range of the file that includes the byte-range the I/O was to be performed on, with an appropriate locktype field of the LOCK operation's arguments (i.e., READ*_LT for a READ operation, WRITE*_LT for a WRITE operation).¶
Note that for UNIX environments that support mandatory byte-range locking, the distinction between advisory and mandatory locking is subtle. In fact, advisory and mandatory byte-range locks are exactly the same as far as the APIs and requirements on implementation. If the mandatory lock attribute is set on the file, the server checks to see if the lock-owner has an appropriate shared (READ_LT) or exclusive (WRITE_LT) byte-range lock on the byte-range it wishes to READ from or WRITE to. If there is no appropriate lock, the server checks if there is a conflicting lock (which can be done by attempting to acquire the conflicting lock on behalf of the lock-owner, and if successful, release the lock after the READ or WRITE operation is done), and if there is, the server returns NFS4ERR_LOCKED.¶
For Windows environments, byte-range locks are always mandatory, so the server always checks for byte-range locks during I/O requests.¶
Thus, the LOCK operation does not need to distinguish between advisory and mandatory byte-range locks. It is the server's processing of the READ and WRITE operations that introduces the distinction.¶
Every stateid that is validly passed to READ, WRITE, or SETATTR, with the exception of special stateid values, defines an access mode for the file (i.e., OPEN4_SHARE_ACCESS_READ, OPEN4_SHARE_ACCESS_WRITE, or OPEN4_SHARE_ACCESS_BOTH).¶
- For stateids associated with opens, this is the mode defined by the original OPEN that caused the allocation of the OPEN stateid and as modified by subsequent OPENs and OPEN_DOWNGRADEs for the same open-owner/file pair.¶
- For stateids returned by byte-range LOCK operations, the appropriate mode is the access mode for the OPEN stateid associated with the lock set represented by the stateid.¶
- For delegation stateids, the access mode is based on the type of delegation.¶
When a READ, WRITE, or SETATTR (that specifies the size attribute) operation is done, the operation is subject to checking against the access mode to verify that the operation is appropriate given the stateid with which the operation is associated.¶
In the case of WRITE-type operations (i.e., WRITEs and SETATTRs that set size), the server MUST verify that the access mode allows writing and MUST return an NFS4ERR_OPENMODE error if it does not. In the case of READ, the server may perform the corresponding check on the access mode, or it may choose to allow READ on OPENs for OPEN4_SHARE_ACCESS_WRITE, to accommodate clients whose WRITE implementation may unavoidably do reads (e.g., due to buffer cache constraints). However, even if READs are allowed in these circumstances, the server MUST still check for locks that conflict with the READ (e.g., another OPEN specified OPEN4_SHARE_DENY_READ or OPEN4_SHARE_DENY_BOTH). Note that a server that does enforce the access mode check on READs need not explicitly check for conflicting share reservations since the existence of OPEN for OPEN4_SHARE_ACCESS_READ guarantees that no conflicting share reservation can exist.¶
The READ bypass special stateid (all bits of "other" and "seqid" set to one) indicates a desire to bypass locking checks. The server MAY allow READ operations to bypass locking checks at the server, when this special stateid is used. However, WRITE operations with this special stateid value MUST NOT bypass locking checks and are treated exactly the same as if a special stateid for anonymous state were used.¶
A lock may not be granted while a READ or WRITE operation using one of the special stateids is being performed and the scope of the lock to be granted would conflict with the READ or WRITE operation. This can occur when:¶
- A mandatory byte-range lock is requested with a byte-range that conflicts with the byte-range of the READ or WRITE operation. For the purposes of this paragraph, a conflict occurs when a shared lock is requested and a WRITE operation is being performed, or an exclusive lock is requested and either a READ or a WRITE operation is being performed.¶
- A share reservation is requested that denies reading and/or writing and the corresponding operation is being performed.¶
- A delegation is to be granted and the delegation type would prevent the I/O operation, i.e., READ and WRITE conflict with an OPEN_DELEGATE_WRITE delegation and WRITE conflicts with an OPEN_DELEGATE_READ delegation.¶
When a client holds a delegation, it needs to ensure that the stateid sent conveys the association of operation with the delegation, to avoid the delegation from being avoidably recalled. When the delegation stateid, a stateid open associated with that delegation, or a stateid representing byte-range locks derived from such an open is used, the server knows that the READ, WRITE, or SETATTR does not conflict with the delegation but is sent under the aegis of the delegation. Even though it is possible for the server to determine from the client ID (via the session ID) that the client does in fact have a delegation, the server is not obliged to check this, so using a special stateid can result in avoidable recall of the delegation.¶
14.2. Lock Ranges
The protocol allows a lock-owner to request a lock with a byte-range and then either upgrade, downgrade, or unlock a sub-range of the initial lock, or a byte-range that overlaps -- fully or partially -- either with that initial lock or a combination of a set of existing locks for the same lock-owner. It is expected that this will be an uncommon type of request. In any case, servers or server file systems may not be able to support sub-range lock semantics. In the event that a server receives a locking request that represents a sub-range of current locking state for the lock-owner, the server is allowed to return the error NFS4ERR_LOCK_RANGE to signify that it does not support sub-range lock operations. Therefore, the client should be prepared to receive this error and, if appropriate, report the error to the requesting application.¶
The client is discouraged from combining multiple independent locking ranges that happen to be adjacent into a single request since the server may not support sub-range requests for reasons related to the recovery of byte-range locking state in the event of server failure. As discussed in Section 13.4.2, the server may employ certain optimizations during recovery that work effectively only when the client's behavior during lock recovery is similar to the client's locking behavior prior to server failure.¶
14.3. Upgrading and Downgrading Locks
If a client has a WRITE_LT lock on a byte-range, it can request an atomic downgrade of the lock to a READ_LT lock via the LOCK operation, by setting the type to READ_LT. If the server supports atomic downgrade, the request will succeed. If not, it will return NFS4ERR_LOCK_NOTSUPP. The client should be prepared to receive this error and, if appropriate, report the error to the requesting application.¶
If a client has a READ_LT lock on a byte-range, it can request an atomic upgrade of the lock to a WRITE_LT lock via the LOCK operation by setting the type to WRITE_LT or WRITEW_LT. If the server does not support atomic upgrade, it will return NFS4ERR_LOCK_NOTSUPP. If the upgrade can be achieved without an existing conflict, the request will succeed. Otherwise, the server will return either NFS4ERR_DENIED or NFS4ERR_DEADLOCK. The error NFS4ERR_DEADLOCK is returned if the client sent the LOCK operation with the type set to WRITEW_LT and the server has detected a deadlock. The client should be prepared to receive such errors and, if appropriate, report the error to the requesting application.¶
14.4. Stateid Seqid Values and Byte-Range Locks
When a LOCK or LOCKU operation is performed, the stateid returned has the same "other" value as the argument's stateid, and a "seqid" value that is incremented (relative to the argument's stateid) to reflect the occurrence of the LOCK or LOCKU operation. The server MUST increment the value of the "seqid" field whenever there is any change to the locking status of any byte offset as described by any of the locks covered by the stateid. A change in locking status includes a change from locked to unlocked or the reverse or a change from being locked for READ_LT to being locked for WRITE_LT or the reverse.¶
When there is no such change, as, for example, when a range already locked for WRITE_LT is locked again for WRITE_LT, the server MAY increment the "seqid" value.¶
14.5. Issues with Multiple Open-Owners
When the same file is opened by multiple open-owners, a client will have multiple OPEN stateids for that file, each associated with a different open-owner. In that case, there can be multiple LOCK and LOCKU requests for the same lock-owner sent using the different OPEN stateids, and so a situation may arise in which there are multiple stateids, each representing byte-range locks on the same file and held by the same lock-owner but each associated with a different open-owner.¶
In such a situation, the locking status of each byte (i.e., whether it is locked, the READ_LT or WRITE_LT type of the lock, and the lock-owner holding the lock) MUST reflect the last LOCK or LOCKU operation done for the lock-owner in question, independent of the stateid through which the request was sent.¶
When a byte is locked by the lock-owner in question, the open-owner to which that byte-range lock is assigned SHOULD be that of the open-owner associated with the stateid through which the last LOCK of that byte was done. When there is a change in the open-owner associated with locks for the stateid through which a LOCK or LOCKU was done, the "seqid" field of the stateid MUST be incremented, even if the locking, in terms of lock-owners has not changed. When there is a change to the set of locked bytes associated with a different stateid for the same lock-owner, i.e., associated with a different open-owner, the "seqid" value for that stateid MUST NOT be incremented.¶
14.6. Blocking Locks
Some clients require the support of blocking locks. While NFSv4.1 provides a callback when a previously unavailable lock becomes available, this is an OPTIONAL feature and clients cannot depend on its presence. Clients need to be prepared to continually poll for the lock. This presents a fairness problem. Two of the lock types, READW_LT and WRITEW_LT, are used to indicate to the server that the client is requesting a blocking lock. When the callback is not used, the server should maintain an ordered list of pending blocking locks. When the conflicting lock is released, the server may wait for the period of time equal to lease_time for the first waiting client to re-request the lock. After the lease period expires, the next waiting client request is allowed the lock. Clients are required to poll at an interval sufficiently small that it is likely to acquire the lock in a timely manner. The server is not required to maintain a list of pending blocked locks as it is used to increase fairness and not correct operation. Because of the unordered nature of crash recovery, storing of lock state to stable storage would be required to guarantee ordered granting of blocking locks.¶
Servers may also note the lock types and delay returning denial of the request to allow extra time for a conflicting lock to be released, allowing a successful return. In this way, clients can avoid the burden of needless frequent polling for blocking locks. The server should take care in the length of delay in the event the client retransmits the request.¶
If a server receives a blocking LOCK operation, denies it, and then later receives a nonblocking request for the same lock, which is also denied, then it should remove the lock in question from its list of pending blocking locks. Clients should use such a nonblocking request to indicate to the server that this is the last time they intend to poll for the lock, as may happen when the process requesting the lock is interrupted. This is a courtesy to the server, to prevent it from unnecessarily waiting a lease period before granting other LOCK operations. However, clients are not required to perform this courtesy, and servers must not depend on them doing so. Also, clients must be prepared for the possibility that this final locking request will be accepted.¶
When a server indicates, via the flag OPEN4_RESULT_MAY_NOTIFY_LOCK, that CB_NOTIFY_LOCK callbacks might be done for the current open file, the client should take notice of this, but, since this is a hint, cannot rely on a CB_NOTIFY_LOCK always being done. A client may reasonably reduce the frequency with which it polls for a denied lock, since the greater latency that might occur is likely to be eliminated given a prompt callback, but it still needs to poll. When it receives a CB_NOTIFY_LOCK, it should promptly try to obtain the lock, but it should be aware that other clients may be polling and that the server is under no obligation to reserve the lock for that particular client.¶
14.8. OPEN/CLOSE Operations
To provide correct share semantics, a client MUST use the OPEN operation to obtain the initial filehandle and indicate the desired access and what access, if any, to deny. Even if the client intends to use a special stateid for anonymous state or READ bypass, it must still obtain the filehandle for the regular file with the OPEN operation so the appropriate share semantics can be applied. Clients that do not have a deny mode built into their programming interfaces for opening a file should request a deny mode of OPEN4_SHARE_DENY_NONE.¶
The OPEN operation with the CREATE flag also subsumes the CREATE operation for regular files as used in previous versions of the NFS protocol. This allows a create with a share to be done atomically.¶
The CLOSE operation removes all share reservations held by the open-owner on that file. If byte-range locks are held, the client SHOULD release all locks before sending a CLOSE operation. The server MAY free all outstanding locks on CLOSE, but some servers may not support the CLOSE of a file that still has byte-range locks held. The server MUST return failure, NFS4ERR_LOCKS_HELD, if any locks would exist after the CLOSE.¶
The LOOKUP operation will return a filehandle without establishing any lock state on the server. Without a valid stateid, the server will assume that the client has the least access. For example, if one client opened a file with OPEN4_SHARE_DENY_BOTH and another client accesses the file via a filehandle obtained through LOOKUP, the second client could only read the file using the special read bypass stateid. The second client could not WRITE the file at all because it would not have a valid stateid from OPEN and the special anonymous stateid would not be allowed access.¶
14.9. Open Upgrade and Downgrade
When an OPEN is done for a file and the open-owner for which the OPEN is being done already has the file open, the result is to upgrade the open file status maintained on the server to include the access and deny bits specified by the new OPEN as well as those for the existing OPEN. The result is that there is one open file, as far as the protocol is concerned, and it includes the union of the access and deny bits for all of the OPEN requests completed. The OPEN is represented by a single stateid whose "other" value matches that of the original open, and whose "seqid" value is incremented to reflect the occurrence of the upgrade. The increment is required in cases in which the "upgrade" results in no change to the open mode (e.g., an OPEN is done for read when the existing open file is opened for OPEN4_SHARE_ACCESS_BOTH). Only a single CLOSE will be done to reset the effects of both OPENs. The client may use the stateid returned by the OPEN effecting the upgrade or with a stateid sharing the same "other" field and a seqid of zero, although care needs to be taken as far as upgrades that happen while the CLOSE is pending. Note that the client, when sending the OPEN, may not know that the same file is in fact being opened. The above only applies if both OPENs result in the OPENed object being designated by the same filehandle.¶
When the server chooses to export multiple filehandles corresponding to the same file object and returns different filehandles on two different OPENs of the same file object, the server MUST NOT "OR" together the access and deny bits and coalesce the two open files. Instead, the server must maintain separate OPENs with separate stateids and will require separate CLOSEs to free them.¶
When multiple open files on the client are merged into a single OPEN file object on the server, the close of one of the open files (on the client) may necessitate change of the access and deny status of the open file on the server. This is because the union of the access and deny bits for the remaining opens may be smaller (i.e., a proper subset) than previously. The OPEN_DOWNGRADE operation is used to make the necessary change and the client should use it to update the server so that share reservation requests by other clients are handled properly. The stateid returned has the same "other" field as that passed to the server. The "seqid" value in the returned stateid MUST be incremented, even in situations in which there is no change to the access and deny bits for the file.¶
14.10. Parallel OPENs
Unlike the case of NFSv4.0, in which OPEN operations for the same open-owner are inherently serialized because of the owner-based seqid, multiple OPENs for the same open-owner may be done in parallel. When clients do this, they may encounter situations in which, because of the existence of hard links, two OPEN operations may turn out to open the same file, with a later OPEN performed being an upgrade of the first, with this fact only visible to the client once the operations complete.¶
In this situation, clients may determine the order in which the OPENs were performed by examining the stateids returned by the OPENs. Stateids that share a common value of the "other" field can be recognized as having opened the same file, with the order of the operations determinable from the order of the "seqid" fields, mod any possible wraparound of the 32-bit field.¶
When the possibility exists that the client will send multiple OPENs for the same open-owner in parallel, it may be the case that an open upgrade may happen without the client knowing beforehand that this could happen. Because of this possibility, CLOSEs and OPEN_DOWNGRADEs should generally be sent with a non-zero seqid in the stateid, to avoid the possibility that the status change associated with an open upgrade is not inadvertently lost.¶
14.11. Reclaim of Open and Byte-Range Locks
Special forms of the LOCK and OPEN operations are provided when it is necessary to re-establish byte-range locks or opens after a server failure.¶
- To reclaim existing opens, an OPEN operation is performed using a CLAIM_PREVIOUS. Because the client, in this type of situation, will have already opened the file and have the filehandle of the target file, this operation requires that the current filehandle be the target file, rather than a directory, and no file name is specified.¶
- To reclaim byte-range locks, a LOCK operation with the reclaim parameter set to true is used.¶
Reclaims of opens associated with delegations are discussed in Section 15.2.1.¶
15. Client-Side Caching
Client-side caching of data, of file attributes, and of file names is essential to providing good performance with the NFS protocol. Providing distributed cache coherence is a difficult problem, and previous versions of the NFS protocol have not attempted it. Instead, several NFS client implementation techniques have been used to reduce the problems that a lack of coherence poses for users. These techniques have not been clearly defined by earlier protocol specifications, and it is often unclear what is valid or invalid client behavior.¶
The NFSv4.1 protocol uses many techniques similar to those that have been used in previous protocol versions. The NFSv4.1 protocol does not provide distributed cache coherence. However, it defines a more limited set of caching guarantees to allow locks and share reservations to be used without destructive interference from client-side caching.¶
In addition, the NFSv4.1 protocol introduces a delegation mechanism, which allows many decisions normally made by the server to be made locally by clients. This mechanism provides efficient support of the common cases where sharing is infrequent or where sharing is read-only.¶
15.1. Performance Challenges for Client-Side Caching
Caching techniques used in previous versions of the NFS protocol have been successful in providing good performance. However, several scalability challenges can arise when those techniques are used with very large numbers of clients. This is particularly true when clients are geographically distributed, which classically increases the latency for cache revalidation requests.¶
The previous versions of the NFS protocol repeat their file data cache validation requests at the time the file is opened. This behavior can have serious performance drawbacks. A common case is one in which a file is only accessed by a single client. Therefore, sharing is infrequent.¶
In this case, repeated references to the server to find that no conflicts exist are expensive. A better option with regards to performance is to allow a client that repeatedly opens a file to do so without reference to the server. This is done until potentially conflicting operations from another client actually occur.¶
A similar situation arises in connection with byte-range locking. Sending LOCK and LOCKU operations as well as the READ and WRITE operations necessary to make data caching consistent with the locking semantics (see Section 15.3.2) can severely limit performance. When locking is used to provide protection against infrequent conflicts, a large penalty is incurred. This penalty may discourage the use of byte-range locking by applications.¶
The NFSv4.1 protocol provides more aggressive caching strategies with the following design goals:¶
- Compatibility with a large range of server semantics.¶
- Providing the same caching benefits as previous versions of the NFS protocol when unable to support the more aggressive model.¶
- Requirements for aggressive caching are organized so that a large portion of the benefit can be obtained even when not all of the requirements can be met.¶
The appropriate requirements for the server are discussed in later sections in which specific forms of caching are covered (see Section 15.4).¶
15.2. Delegation and Callbacks
Recallable delegation of server responsibilities for a file to a client improves performance by avoiding repeated requests to the server in the absence of inter-client conflict. With the use of a "callback" RPC from server to client, a server recalls delegated responsibilities when another client engages in sharing of a delegated file.¶
A delegation is passed from the server to the client, specifying the object of the delegation and the type of delegation. There are different types of delegations, but each type contains a stateid to be used to represent the delegation when performing operations that depend on the delegation. This stateid is similar to those associated with locks and share reservations but differs in that the stateid for a delegation is associated with a client ID and may be used on behalf of all the open-owners for the given client. A delegation is made to the client as a whole and not to any specific process or thread of control within it.¶
The backchannel is established by CREATE_SESSION and BIND_CONN_TO_SESSION, and the client is required to maintain it. Because the backchannel may be down, even temporarily, correct protocol operation does not depend on them. Preliminary testing of backchannel functionality by means of a CB_COMPOUND procedure with a single operation, CB_SEQUENCE, can be used to check the continuity of the backchannel. A server avoids delegating responsibilities until it has determined that the backchannel exists. Because the granting of a delegation is always conditional upon the absence of conflicting access, clients MUST NOT assume that a delegation will be granted and they MUST always be prepared for OPENs, WANT_DELEGATIONs, and GET_DIR_DELEGATIONs to be processed without any delegations being granted.¶
Unlike locks, an operation by a second client to a delegated file will cause the server to recall a delegation through a callback. For individual operations, we will describe, under IMPLEMENTATION, when such operations are required to effect a recall. A number of points should be noted, however.¶
- The server is free to recall a delegation whenever it feels it is desirable and may do so even if no operations requiring recall are being done.¶
- Operations done outside the NFSv4.1 protocol, due to, for example, access by other protocols including other minor version of NFSv4, or by local access, also need to result in delegation recall when they make analogous changes to file system data, including the delegated file's contents, its attributes and the set of names linked to that file. What is crucial is if the change would invalidate the guarantees provided by the delegation. When this is possible, the delegation needs to be recalled and MUST be returned or revoked before allowing the operation to proceed.¶
- The semantics of the file system are crucial in defining when delegation recall is required. If a particular change within a specific implementation causes change to a file attribute, then delegation recall is required, whether that operation has been specifically listed as requiring delegation recall. Again, what is critical is whether the guarantees provided by the delegation are being invalidated.¶
Despite those caveats, the implementation sections for a number of operations describe situations in which delegation recall would be required under some common circumstances:¶
- For GETATTR, see Section 23.7.4.¶
- For LINK, see Section 23.9.4.¶
- For OPEN, see Section 23.16.4.¶
- For READ, see Section 23.22.4.¶
- For REMOVE, see Section 23.25.4.¶
- For RENAME, see Section 23.26.4.¶
- For SETATTR, see Section 23.30.4.¶
- For WRITE, see Section 23.32.4.¶
On recall, the client holding the delegation needs to flush modified state (such as modified data) to the server and return the delegation. The conflicting request will not be acted on until the recall is complete. The recall is considered complete when the client returns the delegation or the server times its wait for the delegation to be returned and revokes the delegation as a result of the timeout. In the interim, the server will either delay responding to conflicting requests or respond to them with NFS4ERR_DELAY. Following the resolution of the recall, the server has the information necessary to grant or deny the second client's request.¶
At the time the client receives a delegation recall, it may have substantial state that needs to be flushed to the server. Therefore, the server should allow sufficient time for the delegation to be returned since it may involve numerous RPCs to the server. If the server is able to determine that the client is diligently flushing state to the server as a result of the recall, the server may extend the usual time allowed for a recall. However, the time allowed for recall completion should not be unbounded.¶
An example of this is when responsibility to mediate opens on a given file is delegated to a client (see Section 15.4). The server will not know what opens are in effect on the client. Without this knowledge, the server will be unable to determine if the access and deny states for the file allow any particular open until the delegation for the file has been returned.¶
A client failure or a network partition can result in failure to respond to a recall callback. In this case, the server will revoke the delegation, which in turn will render useless any modified state still on the client.¶
15.2.1. Delegation Recovery
There are three situations that delegation recovery needs to deal with:¶
In the event the client restarts, establishment of a new clientid associated with the new client instance or failure to renew the lease will result in the revocation of byte-range locks and share reservations. Delegations, however, may be treated somewhat differently. It is also possible for the same sorts of revocation to occur as a result of lease non-renewal.¶
There will be situations in which delegations will need to be re-established after a client restarts. The reason for this is that the client may have file data stored locally and this data was associated with the previously held delegations. The client will need to re-establish the appropriate file state on the server.¶
To allow for this type of client recovery, the server MAY provide a special period to allow the clients to recover the delegations obtained before the restart. This special period will often be longer the typical lease expiration period. As a result, requests from other clients that conflict with these delegations would need to wait. Because the normal recall process may require significant time for the client to flush changed state to the server, other clients need be prepared for delays that occur because of a conflicting delegation. Such a longer interval would increase the window for clients to restart and consult stable storage so that the delegations can be returned after the data is appropriately flushed to the server.¶
This special period, although analogous to the grace period used after server restart, is distinct from it. For OPEN delegations, such delegations are reclaimed using OPEN with a claim type of CLAIM_DELEGATE_PREV or CLAIM_DELEG_PREV_FH (see Sections 15.5 and 23.16 f or discussion of OPEN delegation and the details of OPEN, respectively). Although these types of OPENs are considered reclaim-type operations they are not, like other sorts of reclaims limited to the grace period. They are intended for use during the special delegation recovery period, and are not directly affected by possible existence of a server grace period.¶
A server MAY support claim types of CLAIM_DELEGATE_PREV and CLAIM_DELEG_PREV_FH, and if it does, it MUST NOT remove delegations upon a CREATE_SESSION that confirm a client ID created by EXCHANGE_ID. Instead, the server MUST, for a period of time no less than that of the value of the lease_time attribute, maintain the client's delegations to allow time for the client to send CLAIM_DELEGATE_PREV and/or CLAIM_DELEG_PREV_FH requests. The server that supports CLAIM_DELEGATE_PREV and/or CLAIM_DELEG_PREV_FH MUST support the DELEGPURGE operation.¶
When the server restarts, delegations are reclaimed (using the OPEN operation with CLAIM_PREVIOUS) in a similar fashion to byte-range locks and share reservations. However, there is a slight semantic difference. In the normal case, if the server decides that a delegation should not be granted, it performs the requested action (e.g., OPEN) without granting any delegation. For reclaim, the server grants the delegation but a special designation is applied so that the client treats the delegation as having been granted but recalled by the server. Because of this, the client has the duty to write all modified state to the server and then return the delegation. This process of handling delegation reclaim reconciles three principles of the NFSv4.1 protocol:¶
- Upon reclaim, a client reporting resources assigned to it by an earlier server instance must be granted those resources.¶
- The server has unquestionable authority to determine whether delegations are to be granted and, once granted, whether they are to be continued.¶
- The use of callbacks should not be depended upon until the client has proven its ability to receive them.¶
When a client needs to reclaim a delegation and there is no associated open, the client may use the CLAIM_PREVIOUS variant of the WANT_DELEGATION operation. However, since the server is not required to support this operation, an alternative is to reclaim via a dummy OPEN together with the delegation using an OPEN of type CLAIM_PREVIOUS. The dummy open file can be released using a CLOSE to re-establish the original state to be reclaimed, a delegation without an associated open.¶
When a client has more than a single open associated with a delegation, state for those additional opens can be established using OPEN operations of type CLAIM_DELEGATE_CUR. When these are used to establish opens associated with reclaimed delegations, the server MUST allow them when made within the grace period.¶
When a network partition occurs, delegations are subject to freeing by the server when the lease renewal period expires. This is similar to the behavior for locks and share reservations. For delegations, however, the server may extend the period in which conflicting requests are held off. Eventually, the occurrence of a conflicting request from another client will cause revocation of the delegation. A loss of the backchannel (e.g., by later network configuration change) will have the same effect. A recall request will fail and revocation of the delegation will result.¶
A client normally finds out about revocation of a delegation when it uses a stateid associated with a delegation and receives one of the errors NFS4ERR_EXPIRED, NFS4ERR_ADMIN_REVOKED, or NFS4ERR_DELEG_REVOKED. It also may find out about delegation revocation after a client restart when it attempts to reclaim a delegation and receives that same error. Note that in the case of a revoked OPEN_DELEGATE_WRITE delegation, there are issues because data may have been modified by the client whose delegation is revoked and separately by other clients. See Section 15.5.1 for a discussion of such issues. Note also that when delegations are revoked, information about the revoked delegation will be written by the server to stable storage (as described in Section 13.4.3). This is done to deal with the case in which a server restarts after revoking a delegation but before the client holding the revoked delegation is notified about the revocation.¶
15.3. Data Caching
When applications share access to a set of files, they need to be implemented so as to take account of the possibility of conflicting access by another application. This is true whether the applications in question execute on different clients or reside on the same client.¶
Share reservations and byte-range locks are the facilities the NFSv4.1 protocol provides to allow applications to coordinate access by using mutual exclusion facilities. The NFSv4.1 protocol's data caching must be implemented such that it does not invalidate the assumptions on which those using these facilities depend.¶
15.3.1. Data Caching and OPENs
In order to avoid invalidating the sharing assumptions on which applications rely, NFSv4.1 clients should not provide cached data to applications or modify it on behalf of an application when it would not be valid to obtain or modify that same data via a READ or WRITE operation.¶
Furthermore, in the absence of an OPEN delegation (see Section 15.4), two additional rules apply. Note that these rules are obeyed in practice by many NFSv3 clients.¶
-
First, cached data present on a client must be revalidated after doing an OPEN. Revalidating means that the client fetches the change attribute from the server, compares it with the cached change attribute, and if different, declares the cached data (as well as the cached attributes) as invalid. This is to ensure that the data for the OPENed file is still correctly reflected in the client's cache. This validation must be done at least when the client's OPEN operation includes a deny of OPEN4_SHARE_DENY_WRITE or OPEN4_SHARE_DENY_BOTH, thus terminating a period in which other clients may have had the opportunity to open the file with OPEN4_SHARE_ACCESS_WRITE/OPEN4_SHARE_ACCESS_BOTH access. Clients may choose to do the revalidation more often (i.e., at OPENs specifying a deny mode of OPEN4_SHARE_DENY_NONE) to parallel the NFSv3 protocol's practice for the benefit of users assuming this degree of cache revalidation.¶
Since the change attribute is updated for data and metadata modifications, some client implementers may be tempted to use the time_modify attribute and not the change attribute to validate cached data, so that metadata changes do not spuriously invalidate clean data. The implementer is cautioned in this approach. The change attribute is guaranteed to change for each update to the file, whereas time_modify is guaranteed to change only at the granularity of the time_delta attribute. Use by the client's data cache validation logic of time_modify and not change runs the risk of the client incorrectly marking stale data as valid. Thus, any cache validation approach by the client MUST include the use of the change attribute.¶
- Second, modified data must be flushed to the server before closing a file OPENed for OPEN4_SHARE_ACCESS_WRITE. This is complementary to the first rule. If the data is not flushed at CLOSE, the revalidation done after the client OPENs a file is unable to achieve its purpose. The other aspect to flushing the data before close is that the data must be committed to stable storage, at the server, before the CLOSE operation is requested by the client. In the case of a server restart and a CLOSEd file, it may not be possible to retransmit the data to be written to the file, hence, this requirement.¶
15.3.2. Data Caching and File Locking
For those applications that choose to use byte-range locking instead of share reservations to exclude inconsistent file access, there is an analogous set of constraints that apply to client-side data caching. These rules are effective only if the byte-range locking is used in a way that matches in an equivalent way the actual READ and WRITE operations executed. This is as opposed to byte-range locking that is based on pure convention. For example, it is possible to manipulate a two-megabyte file by dividing the file into two one-megabyte ranges and protecting access to the two byte-ranges by byte-range locks on bytes zero and one. A WRITE_LT lock on byte zero of the file would represent the right to perform READ and WRITE operations on the first byte-range. A WRITE_LT lock on byte one of the file would represent the right to perform READ and WRITE operations on the second byte-range. As long as all applications manipulating the file obey this convention, they will work on a local file system. However, they may not work with the NFSv4.1 protocol unless clients refrain from data caching.¶
The rules for data caching in the byte-range locking environment are:¶
- First, when a client obtains a byte-range lock for a particular byte-range, the data cache corresponding to that byte-range (if any cache data exists) must be revalidated. If the change attribute indicates that the file may have been updated since the cached data was obtained, the client must flush or invalidate the cached data for the newly locked byte-range. A client might choose to invalidate all of the non-modified cached data that it has for the file, but the only requirement for correct operation is to invalidate all of the data in the newly locked byte-range.¶
- Second, before releasing a WRITE_LT lock for a byte-range, all modified data for that byte-range must be flushed to the server. The modified data must also be written to stable storage.¶
Note that flushing data to the server and the invalidation of cached data must reflect the actual byte-ranges locked or unlocked. Rounding these up or down to reflect client cache block boundaries will cause problems if not carefully done. For example, writing a modified block when only half of that block is within an area being unlocked may cause invalid modification to the byte-range outside the unlocked area. This, in turn, may be part of a byte-range locked by another client. Clients can avoid this situation by synchronously performing portions of WRITE operations that overlap that portion (initial or final) that is not a full block. Similarly, invalidating a locked area that is not an integral number of full buffer blocks would require the client to read one or two partial blocks from the server if the revalidation procedure shows that the data that the client possesses may not be valid.¶
The data that is written to the server as a prerequisite to the unlocking of a byte-range must be written, at the server, to stable storage. The client may accomplish this either with synchronous writes or by following asynchronous writes with a COMMIT operation. This is required because retransmission of the modified data after a server restart might conflict with a lock held by another client.¶
A client implementation may choose to accommodate applications that use byte-range locking in non-standard ways (e.g., using a byte-range lock as a global semaphore) by flushing to the server more data upon a LOCKU than is covered by the locked range. This may include modified data within files other than the one for which the unlocks are being done. In such cases, the client must not interfere with applications whose READs and WRITEs are being done only within the bounds of byte-range locks that the application holds. For example, an application locks a single byte of a file and proceeds to write that single byte. A client that chose to handle a LOCKU by flushing all modified data to the server could validly write that single byte in response to an unrelated LOCKU operation. However, it would not be valid to write the entire block in which that single written byte was located since it includes an area that is not locked and might be locked by another client. Client implementations can avoid this problem by dividing files with modified data into those for which all modifications are done to areas covered by an appropriate byte-range lock and those for which there are modifications not covered by a byte-range lock. Any writes done for the former class of files must not include areas not locked and thus not modified on the client.¶
15.3.3. Data Caching and Mandatory File Locking
Client-side data caching needs to respect mandatory byte-range locking when it is in effect. The presence of mandatory byte-range locking for a given file is indicated when the client gets back NFS4ERR_LOCKED from a READ or WRITE operation on a file for which it has an appropriate share reservation. When mandatory locking is in effect for a file, the client must check for an appropriate byte-range lock for data being read or written. If a byte-range lock exists for the range being read or written, the client may satisfy the request using the client's validated cache. If an appropriate byte-range lock is not held for the range of the read or write, the read or write request must not be satisfied by the client's cache and the request must be sent to the server for processing. When a read or write request partially overlaps a locked byte-range, the request should be subdivided into multiple pieces with each byte-range (locked or not) treated appropriately.¶
15.3.4. Data Caching and File Identity
When clients cache data, the file data needs to be organized according to the file system object to which the data belongs. For NFSv3 clients, the typical practice has been to assume for the purpose of caching that distinct filehandles represent distinct file system objects. The client then has the choice to organize and maintain the data cache on this basis.¶
In the NFSv4.1 protocol, there is now the possibility to have significant deviations from a "one filehandle per object" model because a filehandle may be constructed on the basis of the object's pathname. Therefore, clients need a reliable method to determine if two filehandles designate the same file system object. If clients were simply to assume that all distinct filehandles denote distinct objects and proceed to do data caching on this basis, caching inconsistencies would arise between the distinct client-side objects that mapped to the same server-side object.¶
By providing a method to differentiate filehandles, the NFSv4.1 protocol alleviates a potential functional regression in comparison with the NFSv3 protocol. Without this method, caching inconsistencies within the same client could occur, and this has not been present in previous versions of the NFS protocol. Note that it is possible to have such inconsistencies with applications executing on multiple clients, but that is not the issue being addressed here.¶
For the purposes of data caching, the following steps allow an NFSv4.1 client to determine whether two distinct filehandles denote the same server-side object:¶
- If GETATTR directed to two filehandles returns different values of the fsid attribute, then the filehandles represent distinct objects.¶
- If GETATTR for any file with an fsid that matches the fsid of the two filehandles in question returns a unique_handles attribute with a value of TRUE, then the two objects are distinct.¶
- If GETATTR directed to the two filehandles does not return the fileid attribute for both of the handles, then it cannot be determined whether the two objects are the same. Therefore, operations that depend on that knowledge (e.g., client-side data caching) cannot be done reliably. Note that if GETATTR does not return the fileid attribute for both filehandles, it will return it for neither of the filehandles, since the fsid for both filehandles is the same.¶
- If GETATTR directed to the two filehandles returns different values for the fileid attribute, then they are distinct objects.¶
- Otherwise, they are the same object.¶
15.4. Open Delegation
When a file is being OPENed, the server may delegate further handling of opens and closes for that file to the opening client. Any such delegation is recallable since the circumstances that allowed for the delegation are subject to change. In particular, if the server receives a conflicting OPEN from another client, the server must recall the delegation before deciding whether the OPEN from the other client may be granted. Making a delegation is up to the server, and clients should not assume that any particular OPEN either will or will not result in an OPEN delegation. The following is a typical set of conditions that servers might use in deciding whether an OPEN should be delegated:¶
- The client must be able to respond to the server's callback requests. If a backchannel has been established, the server will send a CB_COMPOUND request, containing a single operation, CB_SEQUENCE, for a test of backchannel availability.¶
- The client must have responded properly to previous recalls.¶
- There must be no current OPEN conflicting with the requested delegation.¶
- There should be no current delegation that conflicts with the delegation being requested.¶
- The probability of future conflicting open requests should be low based on the recent history of the file.¶
- The existence of any server-specific semantics of OPEN/CLOSE that would make the required handling incompatible with the prescribed handling that the delegated client would apply (see below).¶
There are two types of OPEN delegations: OPEN_DELEGATE_READ and OPEN_DELEGATE_WRITE. An OPEN_DELEGATE_READ delegation allows a client to handle, on its own, requests to open a file for reading that do not deny OPEN4_SHARE_ACCESS_READ access to others. Multiple OPEN_DELEGATE_READ delegations may be outstanding simultaneously and do not conflict. An OPEN_DELEGATE_WRITE delegation allows the client to handle, on its own, all opens. Only one OPEN_DELEGATE_WRITE delegation may exist for a given file at a given time, and it is inconsistent with any OPEN_DELEGATE_READ delegations.¶
When a client has either type of open delegation, it is assured that neither the contents, the attributes (with the exception of time_access), nor the names of any links to the file will change without its knowledge, so long as the delegation is held. When a client has an OPEN_DELEGATE_WRITE delegation, it may modify the file data locally since no other client will be accessing the file's data. The client holding an OPEN_DELEGATE_WRITE delegation may only locally affect file attributes that are intimately connected with the file data: size, change, time_access, time_metadata, and time_modify. All other attributes must be reflected on the server.¶
When a client has an OPEN delegation, it does not need to send OPENs or CLOSEs to the server. Instead, the client may update the appropriate status internally. For an OPEN_DELEGATE_READ delegation, opens that cannot be handled locally (opens that are for OPEN4_SHARE_ACCESS_WRITE/OPEN4_SHARE_ACCESS_BOTH or that deny OPEN4_SHARE_ACCESS_READ access) must be sent to the server.¶
When an OPEN delegation is made, the reply to the OPEN contains an OPEN delegation structure that specifies the following:¶
- the type of delegation (OPEN_DELEGATE_READ or OPEN_DELEGATE_WRITE).¶
- space limitation information to control flushing of data on close (OPEN_DELEGATE_WRITE delegation only; see Section 15.4.1)¶
- an nfsace4 specifying read and write permissions¶
- a stateid to represent the delegation¶
The delegation stateid is separate and distinct from the stateid for the OPEN proper. The standard stateid, unlike the delegation stateid, is associated with a particular lock-owner and will continue to be valid after the delegation is recalled and the file remains open.¶
When a request internal to the client is made to open a file and an OPEN delegation is in effect, it will be accepted or rejected solely on the basis of the following conditions. Any requirement for other checks to be made by the delegate should result in the OPEN delegation being denied so that the checks can be made by the server itself.¶
- The access and deny bits for the request and the file as described in Section 14.7.¶
- The read and write permissions as determined below.¶
The nfsace4 passed with delegation can be used to avoid frequent ACCESS calls. The permission check should be as follows:¶
- If the nfsace4 indicates that the open may be done, then it should be granted without reference to the server.¶
- If the nfsace4 indicates that the open may not be done, then an ACCESS request must be sent to the server to obtain the definitive answer.¶
The server may return an nfsace4 that is more restrictive than the actual ACL of the file. This includes an nfsace4 that specifies denial of all access. Note that some common practices such as mapping the traditional user "root" to the user "nobody" (see Section 11.13) may make it incorrect to return the actual ACL of the file in the delegation response.¶
The use of a delegation together with various other forms of caching creates the possibility that no server authentication and authorization will ever be performed for a given user since all of the user's requests might be satisfied locally. Where the client is depending on the server for authentication and authorization, the client should be sure authentication and authorization occurs for each user by use of the ACCESS operation. This should be the case even if an ACCESS operation would not be required otherwise. As mentioned before, the server may enforce frequent authentication by returning an nfsace4 denying all access with every OPEN delegation.¶
15.4.1. Open Delegation and Data Caching
An OPEN delegation allows much of the message overhead associated with the opening and closing files to be eliminated. An open when an OPEN delegation is in effect does not require that a validation message be sent to the server. The continued endurance of the "OPEN_DELEGATE_READ delegation" provides a guarantee that no OPEN for OPEN4_SHARE_ACCESS_WRITE/OPEN4_SHARE_ACCESS_BOTH, and thus no write, has occurred. Similarly, when closing a file opened for OPEN4_SHARE_ACCESS_WRITE/OPEN4_SHARE_ACCESS_BOTH and if an OPEN_DELEGATE_WRITE delegation is in effect, the data written does not have to be written to the server until the OPEN delegation is recalled. The continued endurance of the OPEN delegation provides a guarantee that no open, and thus no READ or WRITE, has been done by another client.¶
For the purposes of OPEN delegation, READs and WRITEs done without an OPEN are treated as the functional equivalents of a corresponding type of OPEN. Although a client SHOULD NOT use special stateids when an open exists, delegation handling on the server can use the client ID associated with the current session to determine if the operation has been done by the holder of the delegation (in which case, no recall is necessary) or by another client (in which case, the delegation must be recalled and I/O not proceed until the delegation is returned or revoked).¶
With delegations, a client is able to avoid writing data to the server when the CLOSE of a file is serviced. The file close system call is the usual point at which the client is notified of a lack of stable storage for the modified file data generated by the application. At the close, file data is written to the server and, through normal accounting, the server is able to determine if the available file system space for the data has been exceeded (i.e., the server returns NFS4ERR_NOSPC or NFS4ERR_DQUOT). This accounting includes quotas. The introduction of delegations requires that an alternative method be in place for the same type of communication to occur between client and server.¶
In the delegation response, the server provides either the limit of the size of the file or the number of modified blocks and associated block size. The server must ensure that the client will be able to write modified data to the server of a size equal to that provided in the original delegation. The server must make this assurance for all outstanding delegations. Therefore, the server must be careful in its management of available space for new or modified data, taking into account available file system space and any applicable quotas. The server can recall delegations as a result of managing the available file system space. The client should abide by the server's state space limits for delegations. If the client exceeds the stated limits for the delegation, the server's behavior is undefined.¶
Based on server conditions, quotas, or available file system space, the server may grant OPEN_DELEGATE_WRITE delegations with very restrictive space limitations. The limitations may be defined in a way that will always force modified data to be flushed to the server on close.¶
With respect to authentication, flushing modified data to the server after a CLOSE has occurred may be problematic. For example, the user of the application may have logged off the client, and unexpired authentication credentials may not be present. In this case, the client may need to take special care to ensure that local unexpired credentials will in fact be available. This may be accomplished by tracking the expiration time of credentials and flushing data well in advance of their expiration or by making private copies of credentials to assure their availability when needed.¶
15.4.2. Open Delegation and File Locks
When a client holds an OPEN_DELEGATE_WRITE delegation, lock operations are performed locally. This includes those required for mandatory byte-range locking. This can be done since the delegation implies that there can be no conflicting locks. Similarly, all of the revalidations that would normally be associated with obtaining locks and the flushing of data associated with the releasing of locks need not be done.¶
When a client holds an OPEN_DELEGATE_READ delegation, lock operations are not performed locally. All lock operations, including those requesting non-exclusive locks, are sent to the server for resolution.¶
15.4.3. Handling of CB_GETATTR
The server needs to employ special handling for a GETATTR where the target is a file that has an OPEN_DELEGATE_WRITE delegation in effect. The reason for this is that the client holding the OPEN_DELEGATE_WRITE delegation may have modified the data, and the server needs to reflect this change to the second client that submitted the GETATTR. Therefore, the client holding the OPEN_DELEGATE_WRITE delegation needs to be interrogated. The server will use the CB_GETATTR operation. The only attributes that the server can reliably query via CB_GETATTR are size and change.¶
Since CB_GETATTR is being used to satisfy another client's GETATTR request, the server only needs to know if the client holding the delegation has a modified version of the file. If the client's copy of the delegated file is not modified (data or size), the server can satisfy the second client's GETATTR request from the attributes stored locally at the server. If the file is modified, the server only needs to know about this modified state. If the server determines that the file is currently modified, it will respond to the second client's GETATTR as if the file had been modified locally at the server.¶
Since the form of the change attribute is determined by the server and is opaque to the client, the client and server need to agree on a method of communicating the modified state of the file. For the size attribute, the client will report its current view of the file size. For the change attribute, the handling is more involved.¶
For the client, the following steps will be taken when receiving an OPEN_DELEGATE_WRITE delegation:¶
- The value of the change attribute will be obtained from the server and cached. Let this value be represented by c.¶
- The client will create a value greater than c that will be used for communicating that modified data is held at the client. Let this value be represented by d.¶
- When the client is queried via CB_GETATTR for the change attribute, it checks to see if it holds modified data. If the file is modified, the value d is returned for the change attribute value. If this file is not currently modified, the client returns the value c for the change attribute.¶
For simplicity of implementation, the client MAY for each CB_GETATTR return the same value d. This is true even if, between successive CB_GETATTR operations, the client again modifies the file's data or metadata in its cache. The client can return the same value because the only requirement is that the client be able to indicate to the server that the client holds modified data. Therefore, the value of d may always be c + 1.¶
While the change attribute is opaque to the client in the sense that it has no idea what units of time, if any, the server is counting change with, it is not opaque in that the client has to treat it as an unsigned integer, and the server has to be able to see the results of the client's changes to that integer. Therefore, the server MUST encode the change attribute in network order when sending it to the client. The client MUST decode it from network order to its native order when receiving it, and the client MUST encode it in network order when sending it to the server. For this reason, change is defined as an unsigned integer rather than an opaque array of bytes.¶
For the server, the following steps will be taken when providing an OPEN_DELEGATE_WRITE delegation:¶
- Upon providing an OPEN_DELEGATE_WRITE delegation, the server will cache a copy of the change attribute in the data structure it uses to record the delegation. Let this value be represented by sc.¶
- When a second client sends a GETATTR operation on the same file to the server, the server obtains the change attribute from the first client. Let this value be cc.¶
- If the value cc is equal to sc, the file is not modified and the server returns the current values for change, time_metadata, and time_modify (for example) to the second client.¶
- If the value cc is NOT equal to sc, the file is currently modified at the first client and most likely will be modified at the server at a future time. The server then uses its current time to construct attribute values for time_metadata and time_modify. A new value of sc, which we will call nsc, is computed by the server, such that nsc >= sc + 1. The server then returns the constructed time_metadata, time_modify, and nsc values to the requester. The server replaces sc in the delegation record with nsc. To prevent the possibility of time_modify, time_metadata, and change from appearing to go backward (which would happen if the client holding the delegation fails to write its modified data to the server before the delegation is revoked or returned), the server SHOULD update the file's metadata record with the constructed attribute values. For reasons of reasonable performance, committing the constructed attribute values to stable storage is OPTIONAL.¶
As discussed earlier in this section, the client MAY return the same cc value on subsequent CB_GETATTR calls, even if the file was modified in the client's cache yet again between successive CB_GETATTR calls. Therefore, the server must assume that the file has been modified yet again, and MUST take care to ensure that the new nsc it constructs and returns is greater than the previous nsc it returned. An example implementation's delegation record would satisfy this mandate by including a boolean field (let us call it "modified") that is set to FALSE when the delegation is granted, and an sc value set at the time of grant to the change attribute value. The modified field would be set to TRUE the first time cc != sc, and would stay TRUE until the delegation is returned or revoked. The processing for constructing nsc, time_modify, and time_metadata would use this pseudo code:¶
if (!modified) { do CB_GETATTR for change and size; if (cc != sc) modified = TRUE; } else { do CB_GETATTR for size; } if (modified) { sc = sc + 1; time_modify = time_metadata = current_time; update sc, time_modify, time_metadata into file's metadata; }¶
This would return to the client (that sent GETATTR) the attributes it requested, but make sure size comes from what CB_GETATTR returned. The server would not update the file's metadata with the client's modified size.¶
In the case that the file attribute size is different than the server's current value, the server treats this as a modification regardless of the value of the change attribute retrieved via CB_GETATTR and responds to the second client as in the last step.¶
This methodology resolves issues of clock differences between client and server and other scenarios where the use of CB_GETATTR break down.¶
It should be noted that the server is under no obligation to use CB_GETATTR, and therefore the server MAY simply recall the delegation to avoid its use.¶
15.4.4. Recall of Open Delegation
The following events necessitate recall of an OPEN delegation:¶
- potentially conflicting OPEN request (or a READ or WRITE operation done with a special stateid)¶
- SETATTR sent by another client¶
- REMOVE request for the file¶
- RENAME request for the file as either the source or target of the RENAME¶
Whether a RENAME of a directory in the path leading to the file results in recall of an OPEN delegation depends on the semantics of the server's file system. If that file system denies such RENAMEs when a file is open, the recall must be performed to determine whether the file in question is, in fact, open.¶
In addition to the situations above, the server may choose to recall OPEN delegations at any time if resource constraints make it advisable to do so. Clients should always be prepared for the possibility of recall.¶
When a client receives a recall for an OPEN delegation, it needs to update state on the server before returning the delegation. These same updates must be done whenever a client chooses to return a delegation voluntarily. The following items of state need to be dealt with:¶
- If the file associated with the delegation is no longer open and no previous CLOSE operation has been sent to the server, a CLOSE operation must be sent to the server.¶
- If a file has other open references at the client, then OPEN operations must be sent to the server. The appropriate stateids will be provided by the server for subsequent use by the client since the delegation stateid will no longer be valid. These OPEN requests are done with the claim type of CLAIM_DELEGATE_CUR. This will allow the presentation of the delegation stateid so that the client can establish the appropriate rights to perform the OPEN. (See Section 23.16, which describes the OPEN operation, for details.)¶
- If there are granted byte-range locks, the corresponding LOCK operations need to be performed. This applies to the OPEN_DELEGATE_WRITE delegation case only.¶
- For an OPEN_DELEGATE_WRITE delegation, if at the time of recall the file is not open for OPEN4_SHARE_ACCESS_WRITE/OPEN4_SHARE_ACCESS_BOTH, all modified data for the file must be flushed to the server. If the delegation had not existed, the client would have done this data flush before the CLOSE operation.¶
- For an OPEN_DELEGATE_WRITE delegation when a file is still open at the time of recall, any modified data for the file needs to be flushed to the server.¶
- With the OPEN_DELEGATE_WRITE delegation in place, it is possible that the file was truncated during the duration of the delegation. For example, the truncation could have occurred as a result of an OPEN UNCHECKED with a size attribute value of zero. Therefore, if a truncation of the file has occurred and this operation has not been propagated to the server, the truncation must occur before any modified data is written to the server.¶
In the case of OPEN_DELEGATE_WRITE delegation, byte-range locking imposes some additional requirements. To precisely maintain the associated invariant, it is required to flush any modified data in any byte-range for which a WRITE_LT lock was released while the OPEN_DELEGATE_WRITE delegation was in effect. However, because the OPEN_DELEGATE_WRITE delegation implies no other locking by other clients, a simpler implementation is to flush all modified data for the file (as described just above) if any WRITE_LT lock has been released while the OPEN_DELEGATE_WRITE delegation was in effect.¶
An implementation need not wait until delegation recall (or the decision to voluntarily return a delegation) to perform any of the above actions, if implementation considerations (e.g., resource availability constraints) make that desirable. Generally, however, the fact that the actual OPEN state of the file may continue to change makes it not worthwhile to send information about opens and closes to the server, except as part of delegation return. An exception is when the client has no more internal opens of the file. In this case, sending a CLOSE is useful because it reduces resource utilization on the client and server. Regardless of the client's choices on scheduling these actions, all must be performed before the delegation is returned, including (when applicable) the close that corresponds to the OPEN that resulted in the delegation. These actions can be performed either in previous requests or in previous operations in the same COMPOUND request.¶
15.4.5. Clients That Fail to Honor Delegation Recalls
A client may fail to respond to a recall for various reasons, such as a failure of the backchannel from server to the client. The client may be unaware of a failure in the backchannel. This lack of awareness could result in the client finding out long after the failure that its delegation has been revoked, and another client has modified the data for which the client had a delegation. This is especially a problem for the client that held an OPEN_DELEGATE_WRITE delegation.¶
Status bits returned by SEQUENCE operations help to provide an alternate way of informing the client of issues regarding the status of the backchannel and of recalled delegations. When the backchannel is not available, the server returns the status bit SEQ4_STATUS_CB_PATH_DOWN on SEQUENCE operations. The client can react by attempting to re-establish the backchannel and by returning recallable objects if a backchannel cannot be successfully re-established.¶
Whether the backchannel is functioning or not, it may be that the recalled delegation is not returned. Note that the client's lease might still be renewed, even though the recalled delegation is not returned. In this situation, servers SHOULD revoke delegations that are not returned in a period of time equal to the lease period. This period of time should allow the client time to note the backchannel-down status and re-establish the backchannel.¶
When delegations are revoked, the server will return with the SEQ4_STATUS_RECALLABLE_STATE_REVOKED status bit set on subsequent SEQUENCE operations. The client should note this and then use TEST_STATEID to find which delegations have been revoked.¶
15.4.6. Delegation Revocation
At the point a delegation is revoked, if there are associated opens on the client, these opens may or may not be revoked. If no byte-range lock or open is granted that is inconsistent with the existing open, the stateid for the open may remain valid and be disconnected from the revoked delegation, just as would be the case if the delegation were returned.¶
For example, if an OPEN for OPEN4_SHARE_ACCESS_BOTH with a deny of OPEN4_SHARE_DENY_NONE is associated with the delegation, granting of another such OPEN to a different client will revoke the delegation but need not revoke the OPEN, since the two OPENs are consistent with each other. On the other hand, if an OPEN denying write access is granted, then the existing OPEN must be revoked.¶
When opens and/or locks are revoked, the applications holding these opens or locks need to be notified. This notification usually occurs by returning errors for READ/WRITE operations or when a close is attempted for the open file.¶
If no opens exist for the file at the point the delegation is revoked, then notification of the revocation is unnecessary. However, if there is modified data present at the client for the file, the user of the application should be notified. Unfortunately, it may not be possible to notify the user since active applications may not be present at the client. See Section 15.5.1 for additional details.¶
15.4.7. Delegations via WANT_DELEGATION
In addition to providing delegations as part of the reply to OPEN operations, servers MAY provide delegations separate from open, via the OPTIONAL WANT_DELEGATION operation. This allows delegations to be obtained in advance of an OPEN that might benefit from them, for objects that are not a valid target of OPEN, or to deal with cases in which a delegation has been recalled and the client wants to make an attempt to re-establish it if the absence of use by other clients allows that.¶
The WANT_DELEGATION operation may be performed on any type of file object other than a directory.¶
When a delegation is obtained using WANT_DELEGATION, any open files for the same filehandle held by that client are to be treated as subordinate to the delegation, just as if they had been created using an OPEN of type CLAIM_DELEGATE_CUR. They are otherwise unchanged as to seqid, access and deny modes, and the relationship with byte-range locks. Similarly, because existing byte-range locks are subordinate to an open, those byte-range locks also become indirectly subordinate to that new delegation.¶
The WANT_DELEGATION operation provides for delivery of delegations via callbacks, when the delegations are not immediately available. When a requested delegation is available, it is delivered to the client via a CB_PUSH_DELEG operation. When this happens, open files for the same filehandle become subordinate to the new delegation at the point at which the delegation is delivered, just as if they had been created using an OPEN of type CLAIM_DELEGATE_CUR. Similarly, this occurs for existing byte-range locks subordinate to an open.¶
15.5. Data Caching and Revocation
When locks and delegations are revoked, the assumptions upon which successful caching depends are no longer guaranteed. For any locks or share reservations that have been revoked, the corresponding state-owner needs to be notified. This notification includes applications with a file open that has a corresponding delegation that has been revoked. Cached data associated with the revocation must be removed from the client. In the case of modified data existing in the client's cache, that data must be removed from the client without being written to the server. As mentioned, the assumptions made by the client are no longer valid at the point when a lock or delegation has been revoked. For example, another client may have been granted a conflicting byte-range lock after the revocation of the byte-range lock at the first client. Therefore, the data within the lock range may have been modified by the other client. Obviously, the first client is unable to guarantee to the application what has occurred to the file in the case of revocation.¶
Notification to a state-owner will in many cases consist of simply returning an error on the next and all subsequent READs/WRITEs to the open file or on the close. Where the methods available to a client make such notification impossible because errors for certain operations may not be returned, more drastic action such as signals or process termination may be appropriate. The justification here is that an invariant on which an application depends may be violated. Depending on how errors are typically treated for the client-operating environment, further levels of notification including logging, console messages, and GUI pop-ups may be appropriate.¶
15.5.1. Revocation Recovery for Write Open Delegation
Revocation recovery for an OPEN_DELEGATE_WRITE delegation poses the special issue of modified data in the client cache while the file is not open. In this situation, any client that does not flush modified data to the server on each close must ensure that the user receives appropriate notification of the failure as a result of the revocation. Since such situations may require human action to correct problems, notification schemes in which the appropriate user or administrator is notified may be necessary. Logging and console messages are typical examples.¶
If there is modified data on the client, it must not be flushed normally to the server. A client may attempt to provide a copy of the file data as modified during the delegation under a different name in the file system namespace to ease recovery. Note that when the client can determine that the file has not been modified by any other client, or when the client has a complete cached copy of the file in question, such a saved copy of the client's view of the file may be of particular value for recovery. In another case, recovery using a copy of the file based partially on the client's cached data and partially on the server's copy as modified by other clients will be anything but straightforward, so clients may avoid saving file contents in these situations or specially mark the results to warn users of possible problems.¶
Saving of such modified data in delegation revocation situations may be limited to files of a certain size or might be used only when sufficient disk space is available within the target file system. Such saving may also be restricted to situations when the client has sufficient buffering resources to keep the cached copy available until it is properly stored to the target file system.¶
15.6. Attribute Caching
This section pertains to the caching of a file's attributes on a client when that client does not hold a delegation on the file.¶
The attributes discussed in this section do not include named attributes. Individual named attributes are analogous to files, and caching of the data for these needs to be handled just as data caching is for ordinary files. Similarly, LOOKUP results from an OPENATTR directory (as well as the directory's contents) are to be cached on the same basis as any other pathnames.¶
Clients may cache file attributes obtained from the server and use them to avoid subsequent GETATTR requests. Such caching is write through in that modification to file attributes is always done by means of requests to the server and should not be done locally and should not be cached. The exception to this are modifications to attributes that are intimately connected with data caching. Therefore, extending a file by writing data to the local data cache is reflected immediately in the size as seen on the client without this change being immediately reflected on the server. Normally, such changes are not propagated directly to the server, but when the modified data is flushed to the server, analogous attribute changes are made on the server. When OPEN delegation is in effect, the modified attributes may be returned to the server in reaction to a CB_RECALL call.¶
The result of local caching of attributes is that the attribute caches maintained on individual clients will not be coherent. Changes made in one order on the server may be seen in a different order on one client and in a third order on another client.¶
The typical file system application programming interfaces do not provide means to atomically modify or interrogate attributes for multiple files at the same time. The following rules provide an environment where the potential incoherencies mentioned above can be reasonably managed. These rules are derived from the practice of previous NFS protocols.¶
- All attributes for a given file (per-fsid attributes excepted) are cached as a unit at the client so that no non-serializability can arise within the context of a single file.¶
- An upper time boundary is maintained on how long a client cache entry can be kept without being refreshed from the server.¶
- When operations are performed that change attributes at the server, the updated attribute set is requested as part of the containing RPC. This includes directory operations that update attributes indirectly. This is accomplished by following the modifying operation with a GETATTR operation and then using the results of the GETATTR to update the client's cached attributes.¶
Note that if the full set of attributes to be cached is requested by READDIR, the results can be cached by the client on the same basis as attributes obtained via GETATTR.¶
A client may validate its cached version of attributes for a file by fetching both the change and time_access attributes and assuming that if the change attribute has the same value as it did when the attributes were cached, then no attributes other than time_access have changed. The reason why time_access is also fetched is because many servers operate in environments where the operation that updates change does not update time_access. For example, POSIX file semantics do not update access time when a file is modified by the write system call [write_atime]. Therefore, the client that wants a current time_access value should fetch it with change during the attribute cache validation processing and update its cached time_access.¶
The client may maintain a cache of modified attributes for those attributes intimately connected with data of modified regular files (size, time_modify, and change). Other than those three attributes, the client MUST NOT maintain a cache of modified attributes. Instead, attribute changes are immediately sent to the server.¶
In some operating environments, the equivalent to time_access is expected to be implicitly updated by each read of the content of the file object. If an NFS client is caching the content of a file object, whether it is a regular file, directory, or symbolic link, the client SHOULD NOT update the time_access attribute (via SETATTR or a small READ or READDIR request) on the server with each read that is satisfied from cache. The reason is that this can defeat the performance benefits of caching content, especially since an explicit SETATTR of time_access may alter the change attribute on the server. If the change attribute changes, clients that are caching the content will think the content has changed, and will re-read unmodified data from the server. Nor is the client encouraged to maintain a modified version of time_access in its cache, since the client either would eventually have to write the access time to the server with bad performance effects or never update the server's time_access, thereby resulting in a situation where an application that caches access time between a close and open of the same file observes the access time oscillating between the past and present. The time_access attribute always means the time of last access to a file by a read that was satisfied by the server. This way clients will tend to see only time_access changes that go forward in time.¶
15.7. Data and Metadata Caching and Memory Mapped Files
Some operating environments include the capability for an application to map a file's content into the application's address space. Each time the application accesses a memory location that corresponds to a block that has not been loaded into the address space, a page fault occurs and the file is read (or if the block does not exist in the file, the block is allocated and then instantiated in the application's address space).¶
As long as each memory-mapped access to the file requires a page fault, the relevant attributes of the file that are used to detect access and modification (time_access, time_metadata, time_modify, and change) will be updated. However, in many operating environments, when page faults are not required, these attributes will not be updated on reads or updates to the file via memory access (regardless of whether the file is local or is accessed remotely). A client or server MAY fail to update attributes of a file that is being accessed via memory-mapped I/O. This has several implications:¶
- If there is an application on the server that has memory mapped a file that a client is also accessing, the client may not be able to get a consistent value of the change attribute to determine whether or not its cache is stale. A server that knows that the file is memory-mapped could always pessimistically return updated values for change so as to force the application to always get the most up-to-date data and metadata for the file. However, due to the negative performance implications of this, such behavior is OPTIONAL.¶
- If the memory-mapped file is not being modified on the server, and instead is just being read by an application via the memory-mapped interface, the client will not see an updated time_access attribute. However, in many operating environments, neither will any process running on the server. Thus, NFS clients are at no disadvantage with respect to local processes.¶
-
If there is another client that is memory mapping the file, and if that client is holding an OPEN_DELEGATE_WRITE delegation, the same set of issues as discussed in the previous two bullet points apply. However, it should be noted that it is very unlikely that such a delegation will be held since it is normally required that the file be open for read to be mapped into memory. Only if the file were not open and accessed using a special stateid could the delegation be retained while the file in question is mapped into another client's memory. For this reason, such use is highly undesirable.¶
In this situation, when a server does a CB_GETATTR to a file that the client has modified in its cache, the reply from CB_GETATTR would not necessarily be accurate, assuming the delegation is not recalled at this point. As discussed earlier, the client's obligation is to report that the file has been modified since the delegation was granted, not whether it has been modified again between successive CB_GETATTR calls, and the server MUST assume that any file the client has modified in cache has been modified again between successive CB_GETATTR calls. Depending on the nature of the client's memory management system, it might not be possible to live up to this weak obligation. A client MAY return stale information in CB_GETATTR whenever the file is memory-mapped, if another client is accessing the file without opening it.¶
15.8. Name and Directory Caching without Directory Delegations
The NFSv4.1 directory delegation facility (described in Section 15.9 below) is OPTIONAL for servers to implement. Even where it is implemented, it may not always be functional because of resource availability issues or other constraints. Thus, it is important to understand how name and directory caching are done in the absence of directory delegations. These topics are discussed in the next two subsections.¶
15.8.1. Name Caching
The results of LOOKUP and READDIR operations may be cached to avoid the cost of subsequent LOOKUP operations. Just as in the case of attribute caching, inconsistencies may arise among the various client caches. To mitigate the effects of these inconsistencies and given the context of typical file system APIs, an upper time boundary is maintained for how long a client name cache entry can be kept without verifying that the entry has not been made invalid by a directory change operation performed by another client.¶
When a client is not making changes to a directory for which there exist name cache entries, the client needs to periodically fetch attributes for that directory to ensure that it is not being modified. After determining that no modification has occurred, the expiration time for the associated name cache entries may be updated to be the current time plus the name cache staleness bound.¶
When a client is making changes to a given directory, it needs to determine whether there have been changes made to the directory by other clients. It does this by using the change attribute as reported before and after the directory operation in the associated change_info4 value returned for the operation. The server is able to communicate to the client whether the change_info4 data is provided atomically with respect to the directory operation. If the change values are provided atomically, the client has a basis for determining, given proper care, whether other clients are modifying the directory in question.¶
The simplest way to enable the client to make this determination is for the client to serialize all changes made to a specific directory. When this is done, and the server provides before and after values of the change attribute atomically, the client can simply compare the after value of the change attribute from one operation on a directory with the before value on the subsequent operation modifying that directory. When these are equal, the client is assured that no other client is modifying the directory in question.¶
When such serialization is not used, and there may be multiple simultaneous outstanding operations modifying a single directory sent from a single client, making this sort of determination can be more complicated. If two such operations complete in a different order than they were actually performed, that might give an appearance consistent with modification being made by another client. Where this appears to happen, the client needs to await the completion of all such modifications that were started previously, to see if the outstanding before and after change numbers can be sorted into a chain such that the before value of one change number matches the after value of a previous one, in a chain consistent with this client being the only one modifying the directory.¶
In either of these cases, the client is able to determine whether the directory is being modified by another client. If the comparison indicates that the directory was updated by another client, the name cache associated with the modified directory is purged from the client. If the comparison indicates no modification, the name cache can be updated on the client to reflect the directory operation and the associated timeout can be extended. The post-operation change value needs to be saved as the basis for future change_info4 comparisons.¶
As demonstrated by the scenario above, name caching requires that the client revalidate name cache data by inspecting the change attribute of a directory at the point when the name cache item was cached. This requires that the server update the change attribute for directories when the contents of the corresponding directory is modified. For a client to use the change_info4 information appropriately and correctly, the server must report the pre- and post-operation change attribute values atomically. When the server is unable to report the before and after values atomically with respect to the directory operation, the server must indicate that fact in the change_info4 return value. When the information is not atomically reported, the client should not assume that other clients have not changed the directory.¶
15.8.2. Directory Caching
The results of READDIR operations may be used to avoid subsequent READDIR operations. Just as in the cases of attribute and name caching, inconsistencies may arise among the various client caches. To mitigate the effects of these inconsistencies, and given the context of typical file system APIs, the following rules should be followed:¶
- Cached READDIR information for a directory that is not obtained in a single READDIR operation must always be a consistent snapshot of directory contents. This is determined by using a GETATTR before the first READDIR and after the last READDIR that contributes to the cache.¶
- An upper time boundary is maintained to indicate the length of time a directory cache entry is considered valid before the client must revalidate the cached information.¶
The revalidation technique parallels that discussed in the case of name caching. When the client is not changing the directory in question, checking the change attribute of the directory with GETATTR is adequate. The lifetime of the cache entry can be extended at these checkpoints. When a client is modifying the directory, the client needs to use the change_info4 data to determine whether there are other clients modifying the directory. If it is determined that no other client modifications are occurring, the client may update its directory cache to reflect its own changes.¶
As demonstrated previously, directory caching requires that the client revalidate directory cache data by inspecting the change attribute of a directory at the point when the directory was cached. This requires that the server update the change attribute for directories when the contents of the corresponding directory is modified. For a client to use the change_info4 information appropriately and correctly, the server must report the pre- and post-operation change attribute values atomically. When the server is unable to report the before and after values atomically with respect to the directory operation, the server must indicate that fact in the change_info4 return value. When the information is not atomically reported, the client should not assume that other clients have not changed the directory.¶
15.9. Directory Delegations and Notifications
15.9.1. Motivation for Directory Delegations
Directory caching for the NFSv4.1 protocol when directory delegations are not available, is similar to file and directory caching in previous versions. Clients typically cache directory information for a duration determined by the client. At the end of that predefined period, the client will query the server to see if the directory has been updated. By caching attributes, clients reduce the number of GETATTR calls made to the server to validate attributes. As a result, frequently accessed files and directories, such as the current working directory, have their attributes cached on the client so that some NFS operations can be performed without making an RPC call. By caching name and attributes information about most recently looked up entries in a Directory Name Lookup Cache (DNLC), clients are able to avoid sending LOOKUP/GETATTR calls to the server every time such files are accessed.¶
This caching approach works reasonably well at reducing network traffic in many environments. However, it does not address environments where there are numerous queries for files that do not exist. In these cases of "misses", the client sends requests to the server in order to provide reasonable application semantics and promptly detect the creation of new directory entries. Examples of high miss activity are compilation in software development environments. The current behavior of NFS limits its potential scalability and wide-area sharing effectiveness in these types of environments.¶
Since, other distributed stateful file system architectures such as AFS and DFS have proven that adding state around directory contents can greatly reduce network traffic in high-miss environments, it is sensible to define and implement such facilities in NFsv4.1.¶
15.9.2. Directory Caching Features
Delegation of directory contents is an OPTIONAL feature of NFSv4.1. Possession of a delegation can be taken advantage of in a number of ways:¶
-
It can be used to provide a recallable assurance that the directory contents have not changed, allowing LOOKUP results (whether successful or not) and READDIR results to be cached, in order to enable these operations to be performed locally.¶
This mode of operation in which directory contents are fixed is often referred to as the "pure recall" model since any change in the directory contents results in the delegation being recalled. This mode of operation is most effectively used on large directories which are infrequently changed.¶
-
The client can request, as part of requesting a delegation, that notifications be provided to update the clients view of the directory contents to match that of the server. See Section 15.9.7 for details. This mode of operation allows directory delegations to be effectively used in handling large directories that experience a significant stream of updates.¶
-
Independently of the mode of operation selected, notifications to inform the client of attribute changes can be requested. See Section 15.9.7 for details.¶
15.9.3. Directory Delegation Mechanics
The GET_DIR_DELEGATION (Section 23.39) operation is used by clients to request directory delegation. The delegation is read-only and the client is not provided any means to make changes to the directory other than by performing NFSv4.1 operations that modify the directory.¶
As part of obtaining a delegation, the client specifies, using the bit numbers within the notify_type4 enum that appears below, its choices regarding notification of events related to the reporting of events affecting the delegation. Some, although not all, directly specify the use of particular notification types, to be used to inform the client of events that could otherwise result in recall of the delegation.¶
It is important to note that this enum is subject to extension and has been extended relative to the set of bits defined in [RFC8881]. The distinction between bits that were defined earlier and those added later is important to enable interoperation between clients and servers when one might have been written based on the earlier specification. Although no implementations based on the earlier specification are known, the possibility of their existence cannot be excluded.¶
/* * Directory notification types and associated flags */ enum notify_type4 { /* * Present in RFCs 5661, 8881 */ NOTIFY4_CHANGE_CHILD_ATTRS = 0, NOTIFY4_CHANGE_DIR_ATTRS = 1, NOTIFY4_REMOVE_ENTRY = 2, NOTIFY4_ADD_ENTRY = 3, NOTIFY4_RENAME_ENTRY = 4, NOTIFY4_CHANGE_COOKIE_VERIFIER = 5, /* * Added in NFSv4.1 bis document */ NOTIFY4_GFLAG_EXTEND = 6, NOTIFY4_AUFLAG_VALID = 7, NOTIFY4_AUFLAG_USER = 8, NOTIFY4_AUFLAG_GROUP = 9, NOTIFY4_AUFLAG_OTHER = 10, NOTIFY4_CHANGE_AUTH = 11, NOTIFY4_CFLAG_ORDER = 12, NOTIFY4_AUFLAG_GANOW = 13, NOTIFY4_AUFLAG_GALATER = 14, NOTIFY4_CHANGE_GA = 15, NOTIFY4_CHANGE_AMASK = 16 };¶
Of the newer bits, only NOTIFY4_GFLAG_EXTEND, NOTIFY4_CHANGE_AUTH, NOTIFY4_CFLAG_ORDER, NOTIFY4_CHANGE_GA, and NOTIFY4_CHANGE_AMASK can appear when requesting a delegation. When any of these are set the server, it is possible that the server is unaware of their existence and will ignore them. If the client sets NOTIFY4_GFLAG_EXTEND in the request and it is returned set in the response, the client and server can interact assuming that each is aware of the newer bits. For more details about dealing with possibility of implementations of multiple versions of this feature interacting, see Section 15.9.6.¶
Of these bits the following subsets should be noted:¶
-
The bits NOTIFY4_CHANGE_CHILD_ATTRS , NOTIFY4_CHANGE_DIR_ATTRS, NOTIFY4_REMOVE_ENTRY, NOTIFY4_ADD_ENTRY, NOTIFY4_RENAME_ENTRY, and NOTIFY4_CHANGE_COOKIE_VERIFIER all have associated notification messages and were defined in [RFC8881]¶
When these bits are set when requesting a delegation, the server is being notified of the client's desire to have the corresponding notification sent rather than recalling the delegation. when the server sets these bits in the response, it is indicating its agreement to provide these notifications.¶
-
The bits NOTIFY4_CHANGE_AUTH, NOTIFY4_CHANGE_GA, and NOTIFY4_CHANGE_AMASK also have associated notification messages.¶
These notifications can be requested as in the case above. However, it is possible that the server is unaware of their existence.¶
-
The bit NOTIFY4_GFLAG_EXTEND denotes a flags to be exchanged as part of requesting a delegation.¶
-
The bits NOTIFY4_CFLAG_ORDER denotes a flag that can be set as part of requesting a delegation but has no role in requests.¶
-
The bits NOTIFY4_AUFLAG_VALID, NOTIFY4_AUFLAG_USER, NOTIFY4_AUFLAG_GROUP, NOTIFY4_AUFLAG_OTHER, NOTIFY4_AUFLAG_GANOW and NOTIFY4_AUFLAG_GALATER, can be set in the response but have no role in requests.¶
Of these bits only NOTIFY4_GFLAG_EXTEND is of general applicability and applies to multiple functions discussed in the subsections below. The other are discussed in more detail as grouped below:¶
-
The bits NOTIFY4_REMOVE_ENTRY, NOTIFY4_ADD_ENTRY, NOTIFY4_RENAME_ENTRY, NOTIFY4_CHANGE_COOKIE_VERIFIER, and NOTIFY4_CFLAG_ORDER concern the maintenance of cached directory contents and are discussed in Section 15.9.7¶
-
The bits NOTIFY4_CHANGE_CHILD_ATTRS, NOTIFY4_CHANGE_DIR_ATTRS, and NOTIFY4_CHANGE_AMASK concern the maintenance of cached file object attributes and are discussed in Section 15.9.8¶
-
The bits NOTIFY4_AUFLAG_VALID, NOTIFY4_AUFLAG_USER, NOTIFY4_AUFLAG_GROUP, NOTIFY4_AUFLAG_OTHER, NOTIFY4_CHANGE_AUTH, NOTIFY4_AUFLAG_GANOW, NOTIFY4_AUFLAG_GALATER, NOTIFY4_CHANGE_GA concern the management of authorizations for the cached use of file contents and file attributes are discussed in Section 15.9.9.¶
The holder is assured of certain thing not being changed while the directory is held, as described below.¶
-
That the set of entries within the directory not be changed without sending a requested notification to the client, informing the client of the change.¶
-
That the order of directory entries or the cookie values associated with specific directory entry with the client being informed (via a NOTIFY4_CHANGE_COOKIE_VERIFIER notification) of the possibility of change.¶
Delegations can be recalled by the server at any time and are always recalled before a directory is removed.¶
15.9.4. Directory Delegation Authorization Requirements
When cached data is used locally in place of LOOKUP, GETATTR, or, READDIR operations, the authorization constraints that would normally by imposed by the server have to be applied by the client. The discussion is complicated by the fact that, while facilities have been designed to accomplish that are described in this document, the treatment in earlier specifications did not provide facilities to help the client do this correctly and had little to say on the issue. As a result, clients were faced with the choice of ignoring difficult authorization issues or burdening the implementation with authorization checking that would undercut the performance benefits of the feature.¶
As a result, we are faced with the issue of how to accommodate implementations that are now known to have troubling problems that were not recognized when the feature was first described in a Proposed Standard. Normally, one tries to accommodate such situations by recommending against approaches now known to be flawed while considering, as a valid reason to bypass the recommendation, the reliance of the implementer on an approved Proposed Standard at the time. In this case we have a different approach, because of the following distinctive factors:¶
-
Unlike the case of implementers being told that use of AUTH_SYS in the clear, is an "OPTIONAL means of authentication" with the implication that such use does not result in potentially unacceptable security vulnerabilities, here there is no direct suggestion that neglecting these difficulties is acceptable. Instead, while lack of attention to security issues might have led people astray, they were not specifically asked to adopt a flawed approach to security but chose to adopt one on their own. As a result, while we will make certain allowances to accommodate such early implementations, there is no known paradigm that could be cited as valid but discouraged.¶
-
The set of implementations involved is likely to be quite small and might be empty or only consist of experimental implementations not widely distributed.¶
The approach we take here is the same one we take to servers that do not support the extensions described in Section 15.9.9 and to clients that interact with such servers. It has the following elements:¶
-
Clients are free in deciding whether to use directory delegations to take account of the problems with earlier approaches in deciding whether to use this feature.¶
-
Neglecting the possibility of authorization failure on GETATTR when directory entry attributes are cached is not to be considered disabling. This includes situation in which the server supports the ACE mask bit ACE4_READ_ATTRIBUTES.¶
-
Use of directory attributes for the clients to do its own authorization needs to be discouraged for a number of reasons.¶
Prime among these is the possibility that the acl attribute might be set for the directory, making it impossible for the client to its own authorization checking.¶
-
Even in the case of a file system on which none of attributes acl, dacl , and sacl is supported, the use of client-side authorization is not justifiable, since the attributes can change subsequently and the potential delay for the update of directory attributes has no upper bounds.¶
-
When clients use ACCESS to do authorization checks, as they should, allowance needs to be made for them to cache positive results, since without that ability, you might as well fetch the data over the wire anyway.¶
In this discussion, we will consider how various pair of implementations have dealt and will deal with this issue using combination of server guarantees and the use over-the-wire ACCESS checks and the potential caching of these results. Some things to note about potential implementations based on earlier specifications:¶
-
It is unlikely that the unfortunate effects of authorization failure were considered at all. Since the issue was introduced by the inclusion of very-rarely-implemented ACE mask bit ACE4_READ_ATTRIBTES, it is likely that this issue was imply ignored.¶
-
It is reasonable to suppose that clients were expected to request authorization checks using ACCESS and that clients were prepared to cache these determinations.¶
-
It is likely that server-based guarantees were never provided.¶
In the current description, the protocol has been extended to address these gaps. As a result when both client and server based on this description, the following apply.¶
-
The server provide a guarantee that GETATTRs can be done locally without concern for the possibility of denial or the need to perform action based on AUDIT or alarm ACEs, and that the client will be notified when the server becomes aware of circumstances making that guarantee inappropriate.¶
That guarantee is trivial to provide for the large set of servers that do not support the ACE mask bit ACE4_READ_ATTRIBUTES.¶
For servers that do support that mask bit, the server could provide the guarantee by a scan of the directory for files with troublesome ACLs. However because of the performance effects of requiring that scan to grant a delegation, the sever is allowed to delay that guarantee until after the delegation is granted.¶
-
Authorization for LOOKUP and READDIR is fundamentally the responsibility of the client to ascertain using ACCESS calls.¶
To reduce the burden of those calls, the server is expected to provide information about various classes of users for which such authorization check are unnecessary.¶
To further reduce the burden of those calls, the result of authorization checks can be cached until the server notifies the need to clear caching for those classes due to a change in authorization-related attributes.¶
The server provides notifications when there are changes in the groups of users for which authorization checks are needed.¶
- ¶
If we expand the discussion to apply to all implementations including potential client and server implementations written based on earlier specifications, the following constraints apply:¶
-
Servers SHOULD provide the services described above.¶
In this context, the only valid reason to bypass the recommendation is the implementer's reliance on an earlier specification in which such authorization-checking assistance was not provided for. This includes cases in which the planning for the implementation was based on an earlier specification.¶
-
Clients SHOULD use these facilities when they are available.¶
In this context, the only valid reason to bypass the recommendation is the implementer's reliance on an earlier specification in which such authorization-checking assistance was not provided for.¶
-
When clients do not use these facilities, they MAY avoid use of directory delegations. However, if they choose to use this cached data they MUST do their own authorization checks, using ACCESS.¶