idnits 2.17.1 draft-ietf-nfsv4-minorversion2-31.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- -- The document has examples using IPv4 documentation addresses according to RFC6890, but does not use any IPv6 documentation addresses. Maybe there should be IPv6 examples, too? Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (March 03, 2015) is 3335 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) -- Looks like a reference, but probably isn't: '0' on line 3899 == Missing Reference: '32K' is mentioned on line 3899, but not defined -- Possible downref: Non-RFC (?) normative reference: ref. 'NFSv42xdr' ** Obsolete normative reference: RFC 5661 (Obsoleted by RFC 8881) == Outdated reference: A later version (-06) exists of draft-ietf-nfsv4-lfs-registry-01 -- Obsolete informational reference (is this intentional?): RFC 2401 (Obsoleted by RFC 4301) -- Obsolete informational reference (is this intentional?): RFC 2616 (Obsoleted by RFC 7230, RFC 7231, RFC 7232, RFC 7233, RFC 7234, RFC 7235) Summary: 1 error (**), 0 flaws (~~), 3 warnings (==), 6 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 NFSv4 T. Haynes 3 Internet-Draft Primary Data 4 Intended status: Standards Track March 03, 2015 5 Expires: September 4, 2015 7 NFS Version 4 Minor Version 2 8 draft-ietf-nfsv4-minorversion2-31.txt 10 Abstract 12 This Internet-Draft describes NFS version 4 minor version two, 13 describing the protocol extensions made from NFS version 4 minor 14 version 1. Major extensions introduced in NFS version 4 minor 15 version two include: Server Side Copy, Application I/O Advise, Space 16 Reservations, Sparse Files, Application Data Blocks, and Labeled NFS. 18 Requirements Language 20 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 21 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 22 document are to be interpreted as described in RFC 2119 [RFC2119]. 24 Status of This Memo 26 This Internet-Draft is submitted in full conformance with the 27 provisions of BCP 78 and BCP 79. 29 Internet-Drafts are working documents of the Internet Engineering 30 Task Force (IETF). Note that other groups may also distribute 31 working documents as Internet-Drafts. The list of current Internet- 32 Drafts is at http://datatracker.ietf.org/drafts/current/. 34 Internet-Drafts are draft documents valid for a maximum of six months 35 and may be updated, replaced, or obsoleted by other documents at any 36 time. It is inappropriate to use Internet-Drafts as reference 37 material or to cite them other than as "work in progress." 39 This Internet-Draft will expire on September 4, 2015. 41 Copyright Notice 43 Copyright (c) 2015 IETF Trust and the persons identified as the 44 document authors. All rights reserved. 46 This document is subject to BCP 78 and the IETF Trust's Legal 47 Provisions Relating to IETF Documents 48 (http://trustee.ietf.org/license-info) in effect on the date of 49 publication of this document. Please review these documents 50 carefully, as they describe your rights and restrictions with respect 51 to this document. Code Components extracted from this document must 52 include Simplified BSD License text as described in Section 4.e of 53 the Trust Legal Provisions and are provided without warranty as 54 described in the Simplified BSD License. 56 Table of Contents 58 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 4 59 1.1. The NFS Version 4 Minor Version 2 Protocol . . . . . . . 4 60 1.2. Scope of This Document . . . . . . . . . . . . . . . . . 5 61 1.3. NFSv4.2 Goals . . . . . . . . . . . . . . . . . . . . . . 5 62 1.4. Overview of NFSv4.2 Features . . . . . . . . . . . . . . 5 63 1.4.1. Server Side Copy . . . . . . . . . . . . . . . . . . 5 64 1.4.2. Application I/O Advise . . . . . . . . . . . . . . . 6 65 1.4.3. Sparse Files . . . . . . . . . . . . . . . . . . . . 6 66 1.4.4. Space Reservation . . . . . . . . . . . . . . . . . . 6 67 1.4.5. Application Data Block (ADB) Support . . . . . . . . 6 68 1.4.6. Labeled NFS . . . . . . . . . . . . . . . . . . . . . 6 69 1.5. Enhancements to Minor Versioning Model . . . . . . . . . 7 70 2. Minor Versioning . . . . . . . . . . . . . . . . . . . . . . 7 71 3. pNFS considerations for New Operations . . . . . . . . . . . 8 72 3.1. Atomicity for ALLOCATE and DEALLOCATE . . . . . . . . . . 8 73 3.2. Sharing of stateids with NFSv4.1 . . . . . . . . . . . . 8 74 3.3. NFSv4.2 as a Storage Protocol in pNFS: the File Layout 75 Type . . . . . . . . . . . . . . . . . . . . . . . . . . 8 76 3.3.1. Operations Sent to NFSv4.2 Data Servers . . . . . . . 8 77 4. Server Side Copy . . . . . . . . . . . . . . . . . . . . . . 9 78 4.1. Introduction . . . . . . . . . . . . . . . . . . . . . . 9 79 4.2. Protocol Overview . . . . . . . . . . . . . . . . . . . . 9 80 4.2.1. Copy Operations . . . . . . . . . . . . . . . . . . . 10 81 4.2.2. Requirements for Operations . . . . . . . . . . . . . 10 82 4.3. Requirements for Inter-Server Copy . . . . . . . . . . . 11 83 4.4. Implementation Considerations . . . . . . . . . . . . . . 11 84 4.4.1. Locking the Files . . . . . . . . . . . . . . . . . . 11 85 4.4.2. Client Caches . . . . . . . . . . . . . . . . . . . . 12 86 4.5. Intra-Server Copy . . . . . . . . . . . . . . . . . . . . 12 87 4.6. Inter-Server Copy . . . . . . . . . . . . . . . . . . . . 13 88 4.7. Server-to-Server Copy Protocol . . . . . . . . . . . . . 17 89 4.7.1. Considerations on Selecting a Copy Protocol . . . . . 17 90 4.7.2. Using NFSv4.x as the Copy Protocol . . . . . . . . . 17 91 4.7.3. Using an Alternative Copy Protocol . . . . . . . . . 17 92 4.8. netloc4 - Network Locations . . . . . . . . . . . . . . . 18 93 4.9. Copy Offload Stateids . . . . . . . . . . . . . . . . . . 19 94 4.10. Security Considerations . . . . . . . . . . . . . . . . . 19 95 4.10.1. Inter-Server Copy Security . . . . . . . . . . . . . 19 96 5. Support for Application IO Hints . . . . . . . . . . . . . . 27 97 6. Sparse Files . . . . . . . . . . . . . . . . . . . . . . . . 28 98 6.1. Introduction . . . . . . . . . . . . . . . . . . . . . . 28 99 6.2. Terminology . . . . . . . . . . . . . . . . . . . . . . . 29 100 6.3. New Operations . . . . . . . . . . . . . . . . . . . . . 29 101 6.3.1. READ_PLUS . . . . . . . . . . . . . . . . . . . . . . 29 102 6.3.2. DEALLOCATE . . . . . . . . . . . . . . . . . . . . . 30 103 7. Space Reservation . . . . . . . . . . . . . . . . . . . . . . 30 104 8. Application Data Block Support . . . . . . . . . . . . . . . 32 105 8.1. Generic Framework . . . . . . . . . . . . . . . . . . . . 33 106 8.1.1. Data Block Representation . . . . . . . . . . . . . . 33 107 8.2. An Example of Detecting Corruption . . . . . . . . . . . 34 108 8.3. Example of READ_PLUS . . . . . . . . . . . . . . . . . . 35 109 8.4. An Example of Zeroing Space . . . . . . . . . . . . . . . 36 110 9. Labeled NFS . . . . . . . . . . . . . . . . . . . . . . . . . 36 111 9.1. Introduction . . . . . . . . . . . . . . . . . . . . . . 36 112 9.2. Definitions . . . . . . . . . . . . . . . . . . . . . . . 37 113 9.3. MAC Security Attribute . . . . . . . . . . . . . . . . . 38 114 9.3.1. Delegations . . . . . . . . . . . . . . . . . . . . . 38 115 9.3.2. Permission Checking . . . . . . . . . . . . . . . . . 39 116 9.3.3. Object Creation . . . . . . . . . . . . . . . . . . . 39 117 9.3.4. Existing Objects . . . . . . . . . . . . . . . . . . 39 118 9.3.5. Label Changes . . . . . . . . . . . . . . . . . . . . 39 119 9.4. pNFS Considerations . . . . . . . . . . . . . . . . . . . 40 120 9.5. Discovery of Server Labeled NFS Support . . . . . . . . . 40 121 9.6. MAC Security NFS Modes of Operation . . . . . . . . . . . 40 122 9.6.1. Full Mode . . . . . . . . . . . . . . . . . . . . . . 41 123 9.6.2. Guest Mode . . . . . . . . . . . . . . . . . . . . . 42 124 9.7. Security Considerations for Labeled NFS . . . . . . . . . 42 125 10. Sharing change attribute implementation details with NFSv4 126 clients . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 127 11. Error Values . . . . . . . . . . . . . . . . . . . . . . . . 43 128 11.1. Error Definitions . . . . . . . . . . . . . . . . . . . 44 129 11.1.1. General Errors . . . . . . . . . . . . . . . . . . . 44 130 11.1.2. Server to Server Copy Errors . . . . . . . . . . . . 44 131 11.1.3. Labeled NFS Errors . . . . . . . . . . . . . . . . . 45 132 11.2. New Operations and Their Valid Errors . . . . . . . . . 45 133 11.3. New Callback Operations and Their Valid Errors . . . . . 49 134 12. New File Attributes . . . . . . . . . . . . . . . . . . . . . 49 135 12.1. New RECOMMENDED Attributes - List and Definition 136 References . . . . . . . . . . . . . . . . . . . . . . . 49 137 12.2. Attribute Definitions . . . . . . . . . . . . . . . . . 50 138 13. Operations: REQUIRED, RECOMMENDED, or OPTIONAL . . . . . . . 53 139 14. Modifications to NFSv4.1 Operations . . . . . . . . . . . . . 56 140 14.1. Operation 42: EXCHANGE_ID - Instantiate Client ID . . . 56 141 14.2. Operation 48: GETDEVICELIST - Get All Device Mappings 142 for a File System . . . . . . . . . . . . . . . . . . . 57 143 15. NFSv4.2 Operations . . . . . . . . . . . . . . . . . . . . . 59 144 15.1. Operation 59: ALLOCATE - Reserve Space in A Region of a 145 File . . . . . . . . . . . . . . . . . . . . . . . . . . 59 146 15.2. Operation 60: COPY - Initiate a server-side copy . . . . 60 147 15.3. Operation 61: COPY_NOTIFY - Notify a source server of a 148 future copy . . . . . . . . . . . . . . . . . . . . . . 65 149 15.4. Operation 62: DEALLOCATE - Unreserve Space in a Region 150 of a File . . . . . . . . . . . . . . . . . . . . . . . 66 151 15.5. Operation 63: IO_ADVISE - Application I/O access pattern 152 hints . . . . . . . . . . . . . . . . . . . . . . . . . 67 153 15.6. Operation 64: LAYOUTERROR - Provide Errors for the 154 Layout . . . . . . . . . . . . . . . . . . . . . . . . . 73 155 15.7. Operation 65: LAYOUTSTATS - Provide Statistics for the 156 Layout . . . . . . . . . . . . . . . . . . . . . . . . . 76 157 15.8. Operation 66: OFFLOAD_CANCEL - Stop an Offloaded 158 Operation . . . . . . . . . . . . . . . . . . . . . . . 78 159 15.9. Operation 67: OFFLOAD_STATUS - Poll for Status of 160 Asynchronous Operation . . . . . . . . . . . . . . . . . 79 161 15.10. Operation 68: READ_PLUS - READ Data or Holes from a File 80 162 15.11. Operation 69: SEEK - Find the Next Data or Hole . . . . 85 163 15.12. Operation 70: WRITE_SAME - WRITE an ADB Multiple Times 164 to a File . . . . . . . . . . . . . . . . . . . . . . . 86 165 16. NFSv4.2 Callback Operations . . . . . . . . . . . . . . . . . 90 166 16.1. Operation 15: CB_OFFLOAD - Report results of an 167 asynchronous operation . . . . . . . . . . . . . . . . . 90 168 17. Security Considerations . . . . . . . . . . . . . . . . . . . 91 169 18. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 91 170 19. References . . . . . . . . . . . . . . . . . . . . . . . . . 91 171 19.1. Normative References . . . . . . . . . . . . . . . . . . 91 172 19.2. Informative References . . . . . . . . . . . . . . . . . 92 173 Appendix A. Acknowledgments . . . . . . . . . . . . . . . . . . 94 174 Appendix B. RFC Editor Notes . . . . . . . . . . . . . . . . . . 95 175 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 95 177 1. Introduction 179 1.1. The NFS Version 4 Minor Version 2 Protocol 181 The NFS version 4 minor version 2 (NFSv4.2) protocol is the third 182 minor version of the NFS version 4 (NFSv4) protocol. The first minor 183 version, NFSv4.0, is described in [I-D.ietf-nfsv4-rfc3530bis] and the 184 second minor version, NFSv4.1, is described in [RFC5661]. 186 As a minor version, NFSv4.2 is consistent with the overall goals for 187 NFSv4, but extends the protocol so as to better meet those goals, 188 based on experiences with NFSv4.1. In addition, NFSv4.2 has adopted 189 some additional goals, which motivate some of the major extensions in 190 NFSv4.2. 192 1.2. Scope of This Document 194 This document describes the NFSv4.2 protocol. With respect to 195 NFSv4.0 and NFSv4.1, this document does not: 197 o describe the NFSv4.0 or NFSv4.1 protocols, except where needed to 198 contrast with NFSv4.2 200 o modify the specification of the NFSv4.0 or NFSv4.1 protocols 202 o clarify the NFSv4.0 or NFSv4.1 protocols. I.e., any 203 clarifications made here apply to NFSv4.2 and neither of the prior 204 protocols 206 The full External Data Representation (XDR) [RFC4506] for NFSv4.2 is 207 presented in [NFSv42xdr]. 209 1.3. NFSv4.2 Goals 211 A major goal of the design of NFSv4.2 is to take common local file 212 system features and offer them remotely. These features might 214 o already be available on the servers, e.g., sparse files 216 o be under development as a new standard, e.g., SEEK pulls in both 217 SEEK_HOLE and SEEK_DATA 219 o be used by clients with the servers via some proprietary means, 220 e.g., Labeled NFS 222 NFSv4.2 provides means for clients to leverage these features on the 223 server in cases in which that had previously not been possible within 224 the confines of the NFS protocol. 226 1.4. Overview of NFSv4.2 Features 228 1.4.1. Server Side Copy 230 A traditional file copy of a remotely accessed, whether from one 231 server to another or between location in the same server, results in 232 the data being put on the network twice - source to client and then 233 client to destination. New operations are introduced to allow 234 unnecessary traffic to be eliminated: 236 The intra-server copy feature allows the client to request the 237 server to perform the copy internally, avoiding unnecessary 238 network traffic. 240 The inter-server copy feature allows the client to authorize the 241 source and destination servers to interact directly. 243 As such copies can be lengthy, asynchronous support is also provided. 245 1.4.2. Application I/O Advise 247 Applications and clients want to advise the server as to expected I/O 248 behavior. Using IO_ADVISE (see Section 15.5) to communicate future I 249 /O behavior such as whether a file will be accessed sequentially or 250 randomly, and whether a file will or will not be accessed in the near 251 future, allows servers to optimize future I/O requests for a file by, 252 for example, prefetching or evicting data. This operation can be 253 used to support the posix_fadvise function. In addition, it may be 254 helpful to applications such as databases and video editors. 256 1.4.3. Sparse Files 258 Sparse files are ones which have unallocated or uninitialized data 259 blocks as holes in the file. Such holes are typically transferred as 260 0s during I/O. READ_PLUS (see Section 15.10) allows a server to send 261 back to the client metadata describing the hole and DEALLOCATE (see 262 Section 15.4) allows the client to punch holes into a file. In 263 addition, SEEK (see Section 15.11) is provided to scan for the next 264 hole or data from a given location. 266 1.4.4. Space Reservation 268 When a file is sparse, one concern applications have is ensuring that 269 there will always be enough data blocks available for the file during 270 future writes. ALLOCATE (see Section 15.1) allows a client to 271 request a guarantee that space will be available. Also DEALLOCATE 272 (see Section 15.4) allows the client to punch a hole into a file, 273 thus releasing a space reservation. 275 1.4.5. Application Data Block (ADB) Support 277 Some applications treat a file as if it were a disk and as such want 278 to initialize (or format) the file image. We introduce WRITE_SAME 279 (see Section 15.12) to send this metadata to the server to allow it 280 to write the block contents. 282 1.4.6. Labeled NFS 284 While both clients and servers can employ Mandatory Access Control 285 (MAC) security models to enforce data access, there has been no 286 protocol support for interoperability. A new file object attribute, 287 sec_label (see Section 12.2.2) allows for the server to store MAC 288 labels on files, which the client retrieves and uses to enforce data 289 access (see Section 9.6.2). The format of the sec_label accommodates 290 any MAC security system. 292 1.5. Enhancements to Minor Versioning Model 294 In NFSv4.1, the only way to introduce new variants of an operation 295 was to introduce a new operation. I.e., READ becomes either READ2 or 296 READ_PLUS. With the use of discriminated unions as parameters to 297 such functions in NFSv4.2, it is possible to add a new arm in a 298 subsequent minor version. And it is also possible to move such an 299 operation from OPTIONAL/RECOMMENDED to REQUIRED. Forcing an 300 implementation to adopt each arm of a discriminated union at such a 301 time does not meet the spirit of the minor versioning rules. As 302 such, new arms of a discriminated union MUST follow the same 303 guidelines for minor versioning as operations in NFSv4.1 - i.e., they 304 may not be made REQUIRED. To support this, a new error code, 305 NFS4ERR_UNION_NOTSUPP, allows the server to communicate to the client 306 that the operation is supported, but the specific arm of the 307 discriminated union is not. 309 2. Minor Versioning 311 NFSv4.2 is a minor version of NFSv4 and is built upon NFSv4.1 as 312 documented in [RFC5661] and [RFC5662]. 314 NFSv4.2 does not modify the rules applicable to the NFSv4 versioning 315 process and follows the rules set out in [RFC5661] or in standard- 316 track documents updating that document (e.g., in an RFC based on 317 [NFSv4-Versioning]). 319 NFSv4.2 only defines extensions to NFSv4.1, each of which may be 320 supported (or not) independently. It does not 322 o introduce infrastructural features 324 o make existing features MANDATORY to NOT implement 326 o change the status of existing features (i.e., by changing their 327 status among OPTIONAL, RECOMMENDED, REQUIRED). 329 The following versioning-related considerations should be noted. 331 o When a new case is added to an existing switch, servers need to 332 report non-support of that new case by returning 333 NFS4ERR_UNION_NOTSUPP. 335 o As regards the potential cross-minor-version transfer of stateids, 336 pNFS implementations of the file mapping type may support of use 337 of an NFSv4.2 metadata sever with NFSv4.1 data servers. In this 338 context, a stateid returned by an NFSv4.2 COMPOUND will be used in 339 an NFSv4.1 COMPOUND directed to the data server (see Sections 3.2 340 and 3.3). 342 3. pNFS considerations for New Operations 344 3.1. Atomicity for ALLOCATE and DEALLOCATE 346 Both ALLOCATE (see Section 15.1) and DEALLOCATE (see Section 15.4) 347 are sent to the metadata server, which is responsible for 348 coordinating the changes onto the storage devices. In particular, 349 both operations must either fully succeed or fail, it cannot be the 350 case that one storage device succeeds whilst another fails. 352 3.2. Sharing of stateids with NFSv4.1 354 A NFSv4.2 metadata server can hand out a layout to a NFSv4.1 storage 355 device. Section 13.9.1 of [RFC5661] discusses how the client gets a 356 stateid from the metadata server to present to a storage device. 358 3.3. NFSv4.2 as a Storage Protocol in pNFS: the File Layout Type 360 A file layout provided by a NFSv4.2 server may refer either to a 361 storage device that only implements NFSv4.1 as specified in 362 [RFC5661], or to a storage device that implements additions from 363 NFSv4.2, in which case the rules in Section 3.3.1 apply. As the File 364 Layout Type does not provide a means for informing the client as to 365 which minor version a particular storage device is providing, it will 366 have to negotiate this via the normal RPC semantics of major and 367 minor version discovery. 369 3.3.1. Operations Sent to NFSv4.2 Data Servers 371 In addition to the commands listed in [RFC5661], NFSv4.2 data servers 372 MAY accept a COMPOUND containing the following additional operations: 373 IO_ADVISE (see Section 15.5), READ_PLUS (see Section 15.10), 374 WRITE_SAME (see Section 15.12), and SEEK (see Section 15.11), which 375 will be treated like the subset specified as "Operations Sent to 376 NFSv4.1 Data Servers" in Section 13.6 of [RFC5661]. 378 Additional details on the implementation of these operations in a 379 pNFS context are documented in the operation specific sections. 381 4. Server Side Copy 383 4.1. Introduction 385 The server-side copy feature provides a mechanism for the NFS client 386 to perform a file copy on a server or between two servers without the 387 data being transmitted back and forth over the network through the 388 NFS client. Without this feature, an NFS client copies data from one 389 location to another by reading the data from the source server over 390 the network, and then writing the data back over the network to the 391 destination server. 393 If the source object and destination object are on different file 394 servers, the file servers will communicate with one another to 395 perform the copy operation. The server-to-server protocol by which 396 this is accomplished is not defined in this document. 398 4.2. Protocol Overview 400 The server-side copy offload operations support both intra-server and 401 inter-server file copies. An intra-server copy is a copy in which 402 the source file and destination file reside on the same server. In 403 an inter-server copy, the source file and destination file are on 404 different servers. In both cases, the copy may be performed 405 synchronously or asynchronously. 407 Throughout the rest of this document, we refer to the NFS server 408 containing the source file as the "source server" and the NFS server 409 to which the file is transferred as the "destination server". In the 410 case of an intra-server copy, the source server and destination 411 server are the same server. Therefore in the context of an intra- 412 server copy, the terms source server and destination server refer to 413 the single server performing the copy. 415 The new operations are designed to copy files. Other file system 416 objects can be copied by building on these operations or using other 417 techniques. For example, if the user wishes to copy a directory, the 418 client can synthesize a directory copy by first creating the 419 destination directory and then copying the source directory's files 420 to the new destination directory. 422 For the inter-server copy, the operations are defined to be 423 compatible with the traditional copy authentication approach. The 424 client and user are authorized at the source for reading. Then they 425 are authorized at the destination for writing. 427 4.2.1. Copy Operations 429 COPY_NOTIFY: Used by the client to notify the source server of a 430 future file copy from a given destination server for the given 431 user. (Section 15.3) 433 COPY: Used by the client to request a file copy. (Section 15.2) 435 OFFLOAD_CANCEL: Used by the client to terminate an asynchronous file 436 copy. (Section 15.8) 438 OFFLOAD_STATUS: Used by the client to poll the status of an 439 asynchronous file copy. (Section 15.9) 441 CB_OFFLOAD: Used by the destination server to report the results of 442 an asynchronous file copy to the client. (Section 16.1) 444 4.2.2. Requirements for Operations 446 The implementation of server-side copy is OPTIONAL by the client and 447 the server. However, in order to successfully copy a file, some 448 operations MUST be supported by the client and/or server. 450 If a client desires an intra-server file copy, then it MUST support 451 the COPY and CB_OFFLOAD operations. If COPY returns a stateid, then 452 the client MAY use the OFFLOAD_CANCEL and OFFLOAD_STATUS operations. 454 If a client desires an inter-server file copy, then it MUST support 455 the COPY, COPY_NOTIFY, and CB_OFFLOAD operations, and MAY use the 456 OFFLOAD_CANCEL operation. If COPY returns a stateid, then the client 457 MAY use the OFFLOAD_CANCEL and OFFLOAD_STATUS operations. 459 If a server supports intra-server copy, then the server MUST support 460 the COPY operation. If a server's COPY operation returns a stateid, 461 then the server MUST also support these operations: CB_OFFLOAD, 462 OFFLOAD_CANCEL, and OFFLOAD_STATUS. 464 If a source server supports inter-server copy, then the source server 465 MUST support all these operations: COPY_NOTIFY and OFFLOAD_CANCEL. 466 If a destination server supports inter-server copy, then the 467 destination server MUST support the COPY operation. If a destination 468 server's COPY operation returns a stateid, then the destination 469 server MUST also support these operations: CB_OFFLOAD, 470 OFFLOAD_CANCEL, COPY_NOTIFY, and OFFLOAD_STATUS. 472 Each operation is performed in the context of the user identified by 473 the ONC RPC credential of its containing COMPOUND or CB_COMPOUND 474 request. For example, an OFFLOAD_CANCEL operation issued by a given 475 user indicates that a specified COPY operation initiated by the same 476 user be canceled. Therefore an OFFLOAD_CANCEL MUST NOT interfere 477 with a copy of the same file initiated by another user. 479 An NFS server MAY allow an administrative user to monitor or cancel 480 copy operations using an implementation specific interface. 482 4.3. Requirements for Inter-Server Copy 484 Inter-server copy is driven by several requirements: 486 o The specification MUST NOT mandate the server-to-server protocol. 488 o The specification MUST provide guidance for using NFSv4.x as a 489 copy protocol. For those source and destination servers willing 490 to use NFSv4.x, there are specific security considerations that 491 this specification MUST address. 493 o The specification MUST NOT mandate preconfiguration between the 494 source and destination server. Requiring that the source and 495 destination first have a "copying relationship" increases the 496 administrative burden. However the specification MUST NOT 497 preclude implementations that require preconfiguration. 499 o The specification MUST NOT mandate a trust relationship between 500 the source and destination server. The NFSv4 security model 501 requires mutual authentication between a principal on an NFS 502 client and a principal on an NFS server. This model MUST continue 503 with the introduction of COPY. 505 4.4. Implementation Considerations 507 4.4.1. Locking the Files 509 Both the source and destination file may need to be locked to protect 510 the content during the copy operations. A client can achieve this by 511 a combination of OPEN and LOCK operations. I.e., either share or 512 byte range locks might be desired. 514 Note that when the client establishes a lock stateid on the source, 515 the context of that stateid is for the client and not the 516 destination. As such, there might already be an outstanding stateid, 517 issued to the destination as client of the source, with the same 518 value as that provided for the lock stateid. The source MUST equate 519 the lock stateid as that of the client, i.e., when the destination 520 presents it in the context of a inter-server copy, it is on behalf of 521 the client. 523 4.4.2. Client Caches 525 In a traditional copy, if the client is in the process of writing to 526 the file before the copy (and perhaps with a write delegation), it 527 will be straightforward to update the destination server. With an 528 inter-server copy, the source has no insight into the changes cached 529 on the client. The client SHOULD write back the data to the source. 530 If it does not do so, it is possible that the destination will 531 receive a corrupt copy of file. 533 4.5. Intra-Server Copy 535 To copy a file on a single server, the client uses a COPY operation. 536 The server may respond to the copy operation with the final results 537 of the copy or it may perform the copy asynchronously and deliver the 538 results using a CB_OFFLOAD operation callback. If the copy is 539 performed asynchronously, the client may poll the status of the copy 540 using OFFLOAD_STATUS or cancel the copy using OFFLOAD_CANCEL. 542 A synchronous intra-server copy is shown in Figure 1. In this 543 example, the NFS server chooses to perform the copy synchronously. 544 The copy operation is completed, either successfully or 545 unsuccessfully, before the server replies to the client's request. 546 The server's reply contains the final result of the operation. 548 Client Server 549 + + 550 | | 551 |--- OPEN ---------------------------->| Client opens 552 |<------------------------------------/| the source file 553 | | 554 |--- OPEN ---------------------------->| Client opens 555 |<------------------------------------/| the destination file 556 | | 557 |--- COPY ---------------------------->| Client requests 558 |<------------------------------------/| a file copy 559 | | 560 |--- CLOSE --------------------------->| Client closes 561 |<------------------------------------/| the destination file 562 | | 563 |--- CLOSE --------------------------->| Client closes 564 |<------------------------------------/| the source file 565 | | 566 | | 568 Figure 1: A synchronous intra-server copy. 570 An asynchronous intra-server copy is shown in Figure 2. In this 571 example, the NFS server performs the copy asynchronously. The 572 server's reply to the copy request indicates that the copy operation 573 was initiated and the final result will be delivered at a later time. 574 The server's reply also contains a copy stateid. The client may use 575 this copy stateid to poll for status information (as shown) or to 576 cancel the copy using an OFFLOAD_CANCEL. When the server completes 577 the copy, the server performs a callback to the client and reports 578 the results. 580 Client Server 581 + + 582 | | 583 |--- OPEN ---------------------------->| Client opens 584 |<------------------------------------/| the source file 585 | | 586 |--- OPEN ---------------------------->| Client opens 587 |<------------------------------------/| the destination file 588 | | 589 |--- COPY ---------------------------->| Client requests 590 |<------------------------------------/| a file copy 591 | | 592 | | 593 |--- OFFLOAD_STATUS ------------------>| Client may poll 594 |<------------------------------------/| for status 595 | | 596 | . | Multiple OFFLOAD_STATUS 597 | . | operations may be sent. 598 | . | 599 | | 600 |<-- CB_OFFLOAD -----------------------| Server reports results 601 |\------------------------------------>| 602 | | 603 |--- CLOSE --------------------------->| Client closes 604 |<------------------------------------/| the destination file 605 | | 606 |--- CLOSE --------------------------->| Client closes 607 |<------------------------------------/| the source file 608 | | 609 | | 611 Figure 2: An asynchronous intra-server copy. 613 4.6. Inter-Server Copy 615 A copy may also be performed between two servers. The copy protocol 616 is designed to accommodate a variety of network topologies. As shown 617 in Figure 3, the client and servers may be connected by multiple 618 networks. In particular, the servers may be connected by a 619 specialized, high speed network (network 192.0.2.0/24 in the diagram) 620 that does not include the client. The protocol allows the client to 621 setup the copy between the servers (over network 203.0.113.0/24 in 622 the diagram) and for the servers to communicate on the high speed 623 network if they choose to do so. 625 192.0.2.0/24 626 +-------------------------------------+ 627 | | 628 | | 629 | 192.0.2.18 | 192.0.2.56 630 +-------+------+ +------+------+ 631 | Source | | Destination | 632 +-------+------+ +------+------+ 633 | 203.0.113.18 | 203.0.113.56 634 | | 635 | | 636 | 203.0.113.0/24 | 637 +------------------+------------------+ 638 | 639 | 640 | 203.0.113.243 641 +-----+-----+ 642 | Client | 643 +-----------+ 645 Figure 3: An example inter-server network topology. 647 For an inter-server copy, the client notifies the source server that 648 a file will be copied by the destination server using a COPY_NOTIFY 649 operation. The client then initiates the copy by sending the COPY 650 operation to the destination server. The destination server may 651 perform the copy synchronously or asynchronously. 653 A synchronous inter-server copy is shown in Figure 4. In this case, 654 the destination server chooses to perform the copy before responding 655 to the client's COPY request. 657 An asynchronous copy is shown in Figure 5. In this case, the 658 destination server chooses to respond to the client's COPY request 659 immediately and then perform the copy asynchronously. 661 Client Source Destination 662 + + + 663 | | | 664 |--- OPEN --->| | Returns os1 665 |<------------------/| | 666 | | | 667 |--- COPY_NOTIFY --->| | 668 |<------------------/| | 669 | | | 670 |--- OPEN ---------------------------->| Returns os2 671 |<------------------------------------/| 672 | | | 673 |--- COPY ---------------------------->| 674 | | | 675 | | | 676 | |<----- read -----| 677 | |\--------------->| 678 | | | 679 | | . | Multiple reads may 680 | | . | be necessary 681 | | . | 682 | | | 683 | | | 684 |<------------------------------------/| Destination replies 685 | | | to COPY 686 | | | 687 |--- CLOSE --------------------------->| Release open state 688 |<------------------------------------/| 689 | | | 690 |--- CLOSE --->| | Release open state 691 |<------------------/| | 693 Figure 4: A synchronous inter-server copy. 695 Client Source Destination 696 + + + 697 | | | 698 |--- OPEN --->| | Returns os1 699 |<------------------/| | 700 | | | 701 |--- LOCK --->| | Optional, could be done 702 |<------------------/| | with a share lock 703 | | | 704 |--- COPY_NOTIFY --->| | Need to pass in 705 |<------------------/| | os1 or lock state 706 | | | 707 | | | 708 | | | 709 |--- OPEN ---------------------------->| Returns os2 710 |<------------------------------------/| 711 | | | 712 |--- LOCK ---------------------------->| Optional ... 713 |<------------------------------------/| 714 | | | 715 |--- COPY ---------------------------->| Need to pass in 716 |<------------------------------------/| os2 or lock state 717 | | | 718 | | | 719 | |<----- read -----| 720 | |\--------------->| 721 | | | 722 | | . | Multiple reads may 723 | | . | be necessary 724 | | . | 725 | | | 726 | | | 727 |--- OFFLOAD_STATUS ------------------>| Client may poll 728 |<------------------------------------/| for status 729 | | | 730 | | . | Multiple OFFLOAD_STATUS 731 | | . | operations may be sent 732 | | . | 733 | | | 734 | | | 735 | | | 736 |<-- CB_OFFLOAD -----------------------| Destination reports 737 |\------------------------------------>| results 738 | | | 739 |--- LOCKU --------------------------->| Only if LOCK was done 740 |<------------------------------------/| 741 | | | 742 |--- CLOSE --------------------------->| Release open state 743 |<------------------------------------/| 744 | | | 745 |--- LOCKU --->| | Only if LOCK was done 746 |<------------------/| | 747 | | | 748 |--- CLOSE --->| | Release open state 749 |<------------------/| | 750 | | | 752 Figure 5: An asynchronous inter-server copy. 754 4.7. Server-to-Server Copy Protocol 756 The choice of what protocol to use in an inter-server copy is 757 ultimately the destination server's decision. However, the 758 destination server has to be cognizant that it is working on behalf 759 of the client. 761 4.7.1. Considerations on Selecting a Copy Protocol 763 The client can have requirements over both the size of transactions 764 and error recovery semantics. It may want to split the copy up such 765 that each chunk is synchronously transferred. It may want the copy 766 protocol to copy the bytes in consecutive order such that upon an 767 error, the client can restart the copy at the last known good offset. 768 If the destination server cannot meet these requirements, the client 769 may prefer the traditional copy mechanism such that it can meet those 770 requirements. 772 4.7.2. Using NFSv4.x as the Copy Protocol 774 The destination server MAY use standard NFSv4.x (where x >= 1) 775 operations to read the data from the source server. If NFSv4.x is 776 used for the server-to-server copy protocol, the destination server 777 can use the source filehandle and ca_src_stateid provided in the COPY 778 request with standard NFSv4.x operations to read data from the source 779 server. 781 4.7.3. Using an Alternative Copy Protocol 783 In a homogeneous environment, the source and destination servers 784 might be able to perform the file copy extremely efficiently using 785 specialized protocols. For example the source and destination 786 servers might be two nodes sharing a common file system format for 787 the source and destination file systems. Thus the source and 788 destination are in an ideal position to efficiently render the image 789 of the source file to the destination file by replicating the file 790 system formats at the block level. Another possibility is that the 791 source and destination might be two nodes sharing a common storage 792 area network, and thus there is no need to copy any data at all, and 793 instead ownership of the file and its contents might simply be re- 794 assigned to the destination. To allow for these possibilities, the 795 destination server is allowed to use a server-to-server copy protocol 796 of its choice. 798 In a heterogeneous environment, using a protocol other than NFSv4.x 799 (e.g., HTTP [RFC2616] or FTP [RFC959]) presents some challenges. In 800 particular, the destination server is presented with the challenge of 801 accessing the source file given only an NFSv4.x filehandle. 803 One option for protocols that identify source files with path names 804 is to use an ASCII hexadecimal representation of the source 805 filehandle as the file name. 807 Another option for the source server is to use URLs to direct the 808 destination server to a specialized service. For example, the 809 response to COPY_NOTIFY could include the URL ftp:// 810 s1.example.com:9999/_FH/0x12345, where 0x12345 is the ASCII 811 hexadecimal representation of the source filehandle. When the 812 destination server receives the source server's URL, it would use 813 "_FH/0x12345" as the file name to pass to the FTP server listening on 814 port 9999 of s1.example.com. On port 9999 there would be a special 815 instance of the FTP service that understands how to convert NFS 816 filehandles to an open file descriptor (in many operating systems, 817 this would require a new system call, one which is the inverse of the 818 makefh() function that the pre-NFSv4 MOUNT service needs). 820 Authenticating and identifying the destination server to the source 821 server is also a challenge. Recommendations for how to accomplish 822 this are given in Section 4.10.1.3. 824 4.8. netloc4 - Network Locations 826 The server-side copy operations specify network locations using the 827 netloc4 data type shown below: 829 831 enum netloc_type4 { 832 NL4_NAME = 0, 833 NL4_URL = 1, 834 NL4_NETADDR = 2 835 }; 836 union netloc4 switch (netloc_type4 nl_type) { 837 case NL4_NAME: utf8str_cis nl_name; 838 case NL4_URL: utf8str_cis nl_url; 839 case NL4_NETADDR: netaddr4 nl_addr; 840 }; 842 844 If the netloc4 is of type NL4_NAME, the nl_name field MUST be 845 specified as a UTF-8 string. The nl_name is expected to be resolved 846 to a network address via DNS, LDAP, NIS, /etc/hosts, or some other 847 means. If the netloc4 is of type NL4_URL, a server URL [RFC3986] 848 appropriate for the server-to-server copy operation is specified as a 849 UTF-8 string. If the netloc4 is of type NL4_NETADDR, the nl_addr 850 field MUST contain a valid netaddr4 as defined in Section 3.3.9 of 851 [RFC5661]. 853 When netloc4 values are used for an inter-server copy as shown in 854 Figure 3, their values may be evaluated on the source server, 855 destination server, and client. The network environment in which 856 these systems operate should be configured so that the netloc4 values 857 are interpreted as intended on each system. 859 4.9. Copy Offload Stateids 861 A server may perform a copy offload operation asynchronously. An 862 asynchronous copy is tracked using a copy offload stateid. Copy 863 offload stateids are included in the COPY, OFFLOAD_CANCEL, 864 OFFLOAD_STATUS, and CB_OFFLOAD operations. 866 A copy offload stateid will be valid until either (A) the client or 867 server restarts or (B) the client returns the resource by issuing a 868 OFFLOAD_CANCEL operation or the client replies to a CB_OFFLOAD 869 operation. 871 A copy offload stateid's seqid MUST NOT be 0. In the context of a 872 copy offload operation, it is ambiguous to indicate the most recent 873 copy offload operation using a stateid with seqid of 0. Therefore a 874 copy offload stateid with seqid of 0 MUST be considered invalid. 876 4.10. Security Considerations 878 The security considerations pertaining to NFSv4.1 [RFC5661] apply to 879 this section. And as such, the standard security mechanisms used by 880 the protocol can be used to secure the server-to-server operations. 882 NFSv4 clients and servers supporting the inter-server copy operations 883 described in this chapter are REQUIRED to implement the mechanism 884 described in Section 4.10.1.1, and to support rejecting COPY_NOTIFY 885 requests that do not use RPCSEC_GSS with privacy. If the server-to- 886 server copy protocol is ONC RPC based, the servers are also REQUIRED 887 to implement [rpcsec_gssv3] including the RPCSEC_GSSv3 copy_to_auth, 888 copy_from_auth, and copy_confirm_auth structured privileges. This 889 requirement to implement is not a requirement to use; for example, a 890 server may depending on configuration also allow COPY_NOTIFY requests 891 that use only AUTH_SYS. 893 4.10.1. Inter-Server Copy Security 894 4.10.1.1. Inter-Server Copy via ONC RPC with RPCSEC_GSSv3 896 When the client sends a COPY_NOTIFY to the source server to expect 897 the destination to attempt to copy data from the source server, it is 898 expected that this copy is being done on behalf of the principal 899 (called the "user principal") that sent the RPC request that encloses 900 the COMPOUND procedure that contains the COPY_NOTIFY operation. The 901 user principal is identified by the RPC credentials. A mechanism 902 that allows the user principal to authorize the destination server to 903 perform the copy, that lets the source server properly authenticate 904 the destination's copy, and does not allow the destination server to 905 exceed this authorization, is necessary. 907 An approach that sends delegated credentials of the client's user 908 principal to the destination server is not used for the following 909 reason. If the client's user delegated its credentials, the 910 destination would authenticate as the user principal. If the 911 destination were using the NFSv4 protocol to perform the copy, then 912 the source server would authenticate the destination server as the 913 user principal, and the file copy would securely proceed. However, 914 this approach would allow the destination server to copy other files. 915 The user principal would have to trust the destination server to not 916 do so. This is counter to the requirements, and therefore is not 917 considered. 919 Instead, a feature of the RPCSEC_GSSv3 [rpcsec_gssv3] protocol can be 920 used: RPC application defined structured privilege assertion. This 921 features allow the destination server to authenticate to the source 922 server as acting on behalf of the user principal, and to authorize 923 the destination server to perform READs of the file to be copied from 924 the source on behalf of the user principal. Once the copy is 925 complete, the client can destroy the RPCSEC_GSSv3 handles to end the 926 authorization of both the source and destination servers to copy. 928 We define three RPCSEC_GSSv3 structured privilege assertions that 929 work in tandem to authorize the copy: 931 copy_from_auth: A user principal is authorizing a source principal 932 ("nfs@") to allow a destination principal 933 ("nfs@") to setup the copy_confirm_auth privilege 934 required to copy a file from the source to the destination on 935 behalf of the user principal. This privilege is established on 936 the source server before the user principal sends a COPY_NOTIFY 937 operation to the source server, and the resultant RPCSEC_GSSv3 938 context is used to secure the COPY_NOTIFY operation. 940 942 struct copy_from_auth_priv { 943 secret4 cfap_shared_secret; 944 netloc4 cfap_destination; 945 /* the NFSv4 user name that the user principal maps to */ 946 utf8str_mixed cfap_username; 947 }; 949 951 cfp_shared_secret is an automatically generated random number 952 secret value. 954 copy_to_auth: A user principal is authorizing a destination 955 principal ("nfs@") to setup a copy_confirm_auth 956 privilege with a source principal ("nfs@") to allow it to 957 copy a file from the source to the destination on behalf of the 958 user principal. This privilege is established on the destination 959 server before the user principal sends a COPY operation to the 960 destination server, and the resultant RPCSEC_GSSv3 context is used 961 to secure the COPY operation. 963 965 struct copy_to_auth_priv { 966 /* equal to cfap_shared_secret */ 967 secret4 ctap_shared_secret; 968 netloc4 ctap_source<>; 969 /* the NFSv4 user name that the user principal maps to */ 970 utf8str_mixed ctap_username; 971 }; 973 975 ctap_shared_secret is the automatically generated secret value 976 used to establish the copy_from_auth privilege with the source 977 principal. See Section 4.10.1.1.1. 979 copy_confirm_auth: A destination principal ("nfs@") is 980 confirming with the source principal ("nfs@") that it is 981 authorized to copy data from the source. This privilege is 982 established on the destination server before the file is copied 983 from the source to the destination. The resultant RPCSEC_GSSv3 984 context is used to secure the READ operations from the source to 985 the destination server. 987 989 struct copy_confirm_auth_priv { 990 /* equal to GSS_GetMIC() of cfap_shared_secret */ 991 opaque ccap_shared_secret_mic<>; 992 /* the NFSv4 user name that the user principal maps to */ 993 utf8str_mixed ccap_username; 994 }; 996 998 4.10.1.1.1. Establishing a Security Context 1000 When the user principal wants to COPY a file between two servers, if 1001 it has not established copy_from_auth and copy_to_auth privileges on 1002 the servers, it establishes them: 1004 o As noted in [rpcsec_gssv3] the client uses an existing 1005 RPCSEC_GSSv3 context termed the "parent" handle to establish and 1006 protect RPCSEC_GSSv3 structured privilege assertion exchanges. 1007 The copy_from_auth privilege will use the context established 1008 between the user principal and the source server used to OPEN the 1009 source file as the RPCSEC_GSSv3 parent handle. The copy_to_auth 1010 privilege will use the context established between the user 1011 principal and the destination server used to OPEN the destination 1012 file as the RPCSEC_GSSv3 parent handle. 1014 o A random number is generated to use as a secret to be shared 1015 between the two servers. This shared secret will be placed in the 1016 cfap_shared_secret and ctap_shared_secret fields of the 1017 appropriate privilege data types, copy_from_auth_priv and 1018 copy_to_auth_priv. Because of this shared_secret the 1019 RPCSEC_GSS3_CREATE control messages for copy_from_auth and 1020 copy_to_auth MUST use a QOP of rpc_gss_svc_privacy. 1022 o An instance of copy_from_auth_priv is filled in with the shared 1023 secret, the destination server, and the NFSv4 user id of the user 1024 principal and is placed in rpc_gss3_create_args 1025 assertions[0].privs.privilege. The string "copy_from_auth" is 1026 placed in assertions[0].privs.name. The source server unwraps the 1027 rpc_gss_svc_privacy RPCSEC_GSS3_CREATE payload and verifies that 1028 the NFSv4 user id being asserted matches the source server's 1029 mapping of the user principal. If it does, the privilege is 1030 established on the source server as: <"copy_from_auth", user id, 1031 destination>. The field "handle" in a successful reply is the 1032 RPCSEC_GSSv3 copy_from_auth "child" handle that the client will 1033 use on COPY_NOTIFY requests to the source server. 1035 o An instance of copy_to_auth_priv is filled in with the shared 1036 secret, the cnr_source_server list returned by COPY_NOTIFY, and 1037 the NFSv4 user id of the user principal. The copy_to_auth_priv 1038 instance is placed in rpc_gss3_create_args 1039 assertions[0].privs.privilege. The string "copy_to_auth" is 1040 placed in assertions[0].privs.name. The destination server 1041 unwraps the rpc_gss_svc_privacy RPCSEC_GSS3_CREATE payload and 1042 verifies that the NFSv4 user id being asserted matches the 1043 destination server's mapping of the user principal. If it does, 1044 the privilege is established on the destination server as: 1045 <"copy_to_auth", user id, source list>. The field "handle" in a 1046 successful reply is the RPCSEC_GSSv3 copy_to_auth "child" handle 1047 that the client will use on COPY requests to the destination 1048 server involving the source server. 1050 As noted in [rpcsec_gssv3] Section 2.3.1 "Create Request", both the 1051 client and the source server should associate the RPCSEC_GSSv3 1052 "child" handle with the parent RPCSEC_GSSv3 handle used to create the 1053 RPCSEC_GSSv3 child handle. 1055 4.10.1.1.2. Starting a Secure Inter-Server Copy 1057 When the client sends a COPY_NOTIFY request to the source server, it 1058 uses the privileged "copy_from_auth" RPCSEC_GSSv3 handle. 1059 cna_destination_server in COPY_NOTIFY MUST be the same as 1060 cfap_destination specified in copy_from_auth_priv. Otherwise, 1061 COPY_NOTIFY will fail with NFS4ERR_ACCESS. The source server 1062 verifies that the privilege <"copy_from_auth", user id, destination> 1063 exists, and annotates it with the source filehandle, if the user 1064 principal has read access to the source file, and if administrative 1065 policies give the user principal and the NFS client read access to 1066 the source file (i.e., if the ACCESS operation would grant read 1067 access). Otherwise, COPY_NOTIFY will fail with NFS4ERR_ACCESS. 1069 When the client sends a COPY request to the destination server, it 1070 uses the privileged "copy_to_auth" RPCSEC_GSSv3 handle. 1071 ca_source_server list in COPY MUST be the same as ctap_source list 1072 specified in copy_to_auth_priv. Otherwise, COPY will fail with 1073 NFS4ERR_ACCESS. The destination server verifies that the privilege 1074 <"copy_to_auth", user id, source list> exists, and annotates it with 1075 the source and destination filehandles. If the COPY returns a 1076 wr_callback_id, then this is an asynchronous copy and the 1077 wr_callback_id must also must be annotated to the copy_to_auth 1078 privilege. If the client has failed to establish the "copy_to_auth" 1079 privilege it will reject the request with NFS4ERR_PARTNER_NO_AUTH. 1081 If either the COPY_NOTIFY, or the COPY operations fail, the 1082 associated "copy_from_auth" and "copy_to_auth" RPCSEC_GSSv3 handles 1083 MUST be destroyed. 1085 4.10.1.1.3. Securing ONC RPC Server-to-Server Copy Protocols 1087 After a destination server has a "copy_to_auth" privilege established 1088 on it, and it receives a COPY request, if it knows it will use an ONC 1089 RPC protocol to copy data, it will establish a "copy_confirm_auth" 1090 privilege on the source server prior to responding to the COPY 1091 operation as follows: 1093 o Before establishing an RPCSEC_GSSv3 context, a parent context 1094 needs to exist between nfs@ as the initiator 1095 principal, and nfs@ as the target principal. If NFS is to 1096 be used as the copy protocol, this means that the destination 1097 server must mount the source server using RPCSEC_GSSv3. 1099 o An instance of copy_confirm_auth_priv is filled in with 1100 information from the established "copy_to_auth" privilege. The 1101 value of the field ccap_shared_secret_mic is a GSS_GetMIC() of the 1102 ctap_shared_secret in the copy_to_auth privilege using the parent 1103 handle context. The field ccap_username is the mapping of the 1104 user principal to an NFSv4 user name ("user"@"domain" form), and 1105 MUST be the same as the ctap_username in the copy_to_auth 1106 privilege. The copy_confirm_auth_priv instance is placed in 1107 rpc_gss3_create_args assertions[0].privs.privilege. The string 1108 "copy_confirm_auth" is placed in assertions[0].privs.name. 1110 o The RPCSEC_GSS3_CREATE copy_from_auth message is sent to the 1111 source server with a QOP of rpc_gss_svc_privacy. The source 1112 server unwraps the rpc_gss_svc_privacy RPCSEC_GSS3_CREATE payload 1113 and verifies the cap_shared_secret_mic by calling GSS_VerifyMIC() 1114 using the parent context on the cfap_shared_secret from the 1115 established "copy_from_auth" privilege, and verifies the that the 1116 ccap_username equals the cfap_username. 1118 o If all verification succeeds, the "copy_confirm_auth" privilege is 1119 established on the source server as < "copy_confirm_auth", 1120 shared_secret_mic, user id> Because the shared secret has been 1121 verified, the resultant copy_confirm_auth RPCSEC_GSSv3 child 1122 handle is noted to be acting on behalf of the user principal. 1124 o If the source server fails to verify the copy_from_auth privilege 1125 the COPY operation will be rejected with NFS4ERR_PARTNER_NO_AUTH, 1126 causing in turn the client to destroy the associated 1127 copy_from_auth and copy_to_auth RPCSEC_GSSv3 structured privilege 1128 assertion handles. 1130 o All subsequent ONC RPC READ requests sent from the destination to 1131 copy data from the source to the destination will use the 1132 RPCSEC_GSSv3 copy_confirm_auth child handle. 1134 Note that the use of the "copy_confirm_auth" privilege accomplishes 1135 the following: 1137 o If a protocol like NFS is being used, with export policies, export 1138 policies can be overridden in case the destination server as-an- 1139 NFS-client is not authorized 1141 o Manual configuration to allow a copy relationship between the 1142 source and destination is not needed. 1144 4.10.1.1.4. Maintaining a Secure Inter-Server Copy 1146 If the client determines that either the copy_from_auth or the 1147 copy_to_auth handle becomes invalid during a copy, then the copy MUST 1148 be aborted by the client sending an OFFLOAD_CANCEL to both the source 1149 and destination servers and destroying the respective copy related 1150 context handles as described in Section 4.10.1.1.5. 1152 4.10.1.1.5. Finishing or Stopping a Secure Inter-Server Copy 1154 Under normal operation, the client MUST destroy the copy_from_auth 1155 and the copy_to_auth RPCSEC_GSSv3 handle once the COPY operation 1156 returns for a synchronous inter-server copy or a CB_OFFLOAD reports 1157 the result of an asynchronous copy. 1159 The copy_confirm_auth privilege constructed from information held by 1160 the copy_to_auth privilege, and MUST be destroyed by the destination 1161 server (via an RPCSEC_GSS3_DESTROY call) when the copy_to_auth 1162 RPCSEC_GSSv3 handle is destroyed. 1164 The copy_confirm_auth RPCSEC_GSS3 handle is associated with a 1165 copy_from_auth RPCSEC_GSS3 handle on the source server via the shared 1166 secret and MUST be locally destroyed (there is no RPCSEC_GSS3_DESTROY 1167 as the source server is not the initiator) when the copy_from_auth 1168 RPCSEC_GSSv3 handle is destroyed. 1170 If the client sends an OFFLOAD_CANCEL to the source server to rescind 1171 the destination server's synchronous copy privilege, it uses the 1172 privileged "copy_from_auth" RPCSEC_GSSv3 handle and the 1173 cra_destination_server in OFFLOAD_CANCEL MUST be the same as the name 1174 of the destination server specified in copy_from_auth_priv. The 1175 source server will then delete the <"copy_from_auth", user id, 1176 destination> privilege and fail any subsequent copy requests sent 1177 under the auspices of this privilege from the destination server. 1178 The client MUST destroy both the "copy_from_auth" and the 1179 "copy_to_auth" RPCSEC_GSSv3 handles. 1181 If the client sends an OFFLOAD_STATUS to the destination server to 1182 check on the status of an asynchronous copy, it uses the privileged 1183 "copy_to_auth" RPCSEC_GSSv3 handle and the osa_stateid in 1184 OFFLOAD_STATUS MUST be the same as the wr_callback_id specified in 1185 the "copy_to_auth" privilege stored on the destination server. 1187 If the client sends an OFFLOAD_CANCEL to the destination server to 1188 cancel an asynchronous copy, it uses the privileged "copy_to_auth" 1189 RPCSEC_GSSv3 handle and the oaa_stateid in OFFLOAD_CANCEL MUST be the 1190 same as the wr_callback_id specified in the "copy_to_auth" privilege 1191 stored on the destination server. The destination server will then 1192 delete the <"copy_to_auth", user id, source list, nounce, nounce MIC, 1193 context handle, handle version> privilege and the associated 1194 "copy_confirm_auth" RPCSEC_GSSv3 handle. The client MUST destroy 1195 both the copy_to_auth and copy_from_auth RPCSEC_GSSv3 handles. 1197 4.10.1.2. Inter-Server Copy via ONC RPC without RPCSEC_GSS 1199 ONC RPC security flavors other than RPCSEC_GSS MAY be used with the 1200 server-side copy offload operations described in this chapter. In 1201 particular, host-based ONC RPC security flavors such as AUTH_NONE and 1202 AUTH_SYS MAY be used. If a host-based security flavor is used, a 1203 minimal level of protection for the server-to-server copy protocol is 1204 possible. 1206 In the absence of a strong security mechanism designed for the 1207 purpose, the challenge is how the source server and destination 1208 server identify themselves to each other, especially in the presence 1209 of multi-homed source and destination servers. In a multi-homed 1210 environment, the destination server might not contact the source 1211 server from the same network address specified by the client in the 1212 COPY_NOTIFY. This can be overcome using the procedure described 1213 below. 1215 When the client sends the source server the COPY_NOTIFY operation, 1216 the source server may reply to the client with a list of target 1217 addresses, names, and/or URLs and assign them to the unique 1218 quadruple: . If the destination uses one of these target netlocs to contact 1220 the source server, the source server will be able to uniquely 1221 identify the destination server, even if the destination server does 1222 not connect from the address specified by the client in COPY_NOTIFY. 1223 The level of assurance in this identification depends on the 1224 unpredictability, strength and secrecy of the random number. 1226 For example, suppose the network topology is as shown in Figure 3. 1227 If the source filehandle is 0x12345, the source server may respond to 1228 a COPY_NOTIFY for destination 203.0.113.56 with the URLs: 1230 nfs://203.0.113.18//_COPY/FvhH1OKbu8VrxvV1erdjvR7N/203.0.113.56/ 1231 _FH/0x12345 1233 nfs://192.0.2.18//_COPY/FvhH1OKbu8VrxvV1erdjvR7N/203.0.113.56/_FH/ 1234 0x12345 1236 The name component after _COPY is 24 characters of base 64, more than 1237 enough to encode a 128 bit random number. 1239 The client will then send these URLs to the destination server in the 1240 COPY operation. Suppose that the 192.0.2.0/24 network is a high 1241 speed network and the destination server decides to transfer the file 1242 over this network. If the destination contacts the source server 1243 from 192.0.2.56 over this network using NFSv4.1, it does the 1244 following: 1246 COMPOUND { PUTROOTFH, LOOKUP "_COPY" ; LOOKUP 1247 "FvhH1OKbu8VrxvV1erdjvR7N" ; LOOKUP "203.0.113.56"; LOOKUP "_FH" ; 1248 OPEN "0x12345" ; GETFH } 1250 Provided that the random number is unpredictable and has been kept 1251 secret by the parties involved, the source server will therefore know 1252 that these NFSv4.x operations are being issued by the destination 1253 server identified in the COPY_NOTIFY. This random number technique 1254 only provides initial authentication of the destination server, and 1255 cannot defend against man-in-the-middle attacks after authentication 1256 or an eavesdropper that observes the random number on the wire. 1257 Other secure communication techniques (e.g., IPsec) are necessary to 1258 block these attacks. 1260 Servers SHOULD reject COPY_NOTIFY requests that do not use RPCSEC_GSS 1261 with privacy, thus ensuring the URL in the COPY_NOTIFY reply is 1262 encrypted. For the same reason, clients SHOULD send COPY requests to 1263 the destination using RPCSEC_GSS with privacy. 1265 4.10.1.3. Inter-Server Copy without ONC RPC 1267 The same techniques as Section 4.10.1.2, using unique URLs for each 1268 destination server, can be used for other protocols (e.g., HTTP 1269 [RFC2616] and FTP [RFC959]) as well. 1271 5. Support for Application IO Hints 1273 Applications can issue client I/O hints via posix_fadvise() 1274 [posix_fadvise] to the NFS client. While this can help the NFS 1275 client optimize I/O and caching for a file, it does not allow the NFS 1276 server and its exported file system to do likewise. We add an 1277 IO_ADVISE procedure (Section 15.5) to communicate the client file 1278 access patterns to the NFS server. The NFS server upon receiving a 1279 IO_ADVISE operation MAY choose to alter its I/O and caching behavior, 1280 but is under no obligation to do so. 1282 Application specific NFS clients such as those used by hypervisors 1283 and databases can also leverage application hints to communicate 1284 their specialized requirements. 1286 6. Sparse Files 1288 6.1. Introduction 1290 A sparse file is a common way of representing a large file without 1291 having to utilize all of the disk space for it. Consequently, a 1292 sparse file uses less physical space than its size indicates. This 1293 means the file contains 'holes', byte ranges within the file that 1294 contain no data. Most modern file systems support sparse files, 1295 including most UNIX file systems and NTFS, but notably not Apple's 1296 HFS+. Common examples of sparse files include Virtual Machine (VM) 1297 OS/disk images, database files, log files, and even checkpoint 1298 recovery files most commonly used by the HPC community. 1300 In addition many modern file systems support the concept of 1301 'unwritten' or 'uninitialized' blocks, which have uninitialized space 1302 allocated to them on disk, but will return zeros until data is 1303 written to them. Such functionality is already present in the data 1304 model of the pNFS Block/Volume Layout (see [RFC5663]). Uninitialized 1305 blocks can thought as holes inside a space reservation window. 1307 If an application reads a hole in a sparse file, the file system must 1308 return all zeros to the application. For local data access there is 1309 little penalty, but with NFS these zeroes must be transferred back to 1310 the client. If an application uses the NFS client to read data into 1311 memory, this wastes time and bandwidth as the application waits for 1312 the zeroes to be transferred. 1314 A sparse file is typically created by initializing the file to be all 1315 zeros - nothing is written to the data in the file, instead the hole 1316 is recorded in the metadata for the file. So a 8G disk image might 1317 be represented initially by a couple hundred bits in the inode and 1318 nothing on the disk. If the VM then writes 100M to a file in the 1319 middle of the image, there would now be two holes represented in the 1320 metadata and 100M in the data. 1322 No new operation is needed to allow the creation of a sparsely 1323 populated file, when a file is created and a write occurs past the 1324 current size of the file, the non-allocated region will either be a 1325 hole or filled with zeros. The choice of behavior is dictated by the 1326 underlying file system and is transparent to the application. What 1327 is needed are the abilities to read sparse files and to punch holes 1328 to reinitialize the contents of a file. 1330 Two new operations DEALLOCATE (Section 15.4) and READ_PLUS 1331 (Section 15.10) are introduced. DEALLOCATE allows for the hole 1332 punching. I.e., an application might want to reset the allocation 1333 and reservation status of a range of the file. READ_PLUS supports 1334 all the features of READ but includes an extension to support sparse 1335 files. READ_PLUS is guaranteed to perform no worse than READ, and 1336 can dramatically improve performance with sparse files. READ_PLUS 1337 does not depend on pNFS protocol features, but can be used by pNFS to 1338 support sparse files. 1340 6.2. Terminology 1342 Regular file: An object of file type NF4REG or NF4NAMEDATTR. 1344 Sparse file: A Regular file that contains one or more holes. 1346 Hole: A byte range within a Sparse file that contains regions of all 1347 zeroes. A hole might or might not have space allocated or 1348 reserved to it. 1350 6.3. New Operations 1352 6.3.1. READ_PLUS 1354 READ_PLUS is a new variant of the NFSv4.1 READ operation [RFC5661]. 1355 Besides being able to support all of the data semantics of the READ 1356 operation, it can also be used by the client and server to 1357 efficiently transfer holes. Note that as the client has no a priori 1358 knowledge of whether a hole is present or not, if the client supports 1359 READ_PLUS and so does the server, then it should always use the 1360 READ_PLUS operation in preference to the READ operation. 1362 READ_PLUS extends the response with a new arm representing holes to 1363 avoid returning data for portions of the file which are initialized 1364 to zero and may or may not contain a backing store. Returning data 1365 blocks of uninitialized data wastes computational and network 1366 resources, thus reducing performance. 1368 When a client sends a READ operation, it is not prepared to accept a 1369 READ_PLUS-style response providing a compact encoding of the scope of 1370 holes. If a READ occurs on a sparse file, then the server must 1371 expand such data to be raw bytes. If a READ occurs in the middle of 1372 a hole, the server can only send back bytes starting from that 1373 offset. By contrast, if a READ_PLUS occurs in the middle of a hole, 1374 the server can send back a range which starts before the offset and 1375 extends past the range. 1377 6.3.2. DEALLOCATE 1379 DEALLOCATE can be used to hole punch, which allows the client to 1380 avoid the transfer of a repetitive pattern of zeros across the 1381 network. 1383 7. Space Reservation 1385 Applications want to be able to reserve space for a file, report the 1386 amount of actual disk space a file occupies, and free-up the backing 1387 space of a file when it is not required. 1389 One example is the posix_fallocate ([posix_fallocate]) which allows 1390 applications to ask for space reservations from the operating system, 1391 usually to provide a better file layout and reduce overhead for 1392 random or slow growing file appending workloads. 1394 Another example is space reservation for virtual disks in a 1395 hypervisor. In virtualized environments, virtual disk files are 1396 often stored on NFS mounted volumes. When a hypervisor creates a 1397 virtual disk file, it often tries to preallocate the space for the 1398 file so that there are no future allocation related errors during the 1399 operation of the virtual machine. Such errors prevent a virtual 1400 machine from continuing execution and result in downtime. 1402 Currently, in order to achieve such a guarantee, applications zero 1403 the entire file. The initial zeroing allocates the backing blocks 1404 and all subsequent writes are overwrites of already allocated blocks. 1405 This approach is not only inefficient in terms of the amount of I/O 1406 done, it is also not guaranteed to work on file systems that are log 1407 structured or deduplicated. An efficient way of guaranteeing space 1408 reservation would be beneficial to such applications. 1410 The new ALLOCATE operation (see Section 15.1) allows a client to 1411 request a guarantee that space will be available. The ALLOCATE 1412 operation guarantees that any future writes to the region it was 1413 successfully called for will not fail with NFS4ERR_NOSPC. 1415 Another useful feature is the ability to report the number of blocks 1416 that would be freed when a file is deleted. Currently, NFS reports 1417 two size attributes: 1419 size The logical file size of the file. 1421 space_used The size in bytes that the file occupies on disk 1422 While these attributes are sufficient for space accounting in 1423 traditional file systems, they prove to be inadequate in modern file 1424 systems that support block sharing. In such file systems, multiple 1425 inodes can point to a single block with a block reference count to 1426 guard against premature freeing. Having a way to tell the number of 1427 blocks that would be freed if the file was deleted would be useful to 1428 applications that wish to migrate files when a volume is low on 1429 space. 1431 Since virtual disks represent a hard drive in a virtual machine, a 1432 virtual disk can be viewed as a file system within a file. Since not 1433 all blocks within a file system are in use, there is an opportunity 1434 to reclaim blocks that are no longer in use. A call to deallocate 1435 blocks could result in better space efficiency. Lesser space MAY be 1436 consumed for backups after block deallocation. 1438 The following operations and attributes can be used to resolve these 1439 issues: 1441 space_freed This attribute specifies the space freed when a file is 1442 deleted, taking block sharing into consideration. 1444 DEALLOCATE This operation delallocates the blocks backing a region 1445 of the file. 1447 If space_used of a file is interpreted to mean the size in bytes of 1448 all disk blocks pointed to by the inode of the file, then shared 1449 blocks get double counted, over-reporting the space utilization. 1450 This also has the adverse effect that the deletion of a file with 1451 shared blocks frees up less than space_used bytes. 1453 On the other hand, if space_used is interpreted to mean the size in 1454 bytes of those disk blocks unique to the inode of the file, then 1455 shared blocks are not counted in any file, resulting in under- 1456 reporting of the space utilization. 1458 For example, two files A and B have 10 blocks each. Let 6 of these 1459 blocks be shared between them. Thus, the combined space utilized by 1460 the two files is 14 * BLOCK_SIZE bytes. In the former case, the 1461 combined space utilization of the two files would be reported as 20 * 1462 BLOCK_SIZE. However, deleting either would only result in 4 * 1463 BLOCK_SIZE being freed. Conversely, the latter interpretation would 1464 report that the space utilization is only 8 * BLOCK_SIZE. 1466 Adding another size attribute, space_freed (see Section 12.2.3), is 1467 helpful in solving this problem. space_freed is the number of blocks 1468 that are allocated to the given file that would be freed on its 1469 deletion. In the example, both A and B would report space_freed as 4 1470 * BLOCK_SIZE and space_used as 10 * BLOCK_SIZE. If A is deleted, B 1471 will report space_freed as 10 * BLOCK_SIZE as the deletion of B would 1472 result in the deallocation of all 10 blocks. 1474 The addition of these attributes does not solve the problem of space 1475 being over-reported. However, over-reporting is better than under- 1476 reporting. 1478 8. Application Data Block Support 1480 At the OS level, files are contained on disk blocks. Applications 1481 are also free to impose structure on the data contained in a file and 1482 we can define an Application Data Block (ADB) to be such a structure. 1483 From the application's viewpoint, it only wants to handle ADBs and 1484 not raw bytes (see [Strohm11]). An ADB is typically comprised of two 1485 sections: header and data. The header describes the characteristics 1486 of the block and can provide a means to detect corruption in the data 1487 payload. The data section is typically initialized to all zeros. 1489 The format of the header is application specific, but there are two 1490 main components typically encountered: 1492 1. An Application Data Block Number (ADBN) which allows the 1493 application to determine which data block is being referenced. 1494 This is useful when the client is not storing the blocks in 1495 contiguous memory, i.e., a logical block number. 1497 2. Fields to describe the state of the ADB and a means to detect 1498 block corruption. For both pieces of data, a useful property is 1499 that allowed values be unique in that if passed across the 1500 network, corruption due to translation between big and little 1501 endian architectures are detectable. For example, 0xF0DEDEF0 has 1502 the same bit pattern in both architectures. 1504 Applications already impose structures on files [Strohm11] and detect 1505 corruption in data blocks [Ashdown08]. What they are not able to do 1506 is efficiently transfer and store ADBs. To initialize a file with 1507 ADBs, the client must send each full ADB to the server and that must 1508 be stored on the server. 1510 In this section, we define a framework for transferring the ADB from 1511 client to server and present one approach to detecting corruption in 1512 a given ADB implementation. 1514 8.1. Generic Framework 1516 We want the representation of the ADB to be flexible enough to 1517 support many different applications. The most basic approach is no 1518 imposition of a block at all, which means we are working with the raw 1519 bytes. Such an approach would be useful for storing holes, punching 1520 holes, etc. In more complex deployments, a server might be 1521 supporting multiple applications, each with their own definition of 1522 the ADB. One might store the ADBN at the start of the block and then 1523 have a guard pattern to detect corruption [McDougall07]. The next 1524 might store the ADBN at an offset of 100 bytes within the block and 1525 have no guard pattern at all, i.e., existing applications might 1526 already have well defined formats for their data blocks. 1528 The guard pattern can be used to represent the state of the block, to 1529 protect against corruption, or both. Again, it needs to be able to 1530 be placed anywhere within the ADB. 1532 We need to be able to represent the starting offset of the block and 1533 the size of the block. Note that nothing prevents the application 1534 from defining different sized blocks in a file. 1536 8.1.1. Data Block Representation 1538 1540 struct app_data_block4 { 1541 offset4 adb_offset; 1542 length4 adb_block_size; 1543 length4 adb_block_count; 1544 length4 adb_reloff_blocknum; 1545 count4 adb_block_num; 1546 length4 adb_reloff_pattern; 1547 opaque adb_pattern<>; 1548 }; 1550 1552 The app_data_block4 structure captures the abstraction presented for 1553 the ADB. The additional fields present are to allow the transmission 1554 of adb_block_count ADBs at one time. We also use adb_block_num to 1555 convey the ADBN of the first block in the sequence. Each ADB will 1556 contain the same adb_pattern string. 1558 As both adb_block_num and adb_pattern are optional, if either 1559 adb_reloff_pattern or adb_reloff_blocknum is set to NFS4_UINT64_MAX, 1560 then the corresponding field is not set in any of the ADB. 1562 8.2. An Example of Detecting Corruption 1564 In this section, we define an ADB format in which corruption can be 1565 detected. Note that this is just one possible format and means to 1566 detect corruption. 1568 Consider a very basic implementation of an operating system's disk 1569 blocks. A block is either data or it is an indirect block which 1570 allows for files to be larger than one block. It is desired to be 1571 able to initialize a block. Lastly, to quickly unlink a file, a 1572 block can be marked invalid. The contents remain intact - which 1573 would enable this OS application to undelete a file. 1575 The application defines 4k sized data blocks, with an 8 byte block 1576 counter occurring at offset 0 in the block, and with the guard 1577 pattern occurring at offset 8 inside the block. Furthermore, the 1578 guard pattern can take one of four states: 1580 0xfeedface - This is the FREE state and indicates that the ADB 1581 format has been applied. 1583 0xcafedead - This is the DATA state and indicates that real data 1584 has been written to this block. 1586 0xe4e5c001 - This is the INDIRECT state and indicates that the 1587 block contains block counter numbers that are chained off of this 1588 block. 1590 0xba1ed4a3 - This is the INVALID state and indicates that the block 1591 contains data whose contents are garbage. 1593 Finally, it also defines an 8 byte checksum [Baira08] starting at 1594 byte 16 which applies to the remaining contents of the block. If the 1595 state is FREE, then that checksum is trivially zero. As such, the 1596 application has no need to transfer the checksum implicitly inside 1597 the ADB - it need not make the transfer layer aware of the fact that 1598 there is a checksum (see [Ashdown08] for an example of checksums used 1599 to detect corruption in application data blocks). 1601 Corruption in each ADB can thus be detected: 1603 o If the guard pattern is anything other than one of the allowed 1604 values, including all zeros. 1606 o If the guard pattern is FREE and any other byte in the remainder 1607 of the ADB is anything other than zero. 1609 o If the guard pattern is anything other than FREE, then if the 1610 stored checksum does not match the computed checksum. 1612 o If the guard pattern is INDIRECT and one of the stored indirect 1613 block numbers has a value greater than the number of ADBs in the 1614 file. 1616 o If the guard pattern is INDIRECT and one of the stored indirect 1617 block numbers is a duplicate of another stored indirect block 1618 number. 1620 As can be seen, the application can detect errors based on the 1621 combination of the guard pattern state and the checksum. But also, 1622 the application can detect corruption based on the state and the 1623 contents of the ADB. This last point is important in validating the 1624 minimum amount of data we incorporated into our generic framework. 1625 I.e., the guard pattern is sufficient in allowing applications to 1626 design their own corruption detection. 1628 Finally, it is important to note that none of these corruption checks 1629 occur in the transport layer. The server and client components are 1630 totally unaware of the file format and might report everything as 1631 being transferred correctly even in the case the application detects 1632 corruption. 1634 8.3. Example of READ_PLUS 1636 The hypothetical application presented in Section 8.2 can be used to 1637 illustrate how READ_PLUS would return an array of results. A file is 1638 created and initialized with 100 4k ADBs in the FREE state with the 1639 WRITE_SAME operation (see Section 15.12): 1641 WRITE_SAME {0, 4k, 100, 0, 0, 8, 0xfeedface} 1643 Further, assume the application writes a single ADB at 16k, changing 1644 the guard pattern to 0xcafedead, we would then have in memory: 1646 0k -> (4k - 1) : 00 00 00 00 fe ed fa ce 00 00 ... 00 00 1647 4k -> (8k - 1) : 00 00 00 01 fe ed fa ce 00 00 ... 00 00 1648 8k -> (12k - 1) : 00 00 00 02 fe ed fa ce 00 00 ... 00 00 1649 12k -> (16k - 1) : 00 00 00 03 fe ed fa ce 00 00 ... 00 00 1650 16k -> (20k - 1) : 00 00 00 04 ca fe de ad 00 00 ... 00 00 1651 20k -> (24k - 1) : 00 00 00 05 fe ed fa ce 00 00 ... 00 00 1652 24k -> (28k - 1) : 00 00 00 06 fe ed fa ce 00 00 ... 00 00 1653 ... 1654 396k -> (400k - 1) : 00 00 00 63 fe ed fa ce 00 00 ... 00 00 1656 And when the client did a READ_PLUS of 64k at the start of the file, 1657 it could get back a result of data: 1659 0k -> (4k - 1) : 00 00 00 00 fe ed fa ce 00 00 ... 00 00 1660 4k -> (8k - 1) : 00 00 00 01 fe ed fa ce 00 00 ... 00 00 1661 8k -> (12k - 1) : 00 00 00 02 fe ed fa ce 00 00 ... 00 00 1662 12k -> (16k - 1) : 00 00 00 03 fe ed fa ce 00 00 ... 00 00 1663 16k -> (20k - 1) : 00 00 00 04 ca fe de ad 00 00 ... 00 00 1664 20k -> (24k - 1) : 00 00 00 05 fe ed fa ce 00 00 ... 00 00 1665 24k -> (24k - 1) : 00 00 00 06 fe ed fa ce 00 00 ... 00 00 1666 ... 1667 62k -> (64k - 1) : 00 00 00 15 fe ed fa ce 00 00 ... 00 00 1669 8.4. An Example of Zeroing Space 1671 A simpler use case for WRITE_SAME are applications that want to 1672 efficiently zero out a file, but do not want to modify space 1673 reservations. This can easily be achieved by a call to WRITE_SAME 1674 without a ADB block numbers and pattern, e.g.: 1676 WRITE_SAME {0, 1k, 10000, 0, 0, 0, 0} 1678 9. Labeled NFS 1680 9.1. Introduction 1682 Access control models such as Unix permissions or Access Control 1683 Lists are commonly referred to as Discretionary Access Control (DAC) 1684 models. These systems base their access decisions on user identity 1685 and resource ownership. In contrast Mandatory Access Control (MAC) 1686 models base their access control decisions on the label on the 1687 subject (usually a process) and the object it wishes to access 1688 [RFC7204]. These labels may contain user identity information but 1689 usually contain additional information. In DAC systems users are 1690 free to specify the access rules for resources that they own. MAC 1691 models base their security decisions on a system wide policy 1692 established by an administrator or organization which the users do 1693 not have the ability to override. In this section, we add a MAC 1694 model to NFSv4.2. 1696 First we provide a method for transporting and storing security label 1697 data on NFSv4 file objects. Security labels have several semantics 1698 that are met by NFSv4 recommended attributes such as the ability to 1699 set the label value upon object creation. Access control on these 1700 attributes are done through a combination of two mechanisms. As with 1701 other recommended attributes on file objects the usual DAC checks 1702 (ACLs and permission bits) will be performed to ensure that proper 1703 file ownership is enforced. In addition a MAC system MAY be employed 1704 on the client, server, or both to enforce additional policy on what 1705 subjects may modify security label information. 1707 Second, we describe a method for the client to determine if an NFSv4 1708 file object security label has changed. A client which needs to know 1709 if a label on a file or set of files is going to change SHOULD 1710 request a delegation on each labeled file. In order to change such a 1711 security label, the server will have to recall delegations on any 1712 file affected by the label change, so informing clients of the label 1713 change. 1715 An additional useful feature would be modification to the RPC layer 1716 used by NFSv4 to allow RPC calls to carry security labels and enable 1717 full mode enforcement as described in Section 9.6.1. Such 1718 modifications are outside the scope of this document (see 1719 [rpcsec_gssv3]). 1721 9.2. Definitions 1723 Label Format Specifier (LFS): is an identifier used by the client to 1724 establish the syntactic format of the security label and the 1725 semantic meaning of its components. These specifiers exist in a 1726 registry associated with documents describing the format and 1727 semantics of the label. 1729 Label Format Registry: is the IANA registry (see [Quigley14]) 1730 containing all registered LFSes along with references to the 1731 documents that describe the syntactic format and semantics of the 1732 security label. 1734 Policy Identifier (PI): is an optional part of the definition of a 1735 Label Format Specifier which allows for clients and server to 1736 identify specific security policies. 1738 Object: is a passive resource within the system that we wish to be 1739 protected. Objects can be entities such as files, directories, 1740 pipes, sockets, and many other system resources relevant to the 1741 protection of the system state. 1743 Subject: is an active entity usually a process which is requesting 1744 access to an object. 1746 MAC-Aware: is a server which can transmit and store object labels. 1748 MAC-Functional: is a client or server which is Labeled NFS enabled. 1749 Such a system can interpret labels and apply policies based on the 1750 security system. 1752 Multi-Level Security (MLS): is a traditional model where objects are 1753 given a sensitivity level (Unclassified, Secret, Top Secret, etc) 1754 and a category set (see [BL73], [RFC1108], and [RFC2401]). 1756 9.3. MAC Security Attribute 1758 MAC models base access decisions on security attributes bound to 1759 subjects and objects. This information can range from a user 1760 identity for an identity based MAC model, sensitivity levels for 1761 Multi-level security, or a type for Type Enforcement. These models 1762 base their decisions on different criteria but the semantics of the 1763 security attribute remain the same. The semantics required by the 1764 security attributes are listed below: 1766 o MUST provide flexibility with respect to the MAC model. 1768 o MUST provide the ability to atomically set security information 1769 upon object creation. 1771 o MUST provide the ability to enforce access control decisions both 1772 on the client and the server. 1774 o MUST NOT expose an object to either the client or server name 1775 space before its security information has been bound to it. 1777 NFSv4 implements the security attribute as a recommended attribute. 1778 These attributes have a fixed format and semantics, which conflicts 1779 with the flexible nature of the security attribute. To resolve this 1780 the security attribute consists of two components. The first 1781 component is a LFS as defined in [Quigley14] to allow for 1782 interoperability between MAC mechanisms. The second component is an 1783 opaque field which is the actual security attribute data. To allow 1784 for various MAC models, NFSv4 should be used solely as a transport 1785 mechanism for the security attribute. It is the responsibility of 1786 the endpoints to consume the security attribute and make access 1787 decisions based on their respective models. In addition, creation of 1788 objects through OPEN and CREATE allows for the security attribute to 1789 be specified upon creation. By providing an atomic create and set 1790 operation for the security attribute it is possible to enforce the 1791 second and fourth requirements. The recommended attribute 1792 FATTR4_SEC_LABEL (see Section 12.2.2) will be used to satisfy this 1793 requirement. 1795 9.3.1. Delegations 1797 In the event that a security attribute is changed on the server while 1798 a client holds a delegation on the file, both the server and the 1799 client MUST follow the NFSv4.1 protocol (see Chapter 10 of [RFC5661]) 1800 with respect to attribute changes. It SHOULD flush all changes back 1801 to the server and relinquish the delegation. 1803 9.3.2. Permission Checking 1805 It is not feasible to enumerate all possible MAC models and even 1806 levels of protection within a subset of these models. This means 1807 that the NFSv4 client and servers cannot be expected to directly make 1808 access control decisions based on the security attribute. Instead 1809 NFSv4 should defer permission checking on this attribute to the host 1810 system. These checks are performed in addition to existing DAC and 1811 ACL checks outlined in the NFSv4 protocol. Section 9.6 gives a 1812 specific example of how the security attribute is handled under a 1813 particular MAC model. 1815 9.3.3. Object Creation 1817 When creating files in NFSv4 the OPEN and CREATE operations are used. 1818 One of the parameters to these operations is an fattr4 structure 1819 containing the attributes the file is to be created with. This 1820 allows NFSv4 to atomically set the security attribute of files upon 1821 creation. When a client is MAC-Functional it must always provide the 1822 initial security attribute upon file creation. In the event that the 1823 server is MAC-Functional as well, it should determine by policy 1824 whether it will accept the attribute from the client or instead make 1825 the determination itself. If the client is not MAC-Functional, then 1826 the MAC-Functional server must decide on a default label. A more in 1827 depth explanation can be found in Section 9.6. 1829 9.3.4. Existing Objects 1831 Note that under the MAC model, all objects must have labels. 1832 Therefore, if an existing server is upgraded to include Labeled NFS 1833 support, then it is the responsibility of the security system to 1834 define the behavior for existing objects. 1836 9.3.5. Label Changes 1838 Consider a guest mode system (Section 9.6.2) in which the clients 1839 enforce MAC checks and the server has only a DAC security system 1840 which stores the labels along with the file data. In this type of 1841 system, a user with the appropriate DAC credentials on a client with 1842 poorly configured or disabled MAC labeling enforcement is allowed 1843 access to the file label (and data) on the server and can change the 1844 label. 1846 Clients which need to know if a label on a file or set of files has 1847 changed SHOULD request a delegation on each labeled file so that a 1848 label change by another client will be known via the process 1849 described in Section 9.3.1 which must be followed: the delegation 1850 will be recalled, which effectively notifies the client of the 1851 change. 1853 Note that the MAC security policies on a client can be such that the 1854 client does not have access to the file unless it has a delegation. 1856 9.4. pNFS Considerations 1858 The new FATTR4_SEC_LABEL attribute is metadata information and as 1859 such the DS is not aware of the value contained on the MDS. 1860 Fortunately, the NFSv4.1 protocol [RFC5661] already has provisions 1861 for doing access level checks from the DS to the MDS. In order for 1862 the DS to validate the subject label presented by the client, it 1863 SHOULD utilize this mechanism. 1865 9.5. Discovery of Server Labeled NFS Support 1867 The server can easily determine that a client supports Labeled NFS 1868 when it queries for the FATTR4_SEC_LABEL label for an object. The 1869 client might need to discover which LFS the server supports. 1871 The following compound MUST NOT be denied by any MAC label check: 1873 PUTROOTFH, GETATTR {FATTR4_SEC_LABEL} 1875 Note that the server might have imposed a security flavor on the root 1876 that precludes such access. I.e., if the server requires kerberized 1877 access and the client presents a compound with AUTH_SYS, then the 1878 server is allowed to return NFS4ERR_WRONGSEC in this case. But if 1879 the client presents a correct security flavor, then the server MUST 1880 return the FATTR4_SEC_LABEL attribute with the supported LFS filled 1881 in. 1883 9.6. MAC Security NFS Modes of Operation 1885 A system using Labeled NFS may operate in two modes. The first mode 1886 provides the most protection and is called "full mode". In this mode 1887 both the client and server implement a MAC model allowing each end to 1888 make an access control decision. The remaining mode is called the 1889 "guest mode" and in this mode one end of the connection is not 1890 implementing a MAC model and thus offers less protection than full 1891 mode. 1893 9.6.1. Full Mode 1895 Full mode environments consist of MAC-Functional NFSv4 servers and 1896 clients and may be composed of mixed MAC models and policies. The 1897 system requires that both the client and server have an opportunity 1898 to perform an access control check based on all relevant information 1899 within the network. The file object security attribute is provided 1900 using the mechanism described in Section 9.3. 1902 Fully MAC-Functional NFSv4 servers are not possible in the absence of 1903 RPCSEC_GSSv3 [rpcsec_gssv3] support for subject label transport. 1904 However, servers may make decisions based on the RPC credential 1905 information available. 1907 9.6.1.1. Initial Labeling and Translation 1909 The ability to create a file is an action that a MAC model may wish 1910 to mediate. The client is given the responsibility to determine the 1911 initial security attribute to be placed on a file. This allows the 1912 client to make a decision as to the acceptable security attributes to 1913 create a file with before sending the request to the server. Once 1914 the server receives the creation request from the client it may 1915 choose to evaluate if the security attribute is acceptable. 1917 Security attributes on the client and server may vary based on MAC 1918 model and policy. To handle this the security attribute field has an 1919 LFS component. This component is a mechanism for the host to 1920 identify the format and meaning of the opaque portion of the security 1921 attribute. A full mode environment may contain hosts operating in 1922 several different LFSes. In this case a mechanism for translating 1923 the opaque portion of the security attribute is needed. The actual 1924 translation function will vary based on MAC model and policy and is 1925 out of the scope of this document. If a translation is unavailable 1926 for a given LFS then the request MUST be denied. Another recourse is 1927 to allow the host to provide a fallback mapping for unknown security 1928 attributes. 1930 9.6.1.2. Policy Enforcement 1932 In full mode access control decisions are made by both the clients 1933 and servers. When a client makes a request it takes the security 1934 attribute from the requesting process and makes an access control 1935 decision based on that attribute and the security attribute of the 1936 object it is trying to access. If the client denies that access an 1937 RPC call to the server is never made. If however the access is 1938 allowed the client will make a call to the NFS server. 1940 When the server receives the request from the client it uses any 1941 credential information conveyed in the RPC request and the attributes 1942 of the object the client is trying to access to make an access 1943 control decision. If the server's policy allows this access it will 1944 fulfill the client's request, otherwise it will return 1945 NFS4ERR_ACCESS. 1947 Future protocol extensions may also allow the server to factor into 1948 the decision a security label extracted from the RPC request. 1950 Implementations MAY validate security attributes supplied over the 1951 network to ensure that they are within a set of attributes permitted 1952 from a specific peer, and if not, reject them. Note that a system 1953 may permit a different set of attributes to be accepted from each 1954 peer. 1956 9.6.1.3. Limited Server 1958 A Limited Server mode (see Section 4.2 of [RFC7204]) consists of a 1959 server which is label aware, but does not enforce policies. Such a 1960 server will store and retrieve all object labels presented by 1961 clients, utilize the methods described in Section 9.3.5 to allow the 1962 clients to detect changing labels, but may not factor the label into 1963 access decisions. Instead, it will expect the clients to enforce all 1964 such access locally. 1966 9.6.2. Guest Mode 1968 Guest mode implies that either the client or the server does not 1969 handle labels. If the client is not Labeled NFS aware, then it will 1970 not offer subject labels to the server. The server is the only 1971 entity enforcing policy, and may selectively provide standard NFS 1972 services to clients based on their authentication credentials and/or 1973 associated network attributes (e.g., IP address, network interface). 1974 The level of trust and access extended to a client in this mode is 1975 configuration-specific. If the server is not Labeled NFS aware, then 1976 it will not return object labels to the client. Clients in this 1977 environment are may consist of groups implementing different MAC 1978 model policies. The system requires that all clients in the 1979 environment be responsible for access control checks. 1981 9.7. Security Considerations for Labeled NFS 1983 This entire chapter deals with security issues. 1985 Depending on the level of protection the MAC system offers there may 1986 be a requirement to tightly bind the security attribute to the data. 1988 When only one of the client or server enforces labels, it is 1989 important to realize that the other side is not enforcing MAC 1990 protections. Alternate methods might be in use to handle the lack of 1991 MAC support and care should be taken to identify and mitigate threats 1992 from possible tampering outside of these methods. 1994 An example of this is that a server that modifies READDIR or LOOKUP 1995 results based on the client's subject label might want to always 1996 construct the same subject label for a client which does not present 1997 one. This will prevent a non-Labeled NFS client from mixing entries 1998 in the directory cache. 2000 10. Sharing change attribute implementation details with NFSv4 clients 2002 Although both the NFSv4 [I-D.ietf-nfsv4-rfc3530bis] and NFSv4.1 2003 protocol [RFC5661], define the change attribute as being mandatory to 2004 implement, there is little in the way of guidance as to its 2005 construction. The only mandated constraint is that the value must 2006 change whenever the file data or metadata change. 2008 While this allows for a wide range of implementations, it also leaves 2009 the client with no way to determine which is the most recent value 2010 for the change attribute in a case where several RPC calls have been 2011 issued in parallel. In other words if two COMPOUNDs, both containing 2012 WRITE and GETATTR requests for the same file, have been issued in 2013 parallel, how does the client determine which of the two change 2014 attribute values returned in the replies to the GETATTR requests 2015 correspond to the most recent state of the file? In some cases, the 2016 only recourse may be to send another COMPOUND containing a third 2017 GETATTR that is fully serialized with the first two. 2019 NFSv4.2 avoids this kind of inefficiency by allowing the server to 2020 share details about how the change attribute is expected to evolve, 2021 so that the client may immediately determine which, out of the 2022 several change attribute values returned by the server, is the most 2023 recent. change_attr_type is defined as a new recommended attribute 2024 (see Section 12.2.1), and is per file system. 2026 11. Error Values 2028 NFS error numbers are assigned to failed operations within a Compound 2029 (COMPOUND or CB_COMPOUND) request. A Compound request contains a 2030 number of NFS operations that have their results encoded in sequence 2031 in a Compound reply. The results of successful operations will 2032 consist of an NFS4_OK status followed by the encoded results of the 2033 operation. If an NFS operation fails, an error status will be 2034 entered in the reply and the Compound request will be terminated. 2036 11.1. Error Definitions 2038 Protocol Error Definitions 2040 +-------------------------+--------+------------------+ 2041 | Error | Number | Description | 2042 +-------------------------+--------+------------------+ 2043 | NFS4ERR_BADLABEL | 10093 | Section 11.1.3.1 | 2044 | NFS4ERR_OFFLOAD_DENIED | 10091 | Section 11.1.2.1 | 2045 | NFS4ERR_OFFLOAD_NO_REQS | 10094 | Section 11.1.2.2 | 2046 | NFS4ERR_PARTNER_NO_AUTH | 10089 | Section 11.1.2.3 | 2047 | NFS4ERR_PARTNER_NOTSUPP | 10088 | Section 11.1.2.4 | 2048 | NFS4ERR_UNION_NOTSUPP | 10090 | Section 11.1.1.1 | 2049 | NFS4ERR_WRONG_LFS | 10092 | Section 11.1.3.2 | 2050 +-------------------------+--------+------------------+ 2052 Table 1 2054 11.1.1. General Errors 2056 This section deals with errors that are applicable to a broad set of 2057 different purposes. 2059 11.1.1.1. NFS4ERR_UNION_NOTSUPP (Error Code 10090) 2061 One of the arguments to the operation is a discriminated union and 2062 while the server supports the given operation, it does not support 2063 the selected arm of the discriminated union. 2065 11.1.2. Server to Server Copy Errors 2067 These errors deal with the interaction between server to server 2068 copies. 2070 11.1.2.1. NFS4ERR_OFFLOAD_DENIED (Error Code 10091) 2072 The copy offload operation is supported by both the source and the 2073 destination, but the destination is not allowing it for this file. 2074 If the client sees this error, it should fall back to the normal copy 2075 semantics. 2077 11.1.2.2. NFS4ERR_OFFLOAD_NO_REQS (Error Code 10094) 2079 The copy offload operation is supported by both the source and the 2080 destination, but the destination can not meet the client requirements 2081 for either consecutive byte copy or synchronous copy. If the client 2082 sees this error, it should either relax the requirements (if any) or 2083 fall back to the normal copy semantics. 2085 11.1.2.3. NFS4ERR_PARTNER_NO_AUTH (Error Code 10089) 2087 The source server does not authorize a server-to-server copy offload 2088 operation. This may be due to the client's failure to send the 2089 COPY_NOTIFY operation to the source server, the source server 2090 receiving a server-to-server copy offload request after the copy 2091 lease time expired, or for some other permission problem. 2093 11.1.2.4. NFS4ERR_PARTNER_NOTSUPP (Error Code 10088) 2095 The remote server does not support the server-to-server copy offload 2096 protocol. 2098 11.1.3. Labeled NFS Errors 2100 These errors are used in Labeled NFS. 2102 11.1.3.1. NFS4ERR_BADLABEL (Error Code 10093) 2104 The label specified is invalid in some manner. 2106 11.1.3.2. NFS4ERR_WRONG_LFS (Error Code 10092) 2108 The LFS specified in the subject label is not compatible with the LFS 2109 in the object label. 2111 11.2. New Operations and Their Valid Errors 2113 This section contains a table that gives the valid error returns for 2114 each new NFSv4.2 protocol operation. The error code NFS4_OK 2115 (indicating no error) is not listed but should be understood to be 2116 returnable by all new operations. The error values for all other 2117 operations are defined in Section 15.2 of [RFC5661]. 2119 Valid Error Returns for Each New Protocol Operation 2121 +----------------+--------------------------------------------------+ 2122 | Operation | Errors | 2123 +----------------+--------------------------------------------------+ 2124 | ALLOCATE | NFS4ERR_ACCESS, NFS4ERR_ADMIN_REVOKED, | 2125 | | NFS4ERR_BADXDR, NFS4ERR_BAD_STATEID, | 2126 | | NFS4ERR_DEADSESSION, NFS4ERR_DELAY, | 2127 | | NFS4ERR_DELEG_REVOKED, NFS4ERR_DQUOT, | 2128 | | NFS4ERR_EXPIRED, NFS4ERR_FBIG, | 2129 | | NFS4ERR_FHEXPIRED, NFS4ERR_GRACE, NFS4ERR_INVAL, | 2130 | | NFS4ERR_IO, NFS4ERR_ISDIR, NFS4ERR_MOVED, | 2131 | | NFS4ERR_NOFILEHANDLE, NFS4ERR_NOSPC, | 2132 | | NFS4ERR_NOTSUPP, NFS4ERR_OLD_STATEID, | 2133 | | NFS4ERR_OPENMODE, NFS4ERR_OP_NOT_IN_SESSION, | 2134 | | NFS4ERR_REP_TOO_BIG, | 2135 | | NFS4ERR_REP_TOO_BIG_TO_CACHE, | 2136 | | NFS4ERR_REQ_TOO_BIG, NFS4ERR_RETRY_UNCACHED_REP, | 2137 | | NFS4ERR_ROFS, NFS4ERR_SERVERFAULT, | 2138 | | NFS4ERR_STALE, NFS4ERR_SYMLINK, | 2139 | | NFS4ERR_TOO_MANY_OPS, NFS4ERR_WRONG_TYPE | 2140 | COPY | NFS4ERR_ACCESS, NFS4ERR_ADMIN_REVOKED, | 2141 | | NFS4ERR_BADXDR, NFS4ERR_BAD_STATEID, | 2142 | | NFS4ERR_DEADSESSION, NFS4ERR_DELAY, | 2143 | | NFS4ERR_DELEG_REVOKED, NFS4ERR_DQUOT, | 2144 | | NFS4ERR_EXPIRED, NFS4ERR_FBIG, | 2145 | | NFS4ERR_FHEXPIRED, NFS4ERR_GRACE, NFS4ERR_INVAL, | 2146 | | NFS4ERR_IO, NFS4ERR_ISDIR, NFS4ERR_LOCKED, | 2147 | | NFS4ERR_METADATA_NOTSUPP, NFS4ERR_MOVED, | 2148 | | NFS4ERR_NOFILEHANDLE, NFS4ERR_NOSPC, | 2149 | | NFS4ERR_OFFLOAD_DENIED, NFS4ERR_OLD_STATEID, | 2150 | | NFS4ERR_OPENMODE, NFS4ERR_OP_NOT_IN_SESSION, | 2151 | | NFS4ERR_PARTNER_NO_AUTH, | 2152 | | NFS4ERR_PARTNER_NOTSUPP, NFS4ERR_PNFS_IO_HOLE, | 2153 | | NFS4ERR_PNFS_NO_LAYOUT, NFS4ERR_REP_TOO_BIG, | 2154 | | NFS4ERR_REP_TOO_BIG_TO_CACHE, | 2155 | | NFS4ERR_REQ_TOO_BIG, NFS4ERR_RETRY_UNCACHED_REP, | 2156 | | NFS4ERR_ROFS, NFS4ERR_SERVERFAULT, | 2157 | | NFS4ERR_STALE, NFS4ERR_SYMLINK, | 2158 | | NFS4ERR_TOO_MANY_OPS, NFS4ERR_WRONG_TYPE | 2159 | COPY_NOTIFY | NFS4ERR_ACCESS, NFS4ERR_ADMIN_REVOKED, | 2160 | | NFS4ERR_BADXDR, NFS4ERR_BAD_STATEID, | 2161 | | NFS4ERR_DEADSESSION, NFS4ERR_DELAY, | 2162 | | NFS4ERR_DELEG_REVOKED, NFS4ERR_EXPIRED, | 2163 | | NFS4ERR_FHEXPIRED, NFS4ERR_GRACE, NFS4ERR_INVAL, | 2164 | | NFS4ERR_ISDIR, NFS4ERR_IO, NFS4ERR_LOCKED, | 2165 | | NFS4ERR_MOVED, NFS4ERR_NOFILEHANDLE, | 2166 | | NFS4ERR_OLD_STATEID, NFS4ERR_OPENMODE, | 2167 | | NFS4ERR_OP_NOT_IN_SESSION, NFS4ERR_PNFS_IO_HOLE, | 2168 | | NFS4ERR_PNFS_NO_LAYOUT, NFS4ERR_REP_TOO_BIG, | 2169 | | NFS4ERR_REP_TOO_BIG_TO_CACHE, | 2170 | | NFS4ERR_REQ_TOO_BIG, NFS4ERR_RETRY_UNCACHED_REP, | 2171 | | NFS4ERR_SERVERFAULT, NFS4ERR_STALE, | 2172 | | NFS4ERR_SYMLINK, NFS4ERR_TOO_MANY_OPS, | 2173 | | NFS4ERR_WRONG_TYPE | 2174 | DEALLOCATE | NFS4ERR_ACCESS, NFS4ERR_ADMIN_REVOKED, | 2175 | | NFS4ERR_BADXDR, NFS4ERR_BAD_STATEID, | 2176 | | NFS4ERR_DEADSESSION, NFS4ERR_DELAY, | 2177 | | NFS4ERR_DELEG_REVOKED, NFS4ERR_EXPIRED, | 2178 | | NFS4ERR_FBIG, NFS4ERR_FHEXPIRED, NFS4ERR_GRACE, | 2179 | | NFS4ERR_INVAL, NFS4ERR_IO, NFS4ERR_ISDIR, | 2180 | | NFS4ERR_MOVED, NFS4ERR_NOFILEHANDLE, | 2181 | | NFS4ERR_NOTSUPP, NFS4ERR_OLD_STATEID, | 2182 | | NFS4ERR_OPENMODE, NFS4ERR_OP_NOT_IN_SESSION, | 2183 | | NFS4ERR_REP_TOO_BIG, | 2184 | | NFS4ERR_REP_TOO_BIG_TO_CACHE, | 2185 | | NFS4ERR_REQ_TOO_BIG, NFS4ERR_RETRY_UNCACHED_REP, | 2186 | | NFS4ERR_ROFS, NFS4ERR_SERVERFAULT, | 2187 | | NFS4ERR_STALE, NFS4ERR_SYMLINK, | 2188 | | NFS4ERR_TOO_MANY_OPS, NFS4ERR_WRONG_TYPE | 2189 | GETDEVICELIST | NFS4ERR_NOTSUPP | 2190 | LAYOUTERROR | NFS4ERR_ADMIN_REVOKED, NFS4ERR_BADXDR, | 2191 | | NFS4ERR_BAD_STATEID, NFS4ERR_DEADSESSION, | 2192 | | NFS4ERR_DELAY, NFS4ERR_DELEG_REVOKED, | 2193 | | NFS4ERR_EXPIRED, NFS4ERR_FHEXPIRED, | 2194 | | NFS4ERR_GRACE, NFS4ERR_INVAL, NFS4ERR_ISDIR, | 2195 | | NFS4ERR_MOVED, NFS4ERR_NOFILEHANDLE, | 2196 | | NFS4ERR_NOTSUPP, NFS4ERR_NO_GRACE, | 2197 | | NFS4ERR_OLD_STATEID, NFS4ERR_OP_NOT_IN_SESSION, | 2198 | | NFS4ERR_REP_TOO_BIG, | 2199 | | NFS4ERR_REP_TOO_BIG_TO_CACHE, | 2200 | | NFS4ERR_REQ_TOO_BIG, NFS4ERR_RETRY_UNCACHED_REP, | 2201 | | NFS4ERR_SERVERFAULT, NFS4ERR_STALE, | 2202 | | NFS4ERR_TOO_MANY_OPS, | 2203 | | NFS4ERR_UNKNOWN_LAYOUTTYPE, NFS4ERR_WRONG_CRED, | 2204 | | NFS4ERR_WRONG_TYPE | 2205 | LAYOUTSTATS | NFS4ERR_ADMIN_REVOKED, NFS4ERR_BADXDR, | 2206 | | NFS4ERR_BAD_STATEID, NFS4ERR_DEADSESSION, | 2207 | | NFS4ERR_DELAY, NFS4ERR_DELEG_REVOKED, | 2208 | | NFS4ERR_EXPIRED, NFS4ERR_FHEXPIRED, | 2209 | | NFS4ERR_GRACE, NFS4ERR_INVAL, NFS4ERR_ISDIR, | 2210 | | NFS4ERR_MOVED, NFS4ERR_NOFILEHANDLE, | 2211 | | NFS4ERR_NOTSUPP, NFS4ERR_NO_GRACE, | 2212 | | NFS4ERR_OLD_STATEID, NFS4ERR_OP_NOT_IN_SESSION, | 2213 | | NFS4ERR_REP_TOO_BIG, | 2214 | | NFS4ERR_REP_TOO_BIG_TO_CACHE, | 2215 | | NFS4ERR_REQ_TOO_BIG, NFS4ERR_RETRY_UNCACHED_REP, | 2216 | | NFS4ERR_SERVERFAULT, NFS4ERR_STALE, | 2217 | | NFS4ERR_TOO_MANY_OPS, | 2218 | | NFS4ERR_UNKNOWN_LAYOUTTYPE, NFS4ERR_WRONG_CRED, | 2219 | | NFS4ERR_WRONG_TYPE | 2220 | OFFLOAD_CANCEL | NFS4ERR_ADMIN_REVOKED, NFS4ERR_BADXDR, | 2221 | | NFS4ERR_BAD_STATEID, NFS4ERR_COMPLETE_ALREADY, | 2222 | | NFS4ERR_DEADSESSION, NFS4ERR_EXPIRED, | 2223 | | NFS4ERR_DELAY, NFS4ERR_GRACE, NFS4ERR_NOTSUPP, | 2224 | | NFS4ERR_OLD_STATEID, NFS4ERR_OP_NOT_IN_SESSION, | 2225 | | NFS4ERR_SERVERFAULT, NFS4ERR_TOO_MANY_OPS | 2226 | OFFLOAD_STATUS | NFS4ERR_ADMIN_REVOKED, NFS4ERR_BADXDR, | 2227 | | NFS4ERR_BAD_STATEID, NFS4ERR_COMPLETE_ALREADY, | 2228 | | NFS4ERR_DEADSESSION, NFS4ERR_EXPIRED, | 2229 | | NFS4ERR_DELAY, NFS4ERR_GRACE, NFS4ERR_NOTSUPP, | 2230 | | NFS4ERR_OLD_STATEID, NFS4ERR_OP_NOT_IN_SESSION, | 2231 | | NFS4ERR_SERVERFAULT, NFS4ERR_TOO_MANY_OPS | 2232 | READ_PLUS | NFS4ERR_ACCESS, NFS4ERR_ADMIN_REVOKED, | 2233 | | NFS4ERR_BADXDR, NFS4ERR_BAD_STATEID, | 2234 | | NFS4ERR_DEADSESSION, NFS4ERR_DELAY, | 2235 | | NFS4ERR_DELEG_REVOKED, NFS4ERR_EXPIRED, | 2236 | | NFS4ERR_FHEXPIRED, NFS4ERR_GRACE, NFS4ERR_INVAL, | 2237 | | NFS4ERR_ISDIR, NFS4ERR_IO, NFS4ERR_LOCKED, | 2238 | | NFS4ERR_MOVED, NFS4ERR_NOFILEHANDLE, | 2239 | | NFS4ERR_NOTSUPP, NFS4ERR_OLD_STATEID, | 2240 | | NFS4ERR_OPENMODE, NFS4ERR_OP_NOT_IN_SESSION, | 2241 | | NFS4ERR_PNFS_IO_HOLE, NFS4ERR_PNFS_NO_LAYOUT, | 2242 | | NFS4ERR_REP_TOO_BIG, | 2243 | | NFS4ERR_REP_TOO_BIG_TO_CACHE, | 2244 | | NFS4ERR_REQ_TOO_BIG, NFS4ERR_RETRY_UNCACHED_REP, | 2245 | | NFS4ERR_SERVERFAULT, NFS4ERR_STALE, | 2246 | | NFS4ERR_SYMLINK, NFS4ERR_TOO_MANY_OPS, | 2247 | | NFS4ERR_WRONG_TYPE | 2248 | SEEK | NFS4ERR_ACCESS, NFS4ERR_ADMIN_REVOKED, | 2249 | | NFS4ERR_BADXDR, NFS4ERR_BAD_STATEID, | 2250 | | NFS4ERR_DEADSESSION, NFS4ERR_DELAY, | 2251 | | NFS4ERR_DELEG_REVOKED, NFS4ERR_EXPIRED, | 2252 | | NFS4ERR_FHEXPIRED, NFS4ERR_GRACE, NFS4ERR_INVAL, | 2253 | | NFS4ERR_ISDIR, NFS4ERR_IO, NFS4ERR_LOCKED, | 2254 | | NFS4ERR_MOVED, NFS4ERR_NOFILEHANDLE, | 2255 | | NFS4ERR_NOTSUPP, NFS4ERR_OLD_STATEID, | 2256 | | NFS4ERR_OPENMODE, NFS4ERR_OP_NOT_IN_SESSION, | 2257 | | NFS4ERR_PNFS_IO_HOLE, NFS4ERR_PNFS_NO_LAYOUT, | 2258 | | NFS4ERR_REP_TOO_BIG, | 2259 | | NFS4ERR_REP_TOO_BIG_TO_CACHE, | 2260 | | NFS4ERR_REQ_TOO_BIG, NFS4ERR_RETRY_UNCACHED_REP, | 2261 | | NFS4ERR_SERVERFAULT, NFS4ERR_STALE, | 2262 | | NFS4ERR_SYMLINK, NFS4ERR_TOO_MANY_OPS, | 2263 | | NFS4ERR_UNION_NOTSUPP, NFS4ERR_WRONG_TYPE | 2264 | WRITE_SAME | NFS4ERR_ACCESS, NFS4ERR_ADMIN_REVOKED, | 2265 | | NFS4ERR_BADXDR, NFS4ERR_BAD_STATEID, | 2266 | | NFS4ERR_DEADSESSION, NFS4ERR_DELAY, | 2267 | | NFS4ERR_DELEG_REVOKED, NFS4ERR_DQUOT, | 2268 | | NFS4ERR_EXPIRED, NFS4ERR_FBIG, | 2269 | | NFS4ERR_FHEXPIRED, NFS4ERR_GRACE, NFS4ERR_INVAL, | 2270 | | NFS4ERR_IO, NFS4ERR_ISDIR, NFS4ERR_LOCKED, | 2271 | | NFS4ERR_MOVED, NFS4ERR_NOFILEHANDLE, | 2272 | | NFS4ERR_NOSPC, NFS4ERR_NOTSUPP, | 2273 | | NFS4ERR_OLD_STATEID, NFS4ERR_OPENMODE, | 2274 | | NFS4ERR_OP_NOT_IN_SESSION, NFS4ERR_PNFS_IO_HOLE, | 2275 | | NFS4ERR_PNFS_NO_LAYOUT, NFS4ERR_REP_TOO_BIG, | 2276 | | NFS4ERR_REP_TOO_BIG_TO_CACHE, | 2277 | | NFS4ERR_REQ_TOO_BIG, NFS4ERR_RETRY_UNCACHED_REP, | 2278 | | NFS4ERR_ROFS, NFS4ERR_SERVERFAULT, | 2279 | | NFS4ERR_STALE, NFS4ERR_SYMLINK, | 2280 | | NFS4ERR_TOO_MANY_OPS, NFS4ERR_WRONG_TYPE | 2281 +----------------+--------------------------------------------------+ 2283 Table 2 2285 11.3. New Callback Operations and Their Valid Errors 2287 This section contains a table that gives the valid error returns for 2288 each new NFSv4.2 callback operation. The error code NFS4_OK 2289 (indicating no error) is not listed but should be understood to be 2290 returnable by all new callback operations. The error values for all 2291 other callback operations are defined in Section 15.3 of [RFC5661]. 2293 Valid Error Returns for Each New Protocol Callback Operation 2295 +------------+------------------------------------------------------+ 2296 | Callback | Errors | 2297 | Operation | | 2298 +------------+------------------------------------------------------+ 2299 | CB_OFFLOAD | NFS4ERR_BADHANDLE, NFS4ERR_BADXDR, | 2300 | | NFS4ERR_BAD_STATEID, NFS4ERR_DELAY, | 2301 | | NFS4ERR_OP_NOT_IN_SESSION, NFS4ERR_REP_TOO_BIG, | 2302 | | NFS4ERR_REP_TOO_BIG_TO_CACHE, NFS4ERR_REQ_TOO_BIG, | 2303 | | NFS4ERR_RETRY_UNCACHED_REP, NFS4ERR_SERVERFAULT, | 2304 | | NFS4ERR_TOO_MANY_OPS | 2305 +------------+------------------------------------------------------+ 2307 Table 3 2309 12. New File Attributes 2311 12.1. New RECOMMENDED Attributes - List and Definition References 2313 The list of new RECOMMENDED attributes appears in Table 4. The 2314 meaning of the columns of the table are: 2316 Name: The name of the attribute. 2318 Id: The number assigned to the attribute. In the event of conflicts 2319 between the assigned number and [NFSv42xdr], the latter is likely 2320 authoritative, but should be resolved with Errata to this document 2321 and/or [NFSv42xdr]. See [IESG08] for the Errata process. 2323 Data Type: The XDR data type of the attribute. 2325 Acc: Access allowed to the attribute. 2327 R means read-only (GETATTR may retrieve, SETATTR may not set). 2329 W means write-only (SETATTR may set, GETATTR may not retrieve). 2331 R W means read/write (GETATTR may retrieve, SETATTR may set). 2333 Defined in: The section of this specification that describes the 2334 attribute. 2336 +------------------+----+-------------------+-----+----------------+ 2337 | Name | Id | Data Type | Acc | Defined in | 2338 +------------------+----+-------------------+-----+----------------+ 2339 | space_freed | 77 | length4 | R | Section 12.2.3 | 2340 | change_attr_type | 78 | change_attr_type4 | R | Section 12.2.1 | 2341 | sec_label | 79 | sec_label4 | R W | Section 12.2.2 | 2342 +------------------+----+-------------------+-----+----------------+ 2344 Table 4 2346 12.2. Attribute Definitions 2348 12.2.1. Attribute 78: change_attr_type 2350 2352 enum change_attr_type4 { 2353 NFS4_CHANGE_TYPE_IS_MONOTONIC_INCR = 0, 2354 NFS4_CHANGE_TYPE_IS_VERSION_COUNTER = 1, 2355 NFS4_CHANGE_TYPE_IS_VERSION_COUNTER_NOPNFS = 2, 2356 NFS4_CHANGE_TYPE_IS_TIME_METADATA = 3, 2357 NFS4_CHANGE_TYPE_IS_UNDEFINED = 4 2358 }; 2360 2362 change_attr_type is a per file system attribute which enables the 2363 NFSv4.2 server to provide additional information about how it expects 2364 the change attribute value to evolve after the file data, or metadata 2365 has changed. While Section 5.4 of [RFC5661] discusses per file 2366 system attributes, it is expected that the value of change_attr_type 2367 not depend on the value of "homogeneous" and only changes in the 2368 event of a migration. 2370 NFS4_CHANGE_TYPE_IS_UNDEFINED: The change attribute does not take 2371 values that fit into any of these categories. 2373 NFS4_CHANGE_TYPE_IS_MONOTONIC_INCR: The change attribute value MUST 2374 monotonically increase for every atomic change to the file 2375 attributes, data, or directory contents. 2377 NFS4_CHANGE_TYPE_IS_VERSION_COUNTER: The change attribute value MUST 2378 be incremented by one unit for every atomic change to the file 2379 attributes, data, or directory contents. This property is 2380 preserved when writing to pNFS data servers. 2382 NFS4_CHANGE_TYPE_IS_VERSION_COUNTER_NOPNFS: The change attribute 2383 value MUST be incremented by one unit for every atomic change to 2384 the file attributes, data, or directory contents. In the case 2385 where the client is writing to pNFS data servers, the number of 2386 increments is not guaranteed to exactly match the number of 2387 writes. 2389 NFS4_CHANGE_TYPE_IS_TIME_METADATA: The change attribute is 2390 implemented as suggested in [I-D.ietf-nfsv4-rfc3530bis] in terms 2391 of the time_metadata attribute. 2393 If either NFS4_CHANGE_TYPE_IS_MONOTONIC_INCR, 2394 NFS4_CHANGE_TYPE_IS_VERSION_COUNTER, or 2395 NFS4_CHANGE_TYPE_IS_TIME_METADATA are set, then the client knows at 2396 the very least that the change attribute is monotonically increasing, 2397 which is sufficient to resolve the question of which value is the 2398 most recent. 2400 If the client sees the value NFS4_CHANGE_TYPE_IS_TIME_METADATA, then 2401 by inspecting the value of the 'time_delta' attribute it additionally 2402 has the option of detecting rogue server implementations that use 2403 time_metadata in violation of the spec. 2405 If the client sees NFS4_CHANGE_TYPE_IS_VERSION_COUNTER, it has the 2406 ability to predict what the resulting change attribute value should 2407 be after a COMPOUND containing a SETATTR, WRITE, or CREATE. This 2408 again allows it to detect changes made in parallel by another client. 2409 The value NFS4_CHANGE_TYPE_IS_VERSION_COUNTER_NOPNFS permits the 2410 same, but only if the client is not doing pNFS WRITEs. 2412 Finally, if the server does not support change_attr_type or if 2413 NFS4_CHANGE_TYPE_IS_UNDEFINED is set, then the server SHOULD make an 2414 effort to implement the change attribute in terms of the 2415 time_metadata attribute. 2417 12.2.2. Attribute 79: sec_label 2419 2421 typedef uint32_t policy4; 2423 struct labelformat_spec4 { 2424 policy4 lfs_lfs; 2425 policy4 lfs_pi; 2426 }; 2428 struct sec_label4 { 2429 labelformat_spec4 slai_lfs; 2430 opaque slai_data<>; 2431 }; 2433 2435 The FATTR4_SEC_LABEL contains an array of two components with the 2436 first component being an LFS. It serves to provide the receiving end 2437 with the information necessary to translate the security attribute 2438 into a form that is usable by the endpoint. Label Formats assigned 2439 an LFS may optionally choose to include a Policy Identifier field to 2440 allow for complex policy deployments. The LFS and Label Format 2441 Registry are described in detail in [Quigley14]. The translation 2442 used to interpret the security attribute is not specified as part of 2443 the protocol as it may depend on various factors. The second 2444 component is an opaque section which contains the data of the 2445 attribute. This component is dependent on the MAC model to interpret 2446 and enforce. 2448 In particular, it is the responsibility of the LFS specification to 2449 define a maximum size for the opaque section, slai_data<>. When 2450 creating or modifying a label for an object, the client needs to be 2451 guaranteed that the server will accept a label that is sized 2452 correctly. By both client and server being part of a specific MAC 2453 model, the client will be aware of the size. 2455 12.2.3. Attribute 77: space_freed 2457 space_freed gives the number of bytes freed if the file is deleted. 2458 This attribute is read only and is of type length4. It is a per file 2459 attribute. 2461 13. Operations: REQUIRED, RECOMMENDED, or OPTIONAL 2463 The following tables summarize the operations of the NFSv4.2 protocol 2464 and the corresponding designation of REQUIRED, RECOMMENDED, and 2465 OPTIONAL to implement or MUST NOT implement. The designation of MUST 2466 NOT implement is reserved for those operations that were defined in 2467 either NFSv4.0 or NFSV4.1 and MUST NOT be implemented in NFSv4.2. 2469 For the most part, the REQUIRED, RECOMMENDED, or OPTIONAL designation 2470 for operations sent by the client is for the server implementation. 2471 The client is generally required to implement the operations needed 2472 for the operating environment for which it serves. For example, a 2473 read-only NFSv4.2 client would have no need to implement the WRITE 2474 operation and is not required to do so. 2476 The REQUIRED or OPTIONAL designation for callback operations sent by 2477 the server is for both the client and server. Generally, the client 2478 has the option of creating the backchannel and sending the operations 2479 on the fore channel that will be a catalyst for the server sending 2480 callback operations. A partial exception is CB_RECALL_SLOT; the only 2481 way the client can avoid supporting this operation is by not creating 2482 a backchannel. 2484 Since this is a summary of the operations and their designation, 2485 there are subtleties that are not presented here. Therefore, if 2486 there is a question of the requirements of implementation, the 2487 operation descriptions themselves must be consulted along with other 2488 relevant explanatory text within this either specification or that of 2489 NFSv4.1 [RFC5661]. 2491 The abbreviations used in the second and third columns of the table 2492 are defined as follows. 2494 REQ: REQUIRED to implement 2496 REC: RECOMMENDED to implement 2498 OPT: OPTIONAL to implement 2500 MNI: MUST NOT implement 2502 For the NFSv4.2 features that are OPTIONAL, the operations that 2503 support those features are OPTIONAL, and the server MUST return 2504 NFS4ERR_NOTSUPP in response to the client's use of those operations, 2505 when those operations are not implemented by the server. If an 2506 OPTIONAL feature is supported, it is possible that a set of 2507 operations related to the feature become REQUIRED to implement. The 2508 third column of the table designates the feature(s) and if the 2509 operation is REQUIRED or OPTIONAL in the presence of support for the 2510 feature. 2512 The OPTIONAL features identified and their abbreviations are as 2513 follows: 2515 pNFS: Parallel NFS 2517 FDELG: File Delegations 2519 DDELG: Directory Delegations 2521 COPYra: Intra-server Server Side Copy 2523 COPYer: Inter-server Server Side Copy 2525 ADB: Application Data Blocks 2527 Operations 2529 +----------------------+---------------------+----------------------+ 2530 | Operation | EOL, REQ, REC, OPT, | Feature (REQ, REC, | 2531 | | or MNI | or OPT) | 2532 +----------------------+---------------------+----------------------+ 2533 | ALLOCATE | OPT | | 2534 | ACCESS | REQ | | 2535 | BACKCHANNEL_CTL | REQ | | 2536 | BIND_CONN_TO_SESSION | REQ | | 2537 | CLOSE | REQ | | 2538 | COMMIT | REQ | | 2539 | COPY | OPT | COPYer (REQ), COPYra | 2540 | | | (REQ) | 2541 | COPY_NOTIFY | OPT | COPYer (REQ) | 2542 | DEALLOCATE | OPT | | 2543 | CREATE | REQ | | 2544 | CREATE_SESSION | REQ | | 2545 | DELEGPURGE | OPT | FDELG (REQ) | 2546 | DELEGRETURN | OPT | FDELG, DDELG, pNFS | 2547 | | | (REQ) | 2548 | DESTROY_CLIENTID | REQ | | 2549 | DESTROY_SESSION | REQ | | 2550 | EXCHANGE_ID | REQ | | 2551 | FREE_STATEID | REQ | | 2552 | GETATTR | REQ | | 2553 | GETDEVICEINFO | OPT | pNFS (REQ) | 2554 | GETDEVICELIST | MNI | pNFS (MNI) | 2555 | GETFH | REQ | | 2556 | GET_DIR_DELEGATION | OPT | DDELG (REQ) | 2557 | LAYOUTCOMMIT | OPT | pNFS (REQ) | 2558 | LAYOUTGET | OPT | pNFS (REQ) | 2559 | LAYOUTRETURN | OPT | pNFS (REQ) | 2560 | LAYOUTERROR | OPT | pNFS (OPT) | 2561 | LAYOUTSTATS | OPT | pNFS (OPT) | 2562 | LINK | OPT | | 2563 | LOCK | REQ | | 2564 | LOCKT | REQ | | 2565 | LOCKU | REQ | | 2566 | LOOKUP | REQ | | 2567 | LOOKUPP | REQ | | 2568 | NVERIFY | REQ | | 2569 | OFFLOAD_CANCEL | OPT | COPYer (REQ), COPYra | 2570 | | | (REQ) | 2571 | OFFLOAD_STATUS | OPT | COPYer (REQ), COPYra | 2572 | | | (REQ) | 2573 | OPEN | REQ | | 2574 | OPENATTR | OPT | | 2575 | OPEN_CONFIRM | MNI | | 2576 | OPEN_DOWNGRADE | REQ | | 2577 | PUTFH | REQ | | 2578 | PUTPUBFH | REQ | | 2579 | PUTROOTFH | REQ | | 2580 | READ | REQ | | 2581 | READDIR | REQ | | 2582 | READLINK | OPT | | 2583 | READ_PLUS | OPT | | 2584 | RECLAIM_COMPLETE | REQ | | 2585 | RELEASE_LOCKOWNER | MNI | | 2586 | REMOVE | REQ | | 2587 | RENAME | REQ | | 2588 | RENEW | MNI | | 2589 | RESTOREFH | REQ | | 2590 | SAVEFH | REQ | | 2591 | SECINFO | REQ | | 2592 | SECINFO_NO_NAME | REC | pNFS file layout | 2593 | | | (REQ) | 2594 | SEEK | OPT | | 2595 | SEQUENCE | REQ | | 2596 | SETATTR | REQ | | 2597 | SETCLIENTID | MNI | | 2598 | SETCLIENTID_CONFIRM | MNI | | 2599 | SET_SSV | REQ | | 2600 | TEST_STATEID | REQ | | 2601 | VERIFY | REQ | | 2602 | WANT_DELEGATION | OPT | FDELG (OPT) | 2603 | WRITE | REQ | | 2604 | WRITE_SAME | OPT | ADB (REQ) | 2605 +----------------------+---------------------+----------------------+ 2607 Callback Operations 2609 +-------------------------+------------------+----------------------+ 2610 | Operation | REQ, REC, OPT, | Feature (REQ, REC, | 2611 | | or MNI | or OPT) | 2612 +-------------------------+------------------+----------------------+ 2613 | CB_OFFLOAD | OPT | COPYer (REQ), COPYra | 2614 | | | (REQ) | 2615 | CB_GETATTR | OPT | FDELG (REQ) | 2616 | CB_LAYOUTRECALL | OPT | pNFS (REQ) | 2617 | CB_NOTIFY | OPT | DDELG (REQ) | 2618 | CB_NOTIFY_DEVICEID | OPT | pNFS (OPT) | 2619 | CB_NOTIFY_LOCK | OPT | | 2620 | CB_PUSH_DELEG | OPT | FDELG (OPT) | 2621 | CB_RECALL | OPT | FDELG, DDELG, pNFS | 2622 | | | (REQ) | 2623 | CB_RECALL_ANY | OPT | FDELG, DDELG, pNFS | 2624 | | | (REQ) | 2625 | CB_RECALL_SLOT | REQ | | 2626 | CB_RECALLABLE_OBJ_AVAIL | OPT | DDELG, pNFS (REQ) | 2627 | CB_SEQUENCE | OPT | FDELG, DDELG, pNFS | 2628 | | | (REQ) | 2629 | CB_WANTS_CANCELLED | OPT | FDELG, DDELG, pNFS | 2630 | | | (REQ) | 2631 +-------------------------+------------------+----------------------+ 2633 14. Modifications to NFSv4.1 Operations 2635 14.1. Operation 42: EXCHANGE_ID - Instantiate Client ID 2637 14.1.1. ARGUMENT 2639 2641 /* new */ 2642 const EXCHGID4_FLAG_SUPP_FENCE_OPS = 0x00000004; 2644 2646 14.1.2. RESULT 2648 Unchanged 2650 14.1.3. MOTIVATION 2652 Enterprise applications require guarantees that an operation has 2653 either aborted or completed. NFSv4.1 provides this guarantee as long 2654 as the session is alive: simply send a SEQUENCE operation on the same 2655 slot with a new sequence number, and the successful return of 2656 SEQUENCE indicates the previous operation has completed. However, if 2657 the session is lost, there is no way to know when any in progress 2658 operations have aborted or completed. In hindsight, the NFSv4.1 2659 specification should have mandated that DESTROY_SESSION either abort 2660 or complete all outstanding operations. 2662 14.1.4. DESCRIPTION 2664 A client SHOULD request the EXCHGID4_FLAG_SUPP_FENCE_OPS capability 2665 when it sends an EXCHANGE_ID operation. The server SHOULD set this 2666 capability in the EXCHANGE_ID reply whether the client requests it or 2667 not. It is the server's return that determines whether this 2668 capability is in effect. When it is in effect, the following will 2669 occur: 2671 o The server will not reply to any DESTROY_SESSION invoked with the 2672 client ID until all operations in progress are completed or 2673 aborted. 2675 o The server will not reply to subsequent EXCHANGE_ID invoked on the 2676 same client owner with a new verifier until all operations in 2677 progress on the client ID's session are completed or aborted. 2679 o In implementations where the NFS server is deployed as a cluster, 2680 it does support client ID trunking, and the 2681 EXCHGID4_FLAG_SUPP_FENCE_OPS capability is enabled, then a session 2682 ID created on one node of the storage cluster MUST be destroyable 2683 via DESTROY_SESSION. In addition, DESTROY_CLIENTID and an 2684 EXCHANGE_ID with a new verifier affects all sessions regardless 2685 what node the sessions were created on. 2687 14.2. Operation 48: GETDEVICELIST - Get All Device Mappings for a File 2688 System 2690 14.2.1. ARGUMENT 2692 2693 struct GETDEVICELIST4args { 2694 /* CURRENT_FH: object belonging to the file system */ 2695 layouttype4 gdla_layout_type; 2697 /* number of deviceIDs to return */ 2698 count4 gdla_maxdevices; 2700 nfs_cookie4 gdla_cookie; 2701 verifier4 gdla_cookieverf; 2702 }; 2704 2706 14.2.2. RESULT 2708 2710 struct GETDEVICELIST4resok { 2711 nfs_cookie4 gdlr_cookie; 2712 verifier4 gdlr_cookieverf; 2713 deviceid4 gdlr_deviceid_list<>; 2714 bool gdlr_eof; 2715 }; 2717 union GETDEVICELIST4res switch (nfsstat4 gdlr_status) { 2718 case NFS4_OK: 2719 GETDEVICELIST4resok gdlr_resok4; 2720 default: 2721 void; 2722 }; 2724 2726 14.2.3. MOTIVATION 2728 The GETDEVICELIST operation was introduced in [RFC5661] specifically 2729 to request a list of devices at filesystem mount time from block 2730 layout type servers. However use of the GETDEVICELIST operation 2731 introduces a race condition versus notification about changes to pNFS 2732 device IDs as provided by CB_NOTIFY_DEVICEID. Implementation 2733 experience with block layout servers has shown there is no need for 2734 GETDEVICELIST. Clients have to be able to request new devices using 2735 GETDEVICEINFO at any time in response either to a new deviceid in 2736 LAYOUTGET results or to the CB_NOTIFY_DEVICEID callback operation. 2738 14.2.4. DESCRIPTION 2740 Clients and servers MUST NOT implement the GETDEVICELIST operation. 2742 15. NFSv4.2 Operations 2744 15.1. Operation 59: ALLOCATE - Reserve Space in A Region of a File 2746 15.1.1. ARGUMENT 2748 2750 struct ALLOCATE4args { 2751 /* CURRENT_FH: file */ 2752 stateid4 aa_stateid; 2753 offset4 aa_offset; 2754 length4 aa_length; 2755 }; 2757 2759 15.1.2. RESULT 2761 2763 struct ALLOCATE4res { 2764 nfsstat4 ar_status; 2765 }; 2767 2769 15.1.3. DESCRIPTION 2771 Whenever a client wishes to reserve space for a region in a file it 2772 calls the ALLOCATE operation with the current filehandle set to the 2773 filehandle of the file in question, and the start offset and length 2774 in bytes of the region set in aa_offset and aa_length respectively. 2776 The server will ensure that backing blocks are reserved to the region 2777 specified by aa_offset and aa_length, and that no future writes into 2778 this region will return NFS4ERR_NOSPC. If the region lies partially 2779 or fully outside the current file size the file size will be set to 2780 aa_offset + aa_length implicitly. If the server cannot guarantee 2781 this, it must return NFS4ERR_NOSPC. 2783 The ALLOCATE operation can also be used to extend the size of a file 2784 if the region specified by aa_offset and aa_length extends beyond the 2785 current file size. In that case any data outside of the previous 2786 file size will return zeroes when read before data is written to it. 2788 It is not required that the server allocate the space to the file 2789 before returning success. The allocation can be deferred, however, 2790 it must be guaranteed that it will not fail for lack of space. The 2791 deferral does not result in an asynchronous reply. 2793 The ALLOCATE operation will result in the space_used attribute and 2794 space_freed attributes being increased by the number of bytes 2795 reserved unless they were previously reserved or written and not 2796 shared. 2798 15.2. Operation 60: COPY - Initiate a server-side copy 2800 15.2.1. ARGUMENT 2802 2804 struct COPY4args { 2805 /* SAVED_FH: source file */ 2806 /* CURRENT_FH: destination file */ 2807 stateid4 ca_src_stateid; 2808 stateid4 ca_dst_stateid; 2809 offset4 ca_src_offset; 2810 offset4 ca_dst_offset; 2811 length4 ca_count; 2812 bool ca_consecutive; 2813 bool ca_synchronous; 2814 netloc4 ca_source_server<>; 2815 }; 2817 2819 15.2.2. RESULT 2821 2823 struct write_response4 { 2824 stateid4 wr_callback_id<1>; 2825 length4 wr_count; 2826 stable_how4 wr_committed; 2827 verifier4 wr_writeverf; 2828 }; 2829 struct COPY4res { 2830 nfsstat4 cr_status; 2831 write_response4 cr_response; 2832 bool cr_consecutive; 2833 bool cr_synchronous; 2834 }; 2836 2838 15.2.3. DESCRIPTION 2840 The COPY operation is used for both intra-server and inter-server 2841 copies. In both cases, the COPY is always sent from the client to 2842 the destination server of the file copy. The COPY operation requests 2843 that a file be copied from the location specified by the SAVED_FH 2844 value to the location specified by the CURRENT_FH. 2846 The SAVED_FH must be a regular file. If SAVED_FH is not a regular 2847 file, the operation MUST fail and return NFS4ERR_WRONG_TYPE. 2849 In order to set SAVED_FH to the source file handle, the compound 2850 procedure requesting the COPY will include a sub-sequence of 2851 operations such as 2853 PUTFH source-fh 2854 SAVEFH 2856 If the request is for an inter-server-to-server copy, the source-fh 2857 is a filehandle from the source server and the compound procedure is 2858 being executed on the destination server. In this case, the source- 2859 fh is a foreign filehandle on the server receiving the COPY request. 2860 If either PUTFH or SAVEFH checked the validity of the filehandle, the 2861 operation would likely fail and return NFS4ERR_STALE. 2863 If a server supports the inter-server-to-server COPY feature, a PUTFH 2864 followed by a SAVEFH MUST NOT return NFS4ERR_STALE for either 2865 operation. These restrictions do not pose substantial difficulties 2866 for servers. The CURRENT_FH and SAVED_FH may be validated in the 2867 context of the operation referencing them and an NFS4ERR_STALE error 2868 returned for an invalid file handle at that point. 2870 For an intra-server copy, both the ca_src_stateid and ca_dst_stateid 2871 MUST refer to either open or locking states provided earlier by the 2872 server. If either stateid is invalid, then the operation MUST fail. 2873 If the request is for a inter-server copy, then the ca_src_stateid 2874 can be ignored. If ca_dst_stateid is invalid, then the operation 2875 MUST fail. 2877 The CURRENT_FH specifies the destination of the copy operation. The 2878 CURRENT_FH MUST be a regular file and not a directory. Note, the 2879 file MUST exist before the COPY operation begins. It is the 2880 responsibility of the client to create the file if necessary, 2881 regardless of the actual copy protocol used. If the file cannot be 2882 created in the destination file system (due to file name 2883 restrictions, such as case or length), the COPY operation MUST NOT be 2884 called. 2886 The ca_src_offset is the offset within the source file from which the 2887 data will be read, the ca_dst_offset is the offset within the 2888 destination file to which the data will be written, and the ca_count 2889 is the number of bytes that will be copied. An offset of 0 (zero) 2890 specifies the start of the file. A count of 0 (zero) requests that 2891 all bytes from ca_src_offset through EOF be copied to the 2892 destination. If concurrent modifications to the source file overlap 2893 with the source file region being copied, the data copied may include 2894 all, some, or none of the modifications. The client can use standard 2895 NFS operations (e.g., OPEN with OPEN4_SHARE_DENY_WRITE or mandatory 2896 byte range locks) to protect against concurrent modifications if the 2897 client is concerned about this. If the source file's end of file is 2898 being modified in parallel with a copy that specifies a count of 0 2899 (zero) bytes, the amount of data copied is implementation dependent 2900 (clients may guard against this case by specifying a non-zero count 2901 value or preventing modification of the source file as mentioned 2902 above). 2904 If the source offset or the source offset plus count is greater than 2905 or equal to the size of the source file, the operation will fail with 2906 NFS4ERR_INVAL. The destination offset or destination offset plus 2907 count may be greater than the size of the destination file. This 2908 allows for the client to issue parallel copies to implement 2909 operations such as 2911 2913 % cat file1 file2 file3 file4 > dest 2915 2917 If the ca_source_server list is specified, then this is an inter- 2918 server copy operation and the source file is on a remote server. The 2919 client is expected to have previously issued a successful COPY_NOTIFY 2920 request to the remote source server. The ca_source_server list MUST 2921 be the same as the COPY_NOTIFY response's cnr_source_server list. If 2922 the client includes the entries from the COPY_NOTIFY response's 2923 cnr_source_server list in the ca_source_server list, the source 2924 server can indicate a specific copy protocol for the destination 2925 server to use by returning a URL, which specifies both a protocol 2926 service and server name. Server-to-server copy protocol 2927 considerations are described in Section 4.7 and Section 4.10.1. 2929 If ca_consecutive is set, then the client has specified that the copy 2930 protocol selected MUST copy bytes in consecutive order from 2931 ca_src_offset to ca_count. If the destination server cannot meet 2932 this requirement, then it MUST return an error of 2933 NFS4ERR_OFFLOAD_NO_REQS and set cr_consecutive to be false. 2934 Likewise, if ca_synchronous is set, then the client has required that 2935 the copy protocol selected MUST perform a synchronous copy. If the 2936 destination server cannot meet this requirement, then it MUST return 2937 an error of NFS4ERR_OFFLOAD_NO_REQS and set cr_synchronous to be 2938 false. 2940 If both are set by the client, then the destination SHOULD try to 2941 determine if it can respond to both requirements at the same time. 2942 If it cannot make that determination, it must set to false the one it 2943 can and set to true the other. The client, upon getting an 2944 NFS4ERR_OFFLOAD_NO_REQS error, has to examine both cr_consecutive and 2945 cr_synchronous against the respective values of ca_consecutive and 2946 ca_synchronous to determine the possible requirement not met. It 2947 MUST be prepared for the destination server not being able to 2948 determine both requirements at the same time. 2950 Upon receiving the NFS4ERR_OFFLOAD_NO_REQS error, the client has to 2951 determine if it wants to either re-request the copy with a relaxed 2952 set of requirements or if it wants to revert to manually copying the 2953 data. If it decides to manually copy the data and this is a remote 2954 copy, then the client is responsible for informing the source that 2955 the earlier COPY_NOTIFY is no longer valid by sending it an 2956 OFFLOAD_CANCEL. 2958 The copying of any and all attributes on the source file is the 2959 responsibility of both the client and the copy protocol. Any 2960 attribute which is both exposed via the NFS protocol on the source 2961 file and set SHOULD be copied to the destination file. Any attribute 2962 supported by the destination server that is not set on the source 2963 file SHOULD be left unset. If the client cannot copy an attribute 2964 from the source to destination, it MAY fail the copy transaction. 2966 Metadata attributes not exposed via the NFS protocol SHOULD be copied 2967 to the destination file where appropriate via the copy protocol. 2968 Note that if the copy protocol is NFSv4.x, then these attributes will 2969 be lost. 2971 The destination file's named attributes are not duplicated from the 2972 source file. After the copy process completes, the client MAY 2973 attempt to duplicate named attributes using standard NFSv4 2974 operations. However, the destination file's named attribute 2975 capabilities MAY be different from the source file's named attribute 2976 capabilities. 2978 If the operation does not result in an immediate failure, the server 2979 will return NFS4_OK, and the CURRENT_FH will remain the destination's 2980 filehandle. 2982 If the wr_callback_id is returned, this indicates that the operation 2983 was initiated and a CB_OFFLOAD callback will deliver the final 2984 results of the operation. The wr_callback_id stateid is termed a 2985 copy stateid in this context. The server is given the option of 2986 returning the results in a callback because the data may require a 2987 relatively long period of time to copy. 2989 If no wr_callback_id is returned, the operation completed 2990 synchronously and no callback will be issued by the server. The 2991 completion status of the operation is indicated by cr_status. 2993 If the copy completes successfully, either synchronously or 2994 asynchronously, the data copied from the source file to the 2995 destination file MUST appear identical to the NFS client. However, 2996 the NFS server's on disk representation of the data in the source 2997 file and destination file MAY differ. For example, the NFS server 2998 might encrypt, compress, deduplicate, or otherwise represent the on 2999 disk data in the source and destination file differently. 3001 If a failure does occur for a synchronous copy, wr_count will be set 3002 to the number of bytes copied to the destination file before the 3003 error occurred. If cr_consecutive is true, then the bytes were 3004 copied in order. If the failure occurred for an asynchronous copy, 3005 then the client will have gotten the notification of the consecutive 3006 copy order when it got the copy stateid. It will be able to 3007 determine the bytes copied from the coa_bytes_copied in the 3008 CB_OFFLOAD argument. 3010 In either case, if cr_consecutive was not true, there is no assurance 3011 as to exactly which bytes in the range were copied. The client MUST 3012 assume that there exists a mixture of the original contents of the 3013 range and the new bytes. If the COPY wrote past the end of the file 3014 on the destination, then the last byte written to will determine the 3015 new file size. The contents of any block not written to and past the 3016 original size of the file will be as if a normal WRITE extended the 3017 file. 3019 15.3. Operation 61: COPY_NOTIFY - Notify a source server of a future 3020 copy 3022 15.3.1. ARGUMENT 3024 3026 struct COPY_NOTIFY4args { 3027 /* CURRENT_FH: source file */ 3028 stateid4 cna_src_stateid; 3029 netloc4 cna_destination_server; 3030 }; 3032 3034 15.3.2. RESULT 3036 3038 struct COPY_NOTIFY4resok { 3039 nfstime4 cnr_lease_time; 3040 netloc4 cnr_source_server<>; 3041 }; 3043 union COPY_NOTIFY4res switch (nfsstat4 cnr_status) { 3044 case NFS4_OK: 3045 COPY_NOTIFY4resok resok4; 3046 default: 3047 void; 3048 }; 3050 3052 15.3.3. DESCRIPTION 3054 This operation is used for an inter-server copy. A client sends this 3055 operation in a COMPOUND request to the source server to authorize a 3056 destination server identified by cna_destination_server to read the 3057 file specified by CURRENT_FH on behalf of the given user. 3059 The cna_src_stateid MUST refer to either open or locking states 3060 provided earlier by the server. If it is invalid, then the operation 3061 MUST fail. 3063 The cna_destination_server MUST be specified using the netloc4 3064 network location format. The server is not required to resolve the 3065 cna_destination_server address before completing this operation. 3067 If this operation succeeds, the source server will allow the 3068 cna_destination_server to copy the specified file on behalf of the 3069 given user as long as both of the following conditions are met: 3071 o The destination server begins reading the source file before the 3072 cnr_lease_time expires. If the cnr_lease_time expires while the 3073 destination server is still reading the source file, the 3074 destination server is allowed to finish reading the file. 3076 o The client has not issued a OFFLOAD_CANCEL for the same 3077 combination of user, filehandle, and destination server. 3079 The cnr_lease_time is chosen by the source server. A cnr_lease_time 3080 of 0 (zero) indicates an infinite lease. To avoid the need for 3081 synchronized clocks, copy lease times are granted by the server as a 3082 time delta. To renew the copy lease time the client should resend 3083 the same copy notification request to the source server. 3085 A successful response will also contain a list of netloc4 network 3086 location formats called cnr_source_server, on which the source is 3087 willing to accept connections from the destination. These might not 3088 be reachable from the client and might be located on networks to 3089 which the client has no connection. 3091 For a copy only involving one server (the source and destination are 3092 on the same server), this operation is unnecessary. 3094 15.4. Operation 62: DEALLOCATE - Unreserve Space in a Region of a File 3096 15.4.1. ARGUMENT 3098 3100 struct DEALLOCATE4args { 3101 /* CURRENT_FH: file */ 3102 stateid4 da_stateid; 3103 offset4 da_offset; 3104 length4 da_length; 3105 }; 3107 3109 15.4.2. RESULT 3111 3113 struct DEALLOCATE4res { 3114 nfsstat4 dr_status; 3115 }; 3117 3119 15.4.3. DESCRIPTION 3121 Whenever a client wishes to unreserve space for a region in a file it 3122 calls the DEALLOCATE operation with the current filehandle set to the 3123 filehandle of the file in question, and the start offset and length 3124 in bytes of the region set in da_offset and da_length respectively. 3125 If no space was allocated or reserved for all or parts of the region, 3126 the DEALLOCATE operation will have no effect for the region that 3127 already is in unreserved state. All further reads from the region 3128 passed to DEALLOCATE MUST return zeros until overwritten. The 3129 filehandle specified must be that of a regular file. 3131 Situations may arise where da_offset and/or da_offset + da_length 3132 will not be aligned to a boundary for which the server does 3133 allocations or deallocations. For most file systems, this is the 3134 block size of the file system. In such a case, the server can 3135 deallocate as many bytes as it can in the region. The blocks that 3136 cannot be deallocated MUST be zeroed. 3138 DEALLOCATE will result in the space_used attribute being decreased by 3139 the number of bytes that were deallocated. The space_freed attribute 3140 may or may not decrease, depending on the support and whether the 3141 blocks backing the specified range were shared or not. The size 3142 attribute will remain unchanged. 3144 15.5. Operation 63: IO_ADVISE - Application I/O access pattern hints 3146 15.5.1. ARGUMENT 3148 3149 enum IO_ADVISE_type4 { 3150 IO_ADVISE4_NORMAL = 0, 3151 IO_ADVISE4_SEQUENTIAL = 1, 3152 IO_ADVISE4_SEQUENTIAL_BACKWARDS = 2, 3153 IO_ADVISE4_RANDOM = 3, 3154 IO_ADVISE4_WILLNEED = 4, 3155 IO_ADVISE4_WILLNEED_OPPORTUNISTIC = 5, 3156 IO_ADVISE4_DONTNEED = 6, 3157 IO_ADVISE4_NOREUSE = 7, 3158 IO_ADVISE4_READ = 8, 3159 IO_ADVISE4_WRITE = 9, 3160 IO_ADVISE4_INIT_PROXIMITY = 10 3161 }; 3163 struct IO_ADVISE4args { 3164 /* CURRENT_FH: file */ 3165 stateid4 iaa_stateid; 3166 offset4 iaa_offset; 3167 length4 iaa_count; 3168 bitmap4 iaa_hints; 3169 }; 3171 3173 15.5.2. RESULT 3175 3177 struct IO_ADVISE4resok { 3178 bitmap4 ior_hints; 3179 }; 3181 union IO_ADVISE4res switch (nfsstat4 ior_status) { 3182 case NFS4_OK: 3183 IO_ADVISE4resok resok4; 3184 default: 3185 void; 3186 }; 3188 3190 15.5.3. DESCRIPTION 3192 The IO_ADVISE operation sends an I/O access pattern hint to the 3193 server for the owner of the stateid for a given byte range specified 3194 by iar_offset and iar_count. The byte range specified by iaa_offset 3195 and iaa_count need not currently exist in the file, but the iaa_hints 3196 will apply to the byte range when it does exist. If iaa_count is 0, 3197 all data following iaa_offset is specified. The server MAY ignore 3198 the advice. 3200 The following are the allowed hints for a stateid holder: 3202 IO_ADVISE4_NORMAL There is no advice to give, this is the default 3203 behavior. 3205 IO_ADVISE4_SEQUENTIAL Expects to access the specified data 3206 sequentially from lower offsets to higher offsets. 3208 IO_ADVISE4_SEQUENTIAL_BACKWARDS Expects to access the specified data 3209 sequentially from higher offsets to lower offsets. 3211 IO_ADVISE4_RANDOM Expects to access the specified data in a random 3212 order. 3214 IO_ADVISE4_WILLNEED Expects to access the specified data in the near 3215 future. 3217 IO_ADVISE4_WILLNEED_OPPORTUNISTIC Expects to possibly access the 3218 data in the near future. This is a speculative hint, and 3219 therefore the server should prefetch data or indirect blocks only 3220 if it can be done at a marginal cost. 3222 IO_ADVISE_DONTNEED Expects that it will not access the specified 3223 data in the near future. 3225 IO_ADVISE_NOREUSE Expects to access the specified data once and then 3226 not reuse it thereafter. 3228 IO_ADVISE4_READ Expects to read the specified data in the near 3229 future. 3231 IO_ADVISE4_WRITE Expects to write the specified data in the near 3232 future. 3234 IO_ADVISE4_INIT_PROXIMITY Informs the server that the data in the 3235 byte range remains important to the client. 3237 Since IO_ADVISE is a hint, a server SHOULD NOT return an error and 3238 invalidate a entire Compound request if one of the sent hints in 3239 iar_hints is not supported by the server. Also, the server MUST NOT 3240 return an error if the client sends contradictory hints to the 3241 server, e.g., IO_ADVISE4_SEQUENTIAL and IO_ADVISE4_RANDOM in a single 3242 IO_ADVISE operation. In these cases, the server MUST return success 3243 and a ior_hints value that indicates the hint it intends to 3244 implement. This may mean simply returning IO_ADVISE4_NORMAL. 3246 The ior_hints returned by the server is primarily for debugging 3247 purposes since the server is under no obligation to carry out the 3248 hints that it describes in the ior_hints result. In addition, while 3249 the server may have intended to implement the hints returned in 3250 ior_hints, as time progresses, the server may need to change its 3251 handling of a given file due to several reasons including, but not 3252 limited to, memory pressure, additional IO_ADVISE hints sent by other 3253 clients, and heuristically detected file access patterns. 3255 The server MAY return different advice than what the client 3256 requested. If it does, then this might be due to one of several 3257 conditions, including, but not limited to another client advising of 3258 a different I/O access pattern; a different I/O access pattern from 3259 another client that that the server has heuristically detected; or 3260 the server is not able to support the requested I/O access pattern, 3261 perhaps due to a temporary resource limitation. 3263 Each issuance of the IO_ADVISE operation overrides all previous 3264 issuances of IO_ADVISE for a given byte range. This effectively 3265 follows a strategy of last hint wins for a given stateid and byte 3266 range. 3268 Clients should assume that hints included in an IO_ADVISE operation 3269 will be forgotten once the file is closed. 3271 15.5.4. IMPLEMENTATION 3273 The NFS client may choose to issue an IO_ADVISE operation to the 3274 server in several different instances. 3276 The most obvious is in direct response to an application's execution 3277 of posix_fadvise(). In this case, IO_ADVISE4_WRITE and 3278 IO_ADVISE4_READ may be set based upon the type of file access 3279 specified when the file was opened. 3281 15.5.5. IO_ADVISE4_INIT_PROXIMITY 3283 The IO_ADVISE4_INIT_PROXIMITY hint is non-posix in origin and can be 3284 used to convey that the client has recently accessed the byte range 3285 in its own cache. I.e., it has not accessed it on the server, but it 3286 has locally. When the server reaches resource exhaustion, knowing 3287 which data is more important allows the server to make better choices 3288 about which data to, for example purge from a cache, or move to 3289 secondary storage. It also informs the server which delegations are 3290 more important, since if delegations are working correctly, once 3291 delegated to a client and the client has read the content for that 3292 byte range, a server might never receive another read request for 3293 that byte range. 3295 The IO_ADVISE4_INIT_PROXIMITY hint can also be used in a pNFS setting 3296 to let the client inform the metadata server as to the I/O statistics 3297 between the client and the storage devices. The metadata server is 3298 then free to use this information about client I/O to optimize the 3299 data storage location. 3301 This hint is also useful in the case of NFS clients which are network 3302 booting from a server. If the first client to be booted sends this 3303 hint, then it keeps the cache warm for the remaining clients. 3305 15.5.6. pNFS File Layout Data Type Considerations 3307 The IO_ADVISE considerations for pNFS are very similar to the COMMIT 3308 considerations for pNFS. That is, as with COMMIT, some NFS server 3309 implementations prefer IO_ADVISE be done on the DS, and some prefer 3310 it be done on the MDS. 3312 For the file's layout type, it is proposed that NFSv4.2 include an 3313 additional hint NFL42_CARE_IO_ADVISE_THRU_MDS which is valid only on 3314 metadata servers running NFSv4.2 or higher. Any file's layout 3315 obtained from a NFSv4.1 metadata server MUST NOT have 3316 NFL42_UFLG_IO_ADVISE_THRU_MDS set. Any file's layout obtained with a 3317 NFSv4.2 metadata server MAY have NFL42_UFLG_IO_ADVISE_THRU_MDS set. 3318 However, if the layout utilizes NFSv4.1 storage devices, the 3319 IO_ADVISE operation cannot be sent to them. 3321 If NFL42_UFLG_IO_ADVISE_THRU_MDS is set, the client MUST send the 3322 IO_ADVISE operation to the MDS in order for it to be honored by the 3323 DS. Once the MDS receives the IO_ADVISE operation, it will 3324 communicate the advice to each DS. 3326 If NFL42_UFLG_IO_ADVISE_THRU_MDS is not set, then the client SHOULD 3327 send an IO_ADVISE operation to the appropriate DS for the specified 3328 byte range. While the client MAY always send IO_ADVISE to the MDS, 3329 if the server has not set NFL42_UFLG_IO_ADVISE_THRU_MDS, the client 3330 should expect that such an IO_ADVISE is futile. Note that a client 3331 SHOULD use the same set of arguments on each IO_ADVISE sent to a DS 3332 for the same open file reference. 3334 The server is not required to support different advice for different 3335 DS's with the same open file reference. 3337 15.5.6.1. Dense and Sparse Packing Considerations 3339 The IO_ADVISE operation MUST use the iar_offset and byte range as 3340 dictated by the presence or absence of NFL4_UFLG_DENSE. 3342 E.g., if NFL4_UFLG_DENSE is present, and a READ or WRITE to the DS 3343 for iaa_offset 0 really means iaa_offset 10000 in the logical file, 3344 then an IO_ADVISE for iaa_offset 0 means iaa_offset 10000. 3346 E.g., if NFL4_UFLG_DENSE is absent, then a READ or WRITE to the DS 3347 for iaa_offset 0 really means iaa_offset 0 in the logical file, then 3348 an IO_ADVISE for iaa_offset 0 means iaa_offset 0 in the logical file. 3350 E.g., if NFL4_UFLG_DENSE is present, the stripe unit is 1000 bytes 3351 and the stripe count is 10, and the dense DS file is serving 3352 iar_offset 0. A READ or WRITE to the DS for iaa_offsets 0, 1000, 3353 2000, and 3000, really mean iaa_offsets 10000, 20000, 30000, and 3354 40000 (implying a stripe count of 10 and a stripe unit of 1000), then 3355 an IO_ADVISE sent to the same DS with an iaa_offset of 500, and an 3356 iaa_count of 3000 means that the IO_ADVISE applies to these byte 3357 ranges of the dense DS file: 3359 - 500 to 999 3360 - 1000 to 1999 3361 - 2000 to 2999 3362 - 3000 to 3499 3364 I.e., the contiguous range 500 to 3499 as specified in IO_ADVISE. 3366 It also applies to these byte ranges of the logical file: 3368 - 10500 to 10999 (500 bytes) 3369 - 20000 to 20999 (1000 bytes) 3370 - 30000 to 30999 (1000 bytes) 3371 - 40000 to 40499 (500 bytes) 3372 (total 3000 bytes) 3374 E.g., if NFL4_UFLG_DENSE is absent, the stripe unit is 250 bytes, the 3375 stripe count is 4, and the sparse DS file is serving iaa_offset 0. 3376 Then a READ or WRITE to the DS for iaa_offsets 0, 1000, 2000, and 3377 3000, really means iaa_offsets 0, 1000, 2000, and 3000 in the logical 3378 file, keeping in mind that on the DS file, byte ranges 250 to 999, 3379 1250 to 1999, 2250 to 2999, and 3250 to 3999 are not accessible. 3380 Then an IO_ADVISE sent to the same DS with an iaa_offset of 500, and 3381 a iaa_count of 3000 means that the IO_ADVISE applies to these byte 3382 ranges of the logical file and the sparse DS file: 3384 - 500 to 999 (500 bytes) - no effect 3385 - 1000 to 1249 (250 bytes) - effective 3386 - 1250 to 1999 (750 bytes) - no effect 3387 - 2000 to 2249 (250 bytes) - effective 3388 - 2250 to 2999 (750 bytes) - no effect 3389 - 3000 to 3249 (250 bytes) - effective 3390 - 3250 to 3499 (250 bytes) - no effect 3391 (subtotal 2250 bytes) - no effect 3392 (subtotal 750 bytes) - effective 3393 (grand total 3000 bytes) - no effect + effective 3395 If neither of the flags NFL42_UFLG_IO_ADVISE_THRU_MDS and 3396 NFL4_UFLG_DENSE are set in the layout, then any IO_ADVISE request 3397 sent to the data server with a byte range that overlaps stripe unit 3398 that the data server does not serve MUST NOT result in the status 3399 NFS4ERR_PNFS_IO_HOLE. Instead, the response SHOULD be successful and 3400 if the server applies IO_ADVISE hints on any stripe units that 3401 overlap with the specified range, those hints SHOULD be indicated in 3402 the response. 3404 15.6. Operation 64: LAYOUTERROR - Provide Errors for the Layout 3406 15.6.1. ARGUMENT 3408 3410 struct device_error4 { 3411 deviceid4 de_deviceid; 3412 nfsstat4 de_status; 3413 nfs_opnum4 de_opnum; 3414 }; 3415 struct LAYOUTERROR4args { 3416 /* CURRENT_FH: file */ 3417 offset4 lea_offset; 3418 length4 lea_length; 3419 stateid4 lea_stateid; 3420 device_error4 lea_errors<>; 3421 }; 3423 3425 15.6.2. RESULT 3427 3429 struct LAYOUTERROR4res { 3430 nfsstat4 ler_status; 3431 }; 3433 3435 15.6.3. DESCRIPTION 3437 The client can use LAYOUTERROR to inform the metadata server about 3438 errors in its interaction with the layout represented by the current 3439 filehandle, client ID (derived from the session ID in the preceding 3440 SEQUENCE operation), byte-range (lea_offset + lea_length), and 3441 lea_stateid. 3443 Each individual device_error4 describes a single error associated 3444 with a storage device, which is identified via de_deviceid. If the 3445 Layout Type supports NFSv4 operations, then the operation which 3446 returned the error is identified via de_opnum. If the Layout Type 3447 does not support NFSv4 operations, then it MAY chose to either map 3448 the operation onto one of the allowed operations which can be sent to 3449 a storage device with the File Layout Type (see Section 3.3) or it 3450 can signal no support for operations by marking de_opnum with the 3451 ILLEGAL operation. Finally the NFS error value (nfsstat4) 3452 encountered is provided via de_status and may consist of the 3453 following error codes: 3455 NFS4ERR_NXIO: The client was unable to establish any communication 3456 with the storage device. 3458 NFS4ERR_*: The client was able to establish communication with the 3459 storage device and is returning one of the allowed error codes for 3460 the operation denoted by de_opnum. 3462 Note that while the metadata server may return an error associated 3463 with the layout stateid or the open file, it MUST NOT return an error 3464 in the processing of the errors. If LAYOUTERROR is in a compound 3465 before LAYOUTRETURN, it MUST NOT introduce an error other than what 3466 LAYOUTRETURN would already encounter. 3468 15.6.4. IMPLEMENTATION 3470 There are two broad classes of errors, transient and persistent. The 3471 client SHOULD strive to only use this new mechanism to report 3472 persistent errors. It MUST be able to deal with transient issues by 3473 itself. Also, while the client might consider an issue to be 3474 persistent, it MUST be prepared for the metadata server to consider 3475 such issues to be transient. A prime example of this is if the 3476 metadata server fences off a client from either a stateid or a 3477 filehandle. The client will get an error from the storage device and 3478 might relay either NFS4ERR_ACCESS or NFS4ERR_BAD_STATEID back to the 3479 metadata server, with the belief that this is a hard error. If the 3480 metadata server is informed by the client that there is an error, it 3481 can safely ignore that. For it, the mission is accomplished in that 3482 the client has returned a layout that the metadata server had most 3483 likely recalled. 3485 The client might also need to inform the metadata server that it 3486 cannot reach one or more of the storage devices. While the metadata 3487 server can detect the connectivity of both of these paths: 3489 o metadata server to storage device 3491 o metadata server to client 3493 it cannot determine if the client and storage device path is working. 3494 As with the case of the storage device passing errors to the client, 3495 it must be prepared for the metadata server to consider such outages 3496 as being transitory. 3498 Clients are expected to tolerate transient storage device errors, and 3499 hence clients SHOULD NOT use the LAYOUTERROR error handling for 3500 device access problems that may be transient. The methods by which a 3501 client decides whether a device access problem is transient vs 3502 persistent are implementation-specific, but may include retrying I/Os 3503 to a data server under appropriate conditions. 3505 When an I/O fails to a storage device, the client SHOULD retry the 3506 failed I/O via the metadata server. In this situation, before 3507 retrying the I/O, the client SHOULD return the layout, or the 3508 affected portion thereof, and SHOULD indicate which storage device or 3509 devices was problematic. The client needs to do this when the 3510 storage device is being unresponsive in order to fence off any failed 3511 write attempts, and ensure that they do not end up overwriting any 3512 later data being written through the metadata server. If the client 3513 does not do this, the metadata server MAY issue a layout recall 3514 callback in order to perform the retried I/O. 3516 The client needs to be cognizant that since this error handling is 3517 optional in the metadata server, the metadata server may silently 3518 ignore this functionality. Also, as the metadata server may consider 3519 some issues the client reports to be expected, the client might find 3520 it difficult to detect a metadata server which has not implemented 3521 error handling via LAYOUTERROR. 3523 If an metadata server is aware that a storage device is proving 3524 problematic to a client, the metadata server SHOULD NOT include that 3525 storage device in any pNFS layouts sent to that client. If the 3526 metadata server is aware that a storage device is affecting many 3527 clients, then the metadata server SHOULD NOT include that storage 3528 device in any pNFS layouts sent out. If a client asks for a new 3529 layout for the file from the metadata server, it MUST be prepared for 3530 the metadata server to return that storage device in the layout. The 3531 metadata server might not have any choice in using the storage 3532 device, i.e., there might only be one possible layout for the system. 3533 Also, in the case of existing files, the metadata server might have 3534 no choice in which storage devices to hand out to clients. 3536 The metadata server is not required to indefinitely retain per-client 3537 storage device error information. An metadata server is also not 3538 required to automatically reinstate use of a previously problematic 3539 storage device; administrative intervention may be required instead. 3541 15.7. Operation 65: LAYOUTSTATS - Provide Statistics for the Layout 3543 15.7.1. ARGUMENT 3545 3547 struct layoutupdate4 { 3548 layouttype4 lou_type; 3549 opaque lou_body<>; 3550 }; 3552 struct io_info4 { 3553 uint32_t ii_count; 3554 uint64_t ii_bytes; 3555 }; 3556 struct LAYOUTSTATS4args { 3557 /* CURRENT_FH: file */ 3558 offset4 lsa_offset; 3559 length4 lsa_length; 3560 stateid4 lsa_stateid; 3561 io_info4 lsa_read; 3562 io_info4 lsa_write; 3563 deviceid4 lsa_deviceid; 3564 layoutupdate4 lsa_layoutupdate; 3565 }; 3567 3569 15.7.2. RESULT 3571 3573 struct LAYOUTSTATS4res { 3574 nfsstat4 lsr_status; 3575 }; 3577 3579 15.7.3. DESCRIPTION 3581 The client can use LAYOUTSTATS to inform the metadata server about 3582 its interaction with the layout represented by the current 3583 filehandle, client ID (derived from the session ID in the preceding 3584 SEQUENCE operation), byte-range (lsa_offset and lsa_length), and 3585 lsa_stateid. lsa_read and lsa_write allow for non-Layout Type 3586 specific statistics to be reported. lsa_deviceid allows the client 3587 to specify to which storage device the statistics apply. The 3588 remaining information the client is presenting is specific to the 3589 Layout Type and presented in the lsa_layoutupdate field. Each Layout 3590 Type MUST define the contents of lsa_layoutupdate in their respective 3591 specifications. 3593 LAYOUTSTATS can be combined with IO_ADVISE (see Section 15.5) to 3594 augment the decision making process of how the metadata server 3595 handles a file. I.e., IO_ADVISE lets the server know that a byte 3596 range has a certain characteristic, but not necessarily the intensity 3597 of that characteristic. 3599 The client MUST reset the statistics after getting a successfully 3600 reply from the metadata server. The first LAYOUTSTATS sent by the 3601 client SHOULD be from the opening of the file. The choice of how 3602 often to update the metadata server is made by the client. 3604 Note that while the metadata server may return an error associated 3605 with the layout stateid or the open file, it MUST NOT return an error 3606 in the processing of the statistics. 3608 15.8. Operation 66: OFFLOAD_CANCEL - Stop an Offloaded Operation 3610 15.8.1. ARGUMENT 3612 3614 struct OFFLOAD_CANCEL4args { 3615 /* CURRENT_FH: file to cancel */ 3616 stateid4 oca_stateid; 3617 }; 3619 3621 15.8.2. RESULT 3623 3625 struct OFFLOAD_CANCEL4res { 3626 nfsstat4 ocr_status; 3627 }; 3629 3631 15.8.3. DESCRIPTION 3633 OFFLOAD_CANCEL is used by the client to terminate an asynchronous 3634 operation, which is identified both by CURRENT_FH and the 3635 oca_stateid. I.e., there can be multiple offloaded operations acting 3636 on the file, the stateid will identify to the server exactly which 3637 one is to be stopped. Currently there are only two operations which 3638 can decide to be asynchronous: COPY and WRITE_SAME. 3640 In the context of server-to-server copy, the client can send 3641 OFFLOAD_CANCEL to either the source or destination server, albeit 3642 with a different stateid. The client uses OFFLOAD_CANCEL to inform 3643 the destination to stop the active transfer and uses the stateid it 3644 got back from the COPY operation. The client uses OFFLOAD_CANCEL and 3645 the stateid it used in the COPY_NOTIFY to inform the source to not 3646 allow any more copying from the destination. 3648 OFFLOAD_CANCEL is also useful in situations in which the source 3649 server granted a very long or infinite lease on the destination 3650 server's ability to read the source file and all copy operations on 3651 the source file have been completed. 3653 15.9. Operation 67: OFFLOAD_STATUS - Poll for Status of Asynchronous 3654 Operation 3656 15.9.1. ARGUMENT 3658 3660 struct OFFLOAD_STATUS4args { 3661 /* CURRENT_FH: destination file */ 3662 stateid4 osa_stateid; 3663 }; 3665 3667 15.9.2. RESULT 3669 3671 struct OFFLOAD_STATUS4resok { 3672 length4 osr_count; 3673 nfsstat4 osr_complete<1>; 3674 }; 3676 union OFFLOAD_STATUS4res switch (nfsstat4 osr_status) { 3677 case NFS4_OK: 3678 OFFLOAD_STATUS4resok osr_resok4; 3679 default: 3680 void; 3681 }; 3683 3685 15.9.3. DESCRIPTION 3687 OFFLOAD_STATUS can be used by the client to query the progress of an 3688 asynchronous operation, which is identified both by CURRENT_FH and 3689 the osa_stateid. If this operation is successful, the number of 3690 bytes processed are returned to the client in the osr_count field. 3692 If the optional osr_complete field is present, the asynchronous 3693 operation has completed. In this case the status value indicates the 3694 result of the asynchronous operation. In all cases, the server will 3695 also deliver the final results of the asynchronous operation in a 3696 CB_OFFLOAD operation. 3698 The failure of this operation does not indicate the result of the 3699 asynchronous operation in any way. 3701 15.10. Operation 68: READ_PLUS - READ Data or Holes from a File 3703 15.10.1. ARGUMENT 3705 3707 struct READ_PLUS4args { 3708 /* CURRENT_FH: file */ 3709 stateid4 rpa_stateid; 3710 offset4 rpa_offset; 3711 count4 rpa_count; 3712 }; 3714 3716 15.10.2. RESULT 3718 3720 enum data_content4 { 3721 NFS4_CONTENT_DATA = 0, 3722 NFS4_CONTENT_HOLE = 1 3723 }; 3725 struct data_info4 { 3726 offset4 di_offset; 3727 length4 di_length; 3728 }; 3730 struct data4 { 3731 offset4 d_offset; 3732 opaque d_data<>; 3733 }; 3734 union read_plus_content switch (data_content4 rpc_content) { 3735 case NFS4_CONTENT_DATA: 3736 data4 rpc_data; 3737 case NFS4_CONTENT_HOLE: 3738 data_info4 rpc_hole; 3739 default: 3740 void; 3741 }; 3743 /* 3744 * Allow a return of an array of contents. 3745 */ 3746 struct read_plus_res4 { 3747 bool rpr_eof; 3748 read_plus_content rpr_contents<>; 3749 }; 3751 union READ_PLUS4res switch (nfsstat4 rp_status) { 3752 case NFS4_OK: 3753 read_plus_res4 rp_resok4; 3754 default: 3755 void; 3756 }; 3758 3760 15.10.3. DESCRIPTION 3762 The READ_PLUS operation is based upon the NFSv4.1 READ operation (see 3763 Section 18.22 of [RFC5661]) and similarly reads data from the regular 3764 file identified by the current filehandle. 3766 The client provides a rpa_offset of where the READ_PLUS is to start 3767 and a rpa_count of how many bytes are to be read. A rpa_offset of 3768 zero means to read data starting at the beginning of the file. If 3769 rpa_offset is greater than or equal to the size of the file, the 3770 status NFS4_OK is returned with di_length (the data length) set to 3771 zero and eof set to TRUE. 3773 The READ_PLUS result is comprised of an array of rpr_contents, each 3774 of which describe a data_content4 type of data. For NFSv4.2, the 3775 allowed values are data and hole. A server MUST support both the 3776 data type and the hole if it uses READ_PLUS. If it does not want to 3777 support a hole, it MUST use READ. The array contents MUST be 3778 contiguous in the file. 3780 Holes SHOULD be returned in their entirety - clients must be prepared 3781 to get more information than they requested. Both the start and the 3782 end of the hole may exceed what was requested. If data to be 3783 returned is comprised entirely of zeros, then the server SHOULD 3784 return that data as a hole instead. 3786 The server may elect to return adjacent elements of the same type. 3787 For example, if the server has a range of data comprised entirely of 3788 zeros and then a hole, it might want to return two adjacent holes to 3789 the client. 3791 If the client specifies a rpa_count value of zero, the READ_PLUS 3792 succeeds and returns zero bytes of data. In all situations, the 3793 server may choose to return fewer bytes than specified by the client. 3794 The client needs to check for this condition and handle the condition 3795 appropriately. 3797 If the client specifies an rpa_offset and rpa_count value that is 3798 entirely contained within a hole of the file, then the di_offset and 3799 di_length returned MAY be for the entire hole. If the the owner has 3800 a locked byte range covering rpa_offset and rpa_count entirely the 3801 di_offset and di_length MUST NOT be extended outside the locked byte 3802 range. This result is considered valid until the file is changed 3803 (detected via the change attribute). The server MUST provide the 3804 same semantics for the hole as if the client read the region and 3805 received zeroes; the implied holes contents lifetime MUST be exactly 3806 the same as any other read data. 3808 If the client specifies an rpa_offset and rpa_count value that begins 3809 in a non-hole of the file but extends into hole the server should 3810 return an array comprised of both data and a hole. The client MUST 3811 be prepared for the server to return a short read describing just the 3812 data. The client will then issue another READ_PLUS for the remaining 3813 bytes, which the server will respond with information about the hole 3814 in the file. 3816 Except when special stateids are used, the stateid value for a 3817 READ_PLUS request represents a value returned from a previous byte- 3818 range lock or share reservation request or the stateid associated 3819 with a delegation. The stateid identifies the associated owners if 3820 any and is used by the server to verify that the associated locks are 3821 still valid (e.g., have not been revoked). 3823 If the read ended at the end-of-file (formally, in a correctly formed 3824 READ_PLUS operation, if rpa_offset + rpa_count is equal to the size 3825 of the file), or the READ_PLUS operation extends beyond the size of 3826 the file (if rpa_offset + rpa_count is greater than the size of the 3827 file), eof is returned as TRUE; otherwise, it is FALSE. A successful 3828 READ_PLUS of an empty file will always return eof as TRUE. 3830 If the current filehandle is not an ordinary file, an error will be 3831 returned to the client. In the case that the current filehandle 3832 represents an object of type NF4DIR, NFS4ERR_ISDIR is returned. If 3833 the current filehandle designates a symbolic link, NFS4ERR_SYMLINK is 3834 returned. In all other cases, NFS4ERR_WRONG_TYPE is returned. 3836 For a READ_PLUS with a stateid value of all bits equal to zero, the 3837 server MAY allow the READ_PLUS to be serviced subject to mandatory 3838 byte-range locks or the current share deny modes for the file. For a 3839 READ_PLUS with a stateid value of all bits equal to one, the server 3840 MAY allow READ_PLUS operations to bypass locking checks at the 3841 server. 3843 On success, the current filehandle retains its value. 3845 15.10.3.1. Note on Client Support of Arms of the Union 3847 It was decided not to add a means for the client to inform the server 3848 as to which arms of READ_PLUS it would support. In a later minor 3849 version, it may become necessary for the introduction of a new 3850 operation which would allow the client to inform the server as to 3851 whether it supported the new arms of the union of data types 3852 available in READ_PLUS. 3854 15.10.4. IMPLEMENTATION 3856 In general, the IMPLEMENTATION notes for READ in Section 18.22.4 of 3857 [RFC5661] also apply to READ_PLUS. 3859 15.10.4.1. Additional pNFS Implementation Information 3861 With pNFS, the semantics of using READ_PLUS remains the same. Any 3862 data server MAY return a hole result for a READ_PLUS request that it 3863 receives. When a data server chooses to return such a result, it has 3864 the option of returning information for the data stored on that data 3865 server (as defined by the data layout), but it MUST NOT return 3866 results for a byte range that includes data managed by another data 3867 server. 3869 If mandatory locking is enforced, then the data server must also 3870 ensure that to return only information that is within the owner's 3871 locked byte range. 3873 15.10.5. READ_PLUS with Sparse Files Example 3875 The following table describes a sparse file. For each byte range, 3876 the file contains either non-zero data or a hole. In addition, the 3877 server in this example will only create a hole if it is greater than 3878 32K. 3880 +-------------+----------+ 3881 | Byte-Range | Contents | 3882 +-------------+----------+ 3883 | 0-15999 | Hole | 3884 | 16K-31999 | Non-Zero | 3885 | 32K-255999 | Hole | 3886 | 256K-287999 | Non-Zero | 3887 | 288K-353999 | Hole | 3888 | 354K-417999 | Non-Zero | 3889 +-------------+----------+ 3891 Table 5 3893 Under the given circumstances, if a client was to read from the file 3894 with a max read size of 64K, the following will be the results for 3895 the given READ_PLUS calls. This assumes the client has already 3896 opened the file, acquired a valid stateid ('s' in the example), and 3897 just needs to issue READ_PLUS requests. 3899 1. READ_PLUS(s, 0, 64K) --> NFS_OK, eof = false, . Since the first hole is less than the server's 3901 minimum hole size, the first 32K of the file is returned as data 3902 and the remaining 32K is returned as a hole which actually 3903 extends to 256K. 3905 2. READ_PLUS(s, 32K, 64K) --> NFS_OK, eof = false, 3906 The requested range was all zeros, and the current hole begins at 3907 offset 32K and is 224K in length. Note that the client should 3908 not have followed up the previous READ_PLUS request with this one 3909 as the hole information from the previous call extended past what 3910 the client was requesting. 3912 3. READ_PLUS(s, 256K, 64K) --> NFS_OK, eof = false, . Returns an array of the 32K data and 3914 the hole which extends to 354K. 3916 4. READ_PLUS(s, 354K, 64K) --> NFS_OK, eof = true, . Returns the final 64K of data and informs the client 3918 there is no more data in the file. 3920 15.11. Operation 69: SEEK - Find the Next Data or Hole 3922 15.11.1. ARGUMENT 3924 3926 enum data_content4 { 3927 NFS4_CONTENT_DATA = 0, 3928 NFS4_CONTENT_HOLE = 1 3929 }; 3931 struct SEEK4args { 3932 /* CURRENT_FH: file */ 3933 stateid4 sa_stateid; 3934 offset4 sa_offset; 3935 data_content4 sa_what; 3936 }; 3938 3940 15.11.2. RESULT 3942 3944 struct seek_res4 { 3945 bool sr_eof; 3946 offset4 sr_offset; 3947 }; 3949 union SEEK4res switch (nfsstat4 sa_status) { 3950 case NFS4_OK: 3951 seek_res4 resok4; 3952 default: 3953 void; 3954 }; 3956 3958 15.11.3. DESCRIPTION 3960 SEEK is an operation that allows a client to determine the location 3961 of the next data_content4 in a file. It allows an implementation of 3962 the emerging extension to lseek(2) to allow clients to determine the 3963 next hole whilst in data or the next data whilst in a hole. 3965 From the given sa_offset, find the next data_content4 of type sa_what 3966 in the file. If the server can not find a corresponding sa_what, 3967 then the status will still be NFS4_OK, but sr_eof would be TRUE. If 3968 the server can find the sa_what, then the sr_offset is the start of 3969 that content. If the sa_offset is beyond the end of the file, then 3970 SEEK MUST return NFS4ERR_NXIO. 3972 All files MUST have a virtual hole at the end of the file. I.e., if 3973 a filesystem does not support sparse files, then a compound with 3974 {SEEK 0 NFS4_CONTENT_HOLE;} would return a result of {SEEK 1 X;} 3975 where 'X' was the size of the file. 3977 SEEK must follow the same rules for stateids as READ_PLUS 3978 (Section 15.10.3). 3980 15.12. Operation 70: WRITE_SAME - WRITE an ADB Multiple Times to a File 3982 15.12.1. ARGUMENT 3984 3986 enum stable_how4 { 3987 UNSTABLE4 = 0, 3988 DATA_SYNC4 = 1, 3989 FILE_SYNC4 = 2 3990 }; 3992 struct app_data_block4 { 3993 offset4 adb_offset; 3994 length4 adb_block_size; 3995 length4 adb_block_count; 3996 length4 adb_reloff_blocknum; 3997 count4 adb_block_num; 3998 length4 adb_reloff_pattern; 3999 opaque adb_pattern<>; 4000 }; 4002 struct WRITE_SAME4args { 4003 /* CURRENT_FH: file */ 4004 stateid4 wsa_stateid; 4005 stable_how4 wsa_stable; 4006 app_data_block4 wsa_adb; 4007 }; 4009 4011 15.12.2. RESULT 4013 4015 struct write_response4 { 4016 stateid4 wr_callback_id<1>; 4017 length4 wr_count; 4018 stable_how4 wr_committed; 4019 verifier4 wr_writeverf; 4020 }; 4022 union WRITE_SAME4res switch (nfsstat4 wsr_status) { 4023 case NFS4_OK: 4024 write_response4 resok4; 4025 default: 4026 void; 4027 }; 4029 4031 15.12.3. DESCRIPTION 4033 The WRITE_SAME operation writes an application data block to the 4034 regular file identified by the current filehandle (see WRITE SAME 4035 (10) in [T10-SBC2]). The target file is specified by the current 4036 filehandle. The data to be written is specified by an 4037 app_data_block4 structure (Section 8.1.1). The client specifies with 4038 the wsa_stable parameter the method of how the data is to be 4039 processed by the server. It is treated like the stable parameter in 4040 the NFSv4.1 WRITE operation (see Section 18.2 of [RFC5661]). 4042 A successful WRITE_SAME will construct a reply for wr_count, 4043 wr_committed, and wr_writeverf as per the NFSv4.1 WRITE operation 4044 results. If wr_callback_id is set, it indicates an asynchronous 4045 reply (see Section 15.12.3.1). 4047 WRITE_SAME has to support all of the errors which are returned by 4048 WRITE plus NFS4ERR_NOTSUPP, i.e., it is an OPTIONAL operation. If 4049 the client supports WRITE_SAME, it MUST support CB_OFFLOAD. 4051 If the server supports ADBs, then it MUST support the WRITE_SAME 4052 operation. The server has no concept of the structure imposed by the 4053 application. It is only when the application writes to a section of 4054 the file does order get imposed. In order to detect corruption even 4055 before the application utilizes the file, the application will want 4056 to initialize a range of ADBs using WRITE_SAME. 4058 When the client invokes the WRITE_SAME operation, it wants to record 4059 the block structure described by the app_data_block4 on to the file. 4061 When the server receives the WRITE_SAME operation, it MUST populate 4062 adb_block_count ADBs in the file starting at adb_offset. The block 4063 size will be given by adb_block_size. The ADBN (if provided) will 4064 start at adb_reloff_blocknum and each block will be monotonically 4065 numbered starting from adb_block_num in the first block. The pattern 4066 (if provided) will be at adb_reloff_pattern of each block and will be 4067 provided in adb_pattern. 4069 The server SHOULD return an asynchronous result if it can determine 4070 the operation will be long running (see Section 15.12.3.1). Once 4071 either the WRITE_SAME finishes synchronously or the server uses 4072 CB_OFFLOAD to inform the client of the asynchronous completion of the 4073 WRITE_SAME, the server MUST return the ADBs to clients as data. 4075 15.12.3.1. Asynchronous Transactions 4077 ADB initialization may lead to server determining to service the 4078 operation asynchronously. If it decides to do so, it sets the 4079 stateid in wr_callback_id to be that of the wsa_stateid. If it does 4080 not set the wr_callback_id, then the result is synchronous. 4082 When the client determines that the reply will be given 4083 asynchronously, it should not assume anything about the contents of 4084 what it wrote until it is informed by the server that the operation 4085 is complete. It can use OFFLOAD_STATUS (Section 15.9) to monitor the 4086 operation and OFFLOAD_CANCEL (Section 15.8) to cancel the operation. 4087 An example of a asynchronous WRITE_SAME is shown in Figure 6. Note 4088 that as with the COPY operation, WRITE_SAME must provide a stateid 4089 for tracking the asynchronous operation. 4091 Client Server 4092 + + 4093 | | 4094 |--- OPEN ---------------------------->| Client opens 4095 |<------------------------------------/| the file 4096 | | 4097 |--- WRITE_SAME ----------------------->| Client initializes 4098 |<------------------------------------/| an ADB 4099 | | 4100 | | 4101 |--- OFFLOAD_STATUS ------------------>| Client may poll 4102 |<------------------------------------/| for status 4103 | | 4104 | . | Multiple OFFLOAD_STATUS 4105 | . | operations may be sent. 4106 | . | 4107 | | 4108 |<-- CB_OFFLOAD -----------------------| Server reports results 4109 |\------------------------------------>| 4110 | | 4111 |--- CLOSE --------------------------->| Client closes 4112 |<------------------------------------/| the file 4113 | | 4114 | | 4116 Figure 6: An asynchronous WRITE_SAME. 4118 When CB_OFFLOAD informs the client of the successful WRITE_SAME, the 4119 write_response4 embedded in the operation will provide the necessary 4120 information that a synchronous WRITE_SAME would have provided. 4122 Regardless of whether the operation is asynchronous or synchronous, 4123 it MUST still support the COMMIT operation semantics as outlined in 4124 Section 18.3 of [RFC5661]. I.e., COMMIT works on one or more WRITE 4125 operations and the WRITE_SAME operation can appear as several WRITE 4126 operations to the server. The client can use locking operations to 4127 control the behavior on the server with respect to long running 4128 asynchronous write operations. 4130 15.12.3.2. Error Handling of a Partially Complete WRITE_SAME 4132 WRITE_SAME will clone adb_block_count copies of the given ADB in 4133 consecutive order in the file starting at adb_offset. An error can 4134 occur after writing the Nth ADB to the file. WRITE_SAME MUST appear 4135 to populate the range of the file as if the client used WRITE to 4136 transfer the instantiated ADBs. I.e., the contents of the range will 4137 be easy for the client to determine in case of a partially complete 4138 WRITE_SAME. 4140 16. NFSv4.2 Callback Operations 4142 16.1. Operation 15: CB_OFFLOAD - Report results of an asynchronous 4143 operation 4145 16.1.1. ARGUMENT 4147 4149 struct write_response4 { 4150 stateid4 wr_callback_id<1>; 4151 length4 wr_count; 4152 stable_how4 wr_committed; 4153 verifier4 wr_writeverf; 4154 }; 4156 union offload_info4 switch (nfsstat4 coa_status) { 4157 case NFS4_OK: 4158 write_response4 coa_resok4; 4159 default: 4160 length4 coa_bytes_copied; 4161 }; 4163 struct CB_OFFLOAD4args { 4164 nfs_fh4 coa_fh; 4165 stateid4 coa_stateid; 4166 offload_info4 coa_offload_info; 4167 }; 4169 4171 16.1.2. RESULT 4173 4175 struct CB_OFFLOAD4res { 4176 nfsstat4 cor_status; 4177 }; 4179 4181 16.1.3. DESCRIPTION 4183 CB_OFFLOAD is used to report to the client the results of an 4184 asynchronous operation, e.g., Server Side Copy or WRITE_SAME. The 4185 coa_fh and coa_stateid identify the transaction and the coa_status 4186 indicates success or failure. The coa_resok4.wr_callback_id MUST NOT 4187 be set. If the transaction failed, then the coa_bytes_copied 4188 contains the number of bytes copied before the failure occurred. The 4189 coa_bytes_copied value indicates the number of bytes copied but not 4190 which specific bytes have been copied. 4192 If the client supports any of the following operations: 4194 COPY: for both intra-server and inter-server asynchronous copies 4196 WRITE_SAME: for ADB initialization 4198 then the client is REQUIRED to support the CB_OFFLOAD operation. 4200 There is a potential race between the reply to the original 4201 transaction on the forechannel and the CB_OFFLOAD callback on the 4202 backchannel. Sections 2.10.6.3 and 20.9.3 of [RFC5661] describe how 4203 to handle this type of issue. 4205 Upon success, the coa_resok4.wr_count presents for each operation: 4207 COPY: the total number of bytes copied 4209 WRITE_SAME: the same information that a synchronous WRITE_SAME would 4210 provide 4212 17. Security Considerations 4214 NFSv4.2 has all of the security concerns present in NFSv4.1 (see 4215 Section 21 of [RFC5661]) and those present in the Server Side Copy 4216 (see Section 4.10) and in Labeled NFS (see Section 9.7). 4218 18. IANA Considerations 4220 The IANA Considerations for Labeled NFS are addressed in [Quigley14]. 4222 19. References 4224 19.1. Normative References 4226 [NFSv42xdr] 4227 Haynes, T., "Network File System (NFS) Version 4 Minor 4228 Version 2 External Data Representation Standard (XDR) 4229 Description", December 2014. 4231 [RFC3986] Berners-Lee, T., Fielding, R., and L. Masinter, "Uniform 4232 Resource Identifier (URI): Generic Syntax", STD 66, RFC 4233 3986, January 2005. 4235 [RFC5661] Shepler, S., Eisler, M., and D. Noveck, "Network File 4236 System (NFS) Version 4 Minor Version 1 Protocol", RFC 4237 5661, January 2010. 4239 [RFC5662] Shepler, S., Eisler, M., and D. Noveck, "Network File 4240 System (NFS) Version 4 Minor Version 1 External Data 4241 Representation Standard (XDR) Description", RFC 5662, 4242 January 2010. 4244 [posix_fadvise] 4245 The Open Group, "Section 'posix_fadvise()' of System 4246 Interfaces of The Open Group Base Specifications Issue 6, 4247 IEEE Std 1003.1, 2004 Edition", 2004. 4249 [posix_fallocate] 4250 The Open Group, "Section 'posix_fallocate()' of System 4251 Interfaces of The Open Group Base Specifications Issue 6, 4252 IEEE Std 1003.1, 2004 Edition", 2004. 4254 [rpcsec_gssv3] 4255 Adamson, W. and N. Williams, "Remote Procedure Call (RPC) 4256 Security Version 3", December 2014. 4258 19.2. Informative References 4260 [Ashdown08] 4261 Ashdown, L., "Chapter 15, Validating Database Files and 4262 Backups, of Oracle Database Backup and Recovery User's 4263 Guide 11g Release 1 (11.1)", August 2008. 4265 [BL73] Bell, D. and L. LaPadula, "Secure Computer Systems: 4266 Mathematical Foundations and Model", Technical Report 4267 M74-244, The MITRE Corporation, Bedford, MA, May 1973. 4269 [Baira08] Bairavasundaram, L., Goodson, G., Schroeder, B., Arpaci- 4270 Dusseau, A., and R. Arpaci-Dusseau, "An Analysis of Data 4271 Corruption in the Storage Stack", Proceedings of the 6th 4272 USENIX Symposium on File and Storage Technologies (FAST 4273 '08) , 2008. 4275 [I-D.ietf-nfsv4-rfc3530bis] 4276 Haynes, T. and D. Noveck, "Network File System (NFS) 4277 version 4 Protocol", draft-ietf-nfsv4-rfc3530bis-35 (Work 4278 In Progress), November 2014. 4280 [IESG08] ISEG, "IESG Processing of RFC Errata for the IETF Stream", 4281 2008. 4283 [McDougall07] 4284 McDougall, R. and J. Mauro, "Section 11.4.3, Detecting 4285 Memory Corruption of Solaris Internals", 2007. 4287 [NFSv4-Versioning] 4288 Haynes, T. and D. Noveck, "NFSv4 Version Management", 4289 November 2014. 4291 [Quigley14] 4292 Quigley, D., Lu, J., and T. Haynes, "Registry 4293 Specification for Mandatory Access Control (MAC) Security 4294 Label Formats", draft-ietf-nfsv4-lfs-registry-01 (work in 4295 progress), September 2014. 4297 [RFC1108] Kent, S., "Security Options for the Internet Protocol", 4298 RFC 1108, November 1991. 4300 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 4301 Requirement Levels", March 1997. 4303 [RFC2401] Kent, S. and R. Atkinson, "Security Architecture for the 4304 Internet Protocol", RFC 2401, November 1998. 4306 [RFC2616] Fielding, R., Gettys, J., Mogul, J., Frystyk, H., 4307 Masinter, L., Leach, P., and T. Berners-Lee, "Hypertext 4308 Transfer Protocol -- HTTP/1.1", RFC 2616, June 1999. 4310 [RFC4506] Eisler, M., "XDR: External Data Representation Standard", 4311 RFC 4506, May 2006. 4313 [RFC5663] Black, D., Fridella, S., and J. Glasgow, "Parallel NFS 4314 (pNFS) Block/Volume Layout", RFC 5663, January 2010. 4316 [RFC7204] Haynes, T., "Requirements for Labeled NFS", RFC 7204, 4317 April 2014. 4319 [RFC959] Postel, J. and J. Reynolds, "File Transfer Protocol", STD 4320 9, RFC 959, October 1985. 4322 [Strohm11] 4323 Strohm, R., "Chapter 2, Data Blocks, Extents, and 4324 Segments, of Oracle Database Concepts 11g Release 1 4325 (11.1)", January 2011. 4327 [T10-SBC2] 4328 Elliott, R., Ed., "ANSI INCITS 405-2005, Information 4329 Technology - SCSI Block Commands - 2 (SBC-2)", November 4330 2004. 4332 Appendix A. Acknowledgments 4334 Tom Haynes would like to thank NetApp, Inc. for its funding of his 4335 time on this project. 4337 For the pNFS Access Permissions Check, the original draft was by 4338 Sorin Faibish, David Black, Mike Eisler, and Jason Glasgow. The work 4339 was influenced by discussions with Benny Halevy and Bruce Fields. A 4340 review was done by Tom Haynes. 4342 For the Sharing change attribute implementation details with NFSv4 4343 clients, the original draft was by Trond Myklebust. 4345 For the NFS Server Side Copy, the original draft was by James 4346 Lentini, Mike Eisler, Deepak Kenchammana, Anshul Madan, and Rahul 4347 Iyer. Tom Talpey co-authored an unpublished version of that 4348 document. It was also was reviewed by a number of individuals: 4349 Pranoop Erasani, Tom Haynes, Arthur Lent, Trond Myklebust, Dave 4350 Noveck, Theresa Lingutla-Raj, Manjunath Shankararao, Satyam Vaghani, 4351 and Nico Williams. Anna Schumaker's early prototyping experience 4352 helped us avoid some traps. 4354 For the NFS space reservation operations, the original draft was by 4355 Mike Eisler, James Lentini, Manjunath Shankararao, and Rahul Iyer. 4357 For the sparse file support, the original draft was by Dean 4358 Hildebrand and Marc Eshel. Valuable input and advice was received 4359 from Sorin Faibish, Bruce Fields, Benny Halevy, Trond Myklebust, and 4360 Richard Scheffenegger. 4362 For the Application IO Hints, the original draft was by Dean 4363 Hildebrand, Mike Eisler, Trond Myklebust, and Sam Falkner. Some 4364 early reviewers included Benny Halevy and Pranoop Erasani. 4366 For Labeled NFS, the original draft was by David Quigley, James 4367 Morris, Jarret Lu, and Tom Haynes. Peter Staubach, Trond Myklebust, 4368 Stephen Smalley, Sorin Faibish, Nico Williams, and David Black also 4369 contributed in the final push to get this accepted. 4371 Christoph Hellwig was very helpful in getting the WRITE_SAME 4372 semantics to model more of what T10 was doing for WRITE SAME (10) 4373 [T10-SBC2]. And he led the push to get space reservations to more 4374 closely model the posix_fallocate. 4376 Andy Adamson picked up the RPCSEC_GSSv3 work, which enabled both 4377 Labeled NFS and Server Side Copy to be present more secure options. 4379 Christoph Hellwig provided the update to GETDEVICELIST. 4381 During the review process, Talia Reyes-Ortiz helped the sessions run 4382 smoothly. While many people contributed here and there, the core 4383 reviewers were Andy Adamson, Pranoop Erasani, Bruce Fields, Chuck 4384 Lever, Trond Myklebust, David Noveck, Peter Staubach, and Mike 4385 Kupfer. 4387 Appendix B. RFC Editor Notes 4389 [RFC Editor: please remove this section prior to publishing this 4390 document as an RFC] 4392 [RFC Editor: prior to publishing this document as an RFC, please 4393 replace all occurrences of NFSv42xdr with RFCxxxx where xxxx is the 4394 RFC number of the companion XDR document] 4396 Author's Address 4398 Thomas Haynes 4399 Primary Data, Inc. 4400 4300 El Camino Real Ste 100 4401 Los Altos, CA 94022 4402 USA 4404 Phone: +1 408 215 1519 4405 Email: thomas.haynes@primarydata.com