idnits 2.17.1 draft-ietf-nfsv4-minorversion2-35.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 30, 2015) is 3314 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 3888 == Missing Reference: '32K' is mentioned on line 3888, 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 'NFSv42xdr' ** Obsolete normative reference: RFC 5661 (Obsoleted by RFC 8881) == Outdated reference: A later version (-06) exists of draft-ietf-nfsv4-lfs-registry-01 -- Obsolete informational reference (is this intentional?): RFC 2401 (Obsoleted by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 30, 2015 5 Expires: October 1, 2015 7 NFS Version 4 Minor Version 2 8 draft-ietf-nfsv4-minorversion2-35.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 October 1, 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 . . . . . . . . . . . . . 20 96 5. Support for Application IO Hints . . . . . . . . . . . . . . 27 97 6. Sparse Files . . . . . . . . . . . . . . . . . . . . . . . . 27 98 6.1. Introduction . . . . . . . . . . . . . . . . . . . . . . 27 99 6.2. Terminology . . . . . . . . . . . . . . . . . . . . . . . 28 100 6.3. New Operations . . . . . . . . . . . . . . . . . . . . . 28 101 6.3.1. READ_PLUS . . . . . . . . . . . . . . . . . . . . . . 28 102 6.3.2. DEALLOCATE . . . . . . . . . . . . . . . . . . . . . 29 103 7. Space Reservation . . . . . . . . . . . . . . . . . . . . . . 29 104 8. Application Data Block Support . . . . . . . . . . . . . . . 31 105 8.1. Generic Framework . . . . . . . . . . . . . . . . . . . . 32 106 8.1.1. Data Block Representation . . . . . . . . . . . . . . 32 107 8.2. An Example of Detecting Corruption . . . . . . . . . . . 33 108 8.3. Example of READ_PLUS . . . . . . . . . . . . . . . . . . 34 109 8.4. An Example of Zeroing Space . . . . . . . . . . . . . . . 35 110 9. Labeled NFS . . . . . . . . . . . . . . . . . . . . . . . . . 35 111 9.1. Introduction . . . . . . . . . . . . . . . . . . . . . . 35 112 9.2. Definitions . . . . . . . . . . . . . . . . . . . . . . . 36 113 9.3. MAC Security Attribute . . . . . . . . . . . . . . . . . 37 114 9.3.1. Delegations . . . . . . . . . . . . . . . . . . . . . 38 115 9.3.2. Permission Checking . . . . . . . . . . . . . . . . . 38 116 9.3.3. Object Creation . . . . . . . . . . . . . . . . . . . 38 117 9.3.4. Existing Objects . . . . . . . . . . . . . . . . . . 38 118 9.3.5. Label Changes . . . . . . . . . . . . . . . . . . . . 39 119 9.4. pNFS Considerations . . . . . . . . . . . . . . . . . . . 39 120 9.5. Discovery of Server Labeled NFS Support . . . . . . . . . 39 121 9.6. MAC Security NFS Modes of Operation . . . . . . . . . . . 40 122 9.6.1. Full Mode . . . . . . . . . . . . . . . . . . . . . . 40 123 9.6.2. Guest Mode . . . . . . . . . . . . . . . . . . . . . 41 124 9.7. Security Considerations for Labeled NFS . . . . . . . . . 42 125 10. Sharing change attribute implementation characteristics with 126 NFSv4 clients . . . . . . . . . . . . . . . . . . . . . . . . 42 127 11. Error Values . . . . . . . . . . . . . . . . . . . . . . . . 43 128 11.1. Error Definitions . . . . . . . . . . . . . . . . . . . 43 129 11.1.1. General Errors . . . . . . . . . . . . . . . . . . . 43 130 11.1.2. Server to Server Copy Errors . . . . . . . . . . . . 43 131 11.1.3. Labeled NFS Errors . . . . . . . . . . . . . . . . . 44 132 11.2. New Operations and Their Valid Errors . . . . . . . . . 44 133 11.3. New Callback Operations and Their Valid Errors . . . . . 48 134 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 . . . . . . . 52 139 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 . . . . . . . . . . . . . . . . . . . . . . 64 149 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 . . . . . . . . . . . . . . . . . . . . . . . . . 68 153 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 15.13. Operation 71: CLONE - Clone a file . . . . . . . . . . . 90 166 16. NFSv4.2 Callback Operations . . . . . . . . . . . . . . . . . 91 167 16.1. Operation 15: CB_OFFLOAD - Report results of an 168 asynchronous operation . . . . . . . . . . . . . . . . . 91 169 17. Security Considerations . . . . . . . . . . . . . . . . . . . 93 170 18. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 93 171 19. References . . . . . . . . . . . . . . . . . . . . . . . . . 93 172 19.1. Normative References . . . . . . . . . . . . . . . . . . 93 173 19.2. Informative References . . . . . . . . . . . . . . . . . 94 174 Appendix A. Acknowledgments . . . . . . . . . . . . . . . . . . 95 175 Appendix B. RFC Editor Notes . . . . . . . . . . . . . . . . . . 97 176 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 97 178 1. Introduction 180 1.1. The NFS Version 4 Minor Version 2 Protocol 182 The NFS version 4 minor version 2 (NFSv4.2) protocol is the third 183 minor version of the NFS version 4 (NFSv4) protocol. The first minor 184 version, NFSv4.0, is described in [RFC7530] and the second minor 185 version, NFSv4.1, is described in [RFC5661]. 187 As a minor version, NFSv4.2 is consistent with the overall goals for 188 NFSv4, but extends the protocol so as to better meet those goals, 189 based on experiences with NFSv4.1. In addition, NFSv4.2 has adopted 190 some additional goals, which motivate some of the major extensions in 191 NFSv4.2. 193 1.2. Scope of This Document 195 This document describes the NFSv4.2 protocol. With respect to 196 NFSv4.0 and NFSv4.1, this document does not: 198 o describe the NFSv4.0 or NFSv4.1 protocols, except where needed to 199 contrast with NFSv4.2 201 o modify the specification of the NFSv4.0 or NFSv4.1 protocols 203 o clarify the NFSv4.0 or NFSv4.1 protocols. I.e., any 204 clarifications made here apply to NFSv4.2 and neither of the prior 205 protocols 207 The full External Data Representation (XDR) [RFC4506] for NFSv4.2 is 208 presented in [NFSv42xdr]. 210 1.3. NFSv4.2 Goals 212 A major goal of the design of NFSv4.2 is to take common local file 213 system features and offer them remotely. These features might 215 o already be available on the servers, e.g., sparse files 217 o be under development as a new standard, e.g., SEEK pulls in both 218 SEEK_HOLE and SEEK_DATA 220 o be used by clients with the servers via some proprietary means, 221 e.g., Labeled NFS 223 NFSv4.2 provides means for clients to leverage these features on the 224 server in cases in which that had previously not been possible within 225 the confines of the NFS protocol. 227 1.4. Overview of NFSv4.2 Features 229 1.4.1. Server Side Copy 231 A traditional file copy of a remotely accessed, whether from one 232 server to another or between location in the same server, results in 233 the data being put on the network twice - source to client and then 234 client to destination. New operations are introduced to allow 235 unnecessary traffic to be eliminated: 237 The intra-server copy feature allows the client to request the 238 server to perform the copy internally, avoiding unnecessary 239 network traffic. 241 The inter-server copy feature allows the client to authorize the 242 source and destination servers to interact directly. 244 As such copies can be lengthy, asynchronous support is also provided. 246 1.4.2. Application I/O Advise 248 Applications and clients want to advise the server as to expected I/O 249 behavior. Using IO_ADVISE (see Section 15.5) to communicate future I 250 /O behavior such as whether a file will be accessed sequentially or 251 randomly, and whether a file will or will not be accessed in the near 252 future, allows servers to optimize future I/O requests for a file by, 253 for example, prefetching or evicting data. This operation can be 254 used to support the posix_fadvise function. In addition, it may be 255 helpful to applications such as databases and video editors. 257 1.4.3. Sparse Files 259 Sparse files are ones which have unallocated or uninitialized data 260 blocks as holes in the file. Such holes are typically transferred as 261 0s during I/O. READ_PLUS (see Section 15.10) allows a server to send 262 back to the client metadata describing the hole and DEALLOCATE (see 263 Section 15.4) allows the client to punch holes into a file. In 264 addition, SEEK (see Section 15.11) is provided to scan for the next 265 hole or data from a given location. 267 1.4.4. Space Reservation 269 When a file is sparse, one concern applications have is ensuring that 270 there will always be enough data blocks available for the file during 271 future writes. ALLOCATE (see Section 15.1) allows a client to 272 request a guarantee that space will be available. Also DEALLOCATE 273 (see Section 15.4) allows the client to punch a hole into a file, 274 thus releasing a space reservation. 276 1.4.5. Application Data Block (ADB) Support 278 Some applications treat a file as if it were a disk and as such want 279 to initialize (or format) the file image. We introduce WRITE_SAME 280 (see Section 15.12) to send this metadata to the server to allow it 281 to write the block contents. 283 1.4.6. Labeled NFS 285 While both clients and servers can employ Mandatory Access Control 286 (MAC) security models to enforce data access, there has been no 287 protocol support for interoperability. A new file object attribute, 288 sec_label (see Section 12.2.4) allows for the server to store MAC 289 labels on files, which the client retrieves and uses to enforce data 290 access (see Section 9.6.2). The format of the sec_label accommodates 291 any MAC security system. 293 1.5. Enhancements to Minor Versioning Model 295 In NFSv4.1, the only way to introduce new variants of an operation 296 was to introduce a new operation. I.e., READ becomes either READ2 or 297 READ_PLUS. With the use of discriminated unions as parameters to 298 such functions in NFSv4.2, it is possible to add a new arm in a 299 subsequent minor version. And it is also possible to move such an 300 operation from OPTIONAL/RECOMMENDED to REQUIRED. Forcing an 301 implementation to adopt each arm of a discriminated union at such a 302 time does not meet the spirit of the minor versioning rules. As 303 such, new arms of a discriminated union MUST follow the same 304 guidelines for minor versioning as operations in NFSv4.1 - i.e., they 305 may not be made REQUIRED. To support this, a new error code, 306 NFS4ERR_UNION_NOTSUPP, allows the server to communicate to the client 307 that the operation is supported, but the specific arm of the 308 discriminated union is not. 310 2. Minor Versioning 312 NFSv4.2 is a minor version of NFSv4 and is built upon NFSv4.1 as 313 documented in [RFC5661] and [RFC5662]. 315 NFSv4.2 does not modify the rules applicable to the NFSv4 versioning 316 process and follows the rules set out in [RFC5661] or in standard- 317 track documents updating that document (e.g., in an RFC based on 318 [NFSv4-Versioning]). 320 NFSv4.2 only defines extensions to NFSv4.1, each of which may be 321 supported (or not) independently. It does not 323 o introduce infrastructural features 325 o make existing features MANDATORY to NOT implement 327 o change the status of existing features (i.e., by changing their 328 status among OPTIONAL, RECOMMENDED, REQUIRED). 330 The following versioning-related considerations should be noted. 332 o When a new case is added to an existing switch, servers need to 333 report non-support of that new case by returning 334 NFS4ERR_UNION_NOTSUPP. 336 o As regards the potential cross-minor-version transfer of stateids, 337 pNFS implementations of the file mapping type may support of use 338 of an NFSv4.2 metadata sever with NFSv4.1 data servers. In this 339 context, a stateid returned by an NFSv4.2 COMPOUND will be used in 340 an NFSv4.1 COMPOUND directed to the data server (see Sections 3.2 341 and 3.3). 343 3. pNFS considerations for New Operations 345 3.1. Atomicity for ALLOCATE and DEALLOCATE 347 Both ALLOCATE (see Section 15.1) and DEALLOCATE (see Section 15.4) 348 are sent to the metadata server, which is responsible for 349 coordinating the changes onto the storage devices. In particular, 350 both operations must either fully succeed or fail, it cannot be the 351 case that one storage device succeeds whilst another fails. 353 3.2. Sharing of stateids with NFSv4.1 355 A NFSv4.2 metadata server can hand out a layout to a NFSv4.1 storage 356 device. Section 13.9.1 of [RFC5661] discusses how the client gets a 357 stateid from the metadata server to present to a storage device. 359 3.3. NFSv4.2 as a Storage Protocol in pNFS: the File Layout Type 361 A file layout provided by a NFSv4.2 server may refer either to a 362 storage device that only implements NFSv4.1 as specified in 363 [RFC5661], or to a storage device that implements additions from 364 NFSv4.2, in which case the rules in Section 3.3.1 apply. As the File 365 Layout Type does not provide a means for informing the client as to 366 which minor version a particular storage device is providing, it will 367 have to negotiate this via the normal RPC semantics of major and 368 minor version discovery. 370 3.3.1. Operations Sent to NFSv4.2 Data Servers 372 In addition to the commands listed in [RFC5661], NFSv4.2 data servers 373 MAY accept a COMPOUND containing the following additional operations: 374 IO_ADVISE (see Section 15.5), READ_PLUS (see Section 15.10), 375 WRITE_SAME (see Section 15.12), and SEEK (see Section 15.11), which 376 will be treated like the subset specified as "Operations Sent to 377 NFSv4.1 Data Servers" in Section 13.6 of [RFC5661]. 379 Additional details on the implementation of these operations in a 380 pNFS context are documented in the operation specific sections. 382 4. Server Side Copy 384 4.1. Introduction 386 The server-side copy feature provides a mechanism for the NFS client 387 to perform a file copy on a server or between two servers without the 388 data being transmitted back and forth over the network through the 389 NFS client. Without this feature, an NFS client copies data from one 390 location to another by reading the data from the source server over 391 the network, and then writing the data back over the network to the 392 destination server. 394 If the source object and destination object are on different file 395 servers, the file servers will communicate with one another to 396 perform the copy operation. The server-to-server protocol by which 397 this is accomplished is not defined in this document. 399 4.2. Protocol Overview 401 The server-side copy offload operations support both intra-server and 402 inter-server file copies. An intra-server copy is a copy in which 403 the source file and destination file reside on the same server. In 404 an inter-server copy, the source file and destination file are on 405 different servers. In both cases, the copy may be performed 406 synchronously or asynchronously. 408 Throughout the rest of this document, we refer to the NFS server 409 containing the source file as the "source server" and the NFS server 410 to which the file is transferred as the "destination server". In the 411 case of an intra-server copy, the source server and destination 412 server are the same server. Therefore in the context of an intra- 413 server copy, the terms source server and destination server refer to 414 the single server performing the copy. 416 The new operations are designed to copy files. Other file system 417 objects can be copied by building on these operations or using other 418 techniques. For example, if the user wishes to copy a directory, the 419 client can synthesize a directory copy by first creating the 420 destination directory and then copying the source directory's files 421 to the new destination directory. 423 For the inter-server copy, the operations are defined to be 424 compatible with the traditional copy authentication approach. The 425 client and user are authorized at the source for reading. Then they 426 are authorized at the destination for writing. 428 4.2.1. Copy Operations 430 COPY_NOTIFY: Used by the client to notify the source server of a 431 future file copy from a given destination server for the given 432 user. (Section 15.3) 434 COPY: Used by the client to request a file copy. (Section 15.2) 436 OFFLOAD_CANCEL: Used by the client to terminate an asynchronous file 437 copy. (Section 15.8) 439 OFFLOAD_STATUS: Used by the client to poll the status of an 440 asynchronous file copy. (Section 15.9) 442 CB_OFFLOAD: Used by the destination server to report the results of 443 an asynchronous file copy to the client. (Section 16.1) 445 4.2.2. Requirements for Operations 447 The implementation of server-side copy is OPTIONAL by the client and 448 the server. However, in order to successfully copy a file, some 449 operations MUST be supported by the client and/or server. 451 If a client desires an intra-server file copy, then it MUST support 452 the COPY and CB_OFFLOAD operations. If COPY returns a stateid, then 453 the client MAY use the OFFLOAD_CANCEL and OFFLOAD_STATUS operations. 455 If a client desires an inter-server file copy, then it MUST support 456 the COPY, COPY_NOTIFY, and CB_OFFLOAD operations, and MAY use the 457 OFFLOAD_CANCEL operation. If COPY returns a stateid, then the client 458 MAY use the OFFLOAD_CANCEL and OFFLOAD_STATUS operations. 460 If a server supports intra-server copy, then the server MUST support 461 the COPY operation. If a server's COPY operation returns a stateid, 462 then the server MUST also support these operations: CB_OFFLOAD, 463 OFFLOAD_CANCEL, and OFFLOAD_STATUS. 465 If a source server supports inter-server copy, then the source server 466 MUST support all these operations: COPY_NOTIFY and OFFLOAD_CANCEL. 467 If a destination server supports inter-server copy, then the 468 destination server MUST support the COPY operation. If a destination 469 server's COPY operation returns a stateid, then the destination 470 server MUST also support these operations: CB_OFFLOAD, 471 OFFLOAD_CANCEL, COPY_NOTIFY, and OFFLOAD_STATUS. 473 Each operation is performed in the context of the user identified by 474 the ONC RPC credential of its containing COMPOUND or CB_COMPOUND 475 request. For example, an OFFLOAD_CANCEL operation issued by a given 476 user indicates that a specified COPY operation initiated by the same 477 user be canceled. Therefore an OFFLOAD_CANCEL MUST NOT interfere 478 with a copy of the same file initiated by another user. 480 An NFS server MAY allow an administrative user to monitor or cancel 481 copy operations using an implementation specific interface. 483 4.3. Requirements for Inter-Server Copy 485 Inter-server copy is driven by several requirements: 487 o The specification MUST NOT mandate the server-to-server protocol. 489 o The specification MUST provide guidance for using NFSv4.x as a 490 copy protocol. For those source and destination servers willing 491 to use NFSv4.x, there are specific security considerations that 492 this specification MUST address. 494 o The specification MUST NOT mandate preconfiguration between the 495 source and destination server. Requiring that the source and 496 destination first have a "copying relationship" increases the 497 administrative burden. However the specification MUST NOT 498 preclude implementations that require preconfiguration. 500 o The specification MUST NOT mandate a trust relationship between 501 the source and destination server. The NFSv4 security model 502 requires mutual authentication between a principal on an NFS 503 client and a principal on an NFS server. This model MUST continue 504 with the introduction of COPY. 506 4.4. Implementation Considerations 508 4.4.1. Locking the Files 510 Both the source and destination file may need to be locked to protect 511 the content during the copy operations. A client can achieve this by 512 a combination of OPEN and LOCK operations. I.e., either share or 513 byte range locks might be desired. 515 Note that when the client establishes a lock stateid on the source, 516 the context of that stateid is for the client and not the 517 destination. As such, there might already be an outstanding stateid, 518 issued to the destination as client of the source, with the same 519 value as that provided for the lock stateid. The source MUST equate 520 the lock stateid as that of the client, i.e., when the destination 521 presents it in the context of a inter-server copy, it is on behalf of 522 the client. 524 4.4.2. Client Caches 526 In a traditional copy, if the client is in the process of writing to 527 the file before the copy (and perhaps with a write delegation), it 528 will be straightforward to update the destination server. With an 529 inter-server copy, the source has no insight into the changes cached 530 on the client. The client SHOULD write back the data to the source. 531 If it does not do so, it is possible that the destination will 532 receive a corrupt copy of file. 534 4.5. Intra-Server Copy 536 To copy a file on a single server, the client uses a COPY operation. 537 The server may respond to the copy operation with the final results 538 of the copy or it may perform the copy asynchronously and deliver the 539 results using a CB_OFFLOAD operation callback. If the copy is 540 performed asynchronously, the client may poll the status of the copy 541 using OFFLOAD_STATUS or cancel the copy using OFFLOAD_CANCEL. 543 A synchronous intra-server copy is shown in Figure 1. In this 544 example, the NFS server chooses to perform the copy synchronously. 545 The copy operation is completed, either successfully or 546 unsuccessfully, before the server replies to the client's request. 547 The server's reply contains the final result of the operation. 549 Client Server 550 + + 551 | | 552 |--- OPEN ---------------------------->| Client opens 553 |<------------------------------------/| the source file 554 | | 555 |--- OPEN ---------------------------->| Client opens 556 |<------------------------------------/| the destination file 557 | | 558 |--- COPY ---------------------------->| Client requests 559 |<------------------------------------/| a file copy 560 | | 561 |--- CLOSE --------------------------->| Client closes 562 |<------------------------------------/| the destination file 563 | | 564 |--- CLOSE --------------------------->| Client closes 565 |<------------------------------------/| the source file 566 | | 567 | | 569 Figure 1: A synchronous intra-server copy. 571 An asynchronous intra-server copy is shown in Figure 2. In this 572 example, the NFS server performs the copy asynchronously. The 573 server's reply to the copy request indicates that the copy operation 574 was initiated and the final result will be delivered at a later time. 575 The server's reply also contains a copy stateid. The client may use 576 this copy stateid to poll for status information (as shown) or to 577 cancel the copy using an OFFLOAD_CANCEL. When the server completes 578 the copy, the server performs a callback to the client and reports 579 the results. 581 Client Server 582 + + 583 | | 584 |--- OPEN ---------------------------->| Client opens 585 |<------------------------------------/| the source file 586 | | 587 |--- OPEN ---------------------------->| Client opens 588 |<------------------------------------/| the destination file 589 | | 590 |--- COPY ---------------------------->| Client requests 591 |<------------------------------------/| a file copy 592 | | 593 | | 594 |--- OFFLOAD_STATUS ------------------>| Client may poll 595 |<------------------------------------/| for status 596 | | 597 | . | Multiple OFFLOAD_STATUS 598 | . | operations may be sent. 599 | . | 600 | | 601 |<-- CB_OFFLOAD -----------------------| Server reports results 602 |\------------------------------------>| 603 | | 604 |--- CLOSE --------------------------->| Client closes 605 |<------------------------------------/| the destination file 606 | | 607 |--- CLOSE --------------------------->| Client closes 608 |<------------------------------------/| the source file 609 | | 610 | | 612 Figure 2: An asynchronous intra-server copy. 614 4.6. Inter-Server Copy 616 A copy may also be performed between two servers. The copy protocol 617 is designed to accommodate a variety of network topologies. As shown 618 in Figure 3, the client and servers may be connected by multiple 619 networks. In particular, the servers may be connected by a 620 specialized, high speed network (network 192.0.2.0/24 in the diagram) 621 that does not include the client. The protocol allows the client to 622 setup the copy between the servers (over network 203.0.113.0/24 in 623 the diagram) and for the servers to communicate on the high speed 624 network if they choose to do so. 626 192.0.2.0/24 627 +-------------------------------------+ 628 | | 629 | | 630 | 192.0.2.18 | 192.0.2.56 631 +-------+------+ +------+------+ 632 | Source | | Destination | 633 +-------+------+ +------+------+ 634 | 203.0.113.18 | 203.0.113.56 635 | | 636 | | 637 | 203.0.113.0/24 | 638 +------------------+------------------+ 639 | 640 | 641 | 203.0.113.243 642 +-----+-----+ 643 | Client | 644 +-----------+ 646 Figure 3: An example inter-server network topology. 648 For an inter-server copy, the client notifies the source server that 649 a file will be copied by the destination server using a COPY_NOTIFY 650 operation. The client then initiates the copy by sending the COPY 651 operation to the destination server. The destination server may 652 perform the copy synchronously or asynchronously. 654 A synchronous inter-server copy is shown in Figure 4. In this case, 655 the destination server chooses to perform the copy before responding 656 to the client's COPY request. 658 An asynchronous copy is shown in Figure 5. In this case, the 659 destination server chooses to respond to the client's COPY request 660 immediately and then perform the copy asynchronously. 662 Client Source Destination 663 + + + 664 | | | 665 |--- OPEN --->| | Returns os1 666 |<------------------/| | 667 | | | 668 |--- COPY_NOTIFY --->| | 669 |<------------------/| | 670 | | | 671 |--- OPEN ---------------------------->| Returns os2 672 |<------------------------------------/| 673 | | | 674 |--- COPY ---------------------------->| 675 | | | 676 | | | 677 | |<----- read -----| 678 | |\--------------->| 679 | | | 680 | | . | Multiple reads may 681 | | . | be necessary 682 | | . | 683 | | | 684 | | | 685 |<------------------------------------/| Destination replies 686 | | | to COPY 687 | | | 688 |--- CLOSE --------------------------->| Release open state 689 |<------------------------------------/| 690 | | | 691 |--- CLOSE --->| | Release open state 692 |<------------------/| | 694 Figure 4: A synchronous inter-server copy. 696 Client Source Destination 697 + + + 698 | | | 699 |--- OPEN --->| | Returns os1 700 |<------------------/| | 701 | | | 702 |--- LOCK --->| | Optional, could be done 703 |<------------------/| | with a share lock 704 | | | 705 |--- COPY_NOTIFY --->| | Need to pass in 706 |<------------------/| | os1 or lock state 707 | | | 708 | | | 709 | | | 710 |--- OPEN ---------------------------->| Returns os2 711 |<------------------------------------/| 712 | | | 713 |--- LOCK ---------------------------->| Optional ... 714 |<------------------------------------/| 715 | | | 716 |--- COPY ---------------------------->| Need to pass in 717 |<------------------------------------/| os2 or lock state 718 | | | 719 | | | 720 | |<----- read -----| 721 | |\--------------->| 722 | | | 723 | | . | Multiple reads may 724 | | . | be necessary 725 | | . | 726 | | | 727 | | | 728 |--- OFFLOAD_STATUS ------------------>| Client may poll 729 |<------------------------------------/| for status 730 | | | 731 | | . | Multiple OFFLOAD_STATUS 732 | | . | operations may be sent 733 | | . | 734 | | | 735 | | | 736 | | | 737 |<-- CB_OFFLOAD -----------------------| Destination reports 738 |\------------------------------------>| results 739 | | | 740 |--- LOCKU --------------------------->| Only if LOCK was done 741 |<------------------------------------/| 742 | | | 743 |--- CLOSE --------------------------->| Release open state 744 |<------------------------------------/| 745 | | | 746 |--- LOCKU --->| | Only if LOCK was done 747 |<------------------/| | 748 | | | 749 |--- CLOSE --->| | Release open state 750 |<------------------/| | 751 | | | 753 Figure 5: An asynchronous inter-server copy. 755 4.7. Server-to-Server Copy Protocol 757 The choice of what protocol to use in an inter-server copy is 758 ultimately the destination server's decision. However, the 759 destination server has to be cognizant that it is working on behalf 760 of the client. 762 4.7.1. Considerations on Selecting a Copy Protocol 764 The client can have requirements over both the size of transactions 765 and error recovery semantics. It may want to split the copy up such 766 that each chunk is synchronously transferred. It may want the copy 767 protocol to copy the bytes in consecutive order such that upon an 768 error, the client can restart the copy at the last known good offset. 769 If the destination server cannot meet these requirements, the client 770 may prefer the traditional copy mechanism such that it can meet those 771 requirements. 773 4.7.2. Using NFSv4.x as the Copy Protocol 775 The destination server MAY use standard NFSv4.x (where x >= 1) 776 operations to read the data from the source server. If NFSv4.x is 777 used for the server-to-server copy protocol, the destination server 778 can use the source filehandle and ca_src_stateid provided in the COPY 779 request with standard NFSv4.x operations to read data from the source 780 server. Note that the ca_src_stateid MUST be the cnr_stateid 781 returned from the source via the COPY_NOTIFY. 783 4.7.3. Using an Alternative Copy Protocol 785 In a homogeneous environment, the source and destination servers 786 might be able to perform the file copy extremely efficiently using 787 specialized protocols. For example the source and destination 788 servers might be two nodes sharing a common file system format for 789 the source and destination file systems. Thus the source and 790 destination are in an ideal position to efficiently render the image 791 of the source file to the destination file by replicating the file 792 system formats at the block level. Another possibility is that the 793 source and destination might be two nodes sharing a common storage 794 area network, and thus there is no need to copy any data at all, and 795 instead ownership of the file and its contents might simply be re- 796 assigned to the destination. To allow for these possibilities, the 797 destination server is allowed to use a server-to-server copy protocol 798 of its choice. 800 In a heterogeneous environment, using a protocol other than NFSv4.x 801 (e.g., HTTP [RFC2616] or FTP [RFC959]) presents some challenges. In 802 particular, the destination server is presented with the challenge of 803 accessing the source file given only an NFSv4.x filehandle. 805 One option for protocols that identify source files with path names 806 is to use an ASCII hexadecimal representation of the source 807 filehandle as the file name. 809 Another option for the source server is to use URLs to direct the 810 destination server to a specialized service. For example, the 811 response to COPY_NOTIFY could include the URL ftp:// 812 s1.example.com:9999/_FH/0x12345, where 0x12345 is the ASCII 813 hexadecimal representation of the source filehandle. When the 814 destination server receives the source server's URL, it would use 815 "_FH/0x12345" as the file name to pass to the FTP server listening on 816 port 9999 of s1.example.com. On port 9999 there would be a special 817 instance of the FTP service that understands how to convert NFS 818 filehandles to an open file descriptor (in many operating systems, 819 this would require a new system call, one which is the inverse of the 820 makefh() function that the pre-NFSv4 MOUNT service needs). 822 Authenticating and identifying the destination server to the source 823 server is also a challenge. Recommendations for how to accomplish 824 this are given in Section 4.10.1.3. 826 4.8. netloc4 - Network Locations 828 The server-side copy operations specify network locations using the 829 netloc4 data type shown below: 831 833 enum netloc_type4 { 834 NL4_NAME = 1, 835 NL4_URL = 2, 836 NL4_NETADDR = 3 837 }; 838 union netloc4 switch (netloc_type4 nl_type) { 839 case NL4_NAME: utf8str_cis nl_name; 840 case NL4_URL: utf8str_cis nl_url; 841 case NL4_NETADDR: netaddr4 nl_addr; 842 }; 844 846 If the netloc4 is of type NL4_NAME, the nl_name field MUST be 847 specified as a UTF-8 string. The nl_name is expected to be resolved 848 to a network address via DNS, LDAP, NIS, /etc/hosts, or some other 849 means. If the netloc4 is of type NL4_URL, a server URL [RFC3986] 850 appropriate for the server-to-server copy operation is specified as a 851 UTF-8 string. If the netloc4 is of type NL4_NETADDR, the nl_addr 852 field MUST contain a valid netaddr4 as defined in Section 3.3.9 of 853 [RFC5661]. 855 When netloc4 values are used for an inter-server copy as shown in 856 Figure 3, their values may be evaluated on the source server, 857 destination server, and client. The network environment in which 858 these systems operate should be configured so that the netloc4 values 859 are interpreted as intended on each system. 861 4.9. Copy Offload Stateids 863 A server may perform a copy offload operation asynchronously. An 864 asynchronous copy is tracked using a copy offload stateid. Copy 865 offload stateids are included in the COPY, OFFLOAD_CANCEL, 866 OFFLOAD_STATUS, and CB_OFFLOAD operations. 868 A copy offload stateid will be valid until either (A) the client or 869 server restarts or (B) the client returns the resource by issuing a 870 OFFLOAD_CANCEL operation or the client replies to a CB_OFFLOAD 871 operation. 873 A copy offload stateid's seqid MUST NOT be 0. In the context of a 874 copy offload operation, it is ambiguous to indicate the most recent 875 copy offload operation using a stateid with seqid of 0. Therefore a 876 copy offload stateid with seqid of 0 MUST be considered invalid. 878 4.10. Security Considerations 880 The security considerations pertaining to NFSv4.1 [RFC5661] apply to 881 this section. And as such, the standard security mechanisms used by 882 the protocol can be used to secure the server-to-server operations. 884 NFSv4 clients and servers supporting the inter-server copy operations 885 described in this chapter are REQUIRED to implement the mechanism 886 described in Section 4.10.1.1, and to support rejecting COPY_NOTIFY 887 requests that do not use RPCSEC_GSS with privacy. If the server-to- 888 server copy protocol is ONC RPC based, the servers are also REQUIRED 889 to implement [rpcsec_gssv3] including the RPCSEC_GSSv3 copy_to_auth, 890 copy_from_auth, and copy_confirm_auth structured privileges. This 891 requirement to implement is not a requirement to use; for example, a 892 server may depending on configuration also allow COPY_NOTIFY requests 893 that use only AUTH_SYS. 895 4.10.1. Inter-Server Copy Security 897 4.10.1.1. Inter-Server Copy via ONC RPC with RPCSEC_GSSv3 899 When the client sends a COPY_NOTIFY to the source server to expect 900 the destination to attempt to copy data from the source server, it is 901 expected that this copy is being done on behalf of the principal 902 (called the "user principal") that sent the RPC request that encloses 903 the COMPOUND procedure that contains the COPY_NOTIFY operation. The 904 user principal is identified by the RPC credentials. A mechanism 905 that allows the user principal to authorize the destination server to 906 perform the copy, that lets the source server properly authenticate 907 the destination's copy, and does not allow the destination server to 908 exceed this authorization, is necessary. 910 An approach that sends delegated credentials of the client's user 911 principal to the destination server is not used for the following 912 reason. If the client's user delegated its credentials, the 913 destination would authenticate as the user principal. If the 914 destination were using the NFSv4 protocol to perform the copy, then 915 the source server would authenticate the destination server as the 916 user principal, and the file copy would securely proceed. However, 917 this approach would allow the destination server to copy other files. 918 The user principal would have to trust the destination server to not 919 do so. This is counter to the requirements, and therefore is not 920 considered. 922 Instead, a feature of the RPCSEC_GSSv3 [rpcsec_gssv3] protocol can be 923 used: RPC application defined structured privilege assertion. This 924 features allow the destination server to authenticate to the source 925 server as acting on behalf of the user principal, and to authorize 926 the destination server to perform READs of the file to be copied from 927 the source on behalf of the user principal. Once the copy is 928 complete, the client can destroy the RPCSEC_GSSv3 handles to end the 929 authorization of both the source and destination servers to copy. 931 We define three RPCSEC_GSSv3 structured privilege assertions that 932 work in tandem to authorize the copy: 934 copy_from_auth: A user principal is authorizing a source principal 935 ("nfs@") to allow a destination principal 936 ("nfs@") to setup the copy_confirm_auth privilege 937 required to copy a file from the source to the destination on 938 behalf of the user principal. This privilege is established on 939 the source server before the user principal sends a COPY_NOTIFY 940 operation to the source server, and the resultant RPCSEC_GSSv3 941 context is used to secure the COPY_NOTIFY operation. 943 945 struct copy_from_auth_priv { 946 secret4 cfap_shared_secret; 947 netloc4 cfap_destination; 948 /* the NFSv4 user name that the user principal maps to */ 949 utf8str_mixed cfap_username; 950 }; 952 954 cfp_shared_secret is an automatically generated random number 955 secret value. 957 copy_to_auth: A user principal is authorizing a destination 958 principal ("nfs@") to setup a copy_confirm_auth 959 privilege with a source principal ("nfs@") to allow it to 960 copy a file from the source to the destination on behalf of the 961 user principal. This privilege is established on the destination 962 server before the user principal sends a COPY operation to the 963 destination server, and the resultant RPCSEC_GSSv3 context is used 964 to secure the COPY operation. 966 968 struct copy_to_auth_priv { 969 /* equal to cfap_shared_secret */ 970 secret4 ctap_shared_secret; 971 netloc4 ctap_source<>; 972 /* the NFSv4 user name that the user principal maps to */ 973 utf8str_mixed ctap_username; 974 }; 976 978 ctap_shared_secret is the automatically generated secret value 979 used to establish the copy_from_auth privilege with the source 980 principal. See Section 4.10.1.1.1. 982 copy_confirm_auth: A destination principal ("nfs@") is 983 confirming with the source principal ("nfs@") that it is 984 authorized to copy data from the source. This privilege is 985 established on the destination server before the file is copied 986 from the source to the destination. The resultant RPCSEC_GSSv3 987 context is used to secure the READ operations from the source to 988 the destination server. 990 992 struct copy_confirm_auth_priv { 993 /* equal to GSS_GetMIC() of cfap_shared_secret */ 994 opaque ccap_shared_secret_mic<>; 995 /* the NFSv4 user name that the user principal maps to */ 996 utf8str_mixed ccap_username; 997 }; 999 1001 4.10.1.1.1. Establishing a Security Context 1003 When the user principal wants to COPY a file between two servers, if 1004 it has not established copy_from_auth and copy_to_auth privileges on 1005 the servers, it establishes them: 1007 o As noted in [rpcsec_gssv3] the client uses an existing 1008 RPCSEC_GSSv3 context termed the "parent" handle to establish and 1009 protect RPCSEC_GSSv3 structured privilege assertion exchanges. 1010 The copy_from_auth privilege will use the context established 1011 between the user principal and the source server used to OPEN the 1012 source file as the RPCSEC_GSSv3 parent handle. The copy_to_auth 1013 privilege will use the context established between the user 1014 principal and the destination server used to OPEN the destination 1015 file as the RPCSEC_GSSv3 parent handle. 1017 o A random number is generated to use as a secret to be shared 1018 between the two servers. This shared secret will be placed in the 1019 cfap_shared_secret and ctap_shared_secret fields of the 1020 appropriate privilege data types, copy_from_auth_priv and 1021 copy_to_auth_priv. Because of this shared_secret the 1022 RPCSEC_GSS3_CREATE control messages for copy_from_auth and 1023 copy_to_auth MUST use a QOP of rpc_gss_svc_privacy. 1025 o An instance of copy_from_auth_priv is filled in with the shared 1026 secret, the destination server, and the NFSv4 user id of the user 1027 principal and is placed in rpc_gss3_create_args 1028 assertions[0].privs.privilege. The string "copy_from_auth" is 1029 placed in assertions[0].privs.name. The source server unwraps the 1030 rpc_gss_svc_privacy RPCSEC_GSS3_CREATE payload and verifies that 1031 the NFSv4 user id being asserted matches the source server's 1032 mapping of the user principal. If it does, the privilege is 1033 established on the source server as: <"copy_from_auth", user id, 1034 destination>. The field "handle" in a successful reply is the 1035 RPCSEC_GSSv3 copy_from_auth "child" handle that the client will 1036 use on COPY_NOTIFY requests to the source server. 1038 o An instance of copy_to_auth_priv is filled in with the shared 1039 secret, the cnr_source_server list returned by COPY_NOTIFY, and 1040 the NFSv4 user id of the user principal. The copy_to_auth_priv 1041 instance is placed in rpc_gss3_create_args 1042 assertions[0].privs.privilege. The string "copy_to_auth" is 1043 placed in assertions[0].privs.name. The destination server 1044 unwraps the rpc_gss_svc_privacy RPCSEC_GSS3_CREATE payload and 1045 verifies that the NFSv4 user id being asserted matches the 1046 destination server's mapping of the user principal. If it does, 1047 the privilege is established on the destination server as: 1048 <"copy_to_auth", user id, source list>. The field "handle" in a 1049 successful reply is the RPCSEC_GSSv3 copy_to_auth "child" handle 1050 that the client will use on COPY requests to the destination 1051 server involving the source server. 1053 As noted in [rpcsec_gssv3] Section 2.3.1 "Create Request", both the 1054 client and the source server should associate the RPCSEC_GSSv3 1055 "child" handle with the parent RPCSEC_GSSv3 handle used to create the 1056 RPCSEC_GSSv3 child handle. 1058 4.10.1.1.2. Starting a Secure Inter-Server Copy 1060 When the client sends a COPY_NOTIFY request to the source server, it 1061 uses the privileged "copy_from_auth" RPCSEC_GSSv3 handle. 1062 cna_destination_server in COPY_NOTIFY MUST be the same as 1063 cfap_destination specified in copy_from_auth_priv. Otherwise, 1064 COPY_NOTIFY will fail with NFS4ERR_ACCESS. The source server 1065 verifies that the privilege <"copy_from_auth", user id, destination> 1066 exists, and annotates it with the source filehandle, if the user 1067 principal has read access to the source file, and if administrative 1068 policies give the user principal and the NFS client read access to 1069 the source file (i.e., if the ACCESS operation would grant read 1070 access). Otherwise, COPY_NOTIFY will fail with NFS4ERR_ACCESS. 1072 When the client sends a COPY request to the destination server, it 1073 uses the privileged "copy_to_auth" RPCSEC_GSSv3 handle. 1074 ca_source_server list in COPY MUST be the same as ctap_source list 1075 specified in copy_to_auth_priv. Otherwise, COPY will fail with 1076 NFS4ERR_ACCESS. The destination server verifies that the privilege 1077 <"copy_to_auth", user id, source list> exists, and annotates it with 1078 the source and destination filehandles. If the COPY returns a 1079 wr_callback_id, then this is an asynchronous copy and the 1080 wr_callback_id must also must be annotated to the copy_to_auth 1081 privilege. If the client has failed to establish the "copy_to_auth" 1082 privilege it will reject the request with NFS4ERR_PARTNER_NO_AUTH. 1084 If either the COPY_NOTIFY, or the COPY operations fail, the 1085 associated "copy_from_auth" and "copy_to_auth" RPCSEC_GSSv3 handles 1086 MUST be destroyed. 1088 4.10.1.1.3. Securing ONC RPC Server-to-Server Copy Protocols 1090 After a destination server has a "copy_to_auth" privilege established 1091 on it, and it receives a COPY request, if it knows it will use an ONC 1092 RPC protocol to copy data, it will establish a "copy_confirm_auth" 1093 privilege on the source server prior to responding to the COPY 1094 operation as follows: 1096 o Before establishing an RPCSEC_GSSv3 context, a parent context 1097 needs to exist between nfs@ as the initiator 1098 principal, and nfs@ as the target principal. If NFS is to 1099 be used as the copy protocol, this means that the destination 1100 server must mount the source server using RPCSEC_GSSv3. 1102 o An instance of copy_confirm_auth_priv is filled in with 1103 information from the established "copy_to_auth" privilege. The 1104 value of the field ccap_shared_secret_mic is a GSS_GetMIC() of the 1105 ctap_shared_secret in the copy_to_auth privilege using the parent 1106 handle context. The field ccap_username is the mapping of the 1107 user principal to an NFSv4 user name ("user"@"domain" form), and 1108 MUST be the same as the ctap_username in the copy_to_auth 1109 privilege. The copy_confirm_auth_priv instance is placed in 1110 rpc_gss3_create_args assertions[0].privs.privilege. The string 1111 "copy_confirm_auth" is placed in assertions[0].privs.name. 1113 o The RPCSEC_GSS3_CREATE copy_from_auth message is sent to the 1114 source server with a QOP of rpc_gss_svc_privacy. The source 1115 server unwraps the rpc_gss_svc_privacy RPCSEC_GSS3_CREATE payload 1116 and verifies the cap_shared_secret_mic by calling GSS_VerifyMIC() 1117 using the parent context on the cfap_shared_secret from the 1118 established "copy_from_auth" privilege, and verifies the that the 1119 ccap_username equals the cfap_username. 1121 o If all verification succeeds, the "copy_confirm_auth" privilege is 1122 established on the source server as < "copy_confirm_auth", 1123 shared_secret_mic, user id> Because the shared secret has been 1124 verified, the resultant copy_confirm_auth RPCSEC_GSSv3 child 1125 handle is noted to be acting on behalf of the user principal. 1127 o If the source server fails to verify the copy_from_auth privilege 1128 the COPY operation will be rejected with NFS4ERR_PARTNER_NO_AUTH, 1129 causing in turn the client to destroy the associated 1130 copy_from_auth and copy_to_auth RPCSEC_GSSv3 structured privilege 1131 assertion handles. 1133 o All subsequent ONC RPC READ requests sent from the destination to 1134 copy data from the source to the destination will use the 1135 RPCSEC_GSSv3 copy_confirm_auth child handle. 1137 Note that the use of the "copy_confirm_auth" privilege accomplishes 1138 the following: 1140 o If a protocol like NFS is being used, with export policies, export 1141 policies can be overridden in case the destination server as-an- 1142 NFS-client is not authorized 1144 o Manual configuration to allow a copy relationship between the 1145 source and destination is not needed. 1147 4.10.1.1.4. Maintaining a Secure Inter-Server Copy 1149 If the client determines that either the copy_from_auth or the 1150 copy_to_auth handle becomes invalid during a copy, then the copy MUST 1151 be aborted by the client sending an OFFLOAD_CANCEL to both the source 1152 and destination servers and destroying the respective copy related 1153 context handles as described in Section 4.10.1.1.5. 1155 4.10.1.1.5. Finishing or Stopping a Secure Inter-Server Copy 1157 Under normal operation, the client MUST destroy the copy_from_auth 1158 and the copy_to_auth RPCSEC_GSSv3 handle once the COPY operation 1159 returns for a synchronous inter-server copy or a CB_OFFLOAD reports 1160 the result of an asynchronous copy. 1162 The copy_confirm_auth privilege constructed from information held by 1163 the copy_to_auth privilege, and MUST be destroyed by the destination 1164 server (via an RPCSEC_GSS3_DESTROY call) when the copy_to_auth 1165 RPCSEC_GSSv3 handle is destroyed. 1167 The copy_confirm_auth RPCSEC_GSS3 handle is associated with a 1168 copy_from_auth RPCSEC_GSS3 handle on the source server via the shared 1169 secret and MUST be locally destroyed (there is no RPCSEC_GSS3_DESTROY 1170 as the source server is not the initiator) when the copy_from_auth 1171 RPCSEC_GSSv3 handle is destroyed. 1173 If the client sends an OFFLOAD_CANCEL to the source server to rescind 1174 the destination server's synchronous copy privilege, it uses the 1175 privileged "copy_from_auth" RPCSEC_GSSv3 handle and the 1176 cra_destination_server in OFFLOAD_CANCEL MUST be the same as the name 1177 of the destination server specified in copy_from_auth_priv. The 1178 source server will then delete the <"copy_from_auth", user id, 1179 destination> privilege and fail any subsequent copy requests sent 1180 under the auspices of this privilege from the destination server. 1181 The client MUST destroy both the "copy_from_auth" and the 1182 "copy_to_auth" RPCSEC_GSSv3 handles. 1184 If the client sends an OFFLOAD_STATUS to the destination server to 1185 check on the status of an asynchronous copy, it uses the privileged 1186 "copy_to_auth" RPCSEC_GSSv3 handle and the osa_stateid in 1187 OFFLOAD_STATUS MUST be the same as the wr_callback_id specified in 1188 the "copy_to_auth" privilege stored on the destination server. 1190 If the client sends an OFFLOAD_CANCEL to the destination server to 1191 cancel an asynchronous copy, it uses the privileged "copy_to_auth" 1192 RPCSEC_GSSv3 handle and the oaa_stateid in OFFLOAD_CANCEL MUST be the 1193 same as the wr_callback_id specified in the "copy_to_auth" privilege 1194 stored on the destination server. The destination server will then 1195 delete the <"copy_to_auth", user id, source list, nounce, nounce MIC, 1196 context handle, handle version> privilege and the associated 1197 "copy_confirm_auth" RPCSEC_GSSv3 handle. The client MUST destroy 1198 both the copy_to_auth and copy_from_auth RPCSEC_GSSv3 handles. 1200 4.10.1.2. Inter-Server Copy via ONC RPC without RPCSEC_GSS 1202 ONC RPC security flavors other than RPCSEC_GSS MAY be used with the 1203 server-side copy offload operations described in this chapter. In 1204 particular, host-based ONC RPC security flavors such as AUTH_NONE and 1205 AUTH_SYS MAY be used. If a host-based security flavor is used, a 1206 minimal level of protection for the server-to-server copy protocol is 1207 possible. 1209 In the absence of a strong security mechanism designed for the 1210 purpose, the challenge is how the source server and destination 1211 server identify themselves to each other, especially in the presence 1212 of multi-homed source and destination servers. In a multi-homed 1213 environment, the destination server might not contact the source 1214 server from the same network address specified by the client in the 1215 COPY_NOTIFY. The cnr_stateid returned from the COPY_NOTIFY can be 1216 used to uiniquely identify the destination server to the source 1217 server. The use of cnr_stateid provides initial authentication of 1218 the destination server, but cannot defend against man-in-the-middle 1219 attacks after authentication or an eavesdropper that observes the 1220 opaque stateid on the wire. Other secure communication techniques 1221 (e.g., IPsec) are necessary to block these attacks. 1223 Servers SHOULD reject COPY_NOTIFY requests that do not use RPCSEC_GSS 1224 with privacy, thus ensuring the cnr_stateid in the COPY_NOTIFY reply 1225 is encrypted. For the same reason, clients SHOULD send COPY requests 1226 to the destination using RPCSEC_GSS with privacy. 1228 4.10.1.3. Inter-Server Copy without ONC RPC 1230 The same techniques as Section 4.10.1.2, using unique URLs for each 1231 destination server, can be used for other protocols (e.g., HTTP 1232 [RFC2616] and FTP [RFC959]) as well. 1234 5. Support for Application IO Hints 1236 Applications can issue client I/O hints via posix_fadvise() 1237 [posix_fadvise] to the NFS client. While this can help the NFS 1238 client optimize I/O and caching for a file, it does not allow the NFS 1239 server and its exported file system to do likewise. We add an 1240 IO_ADVISE procedure (Section 15.5) to communicate the client file 1241 access patterns to the NFS server. The NFS server upon receiving a 1242 IO_ADVISE operation MAY choose to alter its I/O and caching behavior, 1243 but is under no obligation to do so. 1245 Application specific NFS clients such as those used by hypervisors 1246 and databases can also leverage application hints to communicate 1247 their specialized requirements. 1249 6. Sparse Files 1251 6.1. Introduction 1253 A sparse file is a common way of representing a large file without 1254 having to utilize all of the disk space for it. Consequently, a 1255 sparse file uses less physical space than its size indicates. This 1256 means the file contains 'holes', byte ranges within the file that 1257 contain no data. Most modern file systems support sparse files, 1258 including most UNIX file systems and NTFS, but notably not Apple's 1259 HFS+. Common examples of sparse files include Virtual Machine (VM) 1260 OS/disk images, database files, log files, and even checkpoint 1261 recovery files most commonly used by the HPC community. 1263 In addition many modern file systems support the concept of 1264 'unwritten' or 'uninitialized' blocks, which have uninitialized space 1265 allocated to them on disk, but will return zeros until data is 1266 written to them. Such functionality is already present in the data 1267 model of the pNFS Block/Volume Layout (see [RFC5663]). Uninitialized 1268 blocks can thought as holes inside a space reservation window. 1270 If an application reads a hole in a sparse file, the file system must 1271 return all zeros to the application. For local data access there is 1272 little penalty, but with NFS these zeroes must be transferred back to 1273 the client. If an application uses the NFS client to read data into 1274 memory, this wastes time and bandwidth as the application waits for 1275 the zeroes to be transferred. 1277 A sparse file is typically created by initializing the file to be all 1278 zeros - nothing is written to the data in the file, instead the hole 1279 is recorded in the metadata for the file. So a 8G disk image might 1280 be represented initially by a couple hundred bits in the inode and 1281 nothing on the disk. If the VM then writes 100M to a file in the 1282 middle of the image, there would now be two holes represented in the 1283 metadata and 100M in the data. 1285 No new operation is needed to allow the creation of a sparsely 1286 populated file, when a file is created and a write occurs past the 1287 current size of the file, the non-allocated region will either be a 1288 hole or filled with zeros. The choice of behavior is dictated by the 1289 underlying file system and is transparent to the application. What 1290 is needed are the abilities to read sparse files and to punch holes 1291 to reinitialize the contents of a file. 1293 Two new operations DEALLOCATE (Section 15.4) and READ_PLUS 1294 (Section 15.10) are introduced. DEALLOCATE allows for the hole 1295 punching. I.e., an application might want to reset the allocation 1296 and reservation status of a range of the file. READ_PLUS supports 1297 all the features of READ but includes an extension to support sparse 1298 files. READ_PLUS is guaranteed to perform no worse than READ, and 1299 can dramatically improve performance with sparse files. READ_PLUS 1300 does not depend on pNFS protocol features, but can be used by pNFS to 1301 support sparse files. 1303 6.2. Terminology 1305 Regular file: An object of file type NF4REG or NF4NAMEDATTR. 1307 Sparse file: A Regular file that contains one or more holes. 1309 Hole: A byte range within a Sparse file that contains regions of all 1310 zeroes. A hole might or might not have space allocated or 1311 reserved to it. 1313 6.3. New Operations 1315 6.3.1. READ_PLUS 1317 READ_PLUS is a new variant of the NFSv4.1 READ operation [RFC5661]. 1318 Besides being able to support all of the data semantics of the READ 1319 operation, it can also be used by the client and server to 1320 efficiently transfer holes. Note that as the client has no a priori 1321 knowledge of whether a hole is present or not, if the client supports 1322 READ_PLUS and so does the server, then it should always use the 1323 READ_PLUS operation in preference to the READ operation. 1325 READ_PLUS extends the response with a new arm representing holes to 1326 avoid returning data for portions of the file which are initialized 1327 to zero and may or may not contain a backing store. Returning data 1328 blocks of uninitialized data wastes computational and network 1329 resources, thus reducing performance. 1331 When a client sends a READ operation, it is not prepared to accept a 1332 READ_PLUS-style response providing a compact encoding of the scope of 1333 holes. If a READ occurs on a sparse file, then the server must 1334 expand such data to be raw bytes. If a READ occurs in the middle of 1335 a hole, the server can only send back bytes starting from that 1336 offset. By contrast, if a READ_PLUS occurs in the middle of a hole, 1337 the server can send back a range which starts before the offset and 1338 extends past the range. 1340 6.3.2. DEALLOCATE 1342 DEALLOCATE can be used to hole punch, which allows the client to 1343 avoid the transfer of a repetitive pattern of zeros across the 1344 network. 1346 7. Space Reservation 1348 Applications want to be able to reserve space for a file, report the 1349 amount of actual disk space a file occupies, and free-up the backing 1350 space of a file when it is not required. 1352 One example is the posix_fallocate ([posix_fallocate]) which allows 1353 applications to ask for space reservations from the operating system, 1354 usually to provide a better file layout and reduce overhead for 1355 random or slow growing file appending workloads. 1357 Another example is space reservation for virtual disks in a 1358 hypervisor. In virtualized environments, virtual disk files are 1359 often stored on NFS mounted volumes. When a hypervisor creates a 1360 virtual disk file, it often tries to preallocate the space for the 1361 file so that there are no future allocation related errors during the 1362 operation of the virtual machine. Such errors prevent a virtual 1363 machine from continuing execution and result in downtime. 1365 Currently, in order to achieve such a guarantee, applications zero 1366 the entire file. The initial zeroing allocates the backing blocks 1367 and all subsequent writes are overwrites of already allocated blocks. 1368 This approach is not only inefficient in terms of the amount of I/O 1369 done, it is also not guaranteed to work on file systems that are log 1370 structured or deduplicated. An efficient way of guaranteeing space 1371 reservation would be beneficial to such applications. 1373 The new ALLOCATE operation (see Section 15.1) allows a client to 1374 request a guarantee that space will be available. The ALLOCATE 1375 operation guarantees that any future writes to the region it was 1376 successfully called for will not fail with NFS4ERR_NOSPC. 1378 Another useful feature is the ability to report the number of blocks 1379 that would be freed when a file is deleted. Currently, NFS reports 1380 two size attributes: 1382 size The logical file size of the file. 1384 space_used The size in bytes that the file occupies on disk 1386 While these attributes are sufficient for space accounting in 1387 traditional file systems, they prove to be inadequate in modern file 1388 systems that support block sharing. In such file systems, multiple 1389 inodes can point to a single block with a block reference count to 1390 guard against premature freeing. Having a way to tell the number of 1391 blocks that would be freed if the file was deleted would be useful to 1392 applications that wish to migrate files when a volume is low on 1393 space. 1395 Since virtual disks represent a hard drive in a virtual machine, a 1396 virtual disk can be viewed as a file system within a file. Since not 1397 all blocks within a file system are in use, there is an opportunity 1398 to reclaim blocks that are no longer in use. A call to deallocate 1399 blocks could result in better space efficiency. Lesser space MAY be 1400 consumed for backups after block deallocation. 1402 The following operations and attributes can be used to resolve these 1403 issues: 1405 space_freed This attribute specifies the space freed when a file is 1406 deleted, taking block sharing into consideration. 1408 DEALLOCATE This operation delallocates the blocks backing a region 1409 of the file. 1411 If space_used of a file is interpreted to mean the size in bytes of 1412 all disk blocks pointed to by the inode of the file, then shared 1413 blocks get double counted, over-reporting the space utilization. 1414 This also has the adverse effect that the deletion of a file with 1415 shared blocks frees up less than space_used bytes. 1417 On the other hand, if space_used is interpreted to mean the size in 1418 bytes of those disk blocks unique to the inode of the file, then 1419 shared blocks are not counted in any file, resulting in under- 1420 reporting of the space utilization. 1422 For example, two files A and B have 10 blocks each. Let 6 of these 1423 blocks be shared between them. Thus, the combined space utilized by 1424 the two files is 14 * BLOCK_SIZE bytes. In the former case, the 1425 combined space utilization of the two files would be reported as 20 * 1426 BLOCK_SIZE. However, deleting either would only result in 4 * 1427 BLOCK_SIZE being freed. Conversely, the latter interpretation would 1428 report that the space utilization is only 8 * BLOCK_SIZE. 1430 Adding another size attribute, space_freed (see Section 12.2.2), is 1431 helpful in solving this problem. space_freed is the number of blocks 1432 that are allocated to the given file that would be freed on its 1433 deletion. In the example, both A and B would report space_freed as 4 1434 * BLOCK_SIZE and space_used as 10 * BLOCK_SIZE. If A is deleted, B 1435 will report space_freed as 10 * BLOCK_SIZE as the deletion of B would 1436 result in the deallocation of all 10 blocks. 1438 The addition of these attributes does not solve the problem of space 1439 being over-reported. However, over-reporting is better than under- 1440 reporting. 1442 8. Application Data Block Support 1444 At the OS level, files are contained on disk blocks. Applications 1445 are also free to impose structure on the data contained in a file and 1446 we can define an Application Data Block (ADB) to be such a structure. 1447 From the application's viewpoint, it only wants to handle ADBs and 1448 not raw bytes (see [Strohm11]). An ADB is typically comprised of two 1449 sections: header and data. The header describes the characteristics 1450 of the block and can provide a means to detect corruption in the data 1451 payload. The data section is typically initialized to all zeros. 1453 The format of the header is application specific, but there are two 1454 main components typically encountered: 1456 1. An Application Data Block Number (ADBN) which allows the 1457 application to determine which data block is being referenced. 1458 This is useful when the client is not storing the blocks in 1459 contiguous memory, i.e., a logical block number. 1461 2. Fields to describe the state of the ADB and a means to detect 1462 block corruption. For both pieces of data, a useful property is 1463 that allowed values be unique in that if passed across the 1464 network, corruption due to translation between big and little 1465 endian architectures are detectable. For example, 0xF0DEDEF0 has 1466 the same bit pattern in both architectures. 1468 Applications already impose structures on files [Strohm11] and detect 1469 corruption in data blocks [Ashdown08]. What they are not able to do 1470 is efficiently transfer and store ADBs. To initialize a file with 1471 ADBs, the client must send each full ADB to the server and that must 1472 be stored on the server. 1474 In this section, we define a framework for transferring the ADB from 1475 client to server and present one approach to detecting corruption in 1476 a given ADB implementation. 1478 8.1. Generic Framework 1480 We want the representation of the ADB to be flexible enough to 1481 support many different applications. The most basic approach is no 1482 imposition of a block at all, which means we are working with the raw 1483 bytes. Such an approach would be useful for storing holes, punching 1484 holes, etc. In more complex deployments, a server might be 1485 supporting multiple applications, each with their own definition of 1486 the ADB. One might store the ADBN at the start of the block and then 1487 have a guard pattern to detect corruption [McDougall07]. The next 1488 might store the ADBN at an offset of 100 bytes within the block and 1489 have no guard pattern at all, i.e., existing applications might 1490 already have well defined formats for their data blocks. 1492 The guard pattern can be used to represent the state of the block, to 1493 protect against corruption, or both. Again, it needs to be able to 1494 be placed anywhere within the ADB. 1496 We need to be able to represent the starting offset of the block and 1497 the size of the block. Note that nothing prevents the application 1498 from defining different sized blocks in a file. 1500 8.1.1. Data Block Representation 1502 1504 struct app_data_block4 { 1505 offset4 adb_offset; 1506 length4 adb_block_size; 1507 length4 adb_block_count; 1508 length4 adb_reloff_blocknum; 1509 count4 adb_block_num; 1510 length4 adb_reloff_pattern; 1511 opaque adb_pattern<>; 1512 }; 1514 1516 The app_data_block4 structure captures the abstraction presented for 1517 the ADB. The additional fields present are to allow the transmission 1518 of adb_block_count ADBs at one time. We also use adb_block_num to 1519 convey the ADBN of the first block in the sequence. Each ADB will 1520 contain the same adb_pattern string. 1522 As both adb_block_num and adb_pattern are optional, if either 1523 adb_reloff_pattern or adb_reloff_blocknum is set to NFS4_UINT64_MAX, 1524 then the corresponding field is not set in any of the ADB. 1526 8.2. An Example of Detecting Corruption 1528 In this section, we define an ADB format in which corruption can be 1529 detected. Note that this is just one possible format and means to 1530 detect corruption. 1532 Consider a very basic implementation of an operating system's disk 1533 blocks. A block is either data or it is an indirect block which 1534 allows for files to be larger than one block. It is desired to be 1535 able to initialize a block. Lastly, to quickly unlink a file, a 1536 block can be marked invalid. The contents remain intact - which 1537 would enable this OS application to undelete a file. 1539 The application defines 4k sized data blocks, with an 8 byte block 1540 counter occurring at offset 0 in the block, and with the guard 1541 pattern occurring at offset 8 inside the block. Furthermore, the 1542 guard pattern can take one of four states: 1544 0xfeedface - This is the FREE state and indicates that the ADB 1545 format has been applied. 1547 0xcafedead - This is the DATA state and indicates that real data 1548 has been written to this block. 1550 0xe4e5c001 - This is the INDIRECT state and indicates that the 1551 block contains block counter numbers that are chained off of this 1552 block. 1554 0xba1ed4a3 - This is the INVALID state and indicates that the block 1555 contains data whose contents are garbage. 1557 Finally, it also defines an 8 byte checksum [Baira08] starting at 1558 byte 16 which applies to the remaining contents of the block. If the 1559 state is FREE, then that checksum is trivially zero. As such, the 1560 application has no need to transfer the checksum implicitly inside 1561 the ADB - it need not make the transfer layer aware of the fact that 1562 there is a checksum (see [Ashdown08] for an example of checksums used 1563 to detect corruption in application data blocks). 1565 Corruption in each ADB can thus be detected: 1567 o If the guard pattern is anything other than one of the allowed 1568 values, including all zeros. 1570 o If the guard pattern is FREE and any other byte in the remainder 1571 of the ADB is anything other than zero. 1573 o If the guard pattern is anything other than FREE, then if the 1574 stored checksum does not match the computed checksum. 1576 o If the guard pattern is INDIRECT and one of the stored indirect 1577 block numbers has a value greater than the number of ADBs in the 1578 file. 1580 o If the guard pattern is INDIRECT and one of the stored indirect 1581 block numbers is a duplicate of another stored indirect block 1582 number. 1584 As can be seen, the application can detect errors based on the 1585 combination of the guard pattern state and the checksum. But also, 1586 the application can detect corruption based on the state and the 1587 contents of the ADB. This last point is important in validating the 1588 minimum amount of data we incorporated into our generic framework. 1589 I.e., the guard pattern is sufficient in allowing applications to 1590 design their own corruption detection. 1592 Finally, it is important to note that none of these corruption checks 1593 occur in the transport layer. The server and client components are 1594 totally unaware of the file format and might report everything as 1595 being transferred correctly even in the case the application detects 1596 corruption. 1598 8.3. Example of READ_PLUS 1600 The hypothetical application presented in Section 8.2 can be used to 1601 illustrate how READ_PLUS would return an array of results. A file is 1602 created and initialized with 100 4k ADBs in the FREE state with the 1603 WRITE_SAME operation (see Section 15.12): 1605 WRITE_SAME {0, 4k, 100, 0, 0, 8, 0xfeedface} 1607 Further, assume the application writes a single ADB at 16k, changing 1608 the guard pattern to 0xcafedead, we would then have in memory: 1610 0k -> (4k - 1) : 00 00 00 00 fe ed fa ce 00 00 ... 00 00 1611 4k -> (8k - 1) : 00 00 00 01 fe ed fa ce 00 00 ... 00 00 1612 8k -> (12k - 1) : 00 00 00 02 fe ed fa ce 00 00 ... 00 00 1613 12k -> (16k - 1) : 00 00 00 03 fe ed fa ce 00 00 ... 00 00 1614 16k -> (20k - 1) : 00 00 00 04 ca fe de ad 00 00 ... 00 00 1615 20k -> (24k - 1) : 00 00 00 05 fe ed fa ce 00 00 ... 00 00 1616 24k -> (28k - 1) : 00 00 00 06 fe ed fa ce 00 00 ... 00 00 1617 ... 1618 396k -> (400k - 1) : 00 00 00 63 fe ed fa ce 00 00 ... 00 00 1620 And when the client did a READ_PLUS of 64k at the start of the file, 1621 it could get back a result of data: 1623 0k -> (4k - 1) : 00 00 00 00 fe ed fa ce 00 00 ... 00 00 1624 4k -> (8k - 1) : 00 00 00 01 fe ed fa ce 00 00 ... 00 00 1625 8k -> (12k - 1) : 00 00 00 02 fe ed fa ce 00 00 ... 00 00 1626 12k -> (16k - 1) : 00 00 00 03 fe ed fa ce 00 00 ... 00 00 1627 16k -> (20k - 1) : 00 00 00 04 ca fe de ad 00 00 ... 00 00 1628 20k -> (24k - 1) : 00 00 00 05 fe ed fa ce 00 00 ... 00 00 1629 24k -> (24k - 1) : 00 00 00 06 fe ed fa ce 00 00 ... 00 00 1630 ... 1631 62k -> (64k - 1) : 00 00 00 15 fe ed fa ce 00 00 ... 00 00 1633 8.4. An Example of Zeroing Space 1635 A simpler use case for WRITE_SAME are applications that want to 1636 efficiently zero out a file, but do not want to modify space 1637 reservations. This can easily be achieved by a call to WRITE_SAME 1638 without a ADB block numbers and pattern, e.g.: 1640 WRITE_SAME {0, 1k, 10000, 0, 0, 0, 0} 1642 9. Labeled NFS 1644 9.1. Introduction 1646 Access control models such as Unix permissions or Access Control 1647 Lists are commonly referred to as Discretionary Access Control (DAC) 1648 models. These systems base their access decisions on user identity 1649 and resource ownership. In contrast Mandatory Access Control (MAC) 1650 models base their access control decisions on the label on the 1651 subject (usually a process) and the object it wishes to access 1652 [RFC7204]. These labels may contain user identity information but 1653 usually contain additional information. In DAC systems users are 1654 free to specify the access rules for resources that they own. MAC 1655 models base their security decisions on a system wide policy 1656 established by an administrator or organization which the users do 1657 not have the ability to override. In this section, we add a MAC 1658 model to NFSv4.2. 1660 First we provide a method for transporting and storing security label 1661 data on NFSv4 file objects. Security labels have several semantics 1662 that are met by NFSv4 recommended attributes such as the ability to 1663 set the label value upon object creation. Access control on these 1664 attributes are done through a combination of two mechanisms. As with 1665 other recommended attributes on file objects the usual DAC checks 1666 (ACLs and permission bits) will be performed to ensure that proper 1667 file ownership is enforced. In addition a MAC system MAY be employed 1668 on the client, server, or both to enforce additional policy on what 1669 subjects may modify security label information. 1671 Second, we describe a method for the client to determine if an NFSv4 1672 file object security label has changed. A client which needs to know 1673 if a label on a file or set of files is going to change SHOULD 1674 request a delegation on each labeled file. In order to change such a 1675 security label, the server will have to recall delegations on any 1676 file affected by the label change, so informing clients of the label 1677 change. 1679 An additional useful feature would be modification to the RPC layer 1680 used by NFSv4 to allow RPC calls to carry security labels and enable 1681 full mode enforcement as described in Section 9.6.1. Such 1682 modifications are outside the scope of this document (see 1683 [rpcsec_gssv3]). 1685 9.2. Definitions 1687 Label Format Specifier (LFS): is an identifier used by the client to 1688 establish the syntactic format of the security label and the 1689 semantic meaning of its components. These specifiers exist in a 1690 registry associated with documents describing the format and 1691 semantics of the label. 1693 Label Format Registry: is the IANA registry (see [Quigley14]) 1694 containing all registered LFSes along with references to the 1695 documents that describe the syntactic format and semantics of the 1696 security label. 1698 Policy Identifier (PI): is an optional part of the definition of a 1699 Label Format Specifier which allows for clients and server to 1700 identify specific security policies. 1702 Object: is a passive resource within the system that we wish to be 1703 protected. Objects can be entities such as files, directories, 1704 pipes, sockets, and many other system resources relevant to the 1705 protection of the system state. 1707 Subject: is an active entity usually a process which is requesting 1708 access to an object. 1710 MAC-Aware: is a server which can transmit and store object labels. 1712 MAC-Functional: is a client or server which is Labeled NFS enabled. 1713 Such a system can interpret labels and apply policies based on the 1714 security system. 1716 Multi-Level Security (MLS): is a traditional model where objects are 1717 given a sensitivity level (Unclassified, Secret, Top Secret, etc) 1718 and a category set (see [BL73], [RFC1108], and [RFC2401]). 1720 9.3. MAC Security Attribute 1722 MAC models base access decisions on security attributes bound to 1723 subjects and objects. This information can range from a user 1724 identity for an identity based MAC model, sensitivity levels for 1725 Multi-level security, or a type for Type Enforcement. These models 1726 base their decisions on different criteria but the semantics of the 1727 security attribute remain the same. The semantics required by the 1728 security attributes are listed below: 1730 o MUST provide flexibility with respect to the MAC model. 1732 o MUST provide the ability to atomically set security information 1733 upon object creation. 1735 o MUST provide the ability to enforce access control decisions both 1736 on the client and the server. 1738 o MUST NOT expose an object to either the client or server name 1739 space before its security information has been bound to it. 1741 NFSv4 implements the security attribute as a recommended attribute. 1742 These attributes have a fixed format and semantics, which conflicts 1743 with the flexible nature of the security attribute. To resolve this 1744 the security attribute consists of two components. The first 1745 component is a LFS as defined in [Quigley14] to allow for 1746 interoperability between MAC mechanisms. The second component is an 1747 opaque field which is the actual security attribute data. To allow 1748 for various MAC models, NFSv4 should be used solely as a transport 1749 mechanism for the security attribute. It is the responsibility of 1750 the endpoints to consume the security attribute and make access 1751 decisions based on their respective models. In addition, creation of 1752 objects through OPEN and CREATE allows for the security attribute to 1753 be specified upon creation. By providing an atomic create and set 1754 operation for the security attribute it is possible to enforce the 1755 second and fourth requirements. The recommended attribute 1756 FATTR4_SEC_LABEL (see Section 12.2.4) will be used to satisfy this 1757 requirement. 1759 9.3.1. Delegations 1761 In the event that a security attribute is changed on the server while 1762 a client holds a delegation on the file, both the server and the 1763 client MUST follow the NFSv4.1 protocol (see Chapter 10 of [RFC5661]) 1764 with respect to attribute changes. It SHOULD flush all changes back 1765 to the server and relinquish the delegation. 1767 9.3.2. Permission Checking 1769 It is not feasible to enumerate all possible MAC models and even 1770 levels of protection within a subset of these models. This means 1771 that the NFSv4 client and servers cannot be expected to directly make 1772 access control decisions based on the security attribute. Instead 1773 NFSv4 should defer permission checking on this attribute to the host 1774 system. These checks are performed in addition to existing DAC and 1775 ACL checks outlined in the NFSv4 protocol. Section 9.6 gives a 1776 specific example of how the security attribute is handled under a 1777 particular MAC model. 1779 9.3.3. Object Creation 1781 When creating files in NFSv4 the OPEN and CREATE operations are used. 1782 One of the parameters to these operations is an fattr4 structure 1783 containing the attributes the file is to be created with. This 1784 allows NFSv4 to atomically set the security attribute of files upon 1785 creation. When a client is MAC-Functional it must always provide the 1786 initial security attribute upon file creation. In the event that the 1787 server is MAC-Functional as well, it should determine by policy 1788 whether it will accept the attribute from the client or instead make 1789 the determination itself. If the client is not MAC-Functional, then 1790 the MAC-Functional server must decide on a default label. A more in 1791 depth explanation can be found in Section 9.6. 1793 9.3.4. Existing Objects 1795 Note that under the MAC model, all objects must have labels. 1796 Therefore, if an existing server is upgraded to include Labeled NFS 1797 support, then it is the responsibility of the security system to 1798 define the behavior for existing objects. 1800 9.3.5. Label Changes 1802 Consider a guest mode system (Section 9.6.2) in which the clients 1803 enforce MAC checks and the server has only a DAC security system 1804 which stores the labels along with the file data. In this type of 1805 system, a user with the appropriate DAC credentials on a client with 1806 poorly configured or disabled MAC labeling enforcement is allowed 1807 access to the file label (and data) on the server and can change the 1808 label. 1810 Clients which need to know if a label on a file or set of files has 1811 changed SHOULD request a delegation on each labeled file so that a 1812 label change by another client will be known via the process 1813 described in Section 9.3.1 which must be followed: the delegation 1814 will be recalled, which effectively notifies the client of the 1815 change. 1817 Note that the MAC security policies on a client can be such that the 1818 client does not have access to the file unless it has a delegation. 1820 9.4. pNFS Considerations 1822 The new FATTR4_SEC_LABEL attribute is metadata information and as 1823 such the DS is not aware of the value contained on the MDS. 1824 Fortunately, the NFSv4.1 protocol [RFC5661] already has provisions 1825 for doing access level checks from the DS to the MDS. In order for 1826 the DS to validate the subject label presented by the client, it 1827 SHOULD utilize this mechanism. 1829 9.5. Discovery of Server Labeled NFS Support 1831 The server can easily determine that a client supports Labeled NFS 1832 when it queries for the FATTR4_SEC_LABEL label for an object. The 1833 client might need to discover which LFS the server supports. 1835 The following compound MUST NOT be denied by any MAC label check: 1837 PUTROOTFH, GETATTR {FATTR4_SEC_LABEL} 1839 Note that the server might have imposed a security flavor on the root 1840 that precludes such access. I.e., if the server requires kerberized 1841 access and the client presents a compound with AUTH_SYS, then the 1842 server is allowed to return NFS4ERR_WRONGSEC in this case. But if 1843 the client presents a correct security flavor, then the server MUST 1844 return the FATTR4_SEC_LABEL attribute with the supported LFS filled 1845 in. 1847 9.6. MAC Security NFS Modes of Operation 1849 A system using Labeled NFS may operate in two modes. The first mode 1850 provides the most protection and is called "full mode". In this mode 1851 both the client and server implement a MAC model allowing each end to 1852 make an access control decision. The remaining mode is called the 1853 "guest mode" and in this mode one end of the connection is not 1854 implementing a MAC model and thus offers less protection than full 1855 mode. 1857 9.6.1. Full Mode 1859 Full mode environments consist of MAC-Functional NFSv4 servers and 1860 clients and may be composed of mixed MAC models and policies. The 1861 system requires that both the client and server have an opportunity 1862 to perform an access control check based on all relevant information 1863 within the network. The file object security attribute is provided 1864 using the mechanism described in Section 9.3. 1866 Fully MAC-Functional NFSv4 servers are not possible in the absence of 1867 RPCSEC_GSSv3 [rpcsec_gssv3] support for subject label transport. 1868 However, servers may make decisions based on the RPC credential 1869 information available. 1871 9.6.1.1. Initial Labeling and Translation 1873 The ability to create a file is an action that a MAC model may wish 1874 to mediate. The client is given the responsibility to determine the 1875 initial security attribute to be placed on a file. This allows the 1876 client to make a decision as to the acceptable security attributes to 1877 create a file with before sending the request to the server. Once 1878 the server receives the creation request from the client it may 1879 choose to evaluate if the security attribute is acceptable. 1881 Security attributes on the client and server may vary based on MAC 1882 model and policy. To handle this the security attribute field has an 1883 LFS component. This component is a mechanism for the host to 1884 identify the format and meaning of the opaque portion of the security 1885 attribute. A full mode environment may contain hosts operating in 1886 several different LFSes. In this case a mechanism for translating 1887 the opaque portion of the security attribute is needed. The actual 1888 translation function will vary based on MAC model and policy and is 1889 out of the scope of this document. If a translation is unavailable 1890 for a given LFS then the request MUST be denied. Another recourse is 1891 to allow the host to provide a fallback mapping for unknown security 1892 attributes. 1894 9.6.1.2. Policy Enforcement 1896 In full mode access control decisions are made by both the clients 1897 and servers. When a client makes a request it takes the security 1898 attribute from the requesting process and makes an access control 1899 decision based on that attribute and the security attribute of the 1900 object it is trying to access. If the client denies that access an 1901 RPC call to the server is never made. If however the access is 1902 allowed the client will make a call to the NFS server. 1904 When the server receives the request from the client it uses any 1905 credential information conveyed in the RPC request and the attributes 1906 of the object the client is trying to access to make an access 1907 control decision. If the server's policy allows this access it will 1908 fulfill the client's request, otherwise it will return 1909 NFS4ERR_ACCESS. 1911 Future protocol extensions may also allow the server to factor into 1912 the decision a security label extracted from the RPC request. 1914 Implementations MAY validate security attributes supplied over the 1915 network to ensure that they are within a set of attributes permitted 1916 from a specific peer, and if not, reject them. Note that a system 1917 may permit a different set of attributes to be accepted from each 1918 peer. 1920 9.6.1.3. Limited Server 1922 A Limited Server mode (see Section 4.2 of [RFC7204]) consists of a 1923 server which is label aware, but does not enforce policies. Such a 1924 server will store and retrieve all object labels presented by 1925 clients, utilize the methods described in Section 9.3.5 to allow the 1926 clients to detect changing labels, but may not factor the label into 1927 access decisions. Instead, it will expect the clients to enforce all 1928 such access locally. 1930 9.6.2. Guest Mode 1932 Guest mode implies that either the client or the server does not 1933 handle labels. If the client is not Labeled NFS aware, then it will 1934 not offer subject labels to the server. The server is the only 1935 entity enforcing policy, and may selectively provide standard NFS 1936 services to clients based on their authentication credentials and/or 1937 associated network attributes (e.g., IP address, network interface). 1938 The level of trust and access extended to a client in this mode is 1939 configuration-specific. If the server is not Labeled NFS aware, then 1940 it will not return object labels to the client. Clients in this 1941 environment are may consist of groups implementing different MAC 1942 model policies. The system requires that all clients in the 1943 environment be responsible for access control checks. 1945 9.7. Security Considerations for Labeled NFS 1947 This entire chapter deals with security issues. 1949 Depending on the level of protection the MAC system offers there may 1950 be a requirement to tightly bind the security attribute to the data. 1952 When only one of the client or server enforces labels, it is 1953 important to realize that the other side is not enforcing MAC 1954 protections. Alternate methods might be in use to handle the lack of 1955 MAC support and care should be taken to identify and mitigate threats 1956 from possible tampering outside of these methods. 1958 An example of this is that a server that modifies READDIR or LOOKUP 1959 results based on the client's subject label might want to always 1960 construct the same subject label for a client which does not present 1961 one. This will prevent a non-Labeled NFS client from mixing entries 1962 in the directory cache. 1964 10. Sharing change attribute implementation characteristics with NFSv4 1965 clients 1967 Although both the NFSv4 [RFC7530] and NFSv4.1 protocol [RFC5661], 1968 define the change attribute as being mandatory to implement, there is 1969 little in the way of guidance as to its construction. The only 1970 mandated constraint is that the value must change whenever the file 1971 data or metadata change. 1973 While this allows for a wide range of implementations, it also leaves 1974 the client with no way to determine which is the most recent value 1975 for the change attribute in a case where several RPC calls have been 1976 issued in parallel. In other words if two COMPOUNDs, both containing 1977 WRITE and GETATTR requests for the same file, have been issued in 1978 parallel, how does the client determine which of the two change 1979 attribute values returned in the replies to the GETATTR requests 1980 correspond to the most recent state of the file? In some cases, the 1981 only recourse may be to send another COMPOUND containing a third 1982 GETATTR that is fully serialized with the first two. 1984 NFSv4.2 avoids this kind of inefficiency by allowing the server to 1985 share details about how the change attribute is expected to evolve, 1986 so that the client may immediately determine which, out of the 1987 several change attribute values returned by the server, is the most 1988 recent. change_attr_type is defined as a new recommended attribute 1989 (see Section 12.2.3), and is per file system. 1991 11. Error Values 1993 NFS error numbers are assigned to failed operations within a Compound 1994 (COMPOUND or CB_COMPOUND) request. A Compound request contains a 1995 number of NFS operations that have their results encoded in sequence 1996 in a Compound reply. The results of successful operations will 1997 consist of an NFS4_OK status followed by the encoded results of the 1998 operation. If an NFS operation fails, an error status will be 1999 entered in the reply and the Compound request will be terminated. 2001 11.1. Error Definitions 2003 Protocol Error Definitions 2005 +-------------------------+--------+------------------+ 2006 | Error | Number | Description | 2007 +-------------------------+--------+------------------+ 2008 | NFS4ERR_BADLABEL | 10093 | Section 11.1.3.1 | 2009 | NFS4ERR_OFFLOAD_DENIED | 10091 | Section 11.1.2.1 | 2010 | NFS4ERR_OFFLOAD_NO_REQS | 10094 | Section 11.1.2.2 | 2011 | NFS4ERR_PARTNER_NO_AUTH | 10089 | Section 11.1.2.3 | 2012 | NFS4ERR_PARTNER_NOTSUPP | 10088 | Section 11.1.2.4 | 2013 | NFS4ERR_UNION_NOTSUPP | 10090 | Section 11.1.1.1 | 2014 | NFS4ERR_WRONG_LFS | 10092 | Section 11.1.3.2 | 2015 +-------------------------+--------+------------------+ 2017 Table 1 2019 11.1.1. General Errors 2021 This section deals with errors that are applicable to a broad set of 2022 different purposes. 2024 11.1.1.1. NFS4ERR_UNION_NOTSUPP (Error Code 10090) 2026 One of the arguments to the operation is a discriminated union and 2027 while the server supports the given operation, it does not support 2028 the selected arm of the discriminated union. 2030 11.1.2. Server to Server Copy Errors 2032 These errors deal with the interaction between server to server 2033 copies. 2035 11.1.2.1. NFS4ERR_OFFLOAD_DENIED (Error Code 10091) 2037 The copy offload operation is supported by both the source and the 2038 destination, but the destination is not allowing it for this file. 2039 If the client sees this error, it should fall back to the normal copy 2040 semantics. 2042 11.1.2.2. NFS4ERR_OFFLOAD_NO_REQS (Error Code 10094) 2044 The copy offload operation is supported by both the source and the 2045 destination, but the destination can not meet the client requirements 2046 for either consecutive byte copy or synchronous copy. If the client 2047 sees this error, it should either relax the requirements (if any) or 2048 fall back to the normal copy semantics. 2050 11.1.2.3. NFS4ERR_PARTNER_NO_AUTH (Error Code 10089) 2052 The source server does not authorize a server-to-server copy offload 2053 operation. This may be due to the client's failure to send the 2054 COPY_NOTIFY operation to the source server, the source server 2055 receiving a server-to-server copy offload request after the copy 2056 lease time expired, or for some other permission problem. 2058 11.1.2.4. NFS4ERR_PARTNER_NOTSUPP (Error Code 10088) 2060 The remote server does not support the server-to-server copy offload 2061 protocol. 2063 11.1.3. Labeled NFS Errors 2065 These errors are used in Labeled NFS. 2067 11.1.3.1. NFS4ERR_BADLABEL (Error Code 10093) 2069 The label specified is invalid in some manner. 2071 11.1.3.2. NFS4ERR_WRONG_LFS (Error Code 10092) 2073 The LFS specified in the subject label is not compatible with the LFS 2074 in the object label. 2076 11.2. New Operations and Their Valid Errors 2078 This section contains a table that gives the valid error returns for 2079 each new NFSv4.2 protocol operation. The error code NFS4_OK 2080 (indicating no error) is not listed but should be understood to be 2081 returnable by all new operations. The error values for all other 2082 operations are defined in Section 15.2 of [RFC5661]. 2084 Valid Error Returns for Each New Protocol Operation 2086 +----------------+--------------------------------------------------+ 2087 | Operation | Errors | 2088 +----------------+--------------------------------------------------+ 2089 | ALLOCATE | NFS4ERR_ACCESS, NFS4ERR_ADMIN_REVOKED, | 2090 | | NFS4ERR_BADXDR, NFS4ERR_BAD_STATEID, | 2091 | | NFS4ERR_DEADSESSION, NFS4ERR_DELAY, | 2092 | | NFS4ERR_DELEG_REVOKED, NFS4ERR_DQUOT, | 2093 | | NFS4ERR_EXPIRED, NFS4ERR_FBIG, | 2094 | | NFS4ERR_FHEXPIRED, NFS4ERR_GRACE, NFS4ERR_INVAL, | 2095 | | NFS4ERR_IO, NFS4ERR_ISDIR, NFS4ERR_MOVED, | 2096 | | NFS4ERR_NOFILEHANDLE, NFS4ERR_NOSPC, | 2097 | | NFS4ERR_NOTSUPP, NFS4ERR_OLD_STATEID, | 2098 | | NFS4ERR_OPENMODE, NFS4ERR_OP_NOT_IN_SESSION, | 2099 | | NFS4ERR_REP_TOO_BIG, | 2100 | | NFS4ERR_REP_TOO_BIG_TO_CACHE, | 2101 | | NFS4ERR_REQ_TOO_BIG, NFS4ERR_RETRY_UNCACHED_REP, | 2102 | | NFS4ERR_ROFS, NFS4ERR_SERVERFAULT, | 2103 | | NFS4ERR_STALE, NFS4ERR_SYMLINK, | 2104 | | NFS4ERR_TOO_MANY_OPS, NFS4ERR_WRONG_TYPE | 2105 | CLONE | NFS4ERR_ACCESS, NFS4ERR_ADMIN_REVOKED, | 2106 | | NFS4ERR_BADXDR, NFS4ERR_BAD_STATEID, | 2107 | | NFS4ERR_DEADSESSION, NFS4ERR_DELAY, | 2108 | | NFS4ERR_DELEG_REVOKED, NFS4ERR_EXPIRED, | 2109 | | NFS4ERR_FBIG, NFS4ERR_FHEXPIRED, NFS4ERR_GRACE, | 2110 | | NFS4ERR_INVAL, NFS4ERR_IO, NFS4ERR_ISDIR, | 2111 | | NFS4ERR_MOVED, NFS4ERR_NOFILEHANDLE, | 2112 | | NFS4ERR_NOTSUPP, NFS4ERR_NOSPC, | 2113 | | NFS4ERR_OLD_STATEID, NFS4ERR_OPENMODE, | 2114 | | NFS4ERR_OP_NOT_IN_SESSION, NFS4ERR_REP_TOO_BIG, | 2115 | | NFS4ERR_REP_TOO_BIG_TO_CACHE, | 2116 | | NFS4ERR_REQ_TOO_BIG, NFS4ERR_RETRY_UNCACHED_REP, | 2117 | | NFS4ERR_ROFS, NFS4ERR_SERVERFAULT, | 2118 | | NFS4ERR_STALE, NFS4ERR_SYMLINK, | 2119 | | NFS4ERR_TOO_MANY_OPS, NFS4ERR_WRONG_TYPE | 2120 | COPY | NFS4ERR_ACCESS, NFS4ERR_ADMIN_REVOKED, | 2121 | | NFS4ERR_BADXDR, NFS4ERR_BAD_STATEID, | 2122 | | NFS4ERR_DEADSESSION, NFS4ERR_DELAY, | 2123 | | NFS4ERR_DELEG_REVOKED, NFS4ERR_DQUOT, | 2124 | | NFS4ERR_EXPIRED, NFS4ERR_FBIG, | 2125 | | NFS4ERR_FHEXPIRED, NFS4ERR_GRACE, NFS4ERR_INVAL, | 2126 | | NFS4ERR_IO, NFS4ERR_ISDIR, NFS4ERR_LOCKED, | 2127 | | NFS4ERR_METADATA_NOTSUPP, NFS4ERR_MOVED, | 2128 | | NFS4ERR_NOFILEHANDLE, NFS4ERR_NOSPC, | 2129 | | NFS4ERR_OFFLOAD_DENIED, NFS4ERR_OLD_STATEID, | 2130 | | NFS4ERR_OPENMODE, NFS4ERR_OP_NOT_IN_SESSION, | 2131 | | NFS4ERR_PARTNER_NO_AUTH, | 2132 | | NFS4ERR_PARTNER_NOTSUPP, NFS4ERR_PNFS_IO_HOLE, | 2133 | | NFS4ERR_PNFS_NO_LAYOUT, NFS4ERR_REP_TOO_BIG, | 2134 | | NFS4ERR_REP_TOO_BIG_TO_CACHE, | 2135 | | NFS4ERR_REQ_TOO_BIG, NFS4ERR_RETRY_UNCACHED_REP, | 2136 | | NFS4ERR_ROFS, NFS4ERR_SERVERFAULT, | 2137 | | NFS4ERR_STALE, NFS4ERR_SYMLINK, | 2138 | | NFS4ERR_TOO_MANY_OPS, NFS4ERR_WRONG_TYPE | 2139 | COPY_NOTIFY | NFS4ERR_ACCESS, NFS4ERR_ADMIN_REVOKED, | 2140 | | NFS4ERR_BADXDR, NFS4ERR_BAD_STATEID, | 2141 | | NFS4ERR_DEADSESSION, NFS4ERR_DELAY, | 2142 | | NFS4ERR_DELEG_REVOKED, NFS4ERR_EXPIRED, | 2143 | | NFS4ERR_FHEXPIRED, NFS4ERR_GRACE, NFS4ERR_INVAL, | 2144 | | NFS4ERR_ISDIR, NFS4ERR_IO, NFS4ERR_LOCKED, | 2145 | | NFS4ERR_MOVED, NFS4ERR_NOFILEHANDLE, | 2146 | | NFS4ERR_OLD_STATEID, NFS4ERR_OPENMODE, | 2147 | | NFS4ERR_OP_NOT_IN_SESSION, NFS4ERR_PNFS_IO_HOLE, | 2148 | | NFS4ERR_PNFS_NO_LAYOUT, NFS4ERR_REP_TOO_BIG, | 2149 | | NFS4ERR_REP_TOO_BIG_TO_CACHE, | 2150 | | NFS4ERR_REQ_TOO_BIG, NFS4ERR_RETRY_UNCACHED_REP, | 2151 | | NFS4ERR_SERVERFAULT, NFS4ERR_STALE, | 2152 | | NFS4ERR_SYMLINK, NFS4ERR_TOO_MANY_OPS, | 2153 | | NFS4ERR_WRONG_TYPE | 2154 | DEALLOCATE | NFS4ERR_ACCESS, NFS4ERR_ADMIN_REVOKED, | 2155 | | NFS4ERR_BADXDR, NFS4ERR_BAD_STATEID, | 2156 | | NFS4ERR_DEADSESSION, NFS4ERR_DELAY, | 2157 | | NFS4ERR_DELEG_REVOKED, NFS4ERR_EXPIRED, | 2158 | | NFS4ERR_FBIG, NFS4ERR_FHEXPIRED, NFS4ERR_GRACE, | 2159 | | NFS4ERR_INVAL, NFS4ERR_IO, NFS4ERR_ISDIR, | 2160 | | NFS4ERR_MOVED, NFS4ERR_NOFILEHANDLE, | 2161 | | NFS4ERR_NOTSUPP, NFS4ERR_OLD_STATEID, | 2162 | | NFS4ERR_OPENMODE, NFS4ERR_OP_NOT_IN_SESSION, | 2163 | | NFS4ERR_REP_TOO_BIG, | 2164 | | NFS4ERR_REP_TOO_BIG_TO_CACHE, | 2165 | | NFS4ERR_REQ_TOO_BIG, NFS4ERR_RETRY_UNCACHED_REP, | 2166 | | NFS4ERR_ROFS, NFS4ERR_SERVERFAULT, | 2167 | | NFS4ERR_STALE, NFS4ERR_SYMLINK, | 2168 | | NFS4ERR_TOO_MANY_OPS, NFS4ERR_WRONG_TYPE | 2169 | GETDEVICELIST | NFS4ERR_NOTSUPP | 2170 | LAYOUTERROR | NFS4ERR_ADMIN_REVOKED, NFS4ERR_BADXDR, | 2171 | | NFS4ERR_BAD_STATEID, NFS4ERR_DEADSESSION, | 2172 | | NFS4ERR_DELAY, NFS4ERR_DELEG_REVOKED, | 2173 | | NFS4ERR_EXPIRED, NFS4ERR_FHEXPIRED, | 2174 | | NFS4ERR_GRACE, NFS4ERR_INVAL, NFS4ERR_ISDIR, | 2175 | | NFS4ERR_MOVED, NFS4ERR_NOFILEHANDLE, | 2176 | | NFS4ERR_NOTSUPP, NFS4ERR_NO_GRACE, | 2177 | | NFS4ERR_OLD_STATEID, NFS4ERR_OP_NOT_IN_SESSION, | 2178 | | NFS4ERR_REP_TOO_BIG, | 2179 | | NFS4ERR_REP_TOO_BIG_TO_CACHE, | 2180 | | NFS4ERR_REQ_TOO_BIG, NFS4ERR_RETRY_UNCACHED_REP, | 2181 | | NFS4ERR_SERVERFAULT, NFS4ERR_STALE, | 2182 | | NFS4ERR_TOO_MANY_OPS, | 2183 | | NFS4ERR_UNKNOWN_LAYOUTTYPE, NFS4ERR_WRONG_CRED, | 2184 | | NFS4ERR_WRONG_TYPE | 2185 | LAYOUTSTATS | NFS4ERR_ADMIN_REVOKED, NFS4ERR_BADXDR, | 2186 | | NFS4ERR_BAD_STATEID, NFS4ERR_DEADSESSION, | 2187 | | NFS4ERR_DELAY, NFS4ERR_DELEG_REVOKED, | 2188 | | NFS4ERR_EXPIRED, NFS4ERR_FHEXPIRED, | 2189 | | NFS4ERR_GRACE, NFS4ERR_INVAL, NFS4ERR_ISDIR, | 2190 | | NFS4ERR_MOVED, NFS4ERR_NOFILEHANDLE, | 2191 | | NFS4ERR_NOTSUPP, NFS4ERR_NO_GRACE, | 2192 | | NFS4ERR_OLD_STATEID, NFS4ERR_OP_NOT_IN_SESSION, | 2193 | | NFS4ERR_REP_TOO_BIG, | 2194 | | NFS4ERR_REP_TOO_BIG_TO_CACHE, | 2195 | | NFS4ERR_REQ_TOO_BIG, NFS4ERR_RETRY_UNCACHED_REP, | 2196 | | NFS4ERR_SERVERFAULT, NFS4ERR_STALE, | 2197 | | NFS4ERR_TOO_MANY_OPS, | 2198 | | NFS4ERR_UNKNOWN_LAYOUTTYPE, NFS4ERR_WRONG_CRED, | 2199 | | NFS4ERR_WRONG_TYPE | 2200 | OFFLOAD_CANCEL | NFS4ERR_ADMIN_REVOKED, NFS4ERR_BADXDR, | 2201 | | NFS4ERR_BAD_STATEID, NFS4ERR_COMPLETE_ALREADY, | 2202 | | NFS4ERR_DEADSESSION, NFS4ERR_EXPIRED, | 2203 | | NFS4ERR_DELAY, NFS4ERR_GRACE, NFS4ERR_NOTSUPP, | 2204 | | NFS4ERR_OLD_STATEID, NFS4ERR_OP_NOT_IN_SESSION, | 2205 | | NFS4ERR_SERVERFAULT, NFS4ERR_TOO_MANY_OPS | 2206 | OFFLOAD_STATUS | NFS4ERR_ADMIN_REVOKED, NFS4ERR_BADXDR, | 2207 | | NFS4ERR_BAD_STATEID, NFS4ERR_COMPLETE_ALREADY, | 2208 | | NFS4ERR_DEADSESSION, NFS4ERR_EXPIRED, | 2209 | | NFS4ERR_DELAY, NFS4ERR_GRACE, NFS4ERR_NOTSUPP, | 2210 | | NFS4ERR_OLD_STATEID, NFS4ERR_OP_NOT_IN_SESSION, | 2211 | | NFS4ERR_SERVERFAULT, NFS4ERR_TOO_MANY_OPS | 2212 | READ_PLUS | NFS4ERR_ACCESS, NFS4ERR_ADMIN_REVOKED, | 2213 | | NFS4ERR_BADXDR, NFS4ERR_BAD_STATEID, | 2214 | | NFS4ERR_DEADSESSION, NFS4ERR_DELAY, | 2215 | | NFS4ERR_DELEG_REVOKED, NFS4ERR_EXPIRED, | 2216 | | NFS4ERR_FHEXPIRED, NFS4ERR_GRACE, NFS4ERR_INVAL, | 2217 | | NFS4ERR_ISDIR, NFS4ERR_IO, NFS4ERR_LOCKED, | 2218 | | NFS4ERR_MOVED, NFS4ERR_NOFILEHANDLE, | 2219 | | NFS4ERR_NOTSUPP, NFS4ERR_OLD_STATEID, | 2220 | | NFS4ERR_OPENMODE, NFS4ERR_OP_NOT_IN_SESSION, | 2221 | | NFS4ERR_PNFS_IO_HOLE, NFS4ERR_PNFS_NO_LAYOUT, | 2222 | | NFS4ERR_REP_TOO_BIG, | 2223 | | NFS4ERR_REP_TOO_BIG_TO_CACHE, | 2224 | | NFS4ERR_REQ_TOO_BIG, NFS4ERR_RETRY_UNCACHED_REP, | 2225 | | NFS4ERR_SERVERFAULT, NFS4ERR_STALE, | 2226 | | NFS4ERR_SYMLINK, NFS4ERR_TOO_MANY_OPS, | 2227 | | NFS4ERR_WRONG_TYPE | 2228 | SEEK | NFS4ERR_ACCESS, NFS4ERR_ADMIN_REVOKED, | 2229 | | NFS4ERR_BADXDR, NFS4ERR_BAD_STATEID, | 2230 | | NFS4ERR_DEADSESSION, NFS4ERR_DELAY, | 2231 | | NFS4ERR_DELEG_REVOKED, NFS4ERR_EXPIRED, | 2232 | | NFS4ERR_FHEXPIRED, NFS4ERR_GRACE, NFS4ERR_INVAL, | 2233 | | NFS4ERR_ISDIR, NFS4ERR_IO, NFS4ERR_LOCKED, | 2234 | | NFS4ERR_MOVED, NFS4ERR_NOFILEHANDLE, | 2235 | | NFS4ERR_NOTSUPP, NFS4ERR_OLD_STATEID, | 2236 | | NFS4ERR_OPENMODE, NFS4ERR_OP_NOT_IN_SESSION, | 2237 | | NFS4ERR_PNFS_IO_HOLE, NFS4ERR_PNFS_NO_LAYOUT, | 2238 | | NFS4ERR_REP_TOO_BIG, | 2239 | | NFS4ERR_REP_TOO_BIG_TO_CACHE, | 2240 | | NFS4ERR_REQ_TOO_BIG, NFS4ERR_RETRY_UNCACHED_REP, | 2241 | | NFS4ERR_SERVERFAULT, NFS4ERR_STALE, | 2242 | | NFS4ERR_SYMLINK, NFS4ERR_TOO_MANY_OPS, | 2243 | | NFS4ERR_UNION_NOTSUPP, NFS4ERR_WRONG_TYPE | 2244 | WRITE_SAME | NFS4ERR_ACCESS, NFS4ERR_ADMIN_REVOKED, | 2245 | | NFS4ERR_BADXDR, NFS4ERR_BAD_STATEID, | 2246 | | NFS4ERR_DEADSESSION, NFS4ERR_DELAY, | 2247 | | NFS4ERR_DELEG_REVOKED, NFS4ERR_DQUOT, | 2248 | | NFS4ERR_EXPIRED, NFS4ERR_FBIG, | 2249 | | NFS4ERR_FHEXPIRED, NFS4ERR_GRACE, NFS4ERR_INVAL, | 2250 | | NFS4ERR_IO, NFS4ERR_ISDIR, NFS4ERR_LOCKED, | 2251 | | NFS4ERR_MOVED, NFS4ERR_NOFILEHANDLE, | 2252 | | NFS4ERR_NOSPC, NFS4ERR_NOTSUPP, | 2253 | | NFS4ERR_OLD_STATEID, NFS4ERR_OPENMODE, | 2254 | | NFS4ERR_OP_NOT_IN_SESSION, NFS4ERR_PNFS_IO_HOLE, | 2255 | | NFS4ERR_PNFS_NO_LAYOUT, NFS4ERR_REP_TOO_BIG, | 2256 | | NFS4ERR_REP_TOO_BIG_TO_CACHE, | 2257 | | NFS4ERR_REQ_TOO_BIG, NFS4ERR_RETRY_UNCACHED_REP, | 2258 | | NFS4ERR_ROFS, NFS4ERR_SERVERFAULT, | 2259 | | NFS4ERR_STALE, NFS4ERR_SYMLINK, | 2260 | | NFS4ERR_TOO_MANY_OPS, NFS4ERR_WRONG_TYPE | 2261 +----------------+--------------------------------------------------+ 2263 Table 2 2265 11.3. New Callback Operations and Their Valid Errors 2267 This section contains a table that gives the valid error returns for 2268 each new NFSv4.2 callback operation. The error code NFS4_OK 2269 (indicating no error) is not listed but should be understood to be 2270 returnable by all new callback operations. The error values for all 2271 other callback operations are defined in Section 15.3 of [RFC5661]. 2273 Valid Error Returns for Each New Protocol Callback Operation 2275 +------------+------------------------------------------------------+ 2276 | Callback | Errors | 2277 | Operation | | 2278 +------------+------------------------------------------------------+ 2279 | CB_OFFLOAD | NFS4ERR_BADHANDLE, NFS4ERR_BADXDR, | 2280 | | NFS4ERR_BAD_STATEID, NFS4ERR_DELAY, | 2281 | | NFS4ERR_OP_NOT_IN_SESSION, NFS4ERR_REP_TOO_BIG, | 2282 | | NFS4ERR_REP_TOO_BIG_TO_CACHE, NFS4ERR_REQ_TOO_BIG, | 2283 | | NFS4ERR_RETRY_UNCACHED_REP, NFS4ERR_SERVERFAULT, | 2284 | | NFS4ERR_TOO_MANY_OPS | 2285 +------------+------------------------------------------------------+ 2287 Table 3 2289 12. New File Attributes 2291 12.1. New RECOMMENDED Attributes - List and Definition References 2293 The list of new RECOMMENDED attributes appears in Table 4. The 2294 meaning of the columns of the table are: 2296 Name: The name of the attribute. 2298 Id: The number assigned to the attribute. In the event of conflicts 2299 between the assigned number and [NFSv42xdr], the latter is likely 2300 authoritative, but should be resolved with Errata to this document 2301 and/or [NFSv42xdr]. See [IESG08] for the Errata process. 2303 Data Type: The XDR data type of the attribute. 2305 Acc: Access allowed to the attribute. 2307 R means read-only (GETATTR may retrieve, SETATTR may not set). 2309 W means write-only (SETATTR may set, GETATTR may not retrieve). 2311 R W means read/write (GETATTR may retrieve, SETATTR may set). 2313 Defined in: The section of this specification that describes the 2314 attribute. 2316 +------------------+----+-------------------+-----+----------------+ 2317 | Name | Id | Data Type | Acc | Defined in | 2318 +------------------+----+-------------------+-----+----------------+ 2319 | clone_blksize | 77 | length4 | R | Section 12.2.1 | 2320 | space_freed | 78 | length4 | R | Section 12.2.2 | 2321 | change_attr_type | 79 | change_attr_type4 | R | Section 12.2.3 | 2322 | sec_label | 80 | sec_label4 | R W | Section 12.2.4 | 2323 +------------------+----+-------------------+-----+----------------+ 2325 Table 4 2327 12.2. Attribute Definitions 2329 12.2.1. Attribute 77: clone_blksize 2331 The clone_blksize attribute indicates the granularity of a CLONE 2332 operation. 2334 12.2.2. Attribute 78: space_freed 2336 space_freed gives the number of bytes freed if the file is deleted. 2337 This attribute is read only and is of type length4. It is a per file 2338 attribute. 2340 12.2.3. Attribute 79: change_attr_type 2342 2344 enum change_attr_type4 { 2345 NFS4_CHANGE_TYPE_IS_MONOTONIC_INCR = 0, 2346 NFS4_CHANGE_TYPE_IS_VERSION_COUNTER = 1, 2347 NFS4_CHANGE_TYPE_IS_VERSION_COUNTER_NOPNFS = 2, 2348 NFS4_CHANGE_TYPE_IS_TIME_METADATA = 3, 2349 NFS4_CHANGE_TYPE_IS_UNDEFINED = 4 2350 }; 2352 2354 change_attr_type is a per file system attribute which enables the 2355 NFSv4.2 server to provide additional information about how it expects 2356 the change attribute value to evolve after the file data, or metadata 2357 has changed. While Section 5.4 of [RFC5661] discusses per file 2358 system attributes, it is expected that the value of change_attr_type 2359 not depend on the value of "homogeneous" and only changes in the 2360 event of a migration. 2362 NFS4_CHANGE_TYPE_IS_UNDEFINED: The change attribute does not take 2363 values that fit into any of these categories. 2365 NFS4_CHANGE_TYPE_IS_MONOTONIC_INCR: The change attribute value MUST 2366 monotonically increase for every atomic change to the file 2367 attributes, data, or directory contents. 2369 NFS4_CHANGE_TYPE_IS_VERSION_COUNTER: The change attribute value MUST 2370 be incremented by one unit for every atomic change to the file 2371 attributes, data, or directory contents. This property is 2372 preserved when writing to pNFS data servers. 2374 NFS4_CHANGE_TYPE_IS_VERSION_COUNTER_NOPNFS: The change attribute 2375 value MUST be incremented by one unit for every atomic change to 2376 the file attributes, data, or directory contents. In the case 2377 where the client is writing to pNFS data servers, the number of 2378 increments is not guaranteed to exactly match the number of 2379 writes. 2381 NFS4_CHANGE_TYPE_IS_TIME_METADATA: The change attribute is 2382 implemented as suggested in [RFC7530] in terms of the 2383 time_metadata attribute. 2385 If either NFS4_CHANGE_TYPE_IS_MONOTONIC_INCR, 2386 NFS4_CHANGE_TYPE_IS_VERSION_COUNTER, or 2387 NFS4_CHANGE_TYPE_IS_TIME_METADATA are set, then the client knows at 2388 the very least that the change attribute is monotonically increasing, 2389 which is sufficient to resolve the question of which value is the 2390 most recent. 2392 If the client sees the value NFS4_CHANGE_TYPE_IS_TIME_METADATA, then 2393 by inspecting the value of the 'time_delta' attribute it additionally 2394 has the option of detecting rogue server implementations that use 2395 time_metadata in violation of the spec. 2397 If the client sees NFS4_CHANGE_TYPE_IS_VERSION_COUNTER, it has the 2398 ability to predict what the resulting change attribute value should 2399 be after a COMPOUND containing a SETATTR, WRITE, or CREATE. This 2400 again allows it to detect changes made in parallel by another client. 2401 The value NFS4_CHANGE_TYPE_IS_VERSION_COUNTER_NOPNFS permits the 2402 same, but only if the client is not doing pNFS WRITEs. 2404 Finally, if the server does not support change_attr_type or if 2405 NFS4_CHANGE_TYPE_IS_UNDEFINED is set, then the server SHOULD make an 2406 effort to implement the change attribute in terms of the 2407 time_metadata attribute. 2409 12.2.4. Attribute 80: sec_label 2411 2413 typedef uint32_t policy4; 2415 struct labelformat_spec4 { 2416 policy4 lfs_lfs; 2417 policy4 lfs_pi; 2418 }; 2420 struct sec_label4 { 2421 labelformat_spec4 slai_lfs; 2422 opaque slai_data<>; 2423 }; 2425 2427 The FATTR4_SEC_LABEL contains an array of two components with the 2428 first component being an LFS. It serves to provide the receiving end 2429 with the information necessary to translate the security attribute 2430 into a form that is usable by the endpoint. Label Formats assigned 2431 an LFS may optionally choose to include a Policy Identifier field to 2432 allow for complex policy deployments. The LFS and Label Format 2433 Registry are described in detail in [Quigley14]. The translation 2434 used to interpret the security attribute is not specified as part of 2435 the protocol as it may depend on various factors. The second 2436 component is an opaque section which contains the data of the 2437 attribute. This component is dependent on the MAC model to interpret 2438 and enforce. 2440 In particular, it is the responsibility of the LFS specification to 2441 define a maximum size for the opaque section, slai_data<>. When 2442 creating or modifying a label for an object, the client needs to be 2443 guaranteed that the server will accept a label that is sized 2444 correctly. By both client and server being part of a specific MAC 2445 model, the client will be aware of the size. 2447 13. Operations: REQUIRED, RECOMMENDED, or OPTIONAL 2449 The following tables summarize the operations of the NFSv4.2 protocol 2450 and the corresponding designation of REQUIRED, RECOMMENDED, and 2451 OPTIONAL to implement or MUST NOT implement. The designation of MUST 2452 NOT implement is reserved for those operations that were defined in 2453 either NFSv4.0 or NFSV4.1 and MUST NOT be implemented in NFSv4.2. 2455 For the most part, the REQUIRED, RECOMMENDED, or OPTIONAL designation 2456 for operations sent by the client is for the server implementation. 2458 The client is generally required to implement the operations needed 2459 for the operating environment for which it serves. For example, a 2460 read-only NFSv4.2 client would have no need to implement the WRITE 2461 operation and is not required to do so. 2463 The REQUIRED or OPTIONAL designation for callback operations sent by 2464 the server is for both the client and server. Generally, the client 2465 has the option of creating the backchannel and sending the operations 2466 on the fore channel that will be a catalyst for the server sending 2467 callback operations. A partial exception is CB_RECALL_SLOT; the only 2468 way the client can avoid supporting this operation is by not creating 2469 a backchannel. 2471 Since this is a summary of the operations and their designation, 2472 there are subtleties that are not presented here. Therefore, if 2473 there is a question of the requirements of implementation, the 2474 operation descriptions themselves must be consulted along with other 2475 relevant explanatory text within this either specification or that of 2476 NFSv4.1 [RFC5661]. 2478 The abbreviations used in the second and third columns of the table 2479 are defined as follows. 2481 REQ: REQUIRED to implement 2483 REC: RECOMMENDED to implement 2485 OPT: OPTIONAL to implement 2487 MNI: MUST NOT implement 2489 For the NFSv4.2 features that are OPTIONAL, the operations that 2490 support those features are OPTIONAL, and the server MUST return 2491 NFS4ERR_NOTSUPP in response to the client's use of those operations, 2492 when those operations are not implemented by the server. If an 2493 OPTIONAL feature is supported, it is possible that a set of 2494 operations related to the feature become REQUIRED to implement. The 2495 third column of the table designates the feature(s) and if the 2496 operation is REQUIRED or OPTIONAL in the presence of support for the 2497 feature. 2499 The OPTIONAL features identified and their abbreviations are as 2500 follows: 2502 pNFS: Parallel NFS 2504 FDELG: File Delegations 2505 DDELG: Directory Delegations 2507 COPYra: Intra-server Server Side Copy 2509 COPYer: Inter-server Server Side Copy 2511 ADB: Application Data Blocks 2513 Operations 2515 +----------------------+---------------------+----------------------+ 2516 | Operation | EOL, REQ, REC, OPT, | Feature (REQ, REC, | 2517 | | or MNI | or OPT) | 2518 +----------------------+---------------------+----------------------+ 2519 | ALLOCATE | OPT | | 2520 | ACCESS | REQ | | 2521 | BACKCHANNEL_CTL | REQ | | 2522 | BIND_CONN_TO_SESSION | REQ | | 2523 | CLONE | OPT | | 2524 | CLOSE | REQ | | 2525 | COMMIT | REQ | | 2526 | COPY | OPT | COPYer (REQ), COPYra | 2527 | | | (REQ) | 2528 | COPY_NOTIFY | OPT | COPYer (REQ) | 2529 | DEALLOCATE | OPT | | 2530 | CREATE | REQ | | 2531 | CREATE_SESSION | REQ | | 2532 | DELEGPURGE | OPT | FDELG (REQ) | 2533 | DELEGRETURN | OPT | FDELG, DDELG, pNFS | 2534 | | | (REQ) | 2535 | DESTROY_CLIENTID | REQ | | 2536 | DESTROY_SESSION | REQ | | 2537 | EXCHANGE_ID | REQ | | 2538 | FREE_STATEID | REQ | | 2539 | GETATTR | REQ | | 2540 | GETDEVICEINFO | OPT | pNFS (REQ) | 2541 | GETDEVICELIST | MNI | pNFS (MNI) | 2542 | GETFH | REQ | | 2543 | GET_DIR_DELEGATION | OPT | DDELG (REQ) | 2544 | LAYOUTCOMMIT | OPT | pNFS (REQ) | 2545 | LAYOUTGET | OPT | pNFS (REQ) | 2546 | LAYOUTRETURN | OPT | pNFS (REQ) | 2547 | LAYOUTERROR | OPT | pNFS (OPT) | 2548 | LAYOUTSTATS | OPT | pNFS (OPT) | 2549 | LINK | OPT | | 2550 | LOCK | REQ | | 2551 | LOCKT | REQ | | 2552 | LOCKU | REQ | | 2553 | LOOKUP | REQ | | 2554 | LOOKUPP | REQ | | 2555 | NVERIFY | REQ | | 2556 | OFFLOAD_CANCEL | OPT | COPYer (REQ), COPYra | 2557 | | | (REQ) | 2558 | OFFLOAD_STATUS | OPT | COPYer (REQ), COPYra | 2559 | | | (REQ) | 2560 | OPEN | REQ | | 2561 | OPENATTR | OPT | | 2562 | OPEN_CONFIRM | MNI | | 2563 | OPEN_DOWNGRADE | REQ | | 2564 | PUTFH | REQ | | 2565 | PUTPUBFH | REQ | | 2566 | PUTROOTFH | REQ | | 2567 | READ | REQ | | 2568 | READDIR | REQ | | 2569 | READLINK | OPT | | 2570 | READ_PLUS | OPT | | 2571 | RECLAIM_COMPLETE | REQ | | 2572 | RELEASE_LOCKOWNER | MNI | | 2573 | REMOVE | REQ | | 2574 | RENAME | REQ | | 2575 | RENEW | MNI | | 2576 | RESTOREFH | REQ | | 2577 | SAVEFH | REQ | | 2578 | SECINFO | REQ | | 2579 | SECINFO_NO_NAME | REC | pNFS file layout | 2580 | | | (REQ) | 2581 | SEEK | OPT | | 2582 | SEQUENCE | REQ | | 2583 | SETATTR | REQ | | 2584 | SETCLIENTID | MNI | | 2585 | SETCLIENTID_CONFIRM | MNI | | 2586 | SET_SSV | REQ | | 2587 | TEST_STATEID | REQ | | 2588 | VERIFY | REQ | | 2589 | WANT_DELEGATION | OPT | FDELG (OPT) | 2590 | WRITE | REQ | | 2591 | WRITE_SAME | OPT | ADB (REQ) | 2592 +----------------------+---------------------+----------------------+ 2593 Callback Operations 2595 +-------------------------+------------------+----------------------+ 2596 | Operation | REQ, REC, OPT, | Feature (REQ, REC, | 2597 | | or MNI | or OPT) | 2598 +-------------------------+------------------+----------------------+ 2599 | CB_OFFLOAD | OPT | COPYer (REQ), COPYra | 2600 | | | (REQ) | 2601 | CB_GETATTR | OPT | FDELG (REQ) | 2602 | CB_LAYOUTRECALL | OPT | pNFS (REQ) | 2603 | CB_NOTIFY | OPT | DDELG (REQ) | 2604 | CB_NOTIFY_DEVICEID | OPT | pNFS (OPT) | 2605 | CB_NOTIFY_LOCK | OPT | | 2606 | CB_PUSH_DELEG | OPT | FDELG (OPT) | 2607 | CB_RECALL | OPT | FDELG, DDELG, pNFS | 2608 | | | (REQ) | 2609 | CB_RECALL_ANY | OPT | FDELG, DDELG, pNFS | 2610 | | | (REQ) | 2611 | CB_RECALL_SLOT | REQ | | 2612 | CB_RECALLABLE_OBJ_AVAIL | OPT | DDELG, pNFS (REQ) | 2613 | CB_SEQUENCE | OPT | FDELG, DDELG, pNFS | 2614 | | | (REQ) | 2615 | CB_WANTS_CANCELLED | OPT | FDELG, DDELG, pNFS | 2616 | | | (REQ) | 2617 +-------------------------+------------------+----------------------+ 2619 14. Modifications to NFSv4.1 Operations 2621 14.1. Operation 42: EXCHANGE_ID - Instantiate Client ID 2623 14.1.1. ARGUMENT 2625 2627 /* new */ 2628 const EXCHGID4_FLAG_SUPP_FENCE_OPS = 0x00000004; 2630 2632 14.1.2. RESULT 2634 Unchanged 2636 14.1.3. MOTIVATION 2638 Enterprise applications require guarantees that an operation has 2639 either aborted or completed. NFSv4.1 provides this guarantee as long 2640 as the session is alive: simply send a SEQUENCE operation on the same 2641 slot with a new sequence number, and the successful return of 2642 SEQUENCE indicates the previous operation has completed. However, if 2643 the session is lost, there is no way to know when any in progress 2644 operations have aborted or completed. In hindsight, the NFSv4.1 2645 specification should have mandated that DESTROY_SESSION either abort 2646 or complete all outstanding operations. 2648 14.1.4. DESCRIPTION 2650 A client SHOULD request the EXCHGID4_FLAG_SUPP_FENCE_OPS capability 2651 when it sends an EXCHANGE_ID operation. The server SHOULD set this 2652 capability in the EXCHANGE_ID reply whether the client requests it or 2653 not. It is the server's return that determines whether this 2654 capability is in effect. When it is in effect, the following will 2655 occur: 2657 o The server will not reply to any DESTROY_SESSION invoked with the 2658 client ID until all operations in progress are completed or 2659 aborted. 2661 o The server will not reply to subsequent EXCHANGE_ID invoked on the 2662 same client owner with a new verifier until all operations in 2663 progress on the client ID's session are completed or aborted. 2665 o In implementations where the NFS server is deployed as a cluster, 2666 it does support client ID trunking, and the 2667 EXCHGID4_FLAG_SUPP_FENCE_OPS capability is enabled, then a session 2668 ID created on one node of the storage cluster MUST be destroyable 2669 via DESTROY_SESSION. In addition, DESTROY_CLIENTID and an 2670 EXCHANGE_ID with a new verifier affects all sessions regardless 2671 what node the sessions were created on. 2673 14.2. Operation 48: GETDEVICELIST - Get All Device Mappings for a File 2674 System 2676 14.2.1. ARGUMENT 2678 2679 struct GETDEVICELIST4args { 2680 /* CURRENT_FH: object belonging to the file system */ 2681 layouttype4 gdla_layout_type; 2683 /* number of deviceIDs to return */ 2684 count4 gdla_maxdevices; 2686 nfs_cookie4 gdla_cookie; 2687 verifier4 gdla_cookieverf; 2688 }; 2690 2692 14.2.2. RESULT 2694 2696 struct GETDEVICELIST4resok { 2697 nfs_cookie4 gdlr_cookie; 2698 verifier4 gdlr_cookieverf; 2699 deviceid4 gdlr_deviceid_list<>; 2700 bool gdlr_eof; 2701 }; 2703 union GETDEVICELIST4res switch (nfsstat4 gdlr_status) { 2704 case NFS4_OK: 2705 GETDEVICELIST4resok gdlr_resok4; 2706 default: 2707 void; 2708 }; 2710 2712 14.2.3. MOTIVATION 2714 The GETDEVICELIST operation was introduced in [RFC5661] specifically 2715 to request a list of devices at filesystem mount time from block 2716 layout type servers. However use of the GETDEVICELIST operation 2717 introduces a race condition versus notification about changes to pNFS 2718 device IDs as provided by CB_NOTIFY_DEVICEID. Implementation 2719 experience with block layout servers has shown there is no need for 2720 GETDEVICELIST. Clients have to be able to request new devices using 2721 GETDEVICEINFO at any time in response either to a new deviceid in 2722 LAYOUTGET results or to the CB_NOTIFY_DEVICEID callback operation. 2724 14.2.4. DESCRIPTION 2726 Clients and servers MUST NOT implement the GETDEVICELIST operation. 2728 15. NFSv4.2 Operations 2730 15.1. Operation 59: ALLOCATE - Reserve Space in A Region of a File 2732 15.1.1. ARGUMENT 2734 2736 struct ALLOCATE4args { 2737 /* CURRENT_FH: file */ 2738 stateid4 aa_stateid; 2739 offset4 aa_offset; 2740 length4 aa_length; 2741 }; 2743 2745 15.1.2. RESULT 2747 2749 struct ALLOCATE4res { 2750 nfsstat4 ar_status; 2751 }; 2753 2755 15.1.3. DESCRIPTION 2757 Whenever a client wishes to reserve space for a region in a file it 2758 calls the ALLOCATE operation with the current filehandle set to the 2759 filehandle of the file in question, and the start offset and length 2760 in bytes of the region set in aa_offset and aa_length respectively. 2762 The server will ensure that backing blocks are reserved to the region 2763 specified by aa_offset and aa_length, and that no future writes into 2764 this region will return NFS4ERR_NOSPC. If the region lies partially 2765 or fully outside the current file size the file size will be set to 2766 aa_offset + aa_length implicitly. If the server cannot guarantee 2767 this, it must return NFS4ERR_NOSPC. 2769 The ALLOCATE operation can also be used to extend the size of a file 2770 if the region specified by aa_offset and aa_length extends beyond the 2771 current file size. In that case any data outside of the previous 2772 file size will return zeroes when read before data is written to it. 2774 It is not required that the server allocate the space to the file 2775 before returning success. The allocation can be deferred, however, 2776 it must be guaranteed that it will not fail for lack of space. The 2777 deferral does not result in an asynchronous reply. 2779 The ALLOCATE operation will result in the space_used attribute and 2780 space_freed attributes being increased by the number of bytes 2781 reserved unless they were previously reserved or written and not 2782 shared. 2784 15.2. Operation 60: COPY - Initiate a server-side copy 2786 15.2.1. ARGUMENT 2788 2790 struct COPY4args { 2791 /* SAVED_FH: source file */ 2792 /* CURRENT_FH: destination file */ 2793 stateid4 ca_src_stateid; 2794 stateid4 ca_dst_stateid; 2795 offset4 ca_src_offset; 2796 offset4 ca_dst_offset; 2797 length4 ca_count; 2798 bool ca_consecutive; 2799 bool ca_synchronous; 2800 netloc4 ca_source_server<>; 2801 }; 2803 2805 15.2.2. RESULT 2807 2809 struct write_response4 { 2810 stateid4 wr_callback_id<1>; 2811 length4 wr_count; 2812 stable_how4 wr_committed; 2813 verifier4 wr_writeverf; 2814 }; 2815 struct copy_requirements4 { 2816 bool cr_consecutive; 2817 bool cr_synchronous; 2818 }; 2820 struct COPY4resok { 2821 write_response4 cr_response; 2822 copy_requirements4 cr_requirements; 2823 }; 2825 union COPY4res switch (nfsstat4 cr_status) { 2826 case NFS4_OK: 2827 COPY4resok cr_resok4; 2828 case NFS4ERR_OFFLOAD_NO_REQS: 2829 copy_requirements4 cr_requirements; 2830 default: 2831 void; 2832 }; 2834 2836 15.2.3. DESCRIPTION 2838 The COPY operation is used for both intra-server and inter-server 2839 copies. In both cases, the COPY is always sent from the client to 2840 the destination server of the file copy. The COPY operation requests 2841 that a range in the file specified by SAVED_FH is copied to a range 2842 in the file specified by CURRENT_FH. 2844 Both SAVED_FH and CURRENT_FH must be regular files. If either 2845 SAVED_FH or CURRENT_FH are not regular files, the operation MUST fail 2846 and return NFS4ERR_WRONG_TYPE. 2848 SAVED_FH and CURRENT_FH must be differet files. If SAVED_FH and 2849 CURRENT_FH refer to the same file, the operation will fail with 2850 NFS4ERR_INVAL. 2852 In order to set SAVED_FH to the source file handle, the compound 2853 procedure requesting the COPY will include a sub-sequence of 2854 operations such as 2856 PUTFH source-fh 2857 SAVEFH 2859 If the request is for an inter-server-to-server copy, the source-fh 2860 is a filehandle from the source server and the compound procedure is 2861 being executed on the destination server. In this case, the source- 2862 fh is a foreign filehandle on the server receiving the COPY request. 2864 If either PUTFH or SAVEFH checked the validity of the filehandle, the 2865 operation would likely fail and return NFS4ERR_STALE. 2867 If a server supports the inter-server-to-server COPY feature, a PUTFH 2868 followed by a SAVEFH MUST NOT return NFS4ERR_STALE for either 2869 operation. These restrictions do not pose substantial difficulties 2870 for servers. CURRENT_FH and SAVED_FH may be validated in the context 2871 of the operation referencing them and an NFS4ERR_STALE error returned 2872 for an invalid file handle at that point. 2874 The ca_dst_stateid MUST refer to a stateid that is valid for a WRITE 2875 operation and follows the rules for stateids in Sections 8.2.5 and 2876 18.32.3 of [RFC5661]. For an inter-server copy, the ca_src_stateid 2877 MUST be the cnr_stateid returned from the earlier COPY_NOTIFY 2878 operation, while for an intra-server copy ca_src_stateid MUST refer 2879 to a stateid that is valid for a READ operations and follows the 2880 rules for stateids in Sections 8.2.5 and 18.22.3 of [RFC5661]. If 2881 either stateid is invalid, then the operation MUST fail. 2883 The ca_src_offset is the offset within the source file from which the 2884 data will be read, the ca_dst_offset is the offset within the 2885 destination file to which the data will be written, and the ca_count 2886 is the number of bytes that will be copied. An offset of 0 (zero) 2887 specifies the start of the file. A count of 0 (zero) requests that 2888 all bytes from ca_src_offset through EOF be copied to the 2889 destination. If concurrent modifications to the source file overlap 2890 with the source file region being copied, the data copied may include 2891 all, some, or none of the modifications. The client can use standard 2892 NFS operations (e.g., OPEN with OPEN4_SHARE_DENY_WRITE or mandatory 2893 byte range locks) to protect against concurrent modifications if the 2894 client is concerned about this. If the source file's end of file is 2895 being modified in parallel with a copy that specifies a count of 0 2896 (zero) bytes, the amount of data copied is implementation dependent 2897 (clients may guard against this case by specifying a non-zero count 2898 value or preventing modification of the source file as mentioned 2899 above). 2901 If the source offset or the source offset plus count is greater than 2902 the size of the source file, the operation will fail with 2903 NFS4ERR_INVAL. The destination offset or destination offset plus 2904 count may be greater than the size of the destination file. This 2905 allows for the client to issue parallel copies to implement 2906 operations such as 2908 2910 % cat file1 file2 file3 file4 > dest 2912 2914 If the ca_source_server list is specified, then this is an inter- 2915 server copy operation and the source file is on a remote server. The 2916 client is expected to have previously issued a successful COPY_NOTIFY 2917 request to the remote source server. The ca_source_server list MUST 2918 be the same as the COPY_NOTIFY response's cnr_source_server list. If 2919 the client includes the entries from the COPY_NOTIFY response's 2920 cnr_source_server list in the ca_source_server list, the source 2921 server can indicate a specific copy protocol for the destination 2922 server to use by returning a URL, which specifies both a protocol 2923 service and server name. Server-to-server copy protocol 2924 considerations are described in Section 4.7 and Section 4.10.1. 2926 If ca_consecutive is set, then the client has specified that the copy 2927 protocol selected MUST copy bytes in consecutive order from 2928 ca_src_offset to ca_count. If the destination server cannot meet 2929 this requirement, then it MUST return an error of 2930 NFS4ERR_OFFLOAD_NO_REQS and set cr_consecutive to be false. 2931 Likewise, if ca_synchronous is set, then the client has required that 2932 the copy protocol selected MUST perform a synchronous copy. If the 2933 destination server cannot meet this requirement, then it MUST return 2934 an error of NFS4ERR_OFFLOAD_NO_REQS and set cr_synchronous to be 2935 false. 2937 If both are set by the client, then the destination SHOULD try to 2938 determine if it can respond to both requirements at the same time. 2939 If it cannot make that determination, it must set to false the one it 2940 can and set to true the other. The client, upon getting an 2941 NFS4ERR_OFFLOAD_NO_REQS error, has to examine both cr_consecutive and 2942 cr_synchronous against the respective values of ca_consecutive and 2943 ca_synchronous to determine the possible requirement not met. It 2944 MUST be prepared for the destination server not being able to 2945 determine both requirements at the same time. 2947 Upon receiving the NFS4ERR_OFFLOAD_NO_REQS error, the client has to 2948 determine if it wants to either re-request the copy with a relaxed 2949 set of requirements or if it wants to revert to manually copying the 2950 data. If it decides to manually copy the data and this is a remote 2951 copy, then the client is responsible for informing the source that 2952 the earlier COPY_NOTIFY is no longer valid by sending it an 2953 OFFLOAD_CANCEL. 2955 If the operation does not result in an immediate failure, the server 2956 will return NFS4_OK. 2958 If the wr_callback_id is returned, this indicates that the operation 2959 was initiated and a CB_OFFLOAD callback will deliver the final 2960 results of the operation. The wr_callback_id stateid is termed a 2961 copy stateid in this context. The server is given the option of 2962 returning the results in a callback because the data may require a 2963 relatively long period of time to copy. 2965 If no wr_callback_id is returned, the operation completed 2966 synchronously and no callback will be issued by the server. The 2967 completion status of the operation is indicated by cr_status. 2969 If the copy completes successfully, either synchronously or 2970 asynchronously, the data copied from the source file to the 2971 destination file MUST appear identical to the NFS client. However, 2972 the NFS server's on disk representation of the data in the source 2973 file and destination file MAY differ. For example, the NFS server 2974 might encrypt, compress, deduplicate, or otherwise represent the on 2975 disk data in the source and destination file differently. 2977 If a failure does occur for a synchronous copy, wr_count will be set 2978 to the number of bytes copied to the destination file before the 2979 error occurred. If cr_consecutive is true, then the bytes were 2980 copied in order. If the failure occurred for an asynchronous copy, 2981 then the client will have gotten the notification of the consecutive 2982 copy order when it got the copy stateid. It will be able to 2983 determine the bytes copied from the coa_bytes_copied in the 2984 CB_OFFLOAD argument. 2986 In either case, if cr_consecutive was not true, there is no assurance 2987 as to exactly which bytes in the range were copied. The client MUST 2988 assume that there exists a mixture of the original contents of the 2989 range and the new bytes. If the COPY wrote past the end of the file 2990 on the destination, then the last byte written to will determine the 2991 new file size. The contents of any block not written to and past the 2992 original size of the file will be as if a normal WRITE extended the 2993 file. 2995 15.3. Operation 61: COPY_NOTIFY - Notify a source server of a future 2996 copy 2998 15.3.1. ARGUMENT 3000 3001 struct COPY_NOTIFY4args { 3002 /* CURRENT_FH: source file */ 3003 stateid4 cna_src_stateid; 3004 netloc4 cna_destination_server; 3005 }; 3007 3009 15.3.2. RESULT 3011 3013 struct COPY_NOTIFY4resok { 3014 nfstime4 cnr_lease_time; 3015 stateid4 cnr_stateid; 3016 netloc4 cnr_source_server<>; 3017 }; 3019 union COPY_NOTIFY4res switch (nfsstat4 cnr_status) { 3020 case NFS4_OK: 3021 COPY_NOTIFY4resok resok4; 3022 default: 3023 void; 3024 }; 3026 3028 15.3.3. DESCRIPTION 3030 This operation is used for an inter-server copy. A client sends this 3031 operation in a COMPOUND request to the source server to authorize a 3032 destination server identified by cna_destination_server to read the 3033 file specified by CURRENT_FH on behalf of the given user. 3035 The cna_src_stateid MUST refer to either open or locking states 3036 provided earlier by the server. If it is invalid, then the operation 3037 MUST fail. 3039 The cna_destination_server MUST be specified using the netloc4 3040 network location format. The server is not required to resolve the 3041 cna_destination_server address before completing this operation. 3043 If this operation succeeds, the source server will allow the 3044 cna_destination_server to copy the specified file on behalf of the 3045 given user as long as both of the following conditions are met: 3047 o The destination server begins reading the source file before the 3048 cnr_lease_time expires. If the cnr_lease_time expires while the 3049 destination server is still reading the source file, the 3050 destination server is allowed to finish reading the file. 3052 o The client has not issued a OFFLOAD_CANCEL for the same 3053 combination of user, filehandle, and destination server. 3055 The cnr_lease_time is chosen by the source server. A cnr_lease_time 3056 of 0 (zero) indicates an infinite lease. To avoid the need for 3057 synchronized clocks, copy lease times are granted by the server as a 3058 time delta. To renew the copy lease time the client should resend 3059 the same copy notification request to the source server. 3061 The cnr_stateid is a copy stateid which uniquely describes the state 3062 needed on the source server to track the proposed copy. As defined 3063 in Section 8.2 of [RFC5661], a stateid is tied to the current 3064 filehandle and if the same stateid is presented by two different 3065 clients, it may refer to different state. As the source does not 3066 know which netloc4 network location the destinaton might use to 3067 establish the copy operation, it can use the cnr_stateid to identify 3068 that the destination is operating on behalf of the client. Thus the 3069 source server MUST construct copy stateids such that they are 3070 distinct from all other stateids handed out to clients. These copy 3071 stateids MUST denote the same set of locks as each of the earlier 3072 delegation, locking, and open states for the client on the given file 3073 (see Section 4.4.1). 3075 A successful response will also contain a list of netloc4 network 3076 location formats called cnr_source_server, on which the source is 3077 willing to accept connections from the destination. These might not 3078 be reachable from the client and might be located on networks to 3079 which the client has no connection. 3081 For a copy only involving one server (the source and destination are 3082 on the same server), this operation is unnecessary. 3084 15.4. Operation 62: DEALLOCATE - Unreserve Space in a Region of a File 3086 15.4.1. ARGUMENT 3088 3089 struct DEALLOCATE4args { 3090 /* CURRENT_FH: file */ 3091 stateid4 da_stateid; 3092 offset4 da_offset; 3093 length4 da_length; 3094 }; 3096 3098 15.4.2. RESULT 3100 3102 struct DEALLOCATE4res { 3103 nfsstat4 dr_status; 3104 }; 3106 3108 15.4.3. DESCRIPTION 3110 Whenever a client wishes to unreserve space for a region in a file it 3111 calls the DEALLOCATE operation with the current filehandle set to the 3112 filehandle of the file in question, and the start offset and length 3113 in bytes of the region set in da_offset and da_length respectively. 3114 If no space was allocated or reserved for all or parts of the region, 3115 the DEALLOCATE operation will have no effect for the region that 3116 already is in unreserved state. All further reads from the region 3117 passed to DEALLOCATE MUST return zeros until overwritten. The 3118 filehandle specified must be that of a regular file. 3120 Situations may arise where da_offset and/or da_offset + da_length 3121 will not be aligned to a boundary for which the server does 3122 allocations or deallocations. For most file systems, this is the 3123 block size of the file system. In such a case, the server can 3124 deallocate as many bytes as it can in the region. The blocks that 3125 cannot be deallocated MUST be zeroed. 3127 DEALLOCATE will result in the space_used attribute being decreased by 3128 the number of bytes that were deallocated. The space_freed attribute 3129 may or may not decrease, depending on the support and whether the 3130 blocks backing the specified range were shared or not. The size 3131 attribute will remain unchanged. 3133 15.5. Operation 63: IO_ADVISE - Application I/O access pattern hints 3135 15.5.1. ARGUMENT 3137 3139 enum IO_ADVISE_type4 { 3140 IO_ADVISE4_NORMAL = 0, 3141 IO_ADVISE4_SEQUENTIAL = 1, 3142 IO_ADVISE4_SEQUENTIAL_BACKWARDS = 2, 3143 IO_ADVISE4_RANDOM = 3, 3144 IO_ADVISE4_WILLNEED = 4, 3145 IO_ADVISE4_WILLNEED_OPPORTUNISTIC = 5, 3146 IO_ADVISE4_DONTNEED = 6, 3147 IO_ADVISE4_NOREUSE = 7, 3148 IO_ADVISE4_READ = 8, 3149 IO_ADVISE4_WRITE = 9, 3150 IO_ADVISE4_INIT_PROXIMITY = 10 3151 }; 3153 struct IO_ADVISE4args { 3154 /* CURRENT_FH: file */ 3155 stateid4 iaa_stateid; 3156 offset4 iaa_offset; 3157 length4 iaa_count; 3158 bitmap4 iaa_hints; 3159 }; 3161 3163 15.5.2. RESULT 3165 3167 struct IO_ADVISE4resok { 3168 bitmap4 ior_hints; 3169 }; 3171 union IO_ADVISE4res switch (nfsstat4 ior_status) { 3172 case NFS4_OK: 3173 IO_ADVISE4resok resok4; 3174 default: 3175 void; 3176 }; 3177 3179 15.5.3. DESCRIPTION 3181 The IO_ADVISE operation sends an I/O access pattern hint to the 3182 server for the owner of the stateid for a given byte range specified 3183 by iar_offset and iar_count. The byte range specified by iaa_offset 3184 and iaa_count need not currently exist in the file, but the iaa_hints 3185 will apply to the byte range when it does exist. If iaa_count is 0, 3186 all data following iaa_offset is specified. The server MAY ignore 3187 the advice. 3189 The following are the allowed hints for a stateid holder: 3191 IO_ADVISE4_NORMAL There is no advice to give, this is the default 3192 behavior. 3194 IO_ADVISE4_SEQUENTIAL Expects to access the specified data 3195 sequentially from lower offsets to higher offsets. 3197 IO_ADVISE4_SEQUENTIAL_BACKWARDS Expects to access the specified data 3198 sequentially from higher offsets to lower offsets. 3200 IO_ADVISE4_RANDOM Expects to access the specified data in a random 3201 order. 3203 IO_ADVISE4_WILLNEED Expects to access the specified data in the near 3204 future. 3206 IO_ADVISE4_WILLNEED_OPPORTUNISTIC Expects to possibly access the 3207 data in the near future. This is a speculative hint, and 3208 therefore the server should prefetch data or indirect blocks only 3209 if it can be done at a marginal cost. 3211 IO_ADVISE_DONTNEED Expects that it will not access the specified 3212 data in the near future. 3214 IO_ADVISE_NOREUSE Expects to access the specified data once and then 3215 not reuse it thereafter. 3217 IO_ADVISE4_READ Expects to read the specified data in the near 3218 future. 3220 IO_ADVISE4_WRITE Expects to write the specified data in the near 3221 future. 3223 IO_ADVISE4_INIT_PROXIMITY Informs the server that the data in the 3224 byte range remains important to the client. 3226 Since IO_ADVISE is a hint, a server SHOULD NOT return an error and 3227 invalidate a entire Compound request if one of the sent hints in 3228 iar_hints is not supported by the server. Also, the server MUST NOT 3229 return an error if the client sends contradictory hints to the 3230 server, e.g., IO_ADVISE4_SEQUENTIAL and IO_ADVISE4_RANDOM in a single 3231 IO_ADVISE operation. In these cases, the server MUST return success 3232 and a ior_hints value that indicates the hint it intends to 3233 implement. This may mean simply returning IO_ADVISE4_NORMAL. 3235 The ior_hints returned by the server is primarily for debugging 3236 purposes since the server is under no obligation to carry out the 3237 hints that it describes in the ior_hints result. In addition, while 3238 the server may have intended to implement the hints returned in 3239 ior_hints, as time progresses, the server may need to change its 3240 handling of a given file due to several reasons including, but not 3241 limited to, memory pressure, additional IO_ADVISE hints sent by other 3242 clients, and heuristically detected file access patterns. 3244 The server MAY return different advice than what the client 3245 requested. If it does, then this might be due to one of several 3246 conditions, including, but not limited to another client advising of 3247 a different I/O access pattern; a different I/O access pattern from 3248 another client that that the server has heuristically detected; or 3249 the server is not able to support the requested I/O access pattern, 3250 perhaps due to a temporary resource limitation. 3252 Each issuance of the IO_ADVISE operation overrides all previous 3253 issuances of IO_ADVISE for a given byte range. This effectively 3254 follows a strategy of last hint wins for a given stateid and byte 3255 range. 3257 Clients should assume that hints included in an IO_ADVISE operation 3258 will be forgotten once the file is closed. 3260 15.5.4. IMPLEMENTATION 3262 The NFS client may choose to issue an IO_ADVISE operation to the 3263 server in several different instances. 3265 The most obvious is in direct response to an application's execution 3266 of posix_fadvise(). In this case, IO_ADVISE4_WRITE and 3267 IO_ADVISE4_READ may be set based upon the type of file access 3268 specified when the file was opened. 3270 15.5.5. IO_ADVISE4_INIT_PROXIMITY 3272 The IO_ADVISE4_INIT_PROXIMITY hint is non-posix in origin and can be 3273 used to convey that the client has recently accessed the byte range 3274 in its own cache. I.e., it has not accessed it on the server, but it 3275 has locally. When the server reaches resource exhaustion, knowing 3276 which data is more important allows the server to make better choices 3277 about which data to, for example purge from a cache, or move to 3278 secondary storage. It also informs the server which delegations are 3279 more important, since if delegations are working correctly, once 3280 delegated to a client and the client has read the content for that 3281 byte range, a server might never receive another read request for 3282 that byte range. 3284 The IO_ADVISE4_INIT_PROXIMITY hint can also be used in a pNFS setting 3285 to let the client inform the metadata server as to the I/O statistics 3286 between the client and the storage devices. The metadata server is 3287 then free to use this information about client I/O to optimize the 3288 data storage location. 3290 This hint is also useful in the case of NFS clients which are network 3291 booting from a server. If the first client to be booted sends this 3292 hint, then it keeps the cache warm for the remaining clients. 3294 15.5.6. pNFS File Layout Data Type Considerations 3296 The IO_ADVISE considerations for pNFS are very similar to the COMMIT 3297 considerations for pNFS. That is, as with COMMIT, some NFS server 3298 implementations prefer IO_ADVISE be done on the DS, and some prefer 3299 it be done on the MDS. 3301 For the file's layout type, it is proposed that NFSv4.2 include an 3302 additional hint NFL42_CARE_IO_ADVISE_THRU_MDS which is valid only on 3303 metadata servers running NFSv4.2 or higher. Any file's layout 3304 obtained from a NFSv4.1 metadata server MUST NOT have 3305 NFL42_UFLG_IO_ADVISE_THRU_MDS set. Any file's layout obtained with a 3306 NFSv4.2 metadata server MAY have NFL42_UFLG_IO_ADVISE_THRU_MDS set. 3307 However, if the layout utilizes NFSv4.1 storage devices, the 3308 IO_ADVISE operation cannot be sent to them. 3310 If NFL42_UFLG_IO_ADVISE_THRU_MDS is set, the client MUST send the 3311 IO_ADVISE operation to the MDS in order for it to be honored by the 3312 DS. Once the MDS receives the IO_ADVISE operation, it will 3313 communicate the advice to each DS. 3315 If NFL42_UFLG_IO_ADVISE_THRU_MDS is not set, then the client SHOULD 3316 send an IO_ADVISE operation to the appropriate DS for the specified 3317 byte range. While the client MAY always send IO_ADVISE to the MDS, 3318 if the server has not set NFL42_UFLG_IO_ADVISE_THRU_MDS, the client 3319 should expect that such an IO_ADVISE is futile. Note that a client 3320 SHOULD use the same set of arguments on each IO_ADVISE sent to a DS 3321 for the same open file reference. 3323 The server is not required to support different advice for different 3324 DS's with the same open file reference. 3326 15.5.6.1. Dense and Sparse Packing Considerations 3328 The IO_ADVISE operation MUST use the iar_offset and byte range as 3329 dictated by the presence or absence of NFL4_UFLG_DENSE. 3331 E.g., if NFL4_UFLG_DENSE is present, and a READ or WRITE to the DS 3332 for iaa_offset 0 really means iaa_offset 10000 in the logical file, 3333 then an IO_ADVISE for iaa_offset 0 means iaa_offset 10000. 3335 E.g., if NFL4_UFLG_DENSE is absent, then a READ or WRITE to the DS 3336 for iaa_offset 0 really means iaa_offset 0 in the logical file, then 3337 an IO_ADVISE for iaa_offset 0 means iaa_offset 0 in the logical file. 3339 E.g., if NFL4_UFLG_DENSE is present, the stripe unit is 1000 bytes 3340 and the stripe count is 10, and the dense DS file is serving 3341 iar_offset 0. A READ or WRITE to the DS for iaa_offsets 0, 1000, 3342 2000, and 3000, really mean iaa_offsets 10000, 20000, 30000, and 3343 40000 (implying a stripe count of 10 and a stripe unit of 1000), then 3344 an IO_ADVISE sent to the same DS with an iaa_offset of 500, and an 3345 iaa_count of 3000 means that the IO_ADVISE applies to these byte 3346 ranges of the dense DS file: 3348 - 500 to 999 3349 - 1000 to 1999 3350 - 2000 to 2999 3351 - 3000 to 3499 3353 I.e., the contiguous range 500 to 3499 as specified in IO_ADVISE. 3355 It also applies to these byte ranges of the logical file: 3357 - 10500 to 10999 (500 bytes) 3358 - 20000 to 20999 (1000 bytes) 3359 - 30000 to 30999 (1000 bytes) 3360 - 40000 to 40499 (500 bytes) 3361 (total 3000 bytes) 3363 E.g., if NFL4_UFLG_DENSE is absent, the stripe unit is 250 bytes, the 3364 stripe count is 4, and the sparse DS file is serving iaa_offset 0. 3365 Then a READ or WRITE to the DS for iaa_offsets 0, 1000, 2000, and 3366 3000, really means iaa_offsets 0, 1000, 2000, and 3000 in the logical 3367 file, keeping in mind that on the DS file, byte ranges 250 to 999, 3368 1250 to 1999, 2250 to 2999, and 3250 to 3999 are not accessible. 3369 Then an IO_ADVISE sent to the same DS with an iaa_offset of 500, and 3370 a iaa_count of 3000 means that the IO_ADVISE applies to these byte 3371 ranges of the logical file and the sparse DS file: 3373 - 500 to 999 (500 bytes) - no effect 3374 - 1000 to 1249 (250 bytes) - effective 3375 - 1250 to 1999 (750 bytes) - no effect 3376 - 2000 to 2249 (250 bytes) - effective 3377 - 2250 to 2999 (750 bytes) - no effect 3378 - 3000 to 3249 (250 bytes) - effective 3379 - 3250 to 3499 (250 bytes) - no effect 3380 (subtotal 2250 bytes) - no effect 3381 (subtotal 750 bytes) - effective 3382 (grand total 3000 bytes) - no effect + effective 3384 If neither of the flags NFL42_UFLG_IO_ADVISE_THRU_MDS and 3385 NFL4_UFLG_DENSE are set in the layout, then any IO_ADVISE request 3386 sent to the data server with a byte range that overlaps stripe unit 3387 that the data server does not serve MUST NOT result in the status 3388 NFS4ERR_PNFS_IO_HOLE. Instead, the response SHOULD be successful and 3389 if the server applies IO_ADVISE hints on any stripe units that 3390 overlap with the specified range, those hints SHOULD be indicated in 3391 the response. 3393 15.6. Operation 64: LAYOUTERROR - Provide Errors for the Layout 3395 15.6.1. ARGUMENT 3397 3399 struct device_error4 { 3400 deviceid4 de_deviceid; 3401 nfsstat4 de_status; 3402 nfs_opnum4 de_opnum; 3403 }; 3404 struct LAYOUTERROR4args { 3405 /* CURRENT_FH: file */ 3406 offset4 lea_offset; 3407 length4 lea_length; 3408 stateid4 lea_stateid; 3409 device_error4 lea_errors<>; 3410 }; 3412 3414 15.6.2. RESULT 3416 3418 struct LAYOUTERROR4res { 3419 nfsstat4 ler_status; 3420 }; 3422 3424 15.6.3. DESCRIPTION 3426 The client can use LAYOUTERROR to inform the metadata server about 3427 errors in its interaction with the layout represented by the current 3428 filehandle, client ID (derived from the session ID in the preceding 3429 SEQUENCE operation), byte-range (lea_offset + lea_length), and 3430 lea_stateid. 3432 Each individual device_error4 describes a single error associated 3433 with a storage device, which is identified via de_deviceid. If the 3434 Layout Type supports NFSv4 operations, then the operation which 3435 returned the error is identified via de_opnum. If the Layout Type 3436 does not support NFSv4 operations, then it MAY chose to either map 3437 the operation onto one of the allowed operations which can be sent to 3438 a storage device with the File Layout Type (see Section 3.3) or it 3439 can signal no support for operations by marking de_opnum with the 3440 ILLEGAL operation. Finally the NFS error value (nfsstat4) 3441 encountered is provided via de_status and may consist of the 3442 following error codes: 3444 NFS4ERR_NXIO: The client was unable to establish any communication 3445 with the storage device. 3447 NFS4ERR_*: The client was able to establish communication with the 3448 storage device and is returning one of the allowed error codes for 3449 the operation denoted by de_opnum. 3451 Note that while the metadata server may return an error associated 3452 with the layout stateid or the open file, it MUST NOT return an error 3453 in the processing of the errors. If LAYOUTERROR is in a compound 3454 before LAYOUTRETURN, it MUST NOT introduce an error other than what 3455 LAYOUTRETURN would already encounter. 3457 15.6.4. IMPLEMENTATION 3459 There are two broad classes of errors, transient and persistent. The 3460 client SHOULD strive to only use this new mechanism to report 3461 persistent errors. It MUST be able to deal with transient issues by 3462 itself. Also, while the client might consider an issue to be 3463 persistent, it MUST be prepared for the metadata server to consider 3464 such issues to be transient. A prime example of this is if the 3465 metadata server fences off a client from either a stateid or a 3466 filehandle. The client will get an error from the storage device and 3467 might relay either NFS4ERR_ACCESS or NFS4ERR_BAD_STATEID back to the 3468 metadata server, with the belief that this is a hard error. If the 3469 metadata server is informed by the client that there is an error, it 3470 can safely ignore that. For it, the mission is accomplished in that 3471 the client has returned a layout that the metadata server had most 3472 likely recalled. 3474 The client might also need to inform the metadata server that it 3475 cannot reach one or more of the storage devices. While the metadata 3476 server can detect the connectivity of both of these paths: 3478 o metadata server to storage device 3480 o metadata server to client 3482 it cannot determine if the client and storage device path is working. 3483 As with the case of the storage device passing errors to the client, 3484 it must be prepared for the metadata server to consider such outages 3485 as being transitory. 3487 Clients are expected to tolerate transient storage device errors, and 3488 hence clients SHOULD NOT use the LAYOUTERROR error handling for 3489 device access problems that may be transient. The methods by which a 3490 client decides whether a device access problem is transient vs 3491 persistent are implementation-specific, but may include retrying I/Os 3492 to a data server under appropriate conditions. 3494 When an I/O fails to a storage device, the client SHOULD retry the 3495 failed I/O via the metadata server. In this situation, before 3496 retrying the I/O, the client SHOULD return the layout, or the 3497 affected portion thereof, and SHOULD indicate which storage device or 3498 devices was problematic. The client needs to do this when the 3499 storage device is being unresponsive in order to fence off any failed 3500 write attempts, and ensure that they do not end up overwriting any 3501 later data being written through the metadata server. If the client 3502 does not do this, the metadata server MAY issue a layout recall 3503 callback in order to perform the retried I/O. 3505 The client needs to be cognizant that since this error handling is 3506 optional in the metadata server, the metadata server may silently 3507 ignore this functionality. Also, as the metadata server may consider 3508 some issues the client reports to be expected, the client might find 3509 it difficult to detect a metadata server which has not implemented 3510 error handling via LAYOUTERROR. 3512 If an metadata server is aware that a storage device is proving 3513 problematic to a client, the metadata server SHOULD NOT include that 3514 storage device in any pNFS layouts sent to that client. If the 3515 metadata server is aware that a storage device is affecting many 3516 clients, then the metadata server SHOULD NOT include that storage 3517 device in any pNFS layouts sent out. If a client asks for a new 3518 layout for the file from the metadata server, it MUST be prepared for 3519 the metadata server to return that storage device in the layout. The 3520 metadata server might not have any choice in using the storage 3521 device, i.e., there might only be one possible layout for the system. 3522 Also, in the case of existing files, the metadata server might have 3523 no choice in which storage devices to hand out to clients. 3525 The metadata server is not required to indefinitely retain per-client 3526 storage device error information. An metadata server is also not 3527 required to automatically reinstate use of a previously problematic 3528 storage device; administrative intervention may be required instead. 3530 15.7. Operation 65: LAYOUTSTATS - Provide Statistics for the Layout 3532 15.7.1. ARGUMENT 3534 3536 struct layoutupdate4 { 3537 layouttype4 lou_type; 3538 opaque lou_body<>; 3539 }; 3541 struct io_info4 { 3542 uint32_t ii_count; 3543 uint64_t ii_bytes; 3544 }; 3545 struct LAYOUTSTATS4args { 3546 /* CURRENT_FH: file */ 3547 offset4 lsa_offset; 3548 length4 lsa_length; 3549 stateid4 lsa_stateid; 3550 io_info4 lsa_read; 3551 io_info4 lsa_write; 3552 deviceid4 lsa_deviceid; 3553 layoutupdate4 lsa_layoutupdate; 3554 }; 3556 3558 15.7.2. RESULT 3560 3562 struct LAYOUTSTATS4res { 3563 nfsstat4 lsr_status; 3564 }; 3566 3568 15.7.3. DESCRIPTION 3570 The client can use LAYOUTSTATS to inform the metadata server about 3571 its interaction with the layout represented by the current 3572 filehandle, client ID (derived from the session ID in the preceding 3573 SEQUENCE operation), byte-range (lsa_offset and lsa_length), and 3574 lsa_stateid. lsa_read and lsa_write allow for non-Layout Type 3575 specific statistics to be reported. lsa_deviceid allows the client 3576 to specify to which storage device the statistics apply. The 3577 remaining information the client is presenting is specific to the 3578 Layout Type and presented in the lsa_layoutupdate field. Each Layout 3579 Type MUST define the contents of lsa_layoutupdate in their respective 3580 specifications. 3582 LAYOUTSTATS can be combined with IO_ADVISE (see Section 15.5) to 3583 augment the decision making process of how the metadata server 3584 handles a file. I.e., IO_ADVISE lets the server know that a byte 3585 range has a certain characteristic, but not necessarily the intensity 3586 of that characteristic. 3588 The client MUST reset the statistics after getting a successfully 3589 reply from the metadata server. The first LAYOUTSTATS sent by the 3590 client SHOULD be from the opening of the file. The choice of how 3591 often to update the metadata server is made by the client. 3593 Note that while the metadata server may return an error associated 3594 with the layout stateid or the open file, it MUST NOT return an error 3595 in the processing of the statistics. 3597 15.8. Operation 66: OFFLOAD_CANCEL - Stop an Offloaded Operation 3599 15.8.1. ARGUMENT 3601 3603 struct OFFLOAD_CANCEL4args { 3604 /* CURRENT_FH: file to cancel */ 3605 stateid4 oca_stateid; 3606 }; 3608 3610 15.8.2. RESULT 3612 3614 struct OFFLOAD_CANCEL4res { 3615 nfsstat4 ocr_status; 3616 }; 3618 3620 15.8.3. DESCRIPTION 3622 OFFLOAD_CANCEL is used by the client to terminate an asynchronous 3623 operation, which is identified both by CURRENT_FH and the 3624 oca_stateid. I.e., there can be multiple offloaded operations acting 3625 on the file, the stateid will identify to the server exactly which 3626 one is to be stopped. Currently there are only two operations which 3627 can decide to be asynchronous: COPY and WRITE_SAME. 3629 In the context of server-to-server copy, the client can send 3630 OFFLOAD_CANCEL to either the source or destination server, albeit 3631 with a different stateid. The client uses OFFLOAD_CANCEL to inform 3632 the destination to stop the active transfer and uses the stateid it 3633 got back from the COPY operation. The client uses OFFLOAD_CANCEL and 3634 the stateid it used in the COPY_NOTIFY to inform the source to not 3635 allow any more copying from the destination. 3637 OFFLOAD_CANCEL is also useful in situations in which the source 3638 server granted a very long or infinite lease on the destination 3639 server's ability to read the source file and all copy operations on 3640 the source file have been completed. 3642 15.9. Operation 67: OFFLOAD_STATUS - Poll for Status of Asynchronous 3643 Operation 3645 15.9.1. ARGUMENT 3647 3649 struct OFFLOAD_STATUS4args { 3650 /* CURRENT_FH: destination file */ 3651 stateid4 osa_stateid; 3652 }; 3654 3656 15.9.2. RESULT 3658 3660 struct OFFLOAD_STATUS4resok { 3661 length4 osr_count; 3662 nfsstat4 osr_complete<1>; 3663 }; 3665 union OFFLOAD_STATUS4res switch (nfsstat4 osr_status) { 3666 case NFS4_OK: 3667 OFFLOAD_STATUS4resok osr_resok4; 3668 default: 3669 void; 3670 }; 3672 3674 15.9.3. DESCRIPTION 3676 OFFLOAD_STATUS can be used by the client to query the progress of an 3677 asynchronous operation, which is identified both by CURRENT_FH and 3678 the osa_stateid. If this operation is successful, the number of 3679 bytes processed are returned to the client in the osr_count field. 3681 If the optional osr_complete field is present, the asynchronous 3682 operation has completed. In this case the status value indicates the 3683 result of the asynchronous operation. In all cases, the server will 3684 also deliver the final results of the asynchronous operation in a 3685 CB_OFFLOAD operation. 3687 The failure of this operation does not indicate the result of the 3688 asynchronous operation in any way. 3690 15.10. Operation 68: READ_PLUS - READ Data or Holes from a File 3692 15.10.1. ARGUMENT 3694 3696 struct READ_PLUS4args { 3697 /* CURRENT_FH: file */ 3698 stateid4 rpa_stateid; 3699 offset4 rpa_offset; 3700 count4 rpa_count; 3701 }; 3703 3705 15.10.2. RESULT 3707 3709 enum data_content4 { 3710 NFS4_CONTENT_DATA = 0, 3711 NFS4_CONTENT_HOLE = 1 3712 }; 3714 struct data_info4 { 3715 offset4 di_offset; 3716 length4 di_length; 3717 }; 3719 struct data4 { 3720 offset4 d_offset; 3721 opaque d_data<>; 3722 }; 3723 union read_plus_content switch (data_content4 rpc_content) { 3724 case NFS4_CONTENT_DATA: 3725 data4 rpc_data; 3726 case NFS4_CONTENT_HOLE: 3727 data_info4 rpc_hole; 3728 default: 3729 void; 3730 }; 3732 /* 3733 * Allow a return of an array of contents. 3734 */ 3735 struct read_plus_res4 { 3736 bool rpr_eof; 3737 read_plus_content rpr_contents<>; 3738 }; 3740 union READ_PLUS4res switch (nfsstat4 rp_status) { 3741 case NFS4_OK: 3742 read_plus_res4 rp_resok4; 3743 default: 3744 void; 3745 }; 3747 3749 15.10.3. DESCRIPTION 3751 The READ_PLUS operation is based upon the NFSv4.1 READ operation (see 3752 Section 18.22 of [RFC5661]) and similarly reads data from the regular 3753 file identified by the current filehandle. 3755 The client provides a rpa_offset of where the READ_PLUS is to start 3756 and a rpa_count of how many bytes are to be read. A rpa_offset of 3757 zero means to read data starting at the beginning of the file. If 3758 rpa_offset is greater than or equal to the size of the file, the 3759 status NFS4_OK is returned with di_length (the data length) set to 3760 zero and eof set to TRUE. 3762 The READ_PLUS result is comprised of an array of rpr_contents, each 3763 of which describe a data_content4 type of data. For NFSv4.2, the 3764 allowed values are data and hole. A server MUST support both the 3765 data type and the hole if it uses READ_PLUS. If it does not want to 3766 support a hole, it MUST use READ. The array contents MUST be 3767 contiguous in the file. 3769 Holes SHOULD be returned in their entirety - clients must be prepared 3770 to get more information than they requested. Both the start and the 3771 end of the hole may exceed what was requested. If data to be 3772 returned is comprised entirely of zeros, then the server SHOULD 3773 return that data as a hole instead. 3775 The server may elect to return adjacent elements of the same type. 3776 For example, if the server has a range of data comprised entirely of 3777 zeros and then a hole, it might want to return two adjacent holes to 3778 the client. 3780 If the client specifies a rpa_count value of zero, the READ_PLUS 3781 succeeds and returns zero bytes of data. In all situations, the 3782 server may choose to return fewer bytes than specified by the client. 3783 The client needs to check for this condition and handle the condition 3784 appropriately. 3786 If the client specifies an rpa_offset and rpa_count value that is 3787 entirely contained within a hole of the file, then the di_offset and 3788 di_length returned MAY be for the entire hole. If the the owner has 3789 a locked byte range covering rpa_offset and rpa_count entirely the 3790 di_offset and di_length MUST NOT be extended outside the locked byte 3791 range. This result is considered valid until the file is changed 3792 (detected via the change attribute). The server MUST provide the 3793 same semantics for the hole as if the client read the region and 3794 received zeroes; the implied holes contents lifetime MUST be exactly 3795 the same as any other read data. 3797 If the client specifies an rpa_offset and rpa_count value that begins 3798 in a non-hole of the file but extends into hole the server should 3799 return an array comprised of both data and a hole. The client MUST 3800 be prepared for the server to return a short read describing just the 3801 data. The client will then issue another READ_PLUS for the remaining 3802 bytes, which the server will respond with information about the hole 3803 in the file. 3805 Except when special stateids are used, the stateid value for a 3806 READ_PLUS request represents a value returned from a previous byte- 3807 range lock or share reservation request or the stateid associated 3808 with a delegation. The stateid identifies the associated owners if 3809 any and is used by the server to verify that the associated locks are 3810 still valid (e.g., have not been revoked). 3812 If the read ended at the end-of-file (formally, in a correctly formed 3813 READ_PLUS operation, if rpa_offset + rpa_count is equal to the size 3814 of the file), or the READ_PLUS operation extends beyond the size of 3815 the file (if rpa_offset + rpa_count is greater than the size of the 3816 file), eof is returned as TRUE; otherwise, it is FALSE. A successful 3817 READ_PLUS of an empty file will always return eof as TRUE. 3819 If the current filehandle is not an ordinary file, an error will be 3820 returned to the client. In the case that the current filehandle 3821 represents an object of type NF4DIR, NFS4ERR_ISDIR is returned. If 3822 the current filehandle designates a symbolic link, NFS4ERR_SYMLINK is 3823 returned. In all other cases, NFS4ERR_WRONG_TYPE is returned. 3825 For a READ_PLUS with a stateid value of all bits equal to zero, the 3826 server MAY allow the READ_PLUS to be serviced subject to mandatory 3827 byte-range locks or the current share deny modes for the file. For a 3828 READ_PLUS with a stateid value of all bits equal to one, the server 3829 MAY allow READ_PLUS operations to bypass locking checks at the 3830 server. 3832 On success, the current filehandle retains its value. 3834 15.10.3.1. Note on Client Support of Arms of the Union 3836 It was decided not to add a means for the client to inform the server 3837 as to which arms of READ_PLUS it would support. In a later minor 3838 version, it may become necessary for the introduction of a new 3839 operation which would allow the client to inform the server as to 3840 whether it supported the new arms of the union of data types 3841 available in READ_PLUS. 3843 15.10.4. IMPLEMENTATION 3845 In general, the IMPLEMENTATION notes for READ in Section 18.22.4 of 3846 [RFC5661] also apply to READ_PLUS. 3848 15.10.4.1. Additional pNFS Implementation Information 3850 With pNFS, the semantics of using READ_PLUS remains the same. Any 3851 data server MAY return a hole result for a READ_PLUS request that it 3852 receives. When a data server chooses to return such a result, it has 3853 the option of returning information for the data stored on that data 3854 server (as defined by the data layout), but it MUST NOT return 3855 results for a byte range that includes data managed by another data 3856 server. 3858 If mandatory locking is enforced, then the data server must also 3859 ensure that to return only information that is within the owner's 3860 locked byte range. 3862 15.10.5. READ_PLUS with Sparse Files Example 3864 The following table describes a sparse file. For each byte range, 3865 the file contains either non-zero data or a hole. In addition, the 3866 server in this example will only create a hole if it is greater than 3867 32K. 3869 +-------------+----------+ 3870 | Byte-Range | Contents | 3871 +-------------+----------+ 3872 | 0-15999 | Hole | 3873 | 16K-31999 | Non-Zero | 3874 | 32K-255999 | Hole | 3875 | 256K-287999 | Non-Zero | 3876 | 288K-353999 | Hole | 3877 | 354K-417999 | Non-Zero | 3878 +-------------+----------+ 3880 Table 5 3882 Under the given circumstances, if a client was to read from the file 3883 with a max read size of 64K, the following will be the results for 3884 the given READ_PLUS calls. This assumes the client has already 3885 opened the file, acquired a valid stateid ('s' in the example), and 3886 just needs to issue READ_PLUS requests. 3888 1. READ_PLUS(s, 0, 64K) --> NFS_OK, eof = false, . Since the first hole is less than the server's 3890 minimum hole size, the first 32K of the file is returned as data 3891 and the remaining 32K is returned as a hole which actually 3892 extends to 256K. 3894 2. READ_PLUS(s, 32K, 64K) --> NFS_OK, eof = false, 3895 The requested range was all zeros, and the current hole begins at 3896 offset 32K and is 224K in length. Note that the client should 3897 not have followed up the previous READ_PLUS request with this one 3898 as the hole information from the previous call extended past what 3899 the client was requesting. 3901 3. READ_PLUS(s, 256K, 64K) --> NFS_OK, eof = false, . Returns an array of the 32K data and 3903 the hole which extends to 354K. 3905 4. READ_PLUS(s, 354K, 64K) --> NFS_OK, eof = true, . Returns the final 64K of data and informs the client 3907 there is no more data in the file. 3909 15.11. Operation 69: SEEK - Find the Next Data or Hole 3911 15.11.1. ARGUMENT 3913 3915 enum data_content4 { 3916 NFS4_CONTENT_DATA = 0, 3917 NFS4_CONTENT_HOLE = 1 3918 }; 3920 struct SEEK4args { 3921 /* CURRENT_FH: file */ 3922 stateid4 sa_stateid; 3923 offset4 sa_offset; 3924 data_content4 sa_what; 3925 }; 3927 3929 15.11.2. RESULT 3931 3933 struct seek_res4 { 3934 bool sr_eof; 3935 offset4 sr_offset; 3936 }; 3938 union SEEK4res switch (nfsstat4 sa_status) { 3939 case NFS4_OK: 3940 seek_res4 resok4; 3941 default: 3942 void; 3943 }; 3945 3947 15.11.3. DESCRIPTION 3949 SEEK is an operation that allows a client to determine the location 3950 of the next data_content4 in a file. It allows an implementation of 3951 the emerging extension to lseek(2) to allow clients to determine the 3952 next hole whilst in data or the next data whilst in a hole. 3954 From the given sa_offset, find the next data_content4 of type sa_what 3955 in the file. If the server can not find a corresponding sa_what, 3956 then the status will still be NFS4_OK, but sr_eof would be TRUE. If 3957 the server can find the sa_what, then the sr_offset is the start of 3958 that content. If the sa_offset is beyond the end of the file, then 3959 SEEK MUST return NFS4ERR_NXIO. 3961 All files MUST have a virtual hole at the end of the file. I.e., if 3962 a filesystem does not support sparse files, then a compound with 3963 {SEEK 0 NFS4_CONTENT_HOLE;} would return a result of {SEEK 1 X;} 3964 where 'X' was the size of the file. 3966 SEEK must follow the same rules for stateids as READ_PLUS 3967 (Section 15.10.3). 3969 15.12. Operation 70: WRITE_SAME - WRITE an ADB Multiple Times to a File 3971 15.12.1. ARGUMENT 3973 3975 enum stable_how4 { 3976 UNSTABLE4 = 0, 3977 DATA_SYNC4 = 1, 3978 FILE_SYNC4 = 2 3979 }; 3981 struct app_data_block4 { 3982 offset4 adb_offset; 3983 length4 adb_block_size; 3984 length4 adb_block_count; 3985 length4 adb_reloff_blocknum; 3986 count4 adb_block_num; 3987 length4 adb_reloff_pattern; 3988 opaque adb_pattern<>; 3989 }; 3991 struct WRITE_SAME4args { 3992 /* CURRENT_FH: file */ 3993 stateid4 wsa_stateid; 3994 stable_how4 wsa_stable; 3995 app_data_block4 wsa_adb; 3996 }; 3998 4000 15.12.2. RESULT 4002 4004 struct write_response4 { 4005 stateid4 wr_callback_id<1>; 4006 length4 wr_count; 4007 stable_how4 wr_committed; 4008 verifier4 wr_writeverf; 4009 }; 4011 union WRITE_SAME4res switch (nfsstat4 wsr_status) { 4012 case NFS4_OK: 4013 write_response4 resok4; 4014 default: 4015 void; 4016 }; 4018 4020 15.12.3. DESCRIPTION 4022 The WRITE_SAME operation writes an application data block to the 4023 regular file identified by the current filehandle (see WRITE SAME 4024 (10) in [T10-SBC2]). The target file is specified by the current 4025 filehandle. The data to be written is specified by an 4026 app_data_block4 structure (Section 8.1.1). The client specifies with 4027 the wsa_stable parameter the method of how the data is to be 4028 processed by the server. It is treated like the stable parameter in 4029 the NFSv4.1 WRITE operation (see Section 18.2 of [RFC5661]). 4031 A successful WRITE_SAME will construct a reply for wr_count, 4032 wr_committed, and wr_writeverf as per the NFSv4.1 WRITE operation 4033 results. If wr_callback_id is set, it indicates an asynchronous 4034 reply (see Section 15.12.3.1). 4036 WRITE_SAME has to support all of the errors which are returned by 4037 WRITE plus NFS4ERR_NOTSUPP, i.e., it is an OPTIONAL operation. If 4038 the client supports WRITE_SAME, it MUST support CB_OFFLOAD. 4040 If the server supports ADBs, then it MUST support the WRITE_SAME 4041 operation. The server has no concept of the structure imposed by the 4042 application. It is only when the application writes to a section of 4043 the file does order get imposed. In order to detect corruption even 4044 before the application utilizes the file, the application will want 4045 to initialize a range of ADBs using WRITE_SAME. 4047 When the client invokes the WRITE_SAME operation, it wants to record 4048 the block structure described by the app_data_block4 on to the file. 4050 When the server receives the WRITE_SAME operation, it MUST populate 4051 adb_block_count ADBs in the file starting at adb_offset. The block 4052 size will be given by adb_block_size. The ADBN (if provided) will 4053 start at adb_reloff_blocknum and each block will be monotonically 4054 numbered starting from adb_block_num in the first block. The pattern 4055 (if provided) will be at adb_reloff_pattern of each block and will be 4056 provided in adb_pattern. 4058 The server SHOULD return an asynchronous result if it can determine 4059 the operation will be long running (see Section 15.12.3.1). Once 4060 either the WRITE_SAME finishes synchronously or the server uses 4061 CB_OFFLOAD to inform the client of the asynchronous completion of the 4062 WRITE_SAME, the server MUST return the ADBs to clients as data. 4064 15.12.3.1. Asynchronous Transactions 4066 ADB initialization may lead to server determining to service the 4067 operation asynchronously. If it decides to do so, it sets the 4068 stateid in wr_callback_id to be that of the wsa_stateid. If it does 4069 not set the wr_callback_id, then the result is synchronous. 4071 When the client determines that the reply will be given 4072 asynchronously, it should not assume anything about the contents of 4073 what it wrote until it is informed by the server that the operation 4074 is complete. It can use OFFLOAD_STATUS (Section 15.9) to monitor the 4075 operation and OFFLOAD_CANCEL (Section 15.8) to cancel the operation. 4076 An example of a asynchronous WRITE_SAME is shown in Figure 6. Note 4077 that as with the COPY operation, WRITE_SAME must provide a stateid 4078 for tracking the asynchronous operation. 4080 Client Server 4081 + + 4082 | | 4083 |--- OPEN ---------------------------->| Client opens 4084 |<------------------------------------/| the file 4085 | | 4086 |--- WRITE_SAME ----------------------->| Client initializes 4087 |<------------------------------------/| an ADB 4088 | | 4089 | | 4090 |--- OFFLOAD_STATUS ------------------>| Client may poll 4091 |<------------------------------------/| for status 4092 | | 4093 | . | Multiple OFFLOAD_STATUS 4094 | . | operations may be sent. 4095 | . | 4096 | | 4097 |<-- CB_OFFLOAD -----------------------| Server reports results 4098 |\------------------------------------>| 4099 | | 4100 |--- CLOSE --------------------------->| Client closes 4101 |<------------------------------------/| the file 4102 | | 4103 | | 4105 Figure 6: An asynchronous WRITE_SAME. 4107 When CB_OFFLOAD informs the client of the successful WRITE_SAME, the 4108 write_response4 embedded in the operation will provide the necessary 4109 information that a synchronous WRITE_SAME would have provided. 4111 Regardless of whether the operation is asynchronous or synchronous, 4112 it MUST still support the COMMIT operation semantics as outlined in 4113 Section 18.3 of [RFC5661]. I.e., COMMIT works on one or more WRITE 4114 operations and the WRITE_SAME operation can appear as several WRITE 4115 operations to the server. The client can use locking operations to 4116 control the behavior on the server with respect to long running 4117 asynchronous write operations. 4119 15.12.3.2. Error Handling of a Partially Complete WRITE_SAME 4121 WRITE_SAME will clone adb_block_count copies of the given ADB in 4122 consecutive order in the file starting at adb_offset. An error can 4123 occur after writing the Nth ADB to the file. WRITE_SAME MUST appear 4124 to populate the range of the file as if the client used WRITE to 4125 transfer the instantiated ADBs. I.e., the contents of the range will 4126 be easy for the client to determine in case of a partially complete 4127 WRITE_SAME. 4129 15.13. Operation 71: CLONE - Clone a file 4131 15.13.1. ARGUMENT 4133 4135 struct CLONE4args { 4136 /* SAVED_FH: source file */ 4137 /* CURRENT_FH: destination file */ 4138 stateid4 cl_src_stateid; 4139 stateid4 cl_dst_stateid; 4140 offset4 cl_src_offset; 4141 offset4 cl_dst_offset; 4142 length4 cl_count; 4143 }; 4145 4147 15.13.2. RESULT 4149 4151 struct CLONE4res { 4152 nfsstat4 cl_status; 4153 }; 4155 4157 15.13.3. DESCRIPTION 4159 The CLONE operation is used to clone file content from a source file 4160 specified by the SAVED_FH value into a destination file specified by 4161 CURRENT_FH without actually copying the data, e.g., by using a copy 4162 on write mechanism. 4164 Both SAVED_FH and CURRENT_FH must be regular files. If either 4165 SAVED_FH or CURRENT_FH are not regular files, the operation MUST fail 4166 and return NFS4ERR_WRONG_TYPE. 4168 SAVED_FH and CURRENT_FH must be different files. If SAVED_FH and 4169 CURRENT_FH refer to the same file, the operation will fail with 4170 NFS4ERR_INVAL. 4172 In order to set SAVED_FH to the source file handle, the compound 4173 procedure requesting the COPY will include a sub-sequence of 4174 operations such as 4175 PUTFH source-fh 4176 SAVEFH 4178 The ca_dst_stateid MUST refer to a stateid that is valid for a WRITE 4179 operation and follows the rules for stateids in Sections 8.2.5 and 4180 18.32.3 of [RFC5661]. The ca_src_stateid MUST refer to a stateid 4181 that is valid for a READ operations and follows the rules for 4182 stateids in Sections 8.2.5 and 18.22.3 of [RFC5661]. If either 4183 stateid is invalid, then the operation MUST fail. 4185 The cl_src_offset is the offset within the source file from which the 4186 data will be read, the cl_dst_offset is the offset within the 4187 destination file to which the data will be written, and the cl_count 4188 is the number of bytes that will be cloned. An offset of 0 (zero) 4189 specifies the start of the file. A count of 0 (zero) requests that 4190 all bytes from ca_src_offset through EOF be cloned to the 4191 destination. Both cl_src_offset and cl_dst_offset must be aligned to 4192 the clone block size Section 12.2.1. cl_count must be aligned to the 4193 clone block size, unless cl_src_offset + cl_count are equal to the 4194 source file size. 4196 If the source offset or the source offset plus count is greater than 4197 the size of the source file, the operation will fail with 4198 NFS4ERR_INVAL. The destination offset or destination offset plus 4199 count may be greater than the size of the destination file. 4201 If the COPY wrote past the end of the file on the destination, then 4202 the last byte written to will determine the new file size. The 4203 contents of any block not written to and past the original size of 4204 the file will be as if a normal WRITE extended the file. 4206 The CLONE operation is atomic, that is either all changes or no 4207 changes are seen by the client or other clients. 4209 The completion status of the operation is indicated by cr_status. 4211 16. NFSv4.2 Callback Operations 4213 16.1. Operation 15: CB_OFFLOAD - Report results of an asynchronous 4214 operation 4216 16.1.1. ARGUMENT 4218 4219 struct write_response4 { 4220 stateid4 wr_callback_id<1>; 4221 length4 wr_count; 4222 stable_how4 wr_committed; 4223 verifier4 wr_writeverf; 4224 }; 4226 union offload_info4 switch (nfsstat4 coa_status) { 4227 case NFS4_OK: 4228 write_response4 coa_resok4; 4229 default: 4230 length4 coa_bytes_copied; 4231 }; 4233 struct CB_OFFLOAD4args { 4234 nfs_fh4 coa_fh; 4235 stateid4 coa_stateid; 4236 offload_info4 coa_offload_info; 4237 }; 4239 4241 16.1.2. RESULT 4243 4245 struct CB_OFFLOAD4res { 4246 nfsstat4 cor_status; 4247 }; 4249 4251 16.1.3. DESCRIPTION 4253 CB_OFFLOAD is used to report to the client the results of an 4254 asynchronous operation, e.g., Server Side Copy or WRITE_SAME. The 4255 coa_fh and coa_stateid identify the transaction and the coa_status 4256 indicates success or failure. The coa_resok4.wr_callback_id MUST NOT 4257 be set. If the transaction failed, then the coa_bytes_copied 4258 contains the number of bytes copied before the failure occurred. The 4259 coa_bytes_copied value indicates the number of bytes copied but not 4260 which specific bytes have been copied. 4262 If the client supports any of the following operations: 4264 COPY: for both intra-server and inter-server asynchronous copies 4266 WRITE_SAME: for ADB initialization 4267 then the client is REQUIRED to support the CB_OFFLOAD operation. 4269 There is a potential race between the reply to the original 4270 transaction on the forechannel and the CB_OFFLOAD callback on the 4271 backchannel. Sections 2.10.6.3 and 20.9.3 of [RFC5661] describe how 4272 to handle this type of issue. 4274 Upon success, the coa_resok4.wr_count presents for each operation: 4276 COPY: the total number of bytes copied 4278 WRITE_SAME: the same information that a synchronous WRITE_SAME would 4279 provide 4281 17. Security Considerations 4283 NFSv4.2 has all of the security concerns present in NFSv4.1 (see 4284 Section 21 of [RFC5661]) and those present in the Server Side Copy 4285 (see Section 4.10) and in Labeled NFS (see Section 9.7). 4287 18. IANA Considerations 4289 The IANA Considerations for Labeled NFS are addressed in [Quigley14]. 4291 19. References 4293 19.1. Normative References 4295 [NFSv42xdr] 4296 Haynes, T., "Network File System (NFS) Version 4 Minor 4297 Version 2 External Data Representation Standard (XDR) 4298 Description", December 2014. 4300 [RFC3986] Berners-Lee, T., Fielding, R., and L. Masinter, "Uniform 4301 Resource Identifier (URI): Generic Syntax", STD 66, RFC 4302 3986, January 2005. 4304 [RFC5661] Shepler, S., Eisler, M., and D. Noveck, "Network File 4305 System (NFS) Version 4 Minor Version 1 Protocol", RFC 4306 5661, January 2010. 4308 [RFC5662] Shepler, S., Eisler, M., and D. Noveck, "Network File 4309 System (NFS) Version 4 Minor Version 1 External Data 4310 Representation Standard (XDR) Description", RFC 5662, 4311 January 2010. 4313 [posix_fadvise] 4314 The Open Group, "Section 'posix_fadvise()' of System 4315 Interfaces of The Open Group Base Specifications Issue 6, 4316 IEEE Std 1003.1, 2004 Edition", 2004. 4318 [posix_fallocate] 4319 The Open Group, "Section 'posix_fallocate()' of System 4320 Interfaces of The Open Group Base Specifications Issue 6, 4321 IEEE Std 1003.1, 2004 Edition", 2004. 4323 [rpcsec_gssv3] 4324 Adamson, W. and N. Williams, "Remote Procedure Call (RPC) 4325 Security Version 3", December 2014. 4327 19.2. Informative References 4329 [Ashdown08] 4330 Ashdown, L., "Chapter 15, Validating Database Files and 4331 Backups, of Oracle Database Backup and Recovery User's 4332 Guide 11g Release 1 (11.1)", August 2008. 4334 [BL73] Bell, D. and L. LaPadula, "Secure Computer Systems: 4335 Mathematical Foundations and Model", Technical Report 4336 M74-244, The MITRE Corporation, Bedford, MA, May 1973. 4338 [Baira08] Bairavasundaram, L., Goodson, G., Schroeder, B., Arpaci- 4339 Dusseau, A., and R. Arpaci-Dusseau, "An Analysis of Data 4340 Corruption in the Storage Stack", Proceedings of the 6th 4341 USENIX Symposium on File and Storage Technologies (FAST 4342 '08) , 2008. 4344 [IESG08] ISEG, "IESG Processing of RFC Errata for the IETF Stream", 4345 2008. 4347 [McDougall07] 4348 McDougall, R. and J. Mauro, "Section 11.4.3, Detecting 4349 Memory Corruption of Solaris Internals", 2007. 4351 [NFSv4-Versioning] 4352 Haynes, T. and D. Noveck, "NFSv4 Version Management", 4353 November 2014. 4355 [Quigley14] 4356 Quigley, D., Lu, J., and T. Haynes, "Registry 4357 Specification for Mandatory Access Control (MAC) Security 4358 Label Formats", draft-ietf-nfsv4-lfs-registry-01 (work in 4359 progress), September 2014. 4361 [RFC1108] Kent, S., "Security Options for the Internet Protocol", 4362 RFC 1108, November 1991. 4364 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 4365 Requirement Levels", March 1997. 4367 [RFC2401] Kent, S. and R. Atkinson, "Security Architecture for the 4368 Internet Protocol", RFC 2401, November 1998. 4370 [RFC2616] Fielding, R., Gettys, J., Mogul, J., Frystyk, H., 4371 Masinter, L., Leach, P., and T. Berners-Lee, "Hypertext 4372 Transfer Protocol -- HTTP/1.1", RFC 2616, June 1999. 4374 [RFC4506] Eisler, M., "XDR: External Data Representation Standard", 4375 RFC 4506, May 2006. 4377 [RFC5663] Black, D., Fridella, S., and J. Glasgow, "Parallel NFS 4378 (pNFS) Block/Volume Layout", RFC 5663, January 2010. 4380 [RFC7204] Haynes, T., "Requirements for Labeled NFS", RFC 7204, 4381 April 2014. 4383 [RFC7530] Haynes, T. and D. Noveck, "Network File System (NFS) 4384 version 4 Protocol", RFC 7530, March 2015. 4386 [RFC959] Postel, J. and J. Reynolds, "File Transfer Protocol", STD 4387 9, RFC 959, October 1985. 4389 [Strohm11] 4390 Strohm, R., "Chapter 2, Data Blocks, Extents, and 4391 Segments, of Oracle Database Concepts 11g Release 1 4392 (11.1)", January 2011. 4394 [T10-SBC2] 4395 Elliott, R., Ed., "ANSI INCITS 405-2005, Information 4396 Technology - SCSI Block Commands - 2 (SBC-2)", November 4397 2004. 4399 Appendix A. Acknowledgments 4401 Tom Haynes would like to thank NetApp, Inc. for its funding of his 4402 time on this project. 4404 For the pNFS Access Permissions Check, the original draft was by 4405 Sorin Faibish, David Black, Mike Eisler, and Jason Glasgow. The work 4406 was influenced by discussions with Benny Halevy and Bruce Fields. A 4407 review was done by Tom Haynes. 4409 For the Sharing change attribute implementation characteristics with 4410 NFSv4 clients, the original draft was by Trond Myklebust. 4412 For the NFS Server Side Copy, the original draft was by James 4413 Lentini, Mike Eisler, Deepak Kenchammana, Anshul Madan, and Rahul 4414 Iyer. Tom Talpey co-authored an unpublished version of that 4415 document. It was also was reviewed by a number of individuals: 4416 Pranoop Erasani, Tom Haynes, Arthur Lent, Trond Myklebust, Dave 4417 Noveck, Theresa Lingutla-Raj, Manjunath Shankararao, Satyam Vaghani, 4418 and Nico Williams. Anna Schumaker's early prototyping experience 4419 helped us avoid some traps. Also, both Olga Kornievskaia and Andy 4420 Adamson brought implementation experience to the use of copy stateids 4421 in inter-server copy. Jorge Mora was able to optimize the handling 4422 of errors for the result of COPY. 4424 For the NFS space reservation operations, the original draft was by 4425 Mike Eisler, James Lentini, Manjunath Shankararao, and Rahul Iyer. 4427 For the sparse file support, the original draft was by Dean 4428 Hildebrand and Marc Eshel. Valuable input and advice was received 4429 from Sorin Faibish, Bruce Fields, Benny Halevy, Trond Myklebust, and 4430 Richard Scheffenegger. 4432 For the Application IO Hints, the original draft was by Dean 4433 Hildebrand, Mike Eisler, Trond Myklebust, and Sam Falkner. Some 4434 early reviewers included Benny Halevy and Pranoop Erasani. 4436 For Labeled NFS, the original draft was by David Quigley, James 4437 Morris, Jarret Lu, and Tom Haynes. Peter Staubach, Trond Myklebust, 4438 Stephen Smalley, Sorin Faibish, Nico Williams, and David Black also 4439 contributed in the final push to get this accepted. 4441 Christoph Hellwig was very helpful in getting the WRITE_SAME 4442 semantics to model more of what T10 was doing for WRITE SAME (10) 4443 [T10-SBC2]. And he led the push to get space reservations to more 4444 closely model the posix_fallocate. 4446 Andy Adamson picked up the RPCSEC_GSSv3 work, which enabled both 4447 Labeled NFS and Server Side Copy to be present more secure options. 4449 Christoph Hellwig provided the update to GETDEVICELIST. 4451 During the review process, Talia Reyes-Ortiz helped the sessions run 4452 smoothly. While many people contributed here and there, the core 4453 reviewers were Andy Adamson, Pranoop Erasani, Bruce Fields, Chuck 4454 Lever, Trond Myklebust, David Noveck, Peter Staubach, and Mike 4455 Kupfer. 4457 Appendix B. RFC Editor Notes 4459 [RFC Editor: please remove this section prior to publishing this 4460 document as an RFC] 4462 [RFC Editor: prior to publishing this document as an RFC, please 4463 replace all occurrences of NFSv42xdr with RFCxxxx where xxxx is the 4464 RFC number of the companion XDR document] 4466 Author's Address 4468 Thomas Haynes 4469 Primary Data, Inc. 4470 4300 El Camino Real Ste 100 4471 Los Altos, CA 94022 4472 USA 4474 Phone: +1 408 215 1519 4475 Email: thomas.haynes@primarydata.com