idnits 2.17.1 draft-ietf-nfsv4-minorversion2-15.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 : ---------------------------------------------------------------------------- == There are 5 instances of lines with non-RFC6890-compliant IPv4 addresses in the document. If these are example addresses, they should be changed. == There are 5 instances of lines with private range IPv4 addresses in the document. If these are generic example addresses, they should be changed to use any of the ranges defined in RFC 6890 (or successor): 192.0.2.x, 198.51.100.x or 203.0.113.x. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year == Line 2794 has weird spacing: '...S4resok res...' == Using lowercase 'not' together with uppercase 'MUST', 'SHALL', 'SHOULD', or 'RECOMMENDED' is not an accepted usage according to RFC 2119. Please use uppercase 'NOT' together with RFC 2119 keywords (if that is what you mean). Found 'MUST not' in this paragraph: o MUST not expose an object to either the client or server name space before its security information has been bound to it. == Using lowercase 'not' together with uppercase 'MUST', 'SHALL', 'SHOULD', or 'RECOMMENDED' is not an accepted usage according to RFC 2119. Please use uppercase 'NOT' together with RFC 2119 keywords (if that is what you mean). Found 'MUST not' in this paragraph: With pNFS, the semantics of using READ_PLUS remains the same. Any data server MAY return a hole or ADH result for a READ_PLUS request that it receives. When a data server chooses to return such a result, it has the option of returning information for the data stored on that data server (as defined by the data layout), but it MUST not return results for a byte range that includes data managed by another data server. == The document seems to contain a disclaimer for pre-RFC5378 work, but was first submitted on or after 10 November 2008. The disclaimer is usually necessary only for documents that revise or obsolete older RFCs, and that take significant amounts of text from those RFCs. If you can contact all authors of the source material and they are willing to grant the BCP78 rights to the IETF Trust, you can and should remove the disclaimer. Otherwise, the disclaimer is needed and you can ignore this comment. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (October 03, 2012) is 4224 days in the past. Is this intentional? -- Found something which looks like a code comment -- if you have code sections in the document, please surround them with '' and '' lines. Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Missing Reference: '0' is mentioned on line 3630, but not defined -- Looks like a reference, but probably isn't: '32K' on line 3630 == Unused Reference: '25' is defined on line 3854, but no explicit reference was found in the text -- Possible downref: Non-RFC (?) normative reference: ref. '1' ** Obsolete normative reference: RFC 5661 (ref. '2') (Obsoleted by RFC 8881) -- Possible downref: Non-RFC (?) normative reference: ref. '3' -- Possible downref: Non-RFC (?) normative reference: ref. '6' == Outdated reference: A later version (-05) exists of draft-ietf-nfsv4-labreqs-00 ** Downref: Normative reference to an Informational draft: draft-ietf-nfsv4-labreqs (ref. '7') == Outdated reference: A later version (-35) exists of draft-ietf-nfsv4-rfc3530bis-09 -- Obsolete informational reference (is this intentional?): RFC 2616 (ref. '13') (Obsoleted by RFC 7230, RFC 7231, RFC 7232, RFC 7233, RFC 7234, RFC 7235) -- Obsolete informational reference (is this intentional?): RFC 5226 (ref. '24') (Obsoleted by RFC 8126) Summary: 2 errors (**), 0 flaws (~~), 11 warnings (==), 8 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 NFSv4 T. Haynes 3 Internet-Draft Editor 4 Intended status: Standards Track October 03, 2012 5 Expires: April 6, 2013 7 NFS Version 4 Minor Version 2 8 draft-ietf-nfsv4-minorversion2-15.txt 10 Abstract 12 This Internet-Draft describes NFS version 4 minor version two, 13 focusing mainly on the protocol extensions made from NFS version 4 14 minor version 0 and NFS version 4 minor version 1. Major extensions 15 introduced in NFS version 4 minor version two include: Server-side 16 Copy, Application I/O Advise, Space Reservations, Sparse Files, 17 Application Data Blocks, and Labeled NFS. 19 Requirements Language 21 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 22 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 23 document are to be interpreted as described in RFC 2119 [1]. 25 Status of this Memo 27 This Internet-Draft is submitted in full conformance with the 28 provisions of BCP 78 and BCP 79. 30 Internet-Drafts are working documents of the Internet Engineering 31 Task Force (IETF). Note that other groups may also distribute 32 working documents as Internet-Drafts. The list of current Internet- 33 Drafts is at http://datatracker.ietf.org/drafts/current/. 35 Internet-Drafts are draft documents valid for a maximum of six months 36 and may be updated, replaced, or obsoleted by other documents at any 37 time. It is inappropriate to use Internet-Drafts as reference 38 material or to cite them other than as "work in progress." 40 This Internet-Draft will expire on April 6, 2013. 42 Copyright Notice 44 Copyright (c) 2012 IETF Trust and the persons identified as the 45 document authors. All rights reserved. 47 This document is subject to BCP 78 and the IETF Trust's Legal 48 Provisions Relating to IETF Documents 49 (http://trustee.ietf.org/license-info) in effect on the date of 50 publication of this document. Please review these documents 51 carefully, as they describe your rights and restrictions with respect 52 to this document. Code Components extracted from this document must 53 include Simplified BSD License text as described in Section 4.e of 54 the Trust Legal Provisions and are provided without warranty as 55 described in the Simplified BSD License. 57 This document may contain material from IETF Documents or IETF 58 Contributions published or made publicly available before November 59 10, 2008. The person(s) controlling the copyright in some of this 60 material may not have granted the IETF Trust the right to allow 61 modifications of such material outside the IETF Standards Process. 62 Without obtaining an adequate license from the person(s) controlling 63 the copyright in such materials, this document may not be modified 64 outside the IETF Standards Process, and derivative works of it may 65 not be created outside the IETF Standards Process, except to format 66 it for publication as an RFC or to translate it into languages other 67 than English. 69 Table of Contents 71 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 5 72 1.1. The NFS Version 4 Minor Version 2 Protocol . . . . . . . 5 73 1.2. Scope of This Document . . . . . . . . . . . . . . . . . 5 74 1.3. NFSv4.2 Goals . . . . . . . . . . . . . . . . . . . . . . 5 75 1.4. Overview of NFSv4.2 Features . . . . . . . . . . . . . . 6 76 1.4.1. Sparse Files . . . . . . . . . . . . . . . . . . . . . 6 77 1.4.2. Application I/O Advise . . . . . . . . . . . . . . . . 6 78 1.5. Differences from NFSv4.1 . . . . . . . . . . . . . . . . 6 79 2. NFS Server-side Copy . . . . . . . . . . . . . . . . . . . . . 6 80 2.1. Introduction . . . . . . . . . . . . . . . . . . . . . . 6 81 2.2. Protocol Overview . . . . . . . . . . . . . . . . . . . . 7 82 2.2.1. Overview of Copy Operations . . . . . . . . . . . . . 7 83 2.2.2. Locking the Files . . . . . . . . . . . . . . . . . . 8 84 2.2.3. Intra-Server Copy . . . . . . . . . . . . . . . . . . 8 85 2.2.4. Inter-Server Copy . . . . . . . . . . . . . . . . . . 10 86 2.2.5. Server-to-Server Copy Protocol . . . . . . . . . . . . 14 87 2.3. Requirements for Operations . . . . . . . . . . . . . . . 15 88 2.3.1. netloc4 - Network Locations . . . . . . . . . . . . . 16 89 2.3.2. Copy Offload Stateids . . . . . . . . . . . . . . . . 16 90 2.4. Security Considerations . . . . . . . . . . . . . . . . . 17 91 2.4.1. Inter-Server Copy Security . . . . . . . . . . . . . . 17 92 3. Support for Application IO Hints . . . . . . . . . . . . . . . 25 93 4. Sparse Files . . . . . . . . . . . . . . . . . . . . . . . . . 25 94 4.1. Introduction . . . . . . . . . . . . . . . . . . . . . . 25 95 4.2. Terminology . . . . . . . . . . . . . . . . . . . . . . . 26 96 5. Space Reservation . . . . . . . . . . . . . . . . . . . . . . 26 97 5.1. Introduction . . . . . . . . . . . . . . . . . . . . . . 27 98 6. Application Data Hole Support . . . . . . . . . . . . . . . . 29 99 6.1. Generic Framework . . . . . . . . . . . . . . . . . . . . 29 100 6.1.1. Data Hole Representation . . . . . . . . . . . . . . . 30 101 6.1.2. Data Content . . . . . . . . . . . . . . . . . . . . . 30 102 6.2. An Example of Detecting Corruption . . . . . . . . . . . 31 103 6.3. Example of READ_PLUS . . . . . . . . . . . . . . . . . . 32 104 7. Labeled NFS . . . . . . . . . . . . . . . . . . . . . . . . . 33 105 7.1. Introduction . . . . . . . . . . . . . . . . . . . . . . 33 106 7.2. Definitions . . . . . . . . . . . . . . . . . . . . . . . 34 107 7.3. MAC Security Attribute . . . . . . . . . . . . . . . . . 34 108 7.3.1. Delegations . . . . . . . . . . . . . . . . . . . . . 35 109 7.3.2. Permission Checking . . . . . . . . . . . . . . . . . 35 110 7.3.3. Object Creation . . . . . . . . . . . . . . . . . . . 36 111 7.3.4. Existing Objects . . . . . . . . . . . . . . . . . . . 36 112 7.3.5. Label Changes . . . . . . . . . . . . . . . . . . . . 36 113 7.4. pNFS Considerations . . . . . . . . . . . . . . . . . . . 37 114 7.5. Discovery of Server Labeled NFS Support . . . . . . . . . 37 115 7.6. MAC Security NFS Modes of Operation . . . . . . . . . . . 37 116 7.6.1. Full Mode . . . . . . . . . . . . . . . . . . . . . . 38 117 7.6.2. Guest Mode . . . . . . . . . . . . . . . . . . . . . . 39 118 7.7. Security Considerations . . . . . . . . . . . . . . . . . 39 119 8. Sharing change attribute implementation details with NFSv4 120 clients . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 121 8.1. Introduction . . . . . . . . . . . . . . . . . . . . . . 40 122 9. Security Considerations . . . . . . . . . . . . . . . . . . . 40 123 10. Error Values . . . . . . . . . . . . . . . . . . . . . . . . . 40 124 10.1. Error Definitions . . . . . . . . . . . . . . . . . . . . 41 125 10.1.1. General Errors . . . . . . . . . . . . . . . . . . . . 41 126 10.1.2. Server to Server Copy Errors . . . . . . . . . . . . . 41 127 10.1.3. Labeled NFS Errors . . . . . . . . . . . . . . . . . . 42 128 11. New File Attributes . . . . . . . . . . . . . . . . . . . . . 42 129 11.1. New RECOMMENDED Attributes - List and Definition 130 References . . . . . . . . . . . . . . . . . . . . . . . 42 131 11.2. Attribute Definitions . . . . . . . . . . . . . . . . . . 43 132 12. Operations: REQUIRED, RECOMMENDED, or OPTIONAL . . . . . . . . 46 133 13. NFSv4.2 Operations . . . . . . . . . . . . . . . . . . . . . . 50 134 13.1. Operation 59: COPY - Initiate a server-side copy . . . . 50 135 13.2. Operation 60: OFFLOAD_ABORT - Cancel a server-side 136 copy . . . . . . . . . . . . . . . . . . . . . . . . . . 58 137 13.3. Operation 61: COPY_NOTIFY - Notify a source server of 138 a future copy . . . . . . . . . . . . . . . . . . . . . . 59 139 13.4. Operation 62: OFFLOAD_REVOKE - Revoke a destination 140 server's copy privileges . . . . . . . . . . . . . . . . 60 141 13.5. Operation 63: OFFLOAD_STATUS - Poll for status of a 142 server-side copy . . . . . . . . . . . . . . . . . . . . 61 143 13.6. Modification to Operation 42: EXCHANGE_ID - 144 Instantiate Client ID . . . . . . . . . . . . . . . . . . 63 145 13.7. Operation 64: INITIALIZE . . . . . . . . . . . . . . . . 64 146 13.8. Operation 67: IO_ADVISE - Application I/O access 147 pattern hints . . . . . . . . . . . . . . . . . . . . . . 67 148 13.9. Changes to Operation 51: LAYOUTRETURN . . . . . . . . . . 73 149 13.10. Operation 65: READ_PLUS . . . . . . . . . . . . . . . . . 76 150 13.11. Operation 66: SEEK . . . . . . . . . . . . . . . . . . . 81 151 14. NFSv4.2 Callback Operations . . . . . . . . . . . . . . . . . 82 152 14.1. Operation 15: CB_COPY - Report results of a 153 server-side copy . . . . . . . . . . . . . . . . . . . . 82 154 15. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 83 155 16. References . . . . . . . . . . . . . . . . . . . . . . . . . . 83 156 16.1. Normative References . . . . . . . . . . . . . . . . . . 83 157 16.2. Informative References . . . . . . . . . . . . . . . . . 84 158 Appendix A. Acknowledgments . . . . . . . . . . . . . . . . . . . 85 159 Appendix B. RFC Editor Notes . . . . . . . . . . . . . . . . . . 86 160 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . . 86 162 1. Introduction 164 1.1. The NFS Version 4 Minor Version 2 Protocol 166 The NFS version 4 minor version 2 (NFSv4.2) protocol is the third 167 minor version of the NFS version 4 (NFSv4) protocol. The first minor 168 version, NFSv4.0, is described in [10] and the second minor version, 169 NFSv4.1, is described in [2]. It follows the guidelines for minor 170 versioning that are listed in Section 11 of [10]. 172 As a minor version, NFSv4.2 is consistent with the overall goals for 173 NFSv4, but extends the protocol so as to better meet those goals, 174 based on experiences with NFSv4.1. In addition, NFSv4.2 has adopted 175 some additional goals, which motivate some of the major extensions in 176 NFSv4.2. 178 1.2. Scope of This Document 180 This document describes the NFSv4.2 protocol. With respect to 181 NFSv4.0 and NFSv4.1, this document does not: 183 o describe the NFSv4.0 or NFSv4.1 protocols, except where needed to 184 contrast with NFSv4.2. 186 o modify the specification of the NFSv4.0 or NFSv4.1 protocols. 188 o clarify the NFSv4.0 or NFSv4.1 protocols. I.e., any 189 clarifications made here apply to NFSv4.2 and neither of the prior 190 protocols. 192 The full XDR for NFSv4.2 is presented in [3]. 194 1.3. NFSv4.2 Goals 196 The goal of the design of NFSv4.2 is to take common local file system 197 features and offer them remotely. These features might 199 o already be available on the servers, e.g., sparse files 201 o be under development as a new standard, e.g., SEEK_HOLE and 202 SEEK_DATA 204 o be used by clients with the servers via some proprietary means, 205 e.g., Labeled NFS 207 but the clients are not able to leverage them on the server within 208 the confines of the NFS protocol. 210 1.4. Overview of NFSv4.2 Features 212 [[Comment.1: This needs fleshing out! --TH]] 214 1.4.1. Sparse Files 216 Two new operations are defined to support the reading of sparse files 217 (READ_PLUS) and the punching of holes to remove backing storage 218 (INITIALIZE). 220 1.4.2. Application I/O Advise 222 We propose a new IO_ADVISE operation for NFSv4.2 that clients can use 223 to communicate expected I/O behavior to the server. By communicating 224 future I/O behavior such as whether a file will be accessed 225 sequentially or randomly, and whether a file will or will not be 226 accessed in the near future, servers can optimize future I/O requests 227 for a file by, for example, prefetching or evicting data. This 228 operation can be used to support the posix_fadvise function as well 229 as other applications such as databases and video editors. 231 1.5. Differences from NFSv4.1 233 In NFSv4.1, the only way to introduce new variants of an operation 234 was to introduce a new operation. I.e., READ becomes either READ2 or 235 READ_PLUS. With the use of discriminated unions as parameters to 236 such functions in NFSv4.2, it is possible to add a new arm in a 237 subsequent minor version. And it is also possible to move such an 238 operation from OPTIONAL/RECOMMENDED to REQUIRED. Forcing an 239 implementation to adopt each arm of a discriminated union at such a 240 time does not meet the spirit of the minor versioning rules. As 241 such, new arms of a discriminated union MUST follow the same 242 guidelines for minor versioning as operations in NFSv4.1 - i.e., they 243 may not be made REQUIRED. To support this, a new error code, 244 NFS4ERR_UNION_NOTSUPP, is introduced which allows the server to 245 communicate to the client that the operation is supported, but the 246 specific arm of the discriminated union is not. 248 2. NFS Server-side Copy 250 2.1. Introduction 252 The server-side copy feature provides a mechanism for the NFS client 253 to perform a file copy on the server without the data being 254 transmitted back and forth over the network. Without this feature, 255 an NFS client copies data from one location to another by reading the 256 data from the server over the network, and then writing the data back 257 over the network to the server. Using this server-side copy 258 operation, the client is able to instruct the server to copy the data 259 locally without the data being sent back and forth over the network 260 unnecessarily. 262 If the source object and destination object are on different file 263 servers, the file servers will communicate with one another to 264 perform the copy operation. The server-to-server protocol by which 265 this is accomplished is not defined in this document. 267 2.2. Protocol Overview 269 The server-side copy offload operations support both intra-server and 270 inter-server file copies. An intra-server copy is a copy in which 271 the source file and destination file reside on the same server. In 272 an inter-server copy, the source file and destination file are on 273 different servers. In both cases, the copy may be performed 274 synchronously or asynchronously. 276 Throughout the rest of this document, we refer to the NFS server 277 containing the source file as the "source server" and the NFS server 278 to which the file is transferred as the "destination server". In the 279 case of an intra-server copy, the source server and destination 280 server are the same server. Therefore in the context of an intra- 281 server copy, the terms source server and destination server refer to 282 the single server performing the copy. 284 The operations described below are designed to copy files. Other 285 file system objects can be copied by building on these operations or 286 using other techniques. For example if the user wishes to copy a 287 directory, the client can synthesize a directory copy by first 288 creating the destination directory and then copying the source 289 directory's files to the new destination directory. If the user 290 wishes to copy a namespace junction [11] [12], the client can use the 291 ONC RPC Federated Filesystem protocol [12] to perform the copy. 292 Specifically the client can determine the source junction's 293 attributes using the FEDFS_LOOKUP_FSN procedure and create a 294 duplicate junction using the FEDFS_CREATE_JUNCTION procedure. 296 For the inter-server copy, the operations are defined to be 297 compatible with the traditional copy authentication approach. The 298 client and user are authorized at the source for reading. Then they 299 are authorized at the destination for writing. 301 2.2.1. Overview of Copy Operations 302 COPY_NOTIFY: For inter-server copies, the client sends this 303 operation to the source server to notify it of a future file copy 304 from a given destination server for the given user. 305 (Section 13.3) 307 OFFLOAD_REVOKE: Also for inter-server copies, the client sends this 308 operation to the source server to revoke permission to copy a file 309 for the given user. (Section 13.4) 311 COPY: Used by the client to request a file copy. (Section 13.1) 313 OFFLOAD_ABORT: Used by the client to abort an asynchronous file 314 copy. (Section 13.2) 316 OFFLOAD_STATUS: Used by the client to poll the status of an 317 asynchronous file copy. (Section 13.5) 319 CB_COPY: Used by the destination server to report the results of an 320 asynchronous file copy to the client. (Section 14.1) 322 2.2.2. Locking the Files 324 Both the source and destination file may need to be locked to protect 325 the content during the copy operations. A client can achieve this by 326 a combination of OPEN and LOCK operations. I.e., either share or 327 byte range locks might be desired. 329 2.2.3. Intra-Server Copy 331 To copy a file on a single server, the client uses a COPY operation. 332 The server may respond to the copy operation with the final results 333 of the copy or it may perform the copy asynchronously and deliver the 334 results using a CB_COPY operation callback. If the copy is performed 335 asynchronously, the client may poll the status of the copy using 336 OFFLOAD_STATUS or cancel the copy using OFFLOAD_ABORT. 338 A synchronous intra-server copy is shown in Figure 1. In this 339 example, the NFS server chooses to perform the copy synchronously. 340 The copy operation is completed, either successfully or 341 unsuccessfully, before the server replies to the client's request. 342 The server's reply contains the final result of the operation. 344 Client Server 345 + + 346 | | 347 |--- OPEN ---------------------------->| Client opens 348 |<------------------------------------/| the source file 349 | | 350 |--- OPEN ---------------------------->| Client opens 351 |<------------------------------------/| the destination file 352 | | 353 |--- COPY ---------------------------->| Client requests 354 |<------------------------------------/| a file copy 355 | | 356 |--- CLOSE --------------------------->| Client closes 357 |<------------------------------------/| the destination file 358 | | 359 |--- CLOSE --------------------------->| Client closes 360 |<------------------------------------/| the source file 361 | | 362 | | 364 Figure 1: A synchronous intra-server copy. 366 An asynchronous intra-server copy is shown in Figure 2. In this 367 example, the NFS server performs the copy asynchronously. The 368 server's reply to the copy request indicates that the copy operation 369 was initiated and the final result will be delivered at a later time. 370 The server's reply also contains a copy stateid. The client may use 371 this copy stateid to poll for status information (as shown) or to 372 cancel the copy using a OFFLOAD_ABORT. When the server completes the 373 copy, the server performs a callback to the client and reports the 374 results. 376 Client Server 377 + + 378 | | 379 |--- OPEN ---------------------------->| Client opens 380 |<------------------------------------/| the source file 381 | | 382 |--- OPEN ---------------------------->| Client opens 383 |<------------------------------------/| the destination file 384 | | 385 |--- COPY ---------------------------->| Client requests 386 |<------------------------------------/| a file copy 387 | | 388 | | 389 |--- OFFLOAD_STATUS ------------------>| Client may poll 390 |<------------------------------------/| for status 391 | | 392 | . | Multiple OFFLOAD_STATUS 393 | . | operations may be sent. 394 | . | 395 | | 396 |<-- CB_COPY --------------------------| Server reports results 397 |\------------------------------------>| 398 | | 399 |--- CLOSE --------------------------->| Client closes 400 |<------------------------------------/| the destination file 401 | | 402 |--- CLOSE --------------------------->| Client closes 403 |<------------------------------------/| the source file 404 | | 405 | | 407 Figure 2: An asynchronous intra-server copy. 409 2.2.4. Inter-Server Copy 411 A copy may also be performed between two servers. The copy protocol 412 is designed to accommodate a variety of network topologies. As shown 413 in Figure 3, the client and servers may be connected by multiple 414 networks. In particular, the servers may be connected by a 415 specialized, high speed network (network 192.168.33.0/24 in the 416 diagram) that does not include the client. The protocol allows the 417 client to setup the copy between the servers (over network 418 10.11.78.0/24 in the diagram) and for the servers to communicate on 419 the high speed network if they choose to do so. 421 192.168.33.0/24 422 +-------------------------------------+ 423 | | 424 | | 425 | 192.168.33.18 | 192.168.33.56 426 +-------+------+ +------+------+ 427 | Source | | Destination | 428 +-------+------+ +------+------+ 429 | 10.11.78.18 | 10.11.78.56 430 | | 431 | | 432 | 10.11.78.0/24 | 433 +------------------+------------------+ 434 | 435 | 436 | 10.11.78.243 437 +-----+-----+ 438 | Client | 439 +-----------+ 441 Figure 3: An example inter-server network topology. 443 For an inter-server copy, the client notifies the source server that 444 a file will be copied by the destination server using a COPY_NOTIFY 445 operation. The client then initiates the copy by sending the COPY 446 operation to the destination server. The destination server may 447 perform the copy synchronously or asynchronously. 449 A synchronous inter-server copy is shown in Figure 4. In this case, 450 the destination server chooses to perform the copy before responding 451 to the client's COPY request. 453 An asynchronous copy is shown in Figure 5. In this case, the 454 destination server chooses to respond to the client's COPY request 455 immediately and then perform the copy asynchronously. 457 Client Source Destination 458 + + + 459 | | | 460 |--- OPEN --->| | Returns os1 461 |<------------------/| | 462 | | | 463 |--- COPY_NOTIFY --->| | 464 |<------------------/| | 465 | | | 466 |--- OPEN ---------------------------->| Returns os2 467 |<------------------------------------/| 468 | | | 469 |--- COPY ---------------------------->| 470 | | | 471 | | | 472 | |<----- read -----| 473 | |\--------------->| 474 | | | 475 | | . | Multiple reads may 476 | | . | be necessary 477 | | . | 478 | | | 479 | | | 480 |<------------------------------------/| Destination replies 481 | | | to COPY 482 | | | 483 |--- CLOSE --------------------------->| Release open state 484 |<------------------------------------/| 485 | | | 486 |--- CLOSE --->| | Release open state 487 |<------------------/| | 489 Figure 4: A synchronous inter-server copy. 491 Client Source Destination 492 + + + 493 | | | 494 |--- OPEN --->| | Returns os1 495 |<------------------/| | 496 | | | 497 |--- LOCK --->| | Optional, could be done 498 |<------------------/| | with a share lock 499 | | | 500 |--- COPY_NOTIFY --->| | Need to pass in 501 |<------------------/| | os1 or lock state 502 | | | 503 | | | 504 | | | 505 |--- OPEN ---------------------------->| Returns os2 506 |<------------------------------------/| 507 | | | 508 |--- LOCK ---------------------------->| Optional ... 509 |<------------------------------------/| 510 | | | 511 |--- COPY ---------------------------->| Need to pass in 512 |<------------------------------------/| os2 or lock state 513 | | | 514 | | | 515 | |<----- read -----| 516 | |\--------------->| 517 | | | 518 | | . | Multiple reads may 519 | | . | be necessary 520 | | . | 521 | | | 522 | | | 523 |--- OFFLOAD_STATUS ------------------>| Client may poll 524 |<------------------------------------/| for status 525 | | | 526 | | . | Multiple OFFLOAD_STATUS 527 | | . | operations may be sent 528 | | . | 529 | | | 530 | | | 531 | | | 532 |<-- CB_COPY --------------------------| Destination reports 533 |\------------------------------------>| results 534 | | | 535 |--- LOCKU --------------------------->| Only if LOCK was done 536 |<------------------------------------/| 537 | | | 538 |--- CLOSE --------------------------->| Release open state 539 |<------------------------------------/| 540 | | | 541 |--- LOCKU --->| | Only if LOCK was done 542 |<------------------/| | 543 | | | 544 |--- CLOSE --->| | Release open state 545 |<------------------/| | 546 | | | 548 Figure 5: An asynchronous inter-server copy. 550 2.2.5. Server-to-Server Copy Protocol 552 The source server and destination server are not required to use a 553 specific protocol to transfer the file data. The choice of what 554 protocol to use is ultimately the destination server's decision. 556 2.2.5.1. Using NFSv4.x as a Server-to-Server Copy Protocol 558 The destination server MAY use standard NFSv4.x (where x >= 1) to 559 read the data from the source server. If NFSv4.x is used for the 560 server-to-server copy protocol, the destination server can use the 561 filehandle contained in the COPY request with standard NFSv4.x 562 operations to read data from the source server. Specifically, the 563 destination server may use the NFSv4.x OPEN operation's CLAIM_FH 564 facility to open the file being copied and obtain an open stateid. 565 Using the stateid, the destination server may then use NFSv4.x READ 566 operations to read the file. 568 2.2.5.2. Using an alternative Server-to-Server Copy Protocol 570 In a homogeneous environment, the source and destination servers 571 might be able to perform the file copy extremely efficiently using 572 specialized protocols. For example the source and destination 573 servers might be two nodes sharing a common file system format for 574 the source and destination file systems. Thus the source and 575 destination are in an ideal position to efficiently render the image 576 of the source file to the destination file by replicating the file 577 system formats at the block level. Another possibility is that the 578 source and destination might be two nodes sharing a common storage 579 area network, and thus there is no need to copy any data at all, and 580 instead ownership of the file and its contents might simply be re- 581 assigned to the destination. To allow for these possibilities, the 582 destination server is allowed to use a server-to-server copy protocol 583 of its choice. 585 In a heterogeneous environment, using a protocol other than NFSv4.x 586 (e.g., HTTP [13] or FTP [14]) presents some challenges. In 587 particular, the destination server is presented with the challenge of 588 accessing the source file given only an NFSv4.x filehandle. 590 One option for protocols that identify source files with path names 591 is to use an ASCII hexadecimal representation of the source 592 filehandle as the file name. 594 Another option for the source server is to use URLs to direct the 595 destination server to a specialized service. For example, the 596 response to COPY_NOTIFY could include the URL 597 ftp://s1.example.com:9999/_FH/0x12345, where 0x12345 is the ASCII 598 hexadecimal representation of the source filehandle. When the 599 destination server receives the source server's URL, it would use 600 "_FH/0x12345" as the file name to pass to the FTP server listening on 601 port 9999 of s1.example.com. On port 9999 there would be a special 602 instance of the FTP service that understands how to convert NFS 603 filehandles to an open file descriptor (in many operating systems, 604 this would require a new system call, one which is the inverse of the 605 makefh() function that the pre-NFSv4 MOUNT service needs). 607 Authenticating and identifying the destination server to the source 608 server is also a challenge. Recommendations for how to accomplish 609 this are given in Section 2.4.1.2.4 and Section 2.4.1.4. 611 2.3. Requirements for Operations 613 The implementation of server-side copy is OPTIONAL by the client and 614 the server. However, in order to successfully copy a file, some 615 operations MUST be supported by the client and/or server. 617 If a client desires an intra-server file copy, then it MUST support 618 the COPY and CB_COPY operations. If COPY returns a stateid, then the 619 client MAY use the OFFLOAD_ABORT and OFFLOAD_STATUS operations. 621 If a client desires an inter-server file copy, then it MUST support 622 the COPY, COPY_NOTICE, and CB_COPY operations, and MAY use the 623 OFFLOAD_REVOKE operation. If COPY returns a stateid, then the client 624 MAY use the OFFLOAD_ABORT and OFFLOAD_STATUS operations. 626 If a server supports intra-server copy, then the server MUST support 627 the COPY operation. If a server's COPY operation returns a stateid, 628 then the server MUST also support these operations: CB_COPY, 629 OFFLOAD_ABORT, and OFFLOAD_STATUS. 631 If a source server supports inter-server copy, then the source server 632 MUST support all these operations: COPY_NOTIFY and OFFLOAD_REVOKE. 633 If a destination server supports inter-server copy, then the 634 destination server MUST support the COPY operation. If a destination 635 server's COPY operation returns a stateid, then the destination 636 server MUST also support these operations: CB_COPY, OFFLOAD_ABORT, 637 COPY_NOTIFY, OFFLOAD_REVOKE, and OFFLOAD_STATUS. 639 Each operation is performed in the context of the user identified by 640 the ONC RPC credential of its containing COMPOUND or CB_COMPOUND 641 request. For example, a OFFLOAD_ABORT operation issued by a given 642 user indicates that a specified COPY operation initiated by the same 643 user be canceled. Therefore a OFFLOAD_ABORT MUST NOT interfere with 644 a copy of the same file initiated by another user. 646 An NFS server MAY allow an administrative user to monitor or cancel 647 copy operations using an implementation specific interface. 649 2.3.1. netloc4 - Network Locations 651 The server-side copy operations specify network locations using the 652 netloc4 data type shown below: 654 enum netloc_type4 { 655 NL4_NAME = 0, 656 NL4_URL = 1, 657 NL4_NETADDR = 2 658 }; 659 union netloc4 switch (netloc_type4 nl_type) { 660 case NL4_NAME: utf8str_cis nl_name; 661 case NL4_URL: utf8str_cis nl_url; 662 case NL4_NETADDR: netaddr4 nl_addr; 663 }; 665 If the netloc4 is of type NL4_NAME, the nl_name field MUST be 666 specified as a UTF-8 string. The nl_name is expected to be resolved 667 to a network address via DNS, LDAP, NIS, /etc/hosts, or some other 668 means. If the netloc4 is of type NL4_URL, a server URL [4] 669 appropriate for the server-to-server copy operation is specified as a 670 UTF-8 string. If the netloc4 is of type NL4_NETADDR, the nl_addr 671 field MUST contain a valid netaddr4 as defined in Section 3.3.9 of 672 [2]. 674 When netloc4 values are used for an inter-server copy as shown in 675 Figure 3, their values may be evaluated on the source server, 676 destination server, and client. The network environment in which 677 these systems operate should be configured so that the netloc4 values 678 are interpreted as intended on each system. 680 2.3.2. Copy Offload Stateids 682 A server may perform a copy offload operation asynchronously. An 683 asynchronous copy is tracked using a copy offload stateid. Copy 684 offload stateids are included in the COPY, OFFLOAD_ABORT, 685 OFFLOAD_STATUS, and CB_COPY operations. 687 Section 8.2.4 of [2] specifies that stateids are valid until either 688 (A) the client or server restart or (B) the client returns the 689 resource. 691 A copy offload stateid will be valid until either (A) the client or 692 server restarts or (B) the client returns the resource by issuing a 693 OFFLOAD_ABORT operation or the client replies to a CB_COPY operation. 695 A copy offload stateid's seqid MUST NOT be 0. In the context of a 696 copy offload operation, it is ambiguous to indicate the most recent 697 copy offload operation using a stateid with seqid of 0. Therefore a 698 copy offload stateid with seqid of 0 MUST be considered invalid. 700 2.4. Security Considerations 702 The security considerations pertaining to NFSv4 [10] apply to this 703 chapter. 705 The standard security mechanisms provide by NFSv4 [10] may be used to 706 secure the protocol described in this chapter. 708 NFSv4 clients and servers supporting the inter-server copy operations 709 described in this chapter are REQUIRED to implement [5], including 710 the RPCSEC_GSSv3 privileges copy_from_auth and copy_to_auth. If the 711 server-to-server copy protocol is ONC RPC based, the servers are also 712 REQUIRED to implement the RPCSEC_GSSv3 privilege copy_confirm_auth. 713 These requirements to implement are not requirements to use. NFSv4 714 clients and servers are RECOMMENDED to use [5] to secure server-side 715 copy operations. 717 2.4.1. Inter-Server Copy Security 719 2.4.1.1. Requirements for Secure Inter-Server Copy 721 Inter-server copy is driven by several requirements: 723 o The specification MUST NOT mandate an inter-server copy protocol. 724 There are many ways to copy data. Some will be more optimal than 725 others depending on the identities of the source server and 726 destination server. For example the source and destination 727 servers might be two nodes sharing a common file system format for 728 the source and destination file systems. Thus the source and 729 destination are in an ideal position to efficiently render the 730 image of the source file to the destination file by replicating 731 the file system formats at the block level. In other cases, the 732 source and destination might be two nodes sharing a common storage 733 area network, and thus there is no need to copy any data at all, 734 and instead ownership of the file and its contents simply gets re- 735 assigned to the destination. 737 o The specification MUST provide guidance for using NFSv4.x as a 738 copy protocol. For those source and destination servers willing 739 to use NFSv4.x there are specific security considerations that 740 this specification can and does address. 742 o The specification MUST NOT mandate pre-configuration between the 743 source and destination server. Requiring that the source and 744 destination first have a "copying relationship" increases the 745 administrative burden. However the specification MUST NOT 746 preclude implementations that require pre-configuration. 748 o The specification MUST NOT mandate a trust relationship between 749 the source and destination server. The NFSv4 security model 750 requires mutual authentication between a principal on an NFS 751 client and a principal on an NFS server. This model MUST continue 752 with the introduction of COPY. 754 2.4.1.2. Inter-Server Copy with RPCSEC_GSSv3 756 When the client sends a COPY_NOTIFY to the source server to expect 757 the destination to attempt to copy data from the source server, it is 758 expected that this copy is being done on behalf of the principal 759 (called the "user principal") that sent the RPC request that encloses 760 the COMPOUND procedure that contains the COPY_NOTIFY operation. The 761 user principal is identified by the RPC credentials. A mechanism 762 that allows the user principal to authorize the destination server to 763 perform the copy in a manner that lets the source server properly 764 authenticate the destination's copy, and without allowing the 765 destination to exceed its authorization is necessary. 767 An approach that sends delegated credentials of the client's user 768 principal to the destination server is not used for the following 769 reasons. If the client's user delegated its credentials, the 770 destination would authenticate as the user principal. If the 771 destination were using the NFSv4 protocol to perform the copy, then 772 the source server would authenticate the destination server as the 773 user principal, and the file copy would securely proceed. However, 774 this approach would allow the destination server to copy other files. 775 The user principal would have to trust the destination server to not 776 do so. This is counter to the requirements, and therefore is not 777 considered. Instead an approach using RPCSEC_GSSv3 [5] privileges is 778 proposed. 780 One of the stated applications of the proposed RPCSEC_GSSv3 protocol 781 is compound client host and user authentication [+ privilege 782 assertion]. For inter-server file copy, we require compound NFS 783 server host and user authentication [+ privilege assertion]. The 784 distinction between the two is one without meaning. 786 RPCSEC_GSSv3 introduces the notion of privileges. We define three 787 privileges: 789 copy_from_auth: A user principal is authorizing a source principal 790 ("nfs@") to allow a destination principal ("nfs@ 791 ") to copy a file from the source to the destination. 792 This privilege is established on the source server before the user 793 principal sends a COPY_NOTIFY operation to the source server. 795 struct copy_from_auth_priv { 796 secret4 cfap_shared_secret; 797 netloc4 cfap_destination; 798 /* the NFSv4 user name that the user principal maps to */ 799 utf8str_mixed cfap_username; 800 /* equal to seq_num of rpc_gss_cred_vers_3_t */ 801 unsigned int cfap_seq_num; 802 }; 804 cfp_shared_secret is a secret value the user principal generates. 806 copy_to_auth: A user principal is authorizing a destination 807 principal ("nfs@") to allow it to copy a file from 808 the source to the destination. This privilege is established on 809 the destination server before the user principal sends a COPY 810 operation to the destination server. 812 struct copy_to_auth_priv { 813 /* equal to cfap_shared_secret */ 814 secret4 ctap_shared_secret; 815 netloc4 ctap_source; 816 /* the NFSv4 user name that the user principal maps to */ 817 utf8str_mixed ctap_username; 818 /* equal to seq_num of rpc_gss_cred_vers_3_t */ 819 unsigned int ctap_seq_num; 820 }; 822 ctap_shared_secret is a secret value the user principal generated 823 and was used to establish the copy_from_auth privilege with the 824 source principal. 826 copy_confirm_auth: A destination principal is confirming with the 827 source principal that it is authorized to copy data from the 828 source on behalf of the user principal. When the inter-server 829 copy protocol is NFSv4, or for that matter, any protocol capable 830 of being secured via RPCSEC_GSSv3 (i.e., any ONC RPC protocol), 831 this privilege is established before the file is copied from the 832 source to the destination. 834 struct copy_confirm_auth_priv { 835 /* equal to GSS_GetMIC() of cfap_shared_secret */ 836 opaque ccap_shared_secret_mic<>; 837 /* the NFSv4 user name that the user principal maps to */ 838 utf8str_mixed ccap_username; 839 /* equal to seq_num of rpc_gss_cred_vers_3_t */ 840 unsigned int ccap_seq_num; 841 }; 843 2.4.1.2.1. Establishing a Security Context 845 When the user principal wants to COPY a file between two servers, if 846 it has not established copy_from_auth and copy_to_auth privileges on 847 the servers, it establishes them: 849 o The user principal generates a secret it will share with the two 850 servers. This shared secret will be placed in the 851 cfap_shared_secret and ctap_shared_secret fields of the 852 appropriate privilege data types, copy_from_auth_priv and 853 copy_to_auth_priv. 855 o An instance of copy_from_auth_priv is filled in with the shared 856 secret, the destination server, and the NFSv4 user id of the user 857 principal. It will be sent with an RPCSEC_GSS3_CREATE procedure, 858 and so cfap_seq_num is set to the seq_num of the credential of the 859 RPCSEC_GSS3_CREATE procedure. Because cfap_shared_secret is a 860 secret, after XDR encoding copy_from_auth_priv, GSS_Wrap() (with 861 privacy) is invoked on copy_from_auth_priv. The 862 RPCSEC_GSS3_CREATE procedure's arguments are: 864 struct { 865 rpc_gss3_gss_binding *compound_binding; 866 rpc_gss3_chan_binding *chan_binding_mic; 867 rpc_gss3_assertion assertions<>; 868 rpc_gss3_extension extensions<>; 869 } rpc_gss3_create_args; 871 The string "copy_from_auth" is placed in assertions[0].privs. The 872 output of GSS_Wrap() is placed in extensions[0].data. The field 873 extensions[0].critical is set to TRUE. The source server calls 874 GSS_Unwrap() on the privilege, and verifies that the seq_num 875 matches the credential. It then verifies that the NFSv4 user id 876 being asserted matches the source server's mapping of the user 877 principal. If it does, the privilege is established on the source 878 server as: <"copy_from_auth", user id, destination>. The 879 successful reply to RPCSEC_GSS3_CREATE has: 881 struct { 882 opaque handle<>; 883 rpc_gss3_chan_binding *chan_binding_mic; 884 rpc_gss3_assertion granted_assertions<>; 885 rpc_gss3_assertion server_assertions<>; 886 rpc_gss3_extension extensions<>; 887 } rpc_gss3_create_res; 889 The field "handle" is the RPCSEC_GSSv3 handle that the client will 890 use on COPY_NOTIFY requests involving the source and destination 891 server. granted_assertions[0].privs will be equal to 892 "copy_from_auth". The server will return a GSS_Wrap() of 893 copy_to_auth_priv. 895 o An instance of copy_to_auth_priv is filled in with the shared 896 secret, the source server, and the NFSv4 user id. It will be sent 897 with an RPCSEC_GSS3_CREATE procedure, and so ctap_seq_num is set 898 to the seq_num of the credential of the RPCSEC_GSS3_CREATE 899 procedure. Because ctap_shared_secret is a secret, after XDR 900 encoding copy_to_auth_priv, GSS_Wrap() is invoked on 901 copy_to_auth_priv. The RPCSEC_GSS3_CREATE procedure's arguments 902 are: 904 struct { 905 rpc_gss3_gss_binding *compound_binding; 906 rpc_gss3_chan_binding *chan_binding_mic; 907 rpc_gss3_assertion assertions<>; 908 rpc_gss3_extension extensions<>; 909 } rpc_gss3_create_args; 911 The string "copy_to_auth" is placed in assertions[0].privs. The 912 output of GSS_Wrap() is placed in extensions[0].data. The field 913 extensions[0].critical is set to TRUE. After unwrapping, 914 verifying the seq_num, and the user principal to NFSv4 user ID 915 mapping, the destination establishes a privilege of 916 <"copy_to_auth", user id, source>. The successful reply to 917 RPCSEC_GSS3_CREATE has: 919 struct { 920 opaque handle<>; 921 rpc_gss3_chan_binding *chan_binding_mic; 922 rpc_gss3_assertion granted_assertions<>; 923 rpc_gss3_assertion server_assertions<>; 924 rpc_gss3_extension extensions<>; 926 } rpc_gss3_create_res; 928 The field "handle" is the RPCSEC_GSSv3 handle that the client will 929 use on COPY requests involving the source and destination server. 930 The field granted_assertions[0].privs will be equal to 931 "copy_to_auth". The server will return a GSS_Wrap() of 932 copy_to_auth_priv. 934 2.4.1.2.2. Starting a Secure Inter-Server Copy 936 When the client sends a COPY_NOTIFY request to the source server, it 937 uses the privileged "copy_from_auth" RPCSEC_GSSv3 handle. 938 cna_destination_server in COPY_NOTIFY MUST be the same as the name of 939 the destination server specified in copy_from_auth_priv. Otherwise, 940 COPY_NOTIFY will fail with NFS4ERR_ACCESS. The source server 941 verifies that the privilege <"copy_from_auth", user id, destination> 942 exists, and annotates it with the source filehandle, if the user 943 principal has read access to the source file, and if administrative 944 policies give the user principal and the NFS client read access to 945 the source file (i.e., if the ACCESS operation would grant read 946 access). Otherwise, COPY_NOTIFY will fail with NFS4ERR_ACCESS. 948 When the client sends a COPY request to the destination server, it 949 uses the privileged "copy_to_auth" RPCSEC_GSSv3 handle. 950 ca_source_server in COPY MUST be the same as the name of the source 951 server specified in copy_to_auth_priv. Otherwise, COPY will fail 952 with NFS4ERR_ACCESS. The destination server verifies that the 953 privilege <"copy_to_auth", user id, source> exists, and annotates it 954 with the source and destination filehandles. If the client has 955 failed to establish the "copy_to_auth" policy it will reject the 956 request with NFS4ERR_PARTNER_NO_AUTH. 958 If the client sends a OFFLOAD_REVOKE to the source server to rescind 959 the destination server's copy privilege, it uses the privileged 960 "copy_from_auth" RPCSEC_GSSv3 handle and the cra_destination_server 961 in OFFLOAD_REVOKE MUST be the same as the name of the destination 962 server specified in copy_from_auth_priv. The source server will then 963 delete the <"copy_from_auth", user id, destination> privilege and 964 fail any subsequent copy requests sent under the auspices of this 965 privilege from the destination server. 967 2.4.1.2.3. Securing ONC RPC Server-to-Server Copy Protocols 969 After a destination server has a "copy_to_auth" privilege established 970 on it, and it receives a COPY request, if it knows it will use an ONC 971 RPC protocol to copy data, it will establish a "copy_confirm_auth" 972 privilege on the source server, using nfs@ as the 973 initiator principal, and nfs@ as the target principal. 975 The value of the field ccap_shared_secret_mic is a GSS_VerifyMIC() of 976 the shared secret passed in the copy_to_auth privilege. The field 977 ccap_username is the mapping of the user principal to an NFSv4 user 978 name ("user"@"domain" form), and MUST be the same as ctap_username 979 and cfap_username. The field ccap_seq_num is the seq_num of the 980 RPCSEC_GSSv3 credential used for the RPCSEC_GSS3_CREATE procedure the 981 destination will send to the source server to establish the 982 privilege. 984 The source server verifies the privilege, and establishes a 985 <"copy_confirm_auth", user id, destination> privilege. If the source 986 server fails to verify the privilege, the COPY operation will be 987 rejected with NFS4ERR_PARTNER_NO_AUTH. All subsequent ONC RPC 988 requests sent from the destination to copy data from the source to 989 the destination will use the RPCSEC_GSSv3 handle returned by the 990 source's RPCSEC_GSS3_CREATE response. 992 Note that the use of the "copy_confirm_auth" privilege accomplishes 993 the following: 995 o if a protocol like NFS is being used, with export policies, export 996 policies can be overridden in case the destination server as-an- 997 NFS-client is not authorized 999 o manual configuration to allow a copy relationship between the 1000 source and destination is not needed. 1002 If the attempt to establish a "copy_confirm_auth" privilege fails, 1003 then when the user principal sends a COPY request to destination, the 1004 destination server will reject it with NFS4ERR_PARTNER_NO_AUTH. 1006 2.4.1.2.4. Securing Non ONC RPC Server-to-Server Copy Protocols 1008 If the destination won't be using ONC RPC to copy the data, then the 1009 source and destination are using an unspecified copy protocol. The 1010 destination could use the shared secret and the NFSv4 user id to 1011 prove to the source server that the user principal has authorized the 1012 copy. 1014 For protocols that authenticate user names with passwords (e.g., HTTP 1015 [13] and FTP [14]), the nfsv4 user id could be used as the user name, 1016 and an ASCII hexadecimal representation of the RPCSEC_GSSv3 shared 1017 secret could be used as the user password or as input into non- 1018 password authentication methods like CHAP [15]. 1020 2.4.1.3. Inter-Server Copy via ONC RPC but without RPCSEC_GSSv3 1022 ONC RPC security flavors other than RPCSEC_GSSv3 MAY be used with the 1023 server-side copy offload operations described in this chapter. In 1024 particular, host-based ONC RPC security flavors such as AUTH_NONE and 1025 AUTH_SYS MAY be used. If a host-based security flavor is used, a 1026 minimal level of protection for the server-to-server copy protocol is 1027 possible. 1029 In the absence of strong security mechanisms such as RPCSEC_GSSv3, 1030 the challenge is how the source server and destination server 1031 identify themselves to each other, especially in the presence of 1032 multi-homed source and destination servers. In a multi-homed 1033 environment, the destination server might not contact the source 1034 server from the same network address specified by the client in the 1035 COPY_NOTIFY. This can be overcome using the procedure described 1036 below. 1038 When the client sends the source server the COPY_NOTIFY operation, 1039 the source server may reply to the client with a list of target 1040 addresses, names, and/or URLs and assign them to the unique 1041 quadruple: . If the destination uses one of these target netlocs to contact 1043 the source server, the source server will be able to uniquely 1044 identify the destination server, even if the destination server does 1045 not connect from the address specified by the client in COPY_NOTIFY. 1046 The level of assurance in this identification depends on the 1047 unpredictability, strength and secrecy of the random number. 1049 For example, suppose the network topology is as shown in Figure 3. 1050 If the source filehandle is 0x12345, the source server may respond to 1051 a COPY_NOTIFY for destination 10.11.78.56 with the URLs: 1053 nfs://10.11.78.18//_COPY/FvhH1OKbu8VrxvV1erdjvR7N/10.11.78.56/_FH/ 1054 0x12345 1056 nfs://192.168.33.18//_COPY/FvhH1OKbu8VrxvV1erdjvR7N/10.11.78.56/ 1057 _FH/0x12345 1059 The name component after _COPY is 24 characters of base 64, more than 1060 enough to encode a 128 bit random number. 1062 The client will then send these URLs to the destination server in the 1063 COPY operation. Suppose that the 192.168.33.0/24 network is a high 1064 speed network and the destination server decides to transfer the file 1065 over this network. If the destination contacts the source server 1066 from 192.168.33.56 over this network using NFSv4.1, it does the 1067 following: 1069 COMPOUND { PUTROOTFH, LOOKUP "_COPY" ; LOOKUP 1070 "FvhH1OKbu8VrxvV1erdjvR7N" ; LOOKUP "10.11.78.56"; LOOKUP "_FH" ; 1071 OPEN "0x12345" ; GETFH } 1073 Provided that the random number is unpredictable and has been kept 1074 secret by the parties involved, the source server will therefore know 1075 that these NFSv4.x operations are being issued by the destination 1076 server identified in the COPY_NOTIFY. This random number technique 1077 only provides initial authentication of the destination server, and 1078 cannot defend against man-in-the-middle attacks after authentication 1079 or an eavesdropper that observes the random number on the wire. 1080 Other secure communication techniques (e.g., IPsec) are necessary to 1081 block these attacks. 1083 2.4.1.4. Inter-Server Copy without ONC RPC and RPCSEC_GSSv3 1085 The same techniques as Section 2.4.1.3, using unique URLs for each 1086 destination server, can be used for other protocols (e.g., HTTP [13] 1087 and FTP [14]) as well. 1089 3. Support for Application IO Hints 1091 Applications can issue client I/O hints via posix_fadvise() [6] to 1092 the NFS client. While this can help the NFS client optimize I/O and 1093 caching for a file, it does not allow the NFS server and its exported 1094 file system to do likewise. We add an IO_ADVISE procedure 1095 (Section 13.8) to communicate the client file access patterns to the 1096 NFS server. The NFS server upon receiving a IO_ADVISE operation MAY 1097 choose to alter its I/O and caching behavior, but is under no 1098 obligation to do so. 1100 Application specific NFS clients such as those used by hypervisors 1101 and databases can also leverage application hints to communicate 1102 their specialized requirements. 1104 4. Sparse Files 1106 4.1. Introduction 1108 A sparse file is a common way of representing a large file without 1109 having to utilize all of the disk space for it. Consequently, a 1110 sparse file uses less physical space than its size indicates. This 1111 means the file contains 'holes', byte ranges within the file that 1112 contain no data. Most modern file systems support sparse files, 1113 including most UNIX file systems and NTFS, but notably not Apple's 1114 HFS+. Common examples of sparse files include Virtual Machine (VM) 1115 OS/disk images, database files, log files, and even checkpoint 1116 recovery files most commonly used by the HPC community. 1118 If an application reads a hole in a sparse file, the file system must 1119 return all zeros to the application. For local data access there is 1120 little penalty, but with NFS these zeroes must be transferred back to 1121 the client. If an application uses the NFS client to read data into 1122 memory, this wastes time and bandwidth as the application waits for 1123 the zeroes to be transferred. 1125 A sparse file is typically created by initializing the file to be all 1126 zeros - nothing is written to the data in the file, instead the hole 1127 is recorded in the metadata for the file. So a 8G disk image might 1128 be represented initially by a couple hundred bits in the inode and 1129 nothing on the disk. If the VM then writes 100M to a file in the 1130 middle of the image, there would now be two holes represented in the 1131 metadata and 100M in the data. 1133 Two new operations INITIALIZE (Section 13.7) and READ_PLUS 1134 (Section 13.10) are introduced. INITIALIZE allows for the creation 1135 of a sparse file and for hole punching. An application might want to 1136 zero out a range of the file. READ_PLUS supports all the features of 1137 READ but includes an extension to support sparse pattern files 1138 (Section 6.1.2). READ_PLUS is guaranteed to perform no worse than 1139 READ, and can dramatically improve performance with sparse files. 1140 READ_PLUS does not depend on pNFS protocol features, but can be used 1141 by pNFS to support sparse files. 1143 4.2. Terminology 1145 Regular file: An object of file type NF4REG or NF4NAMEDATTR. 1147 Sparse file: A Regular file that contains one or more Holes. 1149 Hole: A byte range within a Sparse file that contains regions of all 1150 zeroes. For block-based file systems, this could also be an 1151 unallocated region of the file. 1153 Hole Threshold: The minimum length of a Hole as determined by the 1154 server. If a server chooses to define a Hole Threshold, then it 1155 would not return hole information about holes with a length 1156 shorter than the Hole Threshold. 1158 5. Space Reservation 1159 5.1. Introduction 1161 This section describes a set of operations that allow applications 1162 such as hypervisors to reserve space for a file, report the amount of 1163 actual disk space a file occupies and freeup the backing space of a 1164 file when it is not required. In virtualized environments, virtual 1165 disk files are often stored on NFS mounted volumes. Since virtual 1166 disk files represent the hard disks of virtual machines, hypervisors 1167 often have to guarantee certain properties for the file. 1169 One such example is space reservation. When a hypervisor creates a 1170 virtual disk file, it often tries to preallocate the space for the 1171 file so that there are no future allocation related errors during the 1172 operation of the virtual machine. Such errors prevent a virtual 1173 machine from continuing execution and result in downtime. 1175 Currently, in order to achieve such a guarantee, applications zero 1176 the entire file. The initial zeroing allocates the backing blocks 1177 and all subsequent writes are overwrites of already allocated blocks. 1178 This approach is not only inefficient in terms of the amount of I/O 1179 done, it is also not guaranteed to work on file systems that are log 1180 structured or deduplicated. An efficient way of guaranteeing space 1181 reservation would be beneficial to such applications. 1183 If the space_reserved attribute (see Section 11.2.4) is set on a 1184 file, it is guaranteed that writes that do not grow the file will not 1185 fail with NFSERR_NOSPC. 1187 Another useful feature would be the ability to report the number of 1188 blocks that would be freed when a file is deleted. Currently, NFS 1189 reports two size attributes: 1191 size The logical file size of the file. 1193 space_used The size in bytes that the file occupies on disk 1195 While these attributes are sufficient for space accounting in 1196 traditional file systems, they prove to be inadequate in modern file 1197 systems that support block sharing. In such file systems, multiple 1198 inodes can point to a single block with a block reference count to 1199 guard against premature freeing. Having a way to tell the number of 1200 blocks that would be freed if the file was deleted would be useful to 1201 applications that wish to migrate files when a volume is low on 1202 space. 1204 Since virtual disks represent a hard drive in a virtual machine, a 1205 virtual disk can be viewed as a file system within a file. Since not 1206 all blocks within a file system are in use, there is an opportunity 1207 to reclaim blocks that are no longer in use. A call to deallocate 1208 blocks could result in better space efficiency. Lesser space MAY be 1209 consumed for backups after block deallocation. 1211 The following operations and attributes can be used to resolve this 1212 issues: 1214 space_reserved This attribute specifies whether the blocks backing 1215 the file have been preallocated. 1217 space_freed This attribute specifies the space freed when a file is 1218 deleted, taking block sharing into consideration. 1220 INITIALIZE This operation zeroes and/or deallocates the blocks 1221 backing a region of the file. 1223 If space_used of a file is interpreted to mean the size in bytes of 1224 all disk blocks pointed to by the inode of the file, then shared 1225 blocks get double counted, over-reporting the space utilization. 1226 This also has the adverse effect that the deletion of a file with 1227 shared blocks frees up less than space_used bytes. 1229 On the other hand, if space_used is interpreted to mean the size in 1230 bytes of those disk blocks unique to the inode of the file, then 1231 shared blocks are not counted in any file, resulting in under- 1232 reporting of the space utilization. 1234 For example, two files A and B have 10 blocks each. Let 6 of these 1235 blocks be shared between them. Thus, the combined space utilized by 1236 the two files is 14 * BLOCK_SIZE bytes. In the former case, the 1237 combined space utilization of the two files would be reported as 20 * 1238 BLOCK_SIZE. However, deleting either would only result in 4 * 1239 BLOCK_SIZE being freed. Conversely, the latter interpretation would 1240 report that the space utilization is only 8 * BLOCK_SIZE. 1242 Adding another size attribute, space_freed (see Section 11.2.5), is 1243 helpful in solving this problem. space_freed is the number of blocks 1244 that are allocated to the given file that would be freed on its 1245 deletion. In the example, both A and B would report space_freed as 4 1246 * BLOCK_SIZE and space_used as 10 * BLOCK_SIZE. If A is deleted, B 1247 will report space_freed as 10 * BLOCK_SIZE as the deletion of B would 1248 result in the deallocation of all 10 blocks. 1250 The addition of this problem doesn't solve the problem of space being 1251 over-reported. However, over-reporting is better than under- 1252 reporting. 1254 6. Application Data Hole Support 1256 At the OS level, files are contained on disk blocks. Applications 1257 are also free to impose structure on the data contained in a file and 1258 we can define an Application Data Block (ADB) to be such a structure. 1259 From the application's viewpoint, it only wants to handle ADBs and 1260 not raw bytes (see [16]). An ADB is typically comprised of two 1261 sections: a header and data. The header describes the 1262 characteristics of the block and can provide a means to detect 1263 corruption in the data payload. The data section is typically 1264 initialized to all zeros. 1266 The format of the header is application specific, but there are two 1267 main components typically encountered: 1269 1. A logical block number which allows the application to determine 1270 which data block is being referenced. This is useful when the 1271 client is not storing the blocks in contiguous memory. 1273 2. Fields to describe the state of the ADB and a means to detect 1274 block corruption. For both pieces of data, a useful property is 1275 that allowed values be unique in that if passed across the 1276 network, corruption due to translation between big and little 1277 endian architectures are detectable. For example, 0xF0DEDEF0 has 1278 the same bit pattern in both architectures. 1280 Applications already impose structures on files [16] and detect 1281 corruption in data blocks [17]. What they are not able to do is 1282 efficiently transfer and store ADBs. To initialize a file with ADBs, 1283 the client must send the full ADB to the server and that must be 1284 stored on the server. 1286 In this section, we are going to define an Application Data Hole 1287 (ADH), which is a generic framework for transfering the ADB, present 1288 one approach to detecting corruption in a given ADH implementation, 1289 and describe the model for how the client and server can support 1290 efficient initialization of ADHs, reading of ADH holes, punching ADH 1291 holes in a file, and space reservation. We define the ADHN to be the 1292 Application Data Hole Number, which is the logical block number 1293 discussed earlier. 1295 6.1. Generic Framework 1297 We want the representation of the ADH to be flexible enough to 1298 support many different applications. The most basic approach is no 1299 imposition of a block at all, which means we are working with the raw 1300 bytes. Such an approach would be useful for storing holes, punching 1301 holes, etc. In more complex deployments, a server might be 1302 supporting multiple applications, each with their own definition of 1303 the ADH. One might store the ADHN at the start of the block and then 1304 have a guard pattern to detect corruption [18]. The next might store 1305 the ADHN at an offset of 100 bytes within the block and have no guard 1306 pattern at all, i.e., existing applications might already have well 1307 defined formats for their data blocks. 1309 The guard pattern can be used to represent the state of the block, to 1310 protect against corruption, or both. Again, it needs to be able to 1311 be placed anywhere within the ADH. 1313 We need to be able to represent the starting offset of the block and 1314 the size of the block. Note that nothing prevents the application 1315 from defining different sized blocks in a file. 1317 6.1.1. Data Hole Representation 1319 struct app_data_hole4 { 1320 offset4 adh_offset; 1321 length4 adh_block_size; 1322 length4 adh_block_count; 1323 length4 adh_reloff_blocknum; 1324 count4 adh_block_num; 1325 length4 adh_reloff_pattern; 1326 opaque adh_pattern<>; 1327 }; 1329 The app_data_hole4 structure captures the abstraction presented for 1330 the ADH. The additional fields present are to allow the transmission 1331 of adh_block_count ADHs at one time. We also use adh_block_num to 1332 convey the ADHN of the first block in the sequence. Each ADH will 1333 contain the same adh_pattern string. 1335 As both adh_block_num and adh_pattern are optional, if either 1336 adh_reloff_pattern or adh_reloff_blocknum is set to NFS4_UINT64_MAX, 1337 then the corresponding field is not set in any of the ADH. 1339 6.1.2. Data Content 1341 /* 1342 * Use an enum such that we can extend new types. 1343 */ 1344 enum data_content4 { 1345 NFS4_CONTENT_DATA = 0, 1346 NFS4_CONTENT_APP_DATA_HOLE = 1, 1347 NFS4_CONTENT_HOLE = 2 1348 }; 1349 New operations might need to differentiate between wanting to access 1350 data versus an ADH. Also, future minor versions might want to 1351 introduce new data formats. This enumeration allows that to occur. 1353 6.2. An Example of Detecting Corruption 1355 In this section, we define an ADH format in which corruption can be 1356 detected. Note that this is just one possible format and means to 1357 detect corruption. 1359 Consider a very basic implementation of an operating system's disk 1360 blocks. A block is either data or it is an indirect block which 1361 allows for files to be larger than one block. It is desired to be 1362 able to initialize a block. Lastly, to quickly unlink a file, a 1363 block can be marked invalid. The contents remain intact - which 1364 would enable this OS application to undelete a file. 1366 The application defines 4k sized data blocks, with an 8 byte block 1367 counter occurring at offset 0 in the block, and with the guard 1368 pattern occurring at offset 8 inside the block. Furthermore, the 1369 guard pattern can take one of four states: 1371 0xfeedface - This is the FREE state and indicates that the ADH 1372 format has been applied. 1374 0xcafedead - This is the DATA state and indicates that real data 1375 has been written to this block. 1377 0xe4e5c001 - This is the INDIRECT state and indicates that the 1378 block contains block counter numbers that are chained off of this 1379 block. 1381 0xba1ed4a3 - This is the INVALID state and indicates that the block 1382 contains data whose contents are garbage. 1384 Finally, it also defines an 8 byte checksum [19] starting at byte 16 1385 which applies to the remaining contents of the block. If the state 1386 is FREE, then that checksum is trivially zero. As such, the 1387 application has no need to transfer the checksum implicitly inside 1388 the ADH - it need not make the transfer layer aware of the fact that 1389 there is a checksum (see [17] for an example of checksums used to 1390 detect corruption in application data blocks). 1392 Corruption in each ADH can be detected thusly: 1394 o If the guard pattern is anything other than one of the allowed 1395 values, including all zeros. 1397 o If the guard pattern is FREE and any other byte in the remainder 1398 of the ADH is anything other than zero. 1400 o If the guard pattern is anything other than FREE, then if the 1401 stored checksum does not match the computed checksum. 1403 o If the guard pattern is INDIRECT and one of the stored indirect 1404 block numbers has a value greater than the number of ADHs in the 1405 file. 1407 o If the guard pattern is INDIRECT and one of the stored indirect 1408 block numbers is a duplicate of another stored indirect block 1409 number. 1411 As can be seen, the application can detect errors based on the 1412 combination of the guard pattern state and the checksum. But also, 1413 the application can detect corruption based on the state and the 1414 contents of the ADH. This last point is important in validating the 1415 minimum amount of data we incorporated into our generic framework. 1416 I.e., the guard pattern is sufficient in allowing applications to 1417 design their own corruption detection. 1419 Finally, it is important to note that none of these corruption checks 1420 occur in the transport layer. The server and client components are 1421 totally unaware of the file format and might report everything as 1422 being transferred correctly even in the case the application detects 1423 corruption. 1425 6.3. Example of READ_PLUS 1427 The hypothetical application presented in Section 6.2 can be used to 1428 illustrate how READ_PLUS would return an array of results. A file is 1429 created and initialized with 100 4k ADHs in the FREE state: 1431 INITIALIZE {0, 4k, 100, 0, 0, 8, 0xfeedface} 1433 Further, assume the application writes a single ADH at 16k, changing 1434 the guard pattern to 0xcafedead, we would then have in memory: 1436 0 -> (16k - 1) : 4k, 4, 0, 0, 8, 0xfeedface 1437 16k -> (20k - 1) : 00 00 00 05 ca fe de ad XX XX ... XX XX 1438 20k -> 400k : 4k, 95, 0, 6, 0xfeedface 1440 And when the client did a READ_PLUS of 64k at the start of the file, 1441 it would get back a result of an ADH, some data, and a final ADH: 1443 ADH {0, 4, 0, 0, 8, 0xfeedface} 1444 data 4k 1445 ADH {20k, 4k, 59, 0, 6, 0xfeedface} 1447 7. Labeled NFS 1449 7.1. Introduction 1451 Access control models such as Unix permissions or Access Control 1452 Lists are commonly referred to as Discretionary Access Control (DAC) 1453 models. These systems base their access decisions on user identity 1454 and resource ownership. In contrast Mandatory Access Control (MAC) 1455 models base their access control decisions on the label on the 1456 subject (usually a process) and the object it wishes to access [7]. 1457 These labels may contain user identity information but usually 1458 contain additional information. In DAC systems users are free to 1459 specify the access rules for resources that they own. MAC models 1460 base their security decisions on a system wide policy established by 1461 an administrator or organization which the users do not have the 1462 ability to override. In this section, we add a MAC model to NFSv4.2. 1464 The first change necessary is to devise a method for transporting and 1465 storing security label data on NFSv4 file objects. Security labels 1466 have several semantics that are met by NFSv4 recommended attributes 1467 such as the ability to set the label value upon object creation. 1468 Access control on these attributes are done through a combination of 1469 two mechanisms. As with other recommended attributes on file objects 1470 the usual DAC checks (ACLs and permission bits) will be performed to 1471 ensure that proper file ownership is enforced. In addition a MAC 1472 system MAY be employed on the client, server, or both to enforce 1473 additional policy on what subjects may modify security label 1474 information. 1476 The second change is to provide methods for the client to determine 1477 if the security label has changed. A client which needs to know if a 1478 label is going to change SHOULD register a delegation on that file. 1479 In order to change the security label, the server will have to recall 1480 all delegations. This will inform the client of the change. If a 1481 client wants to detect if the label has changed, it MAY use VERIFY 1482 and NVERIFY on FATTR4_CHANGE_SEC_LABEL to detect that the 1483 FATTR4_SEC_LABEL has been modified. 1485 The final change necessary is a modification to the RPC layer used in 1486 NFSv4 in the form of a new version of the RPCSEC_GSS [8] framework. 1487 In order for an NFSv4 server to apply MAC checks it must obtain 1488 additional information from the client. Several methods were 1489 explored for performing this and it was decided that the best 1490 approach was to incorporate the ability to make security attribute 1491 assertions through the RPC mechanism. RPCSECGSSv3 [5] outlines a 1492 method to assert additional security information such as security 1493 labels on gss context creation and have that data bound to all RPC 1494 requests that make use of that context. 1496 7.2. Definitions 1498 Label Format Specifier (LFS): is an identifier used by the client to 1499 establish the syntactic format of the security label and the 1500 semantic meaning of its components. These specifiers exist in a 1501 registry associated with documents describing the format and 1502 semantics of the label. 1504 Label Format Registry: is the IANA registry containing all 1505 registered LFS along with references to the documents that 1506 describe the syntactic format and semantics of the security label. 1508 Policy Identifier (PI): is an optional part of the definition of a 1509 Label Format Specifier which allows for clients and server to 1510 identify specific security policies. 1512 Object: is a passive resource within the system that we wish to be 1513 protected. Objects can be entities such as files, directories, 1514 pipes, sockets, and many other system resources relevant to the 1515 protection of the system state. 1517 Subject: is an active entity usually a process which is requesting 1518 access to an object. 1520 MAC-Aware: is a server which can transmit and store object labels. 1522 MAC-Functional: is a client or server which is Labeled NFS enabled. 1523 Such a system can interpret labels and apply policies based on the 1524 security system. 1526 Multi-Level Security (MLS): is a traditional model where objects are 1527 given a sensitivity level (Unclassified, Secret, Top Secret, etc) 1528 and a category set [20]. 1530 7.3. MAC Security Attribute 1532 MAC models base access decisions on security attributes bound to 1533 subjects and objects. This information can range from a user 1534 identity for an identity based MAC model, sensitivity levels for 1535 Multi-level security, or a type for Type Enforcement. These models 1536 base their decisions on different criteria but the semantics of the 1537 security attribute remain the same. The semantics required by the 1538 security attributes are listed below: 1540 o MUST provide flexibility with respect to the MAC model. 1542 o MUST provide the ability to atomically set security information 1543 upon object creation. 1545 o MUST provide the ability to enforce access control decisions both 1546 on the client and the server. 1548 o MUST not expose an object to either the client or server name 1549 space before its security information has been bound to it. 1551 NFSv4 implements the security attribute as a recommended attribute. 1552 These attributes have a fixed format and semantics, which conflicts 1553 with the flexible nature of the security attribute. To resolve this 1554 the security attribute consists of two components. The first 1555 component is a LFS as defined in [21] to allow for interoperability 1556 between MAC mechanisms. The second component is an opaque field 1557 which is the actual security attribute data. To allow for various 1558 MAC models, NFSv4 should be used solely as a transport mechanism for 1559 the security attribute. It is the responsibility of the endpoints to 1560 consume the security attribute and make access decisions based on 1561 their respective models. In addition, creation of objects through 1562 OPEN and CREATE allows for the security attribute to be specified 1563 upon creation. By providing an atomic create and set operation for 1564 the security attribute it is possible to enforce the second and 1565 fourth requirements. The recommended attribute FATTR4_SEC_LABEL (see 1566 Section 11.2.2) will be used to satisfy this requirement. 1568 7.3.1. Delegations 1570 In the event that a security attribute is changed on the server while 1571 a client holds a delegation on the file, both the server and the 1572 client MUST follow the NFSv4.1 protocol (see Chapter 10 of [2]) with 1573 respect to attribute changes. It SHOULD flush all changes back to 1574 the server and relinquish the delegation. 1576 7.3.2. Permission Checking 1578 It is not feasible to enumerate all possible MAC models and even 1579 levels of protection within a subset of these models. This means 1580 that the NFSv4 client and servers cannot be expected to directly make 1581 access control decisions based on the security attribute. Instead 1582 NFSv4 should defer permission checking on this attribute to the host 1583 system. These checks are performed in addition to existing DAC and 1584 ACL checks outlined in the NFSv4 protocol. Section 7.6 gives a 1585 specific example of how the security attribute is handled under a 1586 particular MAC model. 1588 7.3.3. Object Creation 1590 When creating files in NFSv4 the OPEN and CREATE operations are used. 1591 One of the parameters to these operations is an fattr4 structure 1592 containing the attributes the file is to be created with. This 1593 allows NFSv4 to atomically set the security attribute of files upon 1594 creation. When a client is MAC-Functional it must always provide the 1595 initial security attribute upon file creation. In the event that the 1596 server is MAC-Functional as well, it should determine by policy 1597 whether it will accept the attribute from the client or instead make 1598 the determination itself. If the client is not MAC-Functional, then 1599 the MAC-Functional server must decide on a default label. A more in 1600 depth explanation can be found in Section 7.6. 1602 7.3.4. Existing Objects 1604 Note that under the MAC model, all objects must have labels. 1605 Therefore, if an existing server is upgraded to include Labeled NFS 1606 support, then it is the responsibility of the security system to 1607 define the behavior for existing objects. 1609 7.3.5. Label Changes 1611 If there are open delegations on the file belonging to client other 1612 than the one making the label change, then the process described in 1613 Section 7.3.1 must be followed. In short, the delegation will be 1614 recalled, which effectively notifies the client of the change. 1616 As the server is always presented with the subject label from the 1617 client, it does not necessarily need to communicate the fact that the 1618 label has changed to the client. In the cases where the change 1619 outright denies the client access, the client will be able to quickly 1620 determine that there is a new label in effect. 1622 Consider a system in which the clients enforce MAC checks and and the 1623 server has a very simple security system which just stores the 1624 labels. In this system, the MAC label check always allows access, 1625 regardless of the subject label. 1627 The way in which MAC labels are enforced is by the client. The 1628 security policies on the client can be such that the client does not 1629 have access to the file unless it has a delegation. The recall of 1630 the delegation will force the client to flush any cached content of 1631 the file. The clients could also be configured to periodically 1632 VERIFY/NVERIFY the FATTR4_CHANGE_SEC_LABEL attribute to determine 1633 when the label has changed. When a change is detected, then the 1634 client could take the costlier action of retrieving the 1635 FATTR4_SEC_LABEL. 1637 7.4. pNFS Considerations 1639 This section examines the issues in deploying Labeled NFS in a pNFS 1640 community of servers. 1642 7.4.1. MAC Label Checks 1644 The new FATTR4_SEC_LABEL attribute is metadata information and as 1645 such the DS is not aware of the value contained on the MDS. 1646 Fortunately, the NFSv4.1 protocol [2] already has provisions for 1647 doing access level checks from the DS to the MDS. In order for the 1648 DS to validate the subject label presented by the client, it SHOULD 1649 utilize this mechanism. 1651 7.5. Discovery of Server Labeled NFS Support 1653 The server can easily determine that a client supports Labeled NFS 1654 when it queries for the FATTR4_SEC_LABEL label for an object. Note 1655 that it cannot assume that the presence of RPCSEC_GSSv3 indicates 1656 Labeled NFS support. The client might need to discover which LFS the 1657 server supports. 1659 A server which supports Labeled NFS MUST allow a client with any 1660 subject label to retrieve the FATTR4_SEC_LABEL attribute for the root 1661 filehandle, ROOTFH. The following compound must always succeed as 1662 far as a MAC label check is concerned: 1664 PUTROOTFH, GETATTR {FATTR4_SEC_LABEL} 1666 Note that the server might have imposed a security flavor on the root 1667 that precludes such access. I.e., if the server requires kerberized 1668 access and the client presents a compound with AUTH_SYS, then the 1669 server is allowed to return NFS4ERR_WRONGSEC in this case. But if 1670 the client presents a correct security flavor, then the server MUST 1671 return the FATTR4_SEC_LABEL attribute with the supported LFS filled 1672 in. 1674 7.6. MAC Security NFS Modes of Operation 1676 A system using Labeled NFS may operate in two modes. The first mode 1677 provides the most protection and is called "full mode". In this mode 1678 both the client and server implement a MAC model allowing each end to 1679 make an access control decision. The remaining mode is called the 1680 "guest mode" and in this mode one end of the connection is not 1681 implementing a MAC model and thus offers less protection than full 1682 mode. 1684 7.6.1. Full Mode 1686 Full mode environments consist of MAC-Functional NFSv4 servers and 1687 clients and may be composed of mixed MAC models and policies. The 1688 system requires that both the client and server have an opportunity 1689 to perform an access control check based on all relevant information 1690 within the network. The file object security attribute is provided 1691 using the mechanism described in Section 7.3. The security attribute 1692 of the subject making the request is transported at the RPC layer 1693 using the mechanism described in RPCSECGSSv3 [5]. 1695 7.6.1.1. Initial Labeling and Translation 1697 The ability to create a file is an action that a MAC model may wish 1698 to mediate. The client is given the responsibility to determine the 1699 initial security attribute to be placed on a file. This allows the 1700 client to make a decision as to the acceptable security attributes to 1701 create a file with before sending the request to the server. Once 1702 the server receives the creation request from the client it may 1703 choose to evaluate if the security attribute is acceptable. 1705 Security attributes on the client and server may vary based on MAC 1706 model and policy. To handle this the security attribute field has an 1707 LFS component. This component is a mechanism for the host to 1708 identify the format and meaning of the opaque portion of the security 1709 attribute. A full mode environment may contain hosts operating in 1710 several different LFSs. In this case a mechanism for translating the 1711 opaque portion of the security attribute is needed. The actual 1712 translation function will vary based on MAC model and policy and is 1713 out of the scope of this document. If a translation is unavailable 1714 for a given LFS then the request MUST be denied. Another recourse is 1715 to allow the host to provide a fallback mapping for unknown security 1716 attributes. 1718 7.6.1.2. Policy Enforcement 1720 In full mode access control decisions are made by both the clients 1721 and servers. When a client makes a request it takes the security 1722 attribute from the requesting process and makes an access control 1723 decision based on that attribute and the security attribute of the 1724 object it is trying to access. If the client denies that access an 1725 RPC call to the server is never made. If however the access is 1726 allowed the client will make a call to the NFS server. 1728 When the server receives the request from the client it extracts the 1729 security attribute conveyed in the RPC request. The server then uses 1730 this security attribute and the attribute of the object the client is 1731 trying to access to make an access control decision. If the server's 1732 policy allows this access it will fulfill the client's request, 1733 otherwise it will return NFS4ERR_ACCESS. 1735 Implementations MAY validate security attributes supplied over the 1736 network to ensure that they are within a set of attributes permitted 1737 from a specific peer, and if not, reject them. Note that a system 1738 may permit a different set of attributes to be accepted from each 1739 peer. 1741 7.6.1.3. Limited Server 1743 A Limited Server mode (see Section 3.5.2 of [7]) consists of a server 1744 which is label aware, but does not enforce policies. Such a server 1745 will store and retrieve all object labels presented by clients, 1746 utililze the methods described in Section 7.3.5 to allow the clients 1747 to detect changing labels,, but will not restrict access via the 1748 subject label. Instead, it will expect the clients to enforce all 1749 such access locally. 1751 7.6.2. Guest Mode 1753 Guest mode implies that either the client or the server does not 1754 handle labels. If the client is not Labeled NFS aware, then it will 1755 not offer subject labels to the server. The server is the only 1756 entity enforcing policy, and may selectively provide standard NFS 1757 services to clients based on their authentication credentials and/or 1758 associated network attributes (e.g., IP address, network interface). 1759 The level of trust and access extended to a client in this mode is 1760 configuration-specific. If the server is not Labeled NFS aware, then 1761 it will not return object labels to the client. Clients in this 1762 environment are may consist of groups implementing different MAC 1763 model policies. The system requires that all clients in the 1764 environment be responsible for access control checks. 1766 7.7. Security Considerations 1768 This entire chapter deals with security issues. 1770 Depending on the level of protection the MAC system offers there may 1771 be a requirement to tightly bind the security attribute to the data. 1773 When only one of the client or server enforces labels, it is 1774 important to realize that the other side is not enforcing MAC 1775 protections. Alternate methods might be in use to handle the lack of 1776 MAC support and care should be taken to identify and mitigate threats 1777 from possible tampering outside of these methods. 1779 An example of this is that a server that modifies READDIR or LOOKUP 1780 results based on the client's subject label might want to always 1781 construct the same subject label for a client which does not present 1782 one. This will prevent a non-Labeled NFS client from mixing entries 1783 in the directory cache. 1785 8. Sharing change attribute implementation details with NFSv4 clients 1787 8.1. Introduction 1789 Although both the NFSv4 [10] and NFSv4.1 protocol [2], define the 1790 change attribute as being mandatory to implement, there is little in 1791 the way of guidance. The only mandated feature is that the value 1792 must change whenever the file data or metadata change. 1794 While this allows for a wide range of implementations, it also leaves 1795 the client with a conundrum: how does it determine which is the most 1796 recent value for the change attribute in a case where several RPC 1797 calls have been issued in parallel? In other words if two COMPOUNDs, 1798 both containing WRITE and GETATTR requests for the same file, have 1799 been issued in parallel, how does the client determine which of the 1800 two change attribute values returned in the replies to the GETATTR 1801 requests correspond to the most recent state of the file? In some 1802 cases, the only recourse may be to send another COMPOUND containing a 1803 third GETATTR that is fully serialised with the first two. 1805 NFSv4.2 avoids this kind of inefficiency by allowing the server to 1806 share details about how the change attribute is expected to evolve, 1807 so that the client may immediately determine which, out of the 1808 several change attribute values returned by the server, is the most 1809 recent. change_attr_type is defined as a new recommended attribute 1810 (see Section 11.2.1), and is per file system. 1812 9. Security Considerations 1814 10. Error Values 1816 NFS error numbers are assigned to failed operations within a Compound 1817 (COMPOUND or CB_COMPOUND) request. A Compound request contains a 1818 number of NFS operations that have their results encoded in sequence 1819 in a Compound reply. The results of successful operations will 1820 consist of an NFS4_OK status followed by the encoded results of the 1821 operation. If an NFS operation fails, an error status will be 1822 entered in the reply and the Compound request will be terminated. 1824 10.1. Error Definitions 1826 Protocol Error Definitions 1828 +--------------------------+--------+------------------+ 1829 | Error | Number | Description | 1830 +--------------------------+--------+------------------+ 1831 | NFS4ERR_BADLABEL | 10093 | Section 10.1.3.1 | 1832 | NFS4ERR_METADATA_NOTSUPP | 10090 | Section 10.1.2.1 | 1833 | NFS4ERR_OFFLOAD_DENIED | 10091 | Section 10.1.2.2 | 1834 | NFS4ERR_PARTNER_NO_AUTH | 10089 | Section 10.1.2.3 | 1835 | NFS4ERR_PARTNER_NOTSUPP | 10088 | Section 10.1.2.4 | 1836 | NFS4ERR_UNION_NOTSUPP | 10094 | Section 10.1.1.1 | 1837 | NFS4ERR_WRONG_LFS | 10092 | Section 10.1.3.2 | 1838 +--------------------------+--------+------------------+ 1840 Table 1 1842 10.1.1. General Errors 1844 This section deals with errors that are applicable to a broad set of 1845 different purposes. 1847 10.1.1.1. NFS4ERR_UNION_NOTSUPP (Error Code 10094) 1849 One of the arguments to the operation is a discriminated union and 1850 while the server supports the given operation, it does not support 1851 the selected arm of the discriminated union. For an example, see 1852 READ_PLUS (Section 13.10). 1854 10.1.2. Server to Server Copy Errors 1856 These errors deal with the interaction between server to server 1857 copies. 1859 10.1.2.1. NFS4ERR_METADATA_NOTSUPP (Error Code 10090) 1861 The destination file cannot support the same metadata as the source 1862 file. 1864 10.1.2.2. NFS4ERR_OFFLOAD_DENIED (Error Code 10091) 1866 The copy offload operation is supported by both the source and the 1867 destination, but the destination is not allowing it for this file. 1868 If the client sees this error, it should fall back to the normal copy 1869 semantics. 1871 10.1.2.3. NFS4ERR_PARTNER_NO_AUTH (Error Code 10089) 1873 The source server does not authorize a server-to-server copy offload 1874 operation. This may be due to the client's failure to send the 1875 COPY_NOTIFY operation to the source server, the source server 1876 receiving a server-to-server copy offload request after the copy 1877 lease time expired, or for some other permission problem. 1879 10.1.2.4. NFS4ERR_PARTNER_NOTSUPP (Error Code 10088) 1881 The remote server does not support the server-to-server copy offload 1882 protocol. 1884 10.1.3. Labeled NFS Errors 1886 These errors are used in Labeled NFS. 1888 10.1.3.1. NFS4ERR_BADLABEL (Error Code 10093) 1890 The label specified is invalid in some manner. 1892 10.1.3.2. NFS4ERR_WRONG_LFS (Error Code 10092) 1894 The LFS specified in the subject label is not compatible with the LFS 1895 in the object label. 1897 11. New File Attributes 1899 11.1. New RECOMMENDED Attributes - List and Definition References 1901 The list of new RECOMMENDED attributes appears in Table 2. The 1902 meaning of the columns of the table are: 1904 Name: The name of the attribute. 1906 Id: The number assigned to the attribute. In the event of conflicts 1907 between the assigned number and [3], the latter is likely 1908 authoritative, but should be resolved with Errata to this document 1909 and/or [3]. See [22] for the Errata process. 1911 Data Type: The XDR data type of the attribute. 1913 Acc: Access allowed to the attribute. 1915 R means read-only (GETATTR may retrieve, SETATTR may not set). 1917 W means write-only (SETATTR may set, GETATTR may not retrieve). 1919 R W means read/write (GETATTR may retrieve, SETATTR may set). 1921 Defined in: The section of this specification that describes the 1922 attribute. 1924 +------------------+----+-------------------+-----+----------------+ 1925 | Name | Id | Data Type | Acc | Defined in | 1926 +------------------+----+-------------------+-----+----------------+ 1927 | change_attr_type | 79 | change_attr_type4 | R | Section 11.2.1 | 1928 | sec_label | 80 | sec_label4 | R W | Section 11.2.2 | 1929 | change_sec_label | 81 | change_sec_label4 | R | Section 11.2.3 | 1930 | space_reserved | 77 | boolean | R W | Section 11.2.4 | 1931 | space_freed | 78 | length4 | R | Section 11.2.5 | 1932 +------------------+----+-------------------+-----+----------------+ 1934 Table 2 1936 11.2. Attribute Definitions 1938 11.2.1. Attribute 79: change_attr_type 1940 enum change_attr_type4 { 1941 NFS4_CHANGE_TYPE_IS_MONOTONIC_INCR = 0, 1942 NFS4_CHANGE_TYPE_IS_VERSION_COUNTER = 1, 1943 NFS4_CHANGE_TYPE_IS_VERSION_COUNTER_NOPNFS = 2, 1944 NFS4_CHANGE_TYPE_IS_TIME_METADATA = 3, 1945 NFS4_CHANGE_TYPE_IS_UNDEFINED = 4 1946 }; 1948 change_attr_type is a per file system attribute which enables the 1949 NFSv4.2 server to provide additional information about how it expects 1950 the change attribute value to evolve after the file data, or metadata 1951 has changed. While Section 5.4 of [2] discusses per file system 1952 attributes, it is expected that the value of change_attr_type not 1953 depend on the value of "homogeneous" and only changes in the event of 1954 a migration. 1956 NFS4_CHANGE_TYPE_IS_UNDEFINED: The change attribute does not take 1957 values that fit into any of these categories. 1959 NFS4_CHANGE_TYPE_IS_MONOTONIC_INCR: The change attribute value MUST 1960 monotonically increase for every atomic change to the file 1961 attributes, data, or directory contents. 1963 NFS4_CHANGE_TYPE_IS_VERSION_COUNTER: The change attribute value MUST 1964 be incremented by one unit for every atomic change to the file 1965 attributes, data, or directory contents. This property is 1966 preserved when writing to pNFS data servers. 1968 NFS4_CHANGE_TYPE_IS_VERSION_COUNTER_NOPNFS: The change attribute 1969 value MUST be incremented by one unit for every atomic change to 1970 the file attributes, data, or directory contents. In the case 1971 where the client is writing to pNFS data servers, the number of 1972 increments is not guaranteed to exactly match the number of 1973 writes. 1975 NFS4_CHANGE_TYPE_IS_TIME_METADATA: The change attribute is 1976 implemented as suggested in the NFSv4 spec [10] in terms of the 1977 time_metadata attribute. 1979 If either NFS4_CHANGE_TYPE_IS_MONOTONIC_INCR, 1980 NFS4_CHANGE_TYPE_IS_VERSION_COUNTER, or 1981 NFS4_CHANGE_TYPE_IS_TIME_METADATA are set, then the client knows at 1982 the very least that the change attribute is monotonically increasing, 1983 which is sufficient to resolve the question of which value is the 1984 most recent. 1986 If the client sees the value NFS4_CHANGE_TYPE_IS_TIME_METADATA, then 1987 by inspecting the value of the 'time_delta' attribute it additionally 1988 has the option of detecting rogue server implementations that use 1989 time_metadata in violation of the spec. 1991 If the client sees NFS4_CHANGE_TYPE_IS_VERSION_COUNTER, it has the 1992 ability to predict what the resulting change attribute value should 1993 be after a COMPOUND containing a SETATTR, WRITE, or CREATE. This 1994 again allows it to detect changes made in parallel by another client. 1995 The value NFS4_CHANGE_TYPE_IS_VERSION_COUNTER_NOPNFS permits the 1996 same, but only if the client is not doing pNFS WRITEs. 1998 Finally, if the server does not support change_attr_type or if 1999 NFS4_CHANGE_TYPE_IS_UNDEFINED is set, then the server SHOULD make an 2000 effort to implement the change attribute in terms of the 2001 time_metadata attribute. 2003 11.2.2. Attribute 80: sec_label 2005 typedef uint32_t policy4; 2006 struct labelformat_spec4 { 2007 policy4 lfs_lfs; 2008 policy4 lfs_pi; 2009 }; 2011 struct sec_label4 { 2012 labelformat_spec4 slai_lfs; 2013 opaque slai_data<>; 2014 }; 2016 The FATTR4_SEC_LABEL contains an array of two components with the 2017 first component being an LFS. It serves to provide the receiving end 2018 with the information necessary to translate the security attribute 2019 into a form that is usable by the endpoint. Label Formats assigned 2020 an LFS may optionally choose to include a Policy Identifier field to 2021 allow for complex policy deployments. The LFS and Label Format 2022 Registry are described in detail in [21]. The translation used to 2023 interpret the security attribute is not specified as part of the 2024 protocol as it may depend on various factors. The second component 2025 is an opaque section which contains the data of the attribute. This 2026 component is dependent on the MAC model to interpret and enforce. 2028 In particular, it is the responsibility of the LFS specification to 2029 define a maximum size for the opaque section, slai_data<>. When 2030 creating or modifying a label for an object, the client needs to be 2031 guaranteed that the server will accept a label that is sized 2032 correctly. By both client and server being part of a specific MAC 2033 model, the client will be aware of the size. 2035 11.2.3. Attribute 81: change_sec_label 2037 typedef uint32_t change_sec_label4; 2039 The change_sec_label attribute is a read-only attribute per file. 2040 When the file is created, the value of change_sec_label is set to 0. 2041 Each time the sec_label is changed, the server MUST increment the 2042 value of change_sec_label by one. As the sec_label is not bounded by 2043 size, this attribute allows for VERIFY and NVERIFY to quickly 2044 determine if the sec_label has been modified. 2046 11.2.4. Attribute 77: space_reserved 2048 The space_reserve attribute is a read/write attribute of type 2049 boolean. It is a per file attribute. When the space_reserved 2050 attribute is set via SETATTR, the server must ensure that there is 2051 disk space to accommodate every byte in the file before it can return 2052 success. If the server cannot guarantee this, it must return 2053 NFS4ERR_NOSPC. 2055 If the client tries to grow a file which has the space_reserved 2056 attribute set, the server must guarantee that there is disk space to 2057 accommodate every byte in the file with the new size before it can 2058 return success. If the server cannot guarantee this, it must return 2059 NFS4ERR_NOSPC. 2061 It is not required that the server allocate the space to the file 2062 before returning success. The allocation can be deferred, however, 2063 it must be guaranteed that it will not fail for lack of space. 2065 The value of space_reserved can be obtained at any time through 2066 GETATTR. 2068 In order to avoid ambiguity, the space_reserve bit cannot be set 2069 along with the size bit in SETATTR. Increasing the size of a file 2070 with space_reserve set will fail if space reservation cannot be 2071 guaranteed for the new size. If the file size is decreased, space 2072 reservation is only guaranteed for the new size and the extra blocks 2073 backing the file can be released. 2075 11.2.5. Attribute 78: space_freed 2077 space_freed gives the number of bytes freed if the file is deleted. 2078 This attribute is read only and is of type length4. It is a per file 2079 attribute. 2081 12. Operations: REQUIRED, RECOMMENDED, or OPTIONAL 2083 The following tables summarize the operations of the NFSv4.2 protocol 2084 and the corresponding designation of REQUIRED, RECOMMENDED, and 2085 OPTIONAL to implement or either OBSOLETE if implemented or MUST NOT 2086 implement. The designation of OBSOLETE if implemented is reserved 2087 for those operations which are defined in either NFSv4.0 or NFSV4.1, 2088 can be implemented in NFSv4.2, and are intended to be MUST NOT be 2089 implemented in NFSv4.3. The designation of MUST NOT implement is 2090 reserved for those operations that were defined in either NFSv4.0 or 2091 NFSV4.1 and MUST NOT be implemented in NFSv4.2. 2093 For the most part, the REQUIRED, RECOMMENDED, or OPTIONAL designation 2094 for operations sent by the client is for the server implementation. 2095 The client is generally required to implement the operations needed 2096 for the operating environment for which it serves. For example, a 2097 read-only NFSv4.2 client would have no need to implement the WRITE 2098 operation and is not required to do so. 2100 The REQUIRED or OPTIONAL designation for callback operations sent by 2101 the server is for both the client and server. Generally, the client 2102 has the option of creating the backchannel and sending the operations 2103 on the fore channel that will be a catalyst for the server sending 2104 callback operations. A partial exception is CB_RECALL_SLOT; the only 2105 way the client can avoid supporting this operation is by not creating 2106 a backchannel. 2108 Since this is a summary of the operations and their designation, 2109 there are subtleties that are not presented here. Therefore, if 2110 there is a question of the requirements of implementation, the 2111 operation descriptions themselves must be consulted along with other 2112 relevant explanatory text within this either specification or that of 2113 NFSv4.1 [2]. 2115 The abbreviations used in the second and third columns of the table 2116 are defined as follows. 2118 REQ REQUIRED to implement 2120 REC RECOMMEND to implement 2122 OPT OPTIONAL to implement 2124 OBS MUST NOT implement 2126 MNI MUST NOT implement 2128 For the NFSv4.2 features that are OPTIONAL, the operations that 2129 support those features are OPTIONAL, and the server would return 2130 NFS4ERR_NOTSUPP in response to the client's use of those operations. 2131 If an OPTIONAL feature is supported, it is possible that a set of 2132 operations related to the feature become REQUIRED to implement. The 2133 third column of the table designates the feature(s) and if the 2134 operation is REQUIRED or OPTIONAL in the presence of support for the 2135 feature. 2137 The OPTIONAL features identified and their abbreviations are as 2138 follows: 2140 pNFS Parallel NFS 2142 FDELG File Delegations 2144 DDELG Directory Delegations 2145 COPY Server Side Copy 2147 ADH Application Data Holes 2149 Operations 2151 +----------------------+--------------------+-----------------------+ 2152 | Operation | REQ, REC, OPT, or | Feature (REQ, REC, or | 2153 | | MNI | OPT) | 2154 +----------------------+--------------------+-----------------------+ 2155 | ACCESS | REQ | | 2156 | BACKCHANNEL_CTL | REQ | | 2157 | BIND_CONN_TO_SESSION | REQ | | 2158 | CLOSE | REQ | | 2159 | COMMIT | REQ | | 2160 | COPY | OPT | COPY (REQ) | 2161 | OFFLOAD_ABORT | OPT | COPY (REQ) | 2162 | COPY_NOTIFY | OPT | COPY (REQ) | 2163 | OFFLOAD_REVOKE | OPT | COPY (REQ) | 2164 | OFFLOAD_STATUS | OPT | COPY (REQ) | 2165 | CREATE | REQ | | 2166 | CREATE_SESSION | REQ | | 2167 | DELEGPURGE | OPT | FDELG (REQ) | 2168 | DELEGRETURN | OPT | FDELG, DDELG, pNFS | 2169 | | | (REQ) | 2170 | DESTROY_CLIENTID | REQ | | 2171 | DESTROY_SESSION | REQ | | 2172 | EXCHANGE_ID | REQ | | 2173 | FREE_STATEID | REQ | | 2174 | GETATTR | REQ | | 2175 | GETDEVICEINFO | OPT | pNFS (REQ) | 2176 | GETDEVICELIST | OPT | pNFS (OPT) | 2177 | GETFH | REQ | | 2178 | INITIALIZE | OPT | ADH (REQ) | 2179 | GET_DIR_DELEGATION | OPT | DDELG (REQ) | 2180 | LAYOUTCOMMIT | OPT | pNFS (REQ) | 2181 | LAYOUTGET | OPT | pNFS (REQ) | 2182 | LAYOUTRETURN | OPT | pNFS (REQ) | 2183 | LINK | OPT | | 2184 | LOCK | REQ | | 2185 | LOCKT | REQ | | 2186 | LOCKU | REQ | | 2187 | LOOKUP | REQ | | 2188 | LOOKUPP | REQ | | 2189 | NVERIFY | REQ | | 2190 | OPEN | REQ | | 2191 | OPENATTR | OPT | | 2192 | OPEN_CONFIRM | MNI | | 2193 | OPEN_DOWNGRADE | REQ | | 2194 | PUTFH | REQ | | 2195 | PUTPUBFH | REQ | | 2196 | PUTROOTFH | REQ | | 2197 | READ | OBS | | 2198 | READDIR | REQ | | 2199 | READLINK | OPT | | 2200 | READ_PLUS | OPT | ADH (REQ) | 2201 | RECLAIM_COMPLETE | REQ | | 2202 | RELEASE_LOCKOWNER | MNI | | 2203 | REMOVE | REQ | | 2204 | RENAME | REQ | | 2205 | RENEW | MNI | | 2206 | RESTOREFH | REQ | | 2207 | SAVEFH | REQ | | 2208 | SECINFO | REQ | | 2209 | SECINFO_NO_NAME | REC | pNFS file layout | 2210 | | | (REQ) | 2211 | SEQUENCE | REQ | | 2212 | SETATTR | REQ | | 2213 | SETCLIENTID | MNI | | 2214 | SETCLIENTID_CONFIRM | MNI | | 2215 | SET_SSV | REQ | | 2216 | TEST_STATEID | REQ | | 2217 | VERIFY | REQ | | 2218 | WANT_DELEGATION | OPT | FDELG (OPT) | 2219 | WRITE | REQ | | 2220 +----------------------+--------------------+-----------------------+ 2221 Callback Operations 2223 +-------------------------+-------------------+---------------------+ 2224 | Operation | REQ, REC, OPT, or | Feature (REQ, REC, | 2225 | | MNI | or OPT) | 2226 +-------------------------+-------------------+---------------------+ 2227 | CB_COPY | OPT | COPY (REQ) | 2228 | CB_GETATTR | OPT | FDELG (REQ) | 2229 | CB_LAYOUTRECALL | OPT | pNFS (REQ) | 2230 | CB_NOTIFY | OPT | DDELG (REQ) | 2231 | CB_NOTIFY_DEVICEID | OPT | pNFS (OPT) | 2232 | CB_NOTIFY_LOCK | OPT | | 2233 | CB_PUSH_DELEG | OPT | FDELG (OPT) | 2234 | CB_RECALL | OPT | FDELG, DDELG, pNFS | 2235 | | | (REQ) | 2236 | CB_RECALL_ANY | OPT | FDELG, DDELG, pNFS | 2237 | | | (REQ) | 2238 | CB_RECALL_SLOT | REQ | | 2239 | CB_RECALLABLE_OBJ_AVAIL | OPT | DDELG, pNFS (REQ) | 2240 | CB_SEQUENCE | OPT | FDELG, DDELG, pNFS | 2241 | | | (REQ) | 2242 | CB_WANTS_CANCELLED | OPT | FDELG, DDELG, pNFS | 2243 | | | (REQ) | 2244 +-------------------------+-------------------+---------------------+ 2246 13. NFSv4.2 Operations 2248 13.1. Operation 59: COPY - Initiate a server-side copy 2250 13.1.1. ARGUMENT 2252 const COPY4_GUARDED = 0x00000001; 2253 const COPY4_METADATA = 0x00000002; 2255 struct COPY4args { 2256 /* SAVED_FH: source file */ 2257 /* CURRENT_FH: destination file or */ 2258 /* directory */ 2259 stateid4 ca_src_stateid; 2260 stateid4 ca_dst_stateid; 2261 offset4 ca_src_offset; 2262 offset4 ca_dst_offset; 2263 length4 ca_count; 2264 uint32_t ca_flags; 2265 component4 ca_destination; 2266 netloc4 ca_source_server<>; 2268 }; 2270 13.1.2. RESULT 2272 union COPY4res switch (nfsstat4 cr_status) { 2273 case NFS4_OK: 2274 stateid4 cr_callback_id<1>; 2275 default: 2276 length4 cr_bytes_copied; 2277 }; 2279 13.1.3. DESCRIPTION 2281 The COPY operation is used for both intra-server and inter-server 2282 copies. In both cases, the COPY is always sent from the client to 2283 the destination server of the file copy. The COPY operation requests 2284 that a file be copied from the location specified by the SAVED_FH 2285 value to the location specified by the combination of CURRENT_FH and 2286 ca_destination. 2288 The SAVED_FH must be a regular file. If SAVED_FH is not a regular 2289 file, the operation MUST fail and return NFS4ERR_WRONG_TYPE. 2291 In order to set SAVED_FH to the source file handle, the compound 2292 procedure requesting the COPY will include a sub-sequence of 2293 operations such as 2295 PUTFH source-fh 2296 SAVEFH 2298 If the request is for a server-to-server copy, the source-fh is a 2299 filehandle from the source server and the compound procedure is being 2300 executed on the destination server. In this case, the source-fh is a 2301 foreign filehandle on the server receiving the COPY request. If 2302 either PUTFH or SAVEFH checked the validity of the filehandle, the 2303 operation would likely fail and return NFS4ERR_STALE. 2305 If a server supports the server-to-server COPY feature, a PUTFH 2306 followed by a SAVEFH MUST NOT return NFS4ERR_STALE for either 2307 operation. These restrictions do not pose substantial difficulties 2308 for servers. The CURRENT_FH and SAVED_FH may be validated in the 2309 context of the operation referencing them and an NFS4ERR_STALE error 2310 returned for an invalid file handle at that point. 2312 For an intra-server copy, both the ca_src_stateid and ca_dst_stateid 2313 MUST refer to either open or locking states provided earlier by the 2314 server. If either stateid is invalid, then the operation MUST fail. 2315 If the request is for a inter-server copy, then the ca_src_stateid 2316 can be ignored. If ca_dst_stateid is invalid, then the operation 2317 MUST fail. 2319 The CURRENT_FH and ca_destination together specify the destination of 2320 the copy operation. If ca_destination is of 0 (zero) length, then 2321 CURRENT_FH specifies the target file. In this case, CURRENT_FH MUST 2322 be a regular file and not a directory. If ca_destination is not of 0 2323 (zero) length, the ca_destination argument specifies the file name to 2324 which the data will be copied within the directory identified by 2325 CURRENT_FH. In this case, CURRENT_FH MUST be a directory and not a 2326 regular file. 2328 If the file named by ca_destination does not exist and the operation 2329 completes successfully, the file will be visible in the file system 2330 namespace. If the file does not exist and the operation fails, the 2331 file MAY be visible in the file system namespace depending on when 2332 the failure occurs and on the implementation of the NFS server 2333 receiving the COPY operation. If the ca_destination name cannot be 2334 created in the destination file system (due to file name 2335 restrictions, such as case or length), the operation MUST fail. 2337 The ca_src_offset is the offset within the source file from which the 2338 data will be read, the ca_dst_offset is the offset within the 2339 destination file to which the data will be written, and the ca_count 2340 is the number of bytes that will be copied. An offset of 0 (zero) 2341 specifies the start of the file. A count of 0 (zero) requests that 2342 all bytes from ca_src_offset through EOF be copied to the 2343 destination. If concurrent modifications to the source file overlap 2344 with the source file region being copied, the data copied may include 2345 all, some, or none of the modifications. The client can use standard 2346 NFS operations (e.g., OPEN with OPEN4_SHARE_DENY_WRITE or mandatory 2347 byte range locks) to protect against concurrent modifications if the 2348 client is concerned about this. If the source file's end of file is 2349 being modified in parallel with a copy that specifies a count of 0 2350 (zero) bytes, the amount of data copied is implementation dependent 2351 (clients may guard against this case by specifying a non-zero count 2352 value or preventing modification of the source file as mentioned 2353 above). 2355 If the source offset or the source offset plus count is greater than 2356 or equal to the size of the source file, the operation will fail with 2357 NFS4ERR_INVAL. The destination offset or destination offset plus 2358 count may be greater than the size of the destination file. This 2359 allows for the client to issue parallel copies to implement 2360 operations such as "cat file1 file2 file3 file4 > dest". 2362 If the destination file is created as a result of this command, the 2363 destination file's size will be equal to the number of bytes 2364 successfully copied. If the destination file already existed, the 2365 destination file's size may increase as a result of this operation 2366 (e.g. if ca_dst_offset plus ca_count is greater than the 2367 destination's initial size). 2369 If the ca_source_server list is specified, then this is an inter- 2370 server copy operation and the source file is on a remote server. The 2371 client is expected to have previously issued a successful COPY_NOTIFY 2372 request to the remote source server. The ca_source_server list MUST 2373 be the same as the COPY_NOTIFY response's cnr_source_server list. If 2374 the client includes the entries from the COPY_NOTIFY response's 2375 cnr_source_server list in the ca_source_server list, the source 2376 server can indicate a specific copy protocol for the destination 2377 server to use by returning a URL, which specifies both a protocol 2378 service and server name. Server-to-server copy protocol 2379 considerations are described in Section 2.2.5 and Section 2.4.1. 2381 The ca_flags argument allows the copy operation to be customized in 2382 the following ways using the guarded flag (COPY4_GUARDED) and the 2383 metadata flag (COPY4_METADATA). 2385 If the guarded flag is set and the destination exists on the server, 2386 this operation will fail with NFS4ERR_EXIST. 2388 If the guarded flag is not set and the destination exists on the 2389 server, the behavior is implementation dependent. 2391 If the metadata flag is set and the client is requesting a whole file 2392 copy (i.e., ca_count is 0 (zero)), a subset of the destination file's 2393 attributes MUST be the same as the source file's corresponding 2394 attributes and a subset of the destination file's attributes SHOULD 2395 be the same as the source file's corresponding attributes. The 2396 attributes in the MUST and SHOULD copy subsets will be defined for 2397 each NFS version. 2399 For NFSv4.2, Table 3 and Table 4 list the REQUIRED and RECOMMENDED 2400 attributes respectively. In the "Copy to destination file?" column, 2401 a "MUST" indicates that the attribute is part of the MUST copy set. 2402 A "SHOULD" indicates that the attribute is part of the SHOULD copy 2403 set. A "no" indicates that the attribute MUST NOT be copied. 2405 REQUIRED attributes 2407 +--------------------+----+---------------------------+ 2408 | Name | Id | Copy to destination file? | 2409 +--------------------+----+---------------------------+ 2410 | supported_attrs | 0 | no | 2411 | type | 1 | MUST | 2412 | fh_expire_type | 2 | no | 2413 | change | 3 | SHOULD | 2414 | size | 4 | MUST | 2415 | link_support | 5 | no | 2416 | symlink_support | 6 | no | 2417 | named_attr | 7 | no | 2418 | fsid | 8 | no | 2419 | unique_handles | 9 | no | 2420 | lease_time | 10 | no | 2421 | rdattr_error | 11 | no | 2422 | filehandle | 19 | no | 2423 | suppattr_exclcreat | 75 | no | 2424 +--------------------+----+---------------------------+ 2426 Table 3 2428 RECOMMENDED attributes 2430 +--------------------+----+---------------------------+ 2431 | Name | Id | Copy to destination file? | 2432 +--------------------+----+---------------------------+ 2433 | acl | 12 | MUST | 2434 | aclsupport | 13 | no | 2435 | archive | 14 | no | 2436 | cansettime | 15 | no | 2437 | case_insensitive | 16 | no | 2438 | case_preserving | 17 | no | 2439 | change_attr_type | 79 | no | 2440 | change_policy | 60 | no | 2441 | chown_restricted | 18 | MUST | 2442 | dacl | 58 | MUST | 2443 | dir_notif_delay | 56 | no | 2444 | dirent_notif_delay | 57 | no | 2445 | fileid | 20 | no | 2446 | files_avail | 21 | no | 2447 | files_free | 22 | no | 2448 | files_total | 23 | no | 2449 | fs_charset_cap | 76 | no | 2450 | fs_layout_type | 62 | no | 2451 | fs_locations | 24 | no | 2452 | fs_locations_info | 67 | no | 2453 | fs_status | 61 | no | 2454 | hidden | 25 | MUST | 2455 | homogeneous | 26 | no | 2456 | layout_alignment | 66 | no | 2457 | layout_blksize | 65 | no | 2458 | layout_hint | 63 | no | 2459 | layout_type | 64 | no | 2460 | maxfilesize | 27 | no | 2461 | maxlink | 28 | no | 2462 | maxname | 29 | no | 2463 | maxread | 30 | no | 2464 | maxwrite | 31 | no | 2465 | mdsthreshold | 68 | no | 2466 | mimetype | 32 | MUST | 2467 | mode | 33 | MUST | 2468 | mode_set_masked | 74 | no | 2469 | mounted_on_fileid | 55 | no | 2470 | no_trunc | 34 | no | 2471 | numlinks | 35 | no | 2472 | owner | 36 | MUST | 2473 | owner_group | 37 | MUST | 2474 | quota_avail_hard | 38 | no | 2475 | quota_avail_soft | 39 | no | 2476 | quota_used | 40 | no | 2477 | rawdev | 41 | no | 2478 | retentevt_get | 71 | MUST | 2479 | retentevt_set | 72 | no | 2480 | retention_get | 69 | MUST | 2481 | retention_hold | 73 | MUST | 2482 | retention_set | 70 | no | 2483 | sacl | 59 | MUST | 2484 | sec_label | 80 | MUST | 2485 | space_avail | 42 | no | 2486 | space_free | 43 | no | 2487 | space_freed | 78 | no | 2488 | space_reserved | 77 | MUST | 2489 | space_total | 44 | no | 2490 | space_used | 45 | no | 2491 | system | 46 | MUST | 2492 | time_access | 47 | MUST | 2493 | time_access_set | 48 | no | 2494 | time_backup | 49 | no | 2495 | time_create | 50 | MUST | 2496 | time_delta | 51 | no | 2497 | time_metadata | 52 | SHOULD | 2498 | time_modify | 53 | MUST | 2499 | time_modify_set | 54 | no | 2500 +--------------------+----+---------------------------+ 2501 Table 4 2503 [NOTE: The source file's attribute values will take precedence over 2504 any attribute values inherited by the destination file.] 2506 In the case of an inter-server copy or an intra-server copy between 2507 file systems, the attributes supported for the source file and 2508 destination file could be different. By definition,the REQUIRED 2509 attributes will be supported in all cases. If the metadata flag is 2510 set and the source file has a RECOMMENDED attribute that is not 2511 supported for the destination file, the copy MUST fail with 2512 NFS4ERR_ATTRNOTSUPP. 2514 Any attribute supported by the destination server that is not set on 2515 the source file SHOULD be left unset. 2517 Metadata attributes not exposed via the NFS protocol SHOULD be copied 2518 to the destination file where appropriate. 2520 The destination file's named attributes are not duplicated from the 2521 source file. After the copy process completes, the client MAY 2522 attempt to duplicate named attributes using standard NFSv4 2523 operations. However, the destination file's named attribute 2524 capabilities MAY be different from the source file's named attribute 2525 capabilities. 2527 If the metadata flag is not set and the client is requesting a whole 2528 file copy (i.e., ca_count is 0 (zero)), the destination file's 2529 metadata is implementation dependent. 2531 If the client is requesting a partial file copy (i.e., ca_count is 2532 not 0 (zero)), the client SHOULD NOT set the metadata flag and the 2533 server MUST ignore the metadata flag. 2535 If the operation does not result in an immediate failure, the server 2536 will return NFS4_OK, and the CURRENT_FH will remain the destination's 2537 filehandle. 2539 If an immediate failure does occur, cr_bytes_copied will be set to 2540 the number of bytes copied to the destination file before the error 2541 occurred. The cr_bytes_copied value indicates the number of bytes 2542 copied but not which specific bytes have been copied. 2544 A return of NFS4_OK indicates that either the operation is complete 2545 or the operation was initiated and a callback will be used to deliver 2546 the final status of the operation. 2548 If the cr_callback_id is returned, this indicates that the operation 2549 was initiated and a CB_COPY callback will deliver the final results 2550 of the operation. The cr_callback_id stateid is termed a copy 2551 stateid in this context. The server is given the option of returning 2552 the results in a callback because the data may require a relatively 2553 long period of time to copy. 2555 If no cr_callback_id is returned, the operation completed 2556 synchronously and no callback will be issued by the server. The 2557 completion status of the operation is indicated by cr_status. 2559 If the copy completes successfully, either synchronously or 2560 asynchronously, the data copied from the source file to the 2561 destination file MUST appear identical to the NFS client. However, 2562 the NFS server's on disk representation of the data in the source 2563 file and destination file MAY differ. For example, the NFS server 2564 might encrypt, compress, deduplicate, or otherwise represent the on 2565 disk data in the source and destination file differently. 2567 In the event of a failure the state of the destination file is 2568 implementation dependent. The COPY operation may fail for the 2569 following reasons (this is a partial list). 2571 o NFS4ERR_MOVED 2573 o NFS4ERR_NOTSUPP 2575 o NFS4ERR_PARTNER_NOTSUPP 2577 o NFS4ERR_OFFLOAD_DENIED 2579 o NFS4ERR_PARTNER_NO_AUTH 2581 o NFS4ERR_FBIG 2583 o NFS4ERR_NOTDIR 2585 o NFS4ERR_WRONG_TYPE 2587 o NFS4ERR_ISDIR 2589 o NFS4ERR_INVAL 2591 o NFS4ERR_DELAY 2593 o NFS4ERR_METADATA_NOTSUPP 2595 o NFS4ERR_WRONGSEC 2597 13.2. Operation 60: OFFLOAD_ABORT - Cancel a server-side copy 2599 13.2.1. ARGUMENT 2601 struct OFFLOAD_ABORT4args { 2602 /* CURRENT_FH: destination file */ 2603 stateid4 oaa_stateid; 2604 }; 2606 13.2.2. RESULT 2608 struct OFFLOAD_ABORT4res { 2609 nfsstat4 oar_status; 2610 }; 2612 13.2.3. DESCRIPTION 2614 OFFLOAD_ABORT is used for both intra- and inter-server asynchronous 2615 copies. The OFFLOAD_ABORT operation allows the client to cancel a 2616 server-side copy operation that it initiated. This operation is sent 2617 in a COMPOUND request from the client to the destination server. 2618 This operation may be used to cancel a copy when the application that 2619 requested the copy exits before the operation is completed or for 2620 some other reason. 2622 The request contains the filehandle and copy stateid cookies that act 2623 as the context for the previously initiated copy operation. 2625 The result's oar_status field indicates whether the cancel was 2626 successful or not. A value of NFS4_OK indicates that the copy 2627 operation was canceled and no callback will be issued by the server. 2628 A copy operation that is successfully canceled may result in none, 2629 some, or all of the data and/or metadata copied. 2631 If the server supports asynchronous copies, the server is REQUIRED to 2632 support the OFFLOAD_ABORT operation. 2634 The OFFLOAD_ABORT operation may fail for the following reasons (this 2635 is a partial list): 2637 o NFS4ERR_NOTSUPP 2639 o NFS4ERR_RETRY 2641 o NFS4ERR_COMPLETE_ALREADY 2642 o NFS4ERR_SERVERFAULT 2644 13.3. Operation 61: COPY_NOTIFY - Notify a source server of a future 2645 copy 2647 13.3.1. ARGUMENT 2649 struct COPY_NOTIFY4args { 2650 /* CURRENT_FH: source file */ 2651 stateid4 cna_src_stateid; 2652 netloc4 cna_destination_server; 2653 }; 2655 13.3.2. RESULT 2657 struct COPY_NOTIFY4resok { 2658 nfstime4 cnr_lease_time; 2659 netloc4 cnr_source_server<>; 2660 }; 2662 union COPY_NOTIFY4res switch (nfsstat4 cnr_status) { 2663 case NFS4_OK: 2664 COPY_NOTIFY4resok resok4; 2665 default: 2666 void; 2667 }; 2669 13.3.3. DESCRIPTION 2671 This operation is used for an inter-server copy. A client sends this 2672 operation in a COMPOUND request to the source server to authorize a 2673 destination server identified by cna_destination_server to read the 2674 file specified by CURRENT_FH on behalf of the given user. 2676 The cna_src_stateid MUST refer to either open or locking states 2677 provided earlier by the server. If it is invalid, then the operation 2678 MUST fail. 2680 The cna_destination_server MUST be specified using the netloc4 2681 network location format. The server is not required to resolve the 2682 cna_destination_server address before completing this operation. 2684 If this operation succeeds, the source server will allow the 2685 cna_destination_server to copy the specified file on behalf of the 2686 given user as long as both of the following conditions are met: 2688 o The destination server begins reading the source file before the 2689 cnr_lease_time expires. If the cnr_lease_time expires while the 2690 destination server is still reading the source file, the 2691 destination server is allowed to finish reading the file. 2693 o The client has not issued a COPY_REVOKE for the same combination 2694 of user, filehandle, and destination server. 2696 The cnr_lease_time is chosen by the source server. A cnr_lease_time 2697 of 0 (zero) indicates an infinite lease. To avoid the need for 2698 synchronized clocks, copy lease times are granted by the server as a 2699 time delta. To renew the copy lease time the client should resend 2700 the same copy notification request to the source server. 2702 A successful response will also contain a list of netloc4 network 2703 location formats called cnr_source_server, on which the source is 2704 willing to accept connections from the destination. These might not 2705 be reachable from the client and might be located on networks to 2706 which the client has no connection. 2708 If the client wishes to perform an inter-server copy, the client MUST 2709 send a COPY_NOTIFY to the source server. Therefore, the source 2710 server MUST support COPY_NOTIFY. 2712 For a copy only involving one server (the source and destination are 2713 on the same server), this operation is unnecessary. 2715 The COPY_NOTIFY operation may fail for the following reasons (this is 2716 a partial list): 2718 o NFS4ERR_MOVED 2720 o NFS4ERR_NOTSUPP 2722 o NFS4ERR_WRONGSEC 2724 13.4. Operation 62: OFFLOAD_REVOKE - Revoke a destination server's copy 2725 privileges 2727 13.4.1. ARGUMENT 2729 struct OFFLOAD_REVOKE4args { 2730 /* CURRENT_FH: source file */ 2731 netloc4 ora_destination_server; 2732 }; 2734 13.4.2. RESULT 2736 struct OFFLOAD_REVOKE4res { 2737 nfsstat4 orr_status; 2738 }; 2740 13.4.3. DESCRIPTION 2742 This operation is used for an inter-server copy. A client sends this 2743 operation in a COMPOUND request to the source server to revoke the 2744 authorization of a destination server identified by 2745 ora_destination_server from reading the file specified by CURRENT_FH 2746 on behalf of given user. If the ora_destination_server has already 2747 begun copying the file, a successful return from this operation 2748 indicates that further access will be prevented. 2750 The ora_destination_server MUST be specified using the netloc4 2751 network location format. The server is not required to resolve the 2752 ora_destination_server address before completing this operation. 2754 The client uses OFFLOAD_ABORT to inform the destination to stop the 2755 active transfer and OFFLOAD_REVOKE to inform the source to not allow 2756 any more copy requests from the destination. The OFFLOAD_REVOKE 2757 operation is also useful in situations in which the source server 2758 granted a very long or infinite lease on the destination server's 2759 ability to read the source file and all copy operations on the source 2760 file have been completed. 2762 For a copy only involving one server (the source and destination are 2763 on the same server), this operation is unnecessary. 2765 If the server supports COPY_NOTIFY, the server is REQUIRED to support 2766 the OFFLOAD_REVOKE operation. 2768 The OFFLOAD_REVOKE operation may fail for the following reasons (this 2769 is a partial list): 2771 o NFS4ERR_MOVED 2773 o NFS4ERR_NOTSUPP 2775 13.5. Operation 63: OFFLOAD_STATUS - Poll for status of a server-side 2776 copy 2778 13.5.1. ARGUMENT 2780 struct OFFLOAD_STATUS4args { 2781 /* CURRENT_FH: destination file */ 2782 stateid4 osa_stateid; 2783 }; 2785 13.5.2. RESULT 2787 struct OFFLOAD_STATUS4resok { 2788 length4 osr_bytes_copied; 2789 nfsstat4 osr_complete<1>; 2790 }; 2792 union OFFLOAD_STATUS4res switch (nfsstat4 osr_status) { 2793 case NFS4_OK: 2794 OFFLOAD_STATUS4resok resok4; 2795 default: 2796 void; 2797 }; 2799 13.5.3. DESCRIPTION 2801 OFFLOAD_STATUS is used for both intra- and inter-server asynchronous 2802 copies. The OFFLOAD_STATUS operation allows the client to poll the 2803 destination server to determine the status of an asynchronous copy 2804 operation. 2806 If this operation is successful, the number of bytes copied are 2807 returned to the client in the osr_bytes_copied field. The 2808 osr_bytes_copied value indicates the number of bytes copied but not 2809 which specific bytes have been copied. 2811 If the optional osr_complete field is present, the copy has 2812 completed. In this case the status value indicates the result of the 2813 asynchronous copy operation. In all cases, the server will also 2814 deliver the final results of the asynchronous copy in a CB_COPY 2815 operation. 2817 The failure of this operation does not indicate the result of the 2818 asynchronous copy in any way. 2820 If the server supports asynchronous copies, the server is REQUIRED to 2821 support the OFFLOAD_STATUS operation. 2823 The OFFLOAD_STATUS operation may fail for the following reasons (this 2824 is a partial list): 2826 o NFS4ERR_NOTSUPP 2828 o NFS4ERR_BAD_STATEID 2830 o NFS4ERR_EXPIRED 2832 13.6. Modification to Operation 42: EXCHANGE_ID - Instantiate Client ID 2834 13.6.1. ARGUMENT 2836 /* new */ 2837 const EXCHGID4_FLAG_SUPP_FENCE_OPS = 0x00000004; 2839 13.6.2. RESULT 2841 Unchanged 2843 13.6.3. MOTIVATION 2845 Enterprise applications require guarantees that an operation has 2846 either aborted or completed. NFSv4.1 provides this guarantee as long 2847 as the session is alive: simply send a SEQUENCE operation on the same 2848 slot with a new sequence number, and the successful return of 2849 SEQUENCE indicates the previous operation has completed. However, if 2850 the session is lost, there is no way to know when any in progress 2851 operations have aborted or completed. In hindsight, the NFSv4.1 2852 specification should have mandated that DESTROY_SESSION either abort 2853 or complete all outstanding operations. 2855 13.6.4. DESCRIPTION 2857 A client SHOULD request the EXCHGID4_FLAG_SUPP_FENCE_OPS capability 2858 when it sends an EXCHANGE_ID operation. The server SHOULD set this 2859 capability in the EXCHANGE_ID reply whether the client requests it or 2860 not. It is the server's return that determines whether this 2861 capability is in effect. When it is in effect, the following will 2862 occur: 2864 o The server will not reply to any DESTROY_SESSION invoked with the 2865 client ID until all operations in progress are completed or 2866 aborted. 2868 o The server will not reply to subsequent EXCHANGE_ID invoked on the 2869 same client owner with a new verifier until all operations in 2870 progress on the client ID's session are completed or aborted. 2872 o The NFS server SHOULD support client ID trunking, and if it does 2873 and the EXCHGID4_FLAG_SUPP_FENCE_OPS capability is enabled, then a 2874 session ID created on one node of the storage cluster MUST be 2875 destroyable via DESTROY_SESSION. In addition, DESTROY_CLIENTID 2876 and an EXCHANGE_ID with a new verifier affects all sessions 2877 regardless what node the sessions were created on. 2879 13.7. Operation 64: INITIALIZE 2881 This operation can be used to initialize the structure imposed by an 2882 application onto a file, i.e., ADHs, and to punch a hole into a file. 2884 13.7.1. ARGUMENT 2886 struct data_info4 { 2887 offset4 di_offset; 2888 length4 di_length; 2889 bool di_allocated; 2890 }; 2892 /* 2893 * We use data_content4 in case we wish to 2894 * extend new types later. Note that we 2895 * are explicitly disallowing data. 2896 */ 2897 union initialize_arg4 switch (data_content4 content) { 2898 case NFS4_CONTENT_APP_DATA_HOLE: 2899 app_data_hole4 ia_adh; 2900 case NFS4_CONTENT_HOLE: 2901 data_info4 ia_hole; 2902 default: 2903 void; 2904 }; 2906 struct INITIALIZE4args { 2907 /* CURRENT_FH: file */ 2908 stateid4 ia_stateid; 2909 stable_how4 ia_stable; 2910 initialize_arg4 ia_data<>; 2911 }; 2913 13.7.2. RESULT 2915 struct INITIALIZE4resok { 2916 count4 ir_count; 2917 stable_how4 ir_committed; 2918 verifier4 ir_writeverf; 2919 data_content4 ir_sparse; 2920 }; 2922 union INITIALIZE4res switch (nfsstat4 status) { 2923 case NFS4_OK: 2924 INITIALIZE4resok resok4; 2925 default: 2926 void; 2927 }; 2929 13.7.3. DESCRIPTION 2931 Using the data_content4 (Section 6.1.2), INITIALIZE can be used 2932 either to punch holes or to impose ADH structure on a file. 2934 13.7.3.1. Hole punching 2936 Whenever a client wishes to zero the blocks backing a particular 2937 region in the file, it calls the INITIALIZE operation with the 2938 current filehandle set to the filehandle of the file in question, and 2939 the equivalent of start offset and length in bytes of the region set 2940 in ia_hole.di_offset and ia_hole.di_length respectively. If the 2941 ia_hole.di_allocated is set to TRUE, then the blocks will be zeroed 2942 and if it is set to FALSE, then they will be deallocated. All 2943 further reads to this region MUST return zeros until overwritten. 2944 The filehandle specified must be that of a regular file. 2946 Situations may arise where di_offset and/or di_offset + di_length 2947 will not be aligned to a boundary that the server does allocations/ 2948 deallocations in. For most file systems, this is the block size of 2949 the file system. In such a case, the server can deallocate as many 2950 bytes as it can in the region. The blocks that cannot be deallocated 2951 MUST be zeroed. Except for the block deallocation and maximum hole 2952 punching capability, a INITIALIZE operation is to be treated similar 2953 to a write of zeroes. 2955 The server is not required to complete deallocating the blocks 2956 specified in the operation before returning. It is acceptable to 2957 have the deallocation be deferred. In fact, INITIALIZE is merely a 2958 hint; it is valid for a server to return success without ever doing 2959 anything towards deallocating the blocks backing the region 2960 specified. However, any future reads to the region MUST return 2961 zeroes. 2963 If used to hole punch, INITIALIZE will result in the space_used 2964 attribute being decreased by the number of bytes that were 2965 deallocated. The space_freed attribute may or may not decrease, 2966 depending on the support and whether the blocks backing the specified 2967 range were shared or not. The size attribute will remain unchanged. 2969 The INITIALIZE operation MUST NOT change the space reservation 2970 guarantee of the file. While the server can deallocate the blocks 2971 specified by di_offset and di_length, future writes to this region 2972 MUST NOT fail with NFSERR_NOSPC. 2974 The INITIALIZE operation may fail for the following reasons (this is 2975 a partial list): 2977 NFS4ERR_NOTSUPP The Hole punch operations are not supported by the 2978 NFS server receiving this request. 2980 NFS4ERR_DIR The current filehandle is of type NF4DIR. 2982 NFS4ERR_SYMLINK The current filehandle is of type NF4LNK. 2984 NFS4ERR_WRONG_TYPE The current filehandle does not designate an 2985 ordinary file. 2987 13.7.3.2. ADHs 2989 If the server supports ADHs, then it MUST support the 2990 NFS4_CONTENT_APP_DATA_HOLE arm of the INITIALIZE operation. The 2991 server has no concept of the structure imposed by the application. 2992 It is only when the application writes to a section of the file does 2993 order get imposed. In order to detect corruption even before the 2994 application utilizes the file, the application will want to 2995 initialize a range of ADHs using INITIALIZE. 2997 For ADHs, when the client invokes the INITIALIZE operation, it has 2998 two desired results: 3000 1. The structure described by the app_data_block4 be imposed on the 3001 file. 3003 2. The contents described by the app_data_block4 be sparse. 3005 If the server supports the INITIALIZE operation, it still might not 3006 support sparse files. So if it receives the INITIALIZE operation, 3007 then it MUST populate the contents of the file with the initialized 3008 ADHs. 3010 If the data was already initialized, there are two interesting 3011 scenarios: 3013 1. The data blocks are allocated. 3015 2. Initializing in the middle of an existing ADH. 3017 If the data blocks were already allocated, then the INITIALIZE is a 3018 hole punch operation. If INITIALIZE supports sparse files, then the 3019 data blocks are to be deallocated. If not, then the data blocks are 3020 to be rewritten in the indicated ADH format. 3022 Since the server has no knowledge of ADHs, it should not report 3023 misaligned creation of ADHs. Even while it can detect them, it 3024 cannot disallow them, as the application might be in the process of 3025 changing the size of the ADHs. Thus the server must be prepared to 3026 handle an INITIALIZE into an existing ADH. 3028 This document does not mandate the manner in which the server stores 3029 ADHs sparsely for a file. However, if an INITIALIZE arrives that 3030 will force a new ADH to start inside an existing ADH then the server 3031 will have three ADHs instead of two. It will have one up to the new 3032 one for the INITIALIZE, one for the INITIALIZE, and one for after the 3033 INITIALIZE. Note that depending on server specific policies for 3034 block allocation, there may also be some physical blocks allocated to 3035 align the boundaries. 3037 13.8. Operation 67: IO_ADVISE - Application I/O access pattern hints 3038 13.8.1. ARGUMENT 3040 enum IO_ADVISE_type4 { 3041 IO_ADVISE4_NORMAL = 0, 3042 IO_ADVISE4_SEQUENTIAL = 1, 3043 IO_ADVISE4_SEQUENTIAL_BACKWARDS = 2, 3044 IO_ADVISE4_RANDOM = 3, 3045 IO_ADVISE4_WILLNEED = 4, 3046 IO_ADVISE4_WILLNEED_OPPORTUNISTIC = 5, 3047 IO_ADVISE4_DONTNEED = 6, 3048 IO_ADVISE4_NOREUSE = 7, 3049 IO_ADVISE4_READ = 8, 3050 IO_ADVISE4_WRITE = 9, 3051 IO_ADVISE4_INIT_PROXIMITY = 10 3052 }; 3054 struct IO_ADVISE4args { 3055 /* CURRENT_FH: file */ 3056 stateid4 iar_stateid; 3057 offset4 iar_offset; 3058 length4 iar_count; 3059 bitmap4 iar_hints; 3060 }; 3062 13.8.2. RESULT 3064 struct IO_ADVISE4resok { 3065 bitmap4 ior_hints; 3066 }; 3068 union IO_ADVISE4res switch (nfsstat4 _status) { 3069 case NFS4_OK: 3070 IO_ADVISE4resok resok4; 3071 default: 3072 void; 3073 }; 3075 13.8.3. DESCRIPTION 3077 The IO_ADVISE operation sends an I/O access pattern hint to the 3078 server for the owner of the stateid for a given byte range specified 3079 by iar_offset and iar_count. The byte range specified by iar_offset 3080 and iar_count need not currently exist in the file, but the iar_hints 3081 will apply to the byte range when it does exist. If iar_count is 0, 3082 all data following iar_offset is specified. The server MAY ignore 3083 the advice. 3085 The following are the allowed hints for a stateid holder: 3087 IO_ADVISE4_NORMAL There is no advice to give, this is the default 3088 behavior. 3090 IO_ADVISE4_SEQUENTIAL Expects to access the specified data 3091 sequentially from lower offsets to higher offsets. 3093 IO_ADVISE4_SEQUENTIAL BACKWARDS Expects to access the specified data 3094 sequentially from higher offsets to lower offsets. 3096 IO_ADVISE4_RANDOM Expects to access the specified data in a random 3097 order. 3099 IO_ADVISE4_WILLNEED Expects to access the specified data in the near 3100 future. 3102 IO_ADVISE4_WILLNEED_OPPORTUNISTIC Expects to possibly access the 3103 data in the near future. This is a speculative hint, and 3104 therefore the server should prefetch data or indirect blocks only 3105 if it can be done at a marginal cost. 3107 IO_ADVISE_DONTNEED Expects that it will not access the specified 3108 data in the near future. 3110 IO_ADVISE_NOREUSE Expects to access the specified data once and then 3111 not reuse it thereafter. 3113 IO_ADVISE4_READ Expects to read the specified data in the near 3114 future. 3116 IO_ADVISE4_WRITE Expects to write the specified data in the near 3117 future. 3119 IO_ADVISE4_INIT_PROXIMITY Informs the server that the data in the 3120 byte range remains important to the client. 3122 Since IO_ADVISE is a hint, a server SHOULD NOT return an error and 3123 invalidate a entire Compound request if one of the sent hints in 3124 iar_hints is not supported by the server. Also, the server MUST NOT 3125 return an error if the client sends contradictory hints to the 3126 server, e.g., IO_ADVISE4_SEQUENTIAL and IO_ADVISE4_RANDOM in a single 3127 IO_ADVISE operation. In these cases, the server MUST return success 3128 and a ior_hints value that indicates the hint it intends to 3129 implement. This may mean simply returning IO_ADVISE4_NORMAL. 3131 The ior_hints returned by the server is primarily for debugging 3132 purposes since the server is under no obligation to carry out the 3133 hints that it describes in the ior_hints result. In addition, while 3134 the server may have intended to implement the hints returned in 3135 ior_hints, as time progresses, the server may need to change its 3136 handling of a given file due to several reasons including, but not 3137 limited to, memory pressure, additional IO_ADVISE hints sent by other 3138 clients, and heuristically detected file access patterns. 3140 The server MAY return different advice than what the client 3141 requested. If it does, then this might be due to one of several 3142 conditions, including, but not limited to another client advising of 3143 a different I/O access pattern; a different I/O access pattern from 3144 another client that that the server has heuristically detected; or 3145 the server is not able to support the requested I/O access pattern, 3146 perhaps due to a temporary resource limitation. 3148 Each issuance of the IO_ADVISE operation overrides all previous 3149 issuances of IO_ADVISE for a given byte range. This effectively 3150 follows a strategy of last hint wins for a given stateid and byte 3151 range. 3153 Clients should assume that hints included in an IO_ADVISE operation 3154 will be forgotten once the file is closed. 3156 13.8.4. IMPLEMENTATION 3158 The NFS client may choose to issue an IO_ADVISE operation to the 3159 server in several different instances. 3161 The most obvious is in direct response to an application's execution 3162 of posix_fadvise(). In this case, IO_ADVISE4_WRITE and 3163 IO_ADVISE4_READ may be set based upon the type of file access 3164 specified when the file was opened. 3166 13.8.5. IO_ADVISE4_INIT_PROXIMITY 3168 The IO_ADVISE4_INIT_PROXIMITY hint is non-posix in origin and conveys 3169 that the client has recently accessed the byte range in its own 3170 cache. I.e., it has not accessed it on the server, but it has 3171 locally. When the server reaches resource exhaustion, knowing which 3172 data is more important allows the server to make better choices about 3173 which data to, for example purge from a cache, or move to secondary 3174 storage. It also informs the server which delegations are more 3175 important, since if delegations are working correctly, once delegated 3176 to a client and the client has read the content for that byte range, 3177 a server might never receive another read request for that byte 3178 range. 3180 This hint is also useful in the case of NFS clients which are network 3181 booting from a server. If the first client to be booted sends this 3182 hint, then it keeps the cache warm for the remaining clients. 3184 13.8.6. pNFS File Layout Data Type Considerations 3186 The IO_ADVISE considerations for pNFS are very similar to the COMMIT 3187 considerations for pNFS. That is, as with COMMIT, some NFS server 3188 implementations prefer IO_ADVISE be done on the DS, and some prefer 3189 it be done on the MDS. 3191 So for the file's layout type, it is proposed that NFSv4.2 include an 3192 additional hint NFL42_CARE_IO_ADVISE_THRU_MDS which is valid only on 3193 NFSv4.2 or higher. Any file's layout obtained with NFSv4.1 MUST NOT 3194 have NFL42_UFLG_IO_ADVISE_THRU_MDS set. Any file's layout obtained 3195 with NFSv4.2 MAY have NFL42_UFLG_IO_ADVISE_THRU_MDS set. If the 3196 client does not implement IO_ADVISE, then it MUST ignore 3197 NFL42_UFLG_IO_ADVISE_THRU_MDS. 3199 If NFL42_UFLG_IO_ADVISE_THRU_MDS is set, the client MUST send the 3200 IO_ADVISE operation to the MDS in order for it to be honored by the 3201 DS. Once the MDS receives the IO_ADVISE operation, it will 3202 communicate the advice to each DS. 3204 If NFL42_UFLG_IO_ADVISE_THRU_MDS is not set, then the client SHOULD 3205 send an IO_ADVISE operation to the appropriate DS for the specified 3206 byte range. While the client MAY always send IO_ADVISE to the MDS, 3207 if the server has not set NFL42_UFLG_IO_ADVISE_THRU_MDS, the client 3208 should expect that such an IO_ADVISE is futile. Note that a client 3209 SHOULD use the same set of arguments on each IO_ADVISE sent to a DS 3210 for the same open file reference. 3212 The server is not required to support different advice for different 3213 DS's with the same open file reference. 3215 13.8.6.1. Dense and Sparse Packing Considerations 3217 The IO_ADVISE operation MUST use the iar_offset and byte range as 3218 dictated by the presence or absence of NFL4_UFLG_DENSE. 3220 E.g., if NFL4_UFLG_DENSE is present, and a READ or WRITE to the DS 3221 for iar_offset 0 really means iar_offset 10000 in the logical file, 3222 then an IO_ADVISE for iar_offset 0 means iar_offset 10000. 3224 E.g., if NFL4_UFLG_DENSE is absent, then a READ or WRITE to the DS 3225 for iar_offset 0 really means iar_offset 0 in the logical file, then 3226 an IO_ADVISE for iar_offset 0 means iar_offset 0 in the logical file. 3228 E.g., if NFL4_UFLG_DENSE is present, the stripe unit is 1000 bytes 3229 and the stripe count is 10, and the dense DS file is serving 3230 iar_offset 0. A READ or WRITE to the DS for iar_offsets 0, 1000, 3231 2000, and 3000, really mean iar_offsets 10000, 20000, 30000, and 3232 40000 (implying a stripe count of 10 and a stripe unit of 1000), then 3233 an IO_ADVISE sent to the same DS with an iar_offset of 500, and a 3234 iar_count of 3000 means that the IO_ADVISE applies to these byte 3235 ranges of the dense DS file: 3237 - 500 to 999 3238 - 1000 to 1999 3239 - 2000 to 2999 3240 - 3000 to 3499 3242 I.e., the contiguous range 500 to 3499 as specified in IO_ADVISE. 3244 It also applies to these byte ranges of the logical file: 3246 - 10500 to 10999 (500 bytes) 3247 - 20000 to 20999 (1000 bytes) 3248 - 30000 to 30999 (1000 bytes) 3249 - 40000 to 40499 (500 bytes) 3250 (total 3000 bytes) 3252 E.g., if NFL4_UFLG_DENSE is absent, the stripe unit is 250 bytes, the 3253 stripe count is 4, and the sparse DS file is serving iar_offset 0. 3254 Then a READ or WRITE to the DS for iar_offsets 0, 1000, 2000, and 3255 3000, really mean iar_offsets 0, 1000, 2000, and 3000 in the logical 3256 file, keeping in mind that on the DS file,. byte ranges 250 to 999, 3257 1250 to 1999, 2250 to 2999, and 3250 to 3999 are not accessible. 3258 Then an IO_ADVISE sent to the same DS with an iar_offset of 500, and 3259 a iar_count of 3000 means that the IO_ADVISE applies to these byte 3260 ranges of the logical file and the sparse DS file: 3262 - 500 to 999 (500 bytes) - no effect 3263 - 1000 to 1249 (250 bytes) - effective 3264 - 1250 to 1999 (750 bytes) - no effect 3265 - 2000 to 2249 (250 bytes) - effective 3266 - 2250 to 2999 (750 bytes) - no effect 3267 - 3000 to 3249 (250 bytes) - effective 3268 - 3250 to 3499 (250 bytes) - no effect 3269 (subtotal 2250 bytes) - no effect 3270 (subtotal 750 bytes) - effective 3271 (grand total 3000 bytes) - no effect + effective 3273 If neither of the flags NFL42_UFLG_IO_ADVISE_THRU_MDS and 3274 NFL4_UFLG_DENSE are set in the layout, then any IO_ADVISE request 3275 sent to the data server with a byte range that overlaps stripe unit 3276 that the data server does not serve MUST NOT result in the status 3277 NFS4ERR_PNFS_IO_HOLE. Instead, the response SHOULD be successful and 3278 if the server applies IO_ADVISE hints on any stripe units that 3279 overlap with the specified range, those hints SHOULD be indicated in 3280 the response. 3282 13.9. Changes to Operation 51: LAYOUTRETURN 3284 13.9.1. Introduction 3286 In the pNFS description provided in [2], the client is not capable to 3287 relay an error code from the DS to the MDS. In the specification of 3288 the Objects-Based Layout protocol [9], use is made of the opaque 3289 lrf_body field of the LAYOUTRETURN argument to do such a relaying of 3290 error codes. In this section, we define a new data structure to 3291 enable the passing of error codes back to the MDS and provide some 3292 guidelines on what both the client and MDS should expect in such 3293 circumstances. 3295 There are two broad classes of errors, transient and persistent. The 3296 client SHOULD strive to only use this new mechanism to report 3297 persistent errors. It MUST be able to deal with transient issues by 3298 itself. Also, while the client might consider an issue to be 3299 persistent, it MUST be prepared for the MDS to consider such issues 3300 to be transient. A prime example of this is if the MDS fences off a 3301 client from either a stateid or a filehandle. The client will get an 3302 error from the DS and might relay either NFS4ERR_ACCESS or 3303 NFS4ERR_BAD_STATEID back to the MDS, with the belief that this is a 3304 hard error. If the MDS is informed by the client that there is an 3305 error, it can safely ignore that. For it, the mission is 3306 accomplished in that the client has returned a layout that the MDS 3307 had most likley recalled. 3309 The client might also need to inform the MDS that it cannot reach one 3310 or more of the DSes. While the MDS can detect the connectivity of 3311 both of these paths: 3313 o MDS to DS 3315 o MDS to client 3317 it cannot determine if the client and DS path is working. As with 3318 the case of the DS passing errors to the client, it must be prepared 3319 for the MDS to consider such outages as being transistory. 3321 The existing LAYOUTRETURN operation is extended by introducing a new 3322 data structure to report errors, layoutreturn_device_error4. Also, 3323 layoutreturn_device_error4 is introduced to enable an array of errors 3324 to be reported. 3326 13.9.2. ARGUMENT 3328 The ARGUMENT specification of the LAYOUTRETURN operation in section 3329 18.44.1 of [2] is augmented by the following XDR code [23]: 3331 struct layoutreturn_device_error4 { 3332 deviceid4 lrde_deviceid; 3333 nfsstat4 lrde_status; 3334 nfs_opnum4 lrde_opnum; 3335 }; 3337 struct layoutreturn_error_report4 { 3338 layoutreturn_device_error4 lrer_errors<>; 3339 }; 3341 13.9.3. RESULT 3343 The RESULT of the LAYOUTRETURN operation is unchanged; see section 3344 18.44.2 of [2]. 3346 13.9.4. DESCRIPTION 3348 The following text is added to the end of the LAYOUTRETURN operation 3349 DESCRIPTION in section 18.44.3 of [2]. 3351 When a client uses LAYOUTRETURN with a type of LAYOUTRETURN4_FILE, 3352 then if the lrf_body field is NULL, it indicates to the MDS that the 3353 client experienced no errors. If lrf_body is non-NULL, then the 3354 field references error information which is layout type specific. 3355 I.e., the Objects-Based Layout protocol can continue to utilize 3356 lrf_body as specified in [9]. For both Files-Based and Block-Based 3357 Layouts, the field references a layoutreturn_device_error4, which 3358 contains an array of layoutreturn_device_error4. 3360 Each individual layoutreturn_device_error4 descibes a single error 3361 associated with a DS, which is identfied via lrde_deviceid. The 3362 operation which returned the error is identified via lrde_opnum. 3363 Finally the NFS error value (nfsstat4) encountered is provided via 3364 lrde_status and may consist of the following error codes: 3366 NFS4ERR_NXIO: The client was unable to establish any communication 3367 with the DS. 3369 NFS4ERR_*: The client was able to establish communication with the 3370 DS and is returning one of the allowed error codes for the 3371 operation denoted by lrde_opnum. 3373 13.9.5. IMPLEMENTATION 3375 The following text is added to the end of the LAYOUTRETURN operation 3376 IMPLEMENTATION in section 18.4.4 of [2]. 3378 Clients are expected to tolerate transient storage device errors, and 3379 hence clients SHOULD NOT use the LAYOUTRETURN error handling for 3380 device access problems that may be transient. The methods by which a 3381 client decides whether a device access problem is transient vs. 3382 persistent are implementation-specific, but may include retrying I/Os 3383 to a data server under appropriate conditions. 3385 When an I/O fails to a storage device, the client SHOULD retry the 3386 failed I/O via the MDS. In this situation, before retrying the I/O, 3387 the client SHOULD return the layout, or the affected portion thereof, 3388 and SHOULD indicate which storage device or devices was problematic. 3389 The client needs to do this when the DS is being unresponsive in 3390 order to fence off any failed write attempts, and ensure that they do 3391 not end up overwriting any later data being written through the MDS. 3392 If the client does not do this, the MDS MAY issue a layout recall 3393 callback in order to perform the retried I/O. 3395 The client needs to be cognizant that since this error handling is 3396 optional in the MDS, the MDS may silently ignore this functionality. 3397 Also, as the MDS may consider some issues the client reports to be 3398 expected (see Section 13.9.1), the client might find it difficult to 3399 detect a MDS which has not implemented error handling via 3400 LAYOUTRETURN. 3402 If an MDS is aware that a storage device is proving problematic to a 3403 client, the MDS SHOULD NOT include that storage device in any pNFS 3404 layouts sent to that client. If the MDS is aware that a storage 3405 device is affecting many clients, then the MDS SHOULD NOT include 3406 that storage device in any pNFS layouts sent out. If a client asks 3407 for a new layout for the file from the MDS, it MUST be prepared for 3408 the MDS to return that storage device in the layout. The MDS might 3409 not have any choice in using the storage device, i.e., there might 3410 only be one possible layout for the system. Also, in the case of 3411 existing files, the MDS might have no choice in which storage devices 3412 to hand out to clients. 3414 The MDS is not required to indefinitely retain per-client storage 3415 device error information. An MDS is also not required to 3416 automatically reinstate use of a previously problematic storage 3417 device; administrative intervention may be required instead. 3419 13.10. Operation 65: READ_PLUS 3421 READ_PLUS is a new variant of the NFSv4.1 READ operation [2]. 3422 Besides being able to support all of the data semantics of READ, it 3423 can also be used by the server to return either holes or ADHs to the 3424 client. For holes, READ_PLUS extends the response to avoid returning 3425 data for portions of the file which are either initialized and 3426 contain no backing store or if the result would appear to be so. 3427 I.e., if the result was a data block composed entirely of zeros, then 3428 it is easier to return a hole. Returning data blocks of 3429 uninitialized data wastes computational and network resources, thus 3430 reducing performance. For ADHs, READ_PLUS is used to return the 3431 metadata describing the portions of the file which are either 3432 initialized and contain no backing store. 3434 If the client sends a READ operation, it is explicitly stating that 3435 it is neither supporting sparse files nor ADHs. So if a READ occurs 3436 on a sparse ADH or file, then the server must expand such data to be 3437 raw bytes. If a READ occurs in the middle of a hole or ADH, the 3438 server can only send back bytes starting from that offset. In 3439 contrast, if a READ_PLUS occurs in the middle of a hole or ADH, the 3440 server can send back a range which starts before the offset and 3441 extends past the range. 3443 READ is inefficient for transfer of sparse sections of the file. As 3444 such, READ is marked as OBSOLETE in NFSv4.2. Instead, a client 3445 should issue READ_PLUS. Note that as the client has no a priori 3446 knowledge of whether either an ADH or a hole is present or not, it 3447 should always use READ_PLUS. 3449 13.10.1. ARGUMENT 3451 struct READ_PLUS4args { 3452 /* CURRENT_FH: file */ 3453 stateid4 rpa_stateid; 3454 offset4 rpa_offset; 3455 count4 rpa_count; 3456 }; 3458 13.10.2. RESULT 3460 union read_plus_content switch (data_content4 content) { 3461 case NFS4_CONTENT_DATA: 3462 opaque rpc_data<>; 3463 case NFS4_CONTENT_APP_DATA_HOLE: 3464 app_data_hole4 rpc_adh; 3465 case NFS4_CONTENT_HOLE: 3466 data_info4 rpc_hole; 3467 default: 3468 void; 3469 }; 3471 /* 3472 * Allow a return of an array of contents. 3473 */ 3474 struct read_plus_res4 { 3475 bool rpr_eof; 3476 read_plus_content rpr_contents<>; 3477 }; 3479 union READ_PLUS4res switch (nfsstat4 status) { 3480 case NFS4_OK: 3481 read_plus_res4 resok4; 3482 default: 3483 void; 3484 }; 3486 13.10.3. DESCRIPTION 3488 The READ_PLUS operation is based upon the NFSv4.1 READ operation [2] 3489 and similarly reads data from the regular file identified by the 3490 current filehandle. 3492 The client provides a rpa_offset of where the READ_PLUS is to start 3493 and a rpa_count of how many bytes are to be read. A rpa_offset of 3494 zero means to read data starting at the beginning of the file. If 3495 rpa_offset is greater than or equal to the size of the file, the 3496 status NFS4_OK is returned with di_length (the data length) set to 3497 zero and eof set to TRUE. 3499 The READ_PLUS result is comprised of an array of rpr_contents, each 3500 of which describe a data_content4 type of data (Section 6.1.2). For 3501 NFSv4.2, the allowed values are data, ADH, and hole. A server is 3502 required to support the data type, but neither ADH nor hole. Both an 3503 ADH and a hole must be returned in its entirety - clients must be 3504 prepared to get more information than they requested. Both the start 3505 and the end of the hole may execeed what was requested. 3507 READ_PLUS has to support all of the errors which are returned by READ 3508 plus NFS4ERR_UNION_NOTSUPP. If the client asks for a hole and the 3509 server does not support that arm of the discriminated union, but does 3510 support one or more additional arms, it can signal to the client that 3511 it supports the operation, but not the arm with 3512 NFS4ERR_UNION_NOTSUPP. 3514 If the data to be returned is comprised entirely of zeros, then the 3515 server may elect to return that data as a hole. The server 3516 differentiates this to the client by setting di_allocated to TRUE in 3517 this case. Note that in such a scenario, the server is not required 3518 to determine the full extent of the "hole" - it does not need to 3519 determine where the zeros start and end. 3521 The server may elect to return adjacent elements of the same type. 3522 For example, the guard pattern or block size of an ADH might change, 3523 which would require adjacent elements of type ADH. Likewise if the 3524 server has a range of data comprised entirely of zeros and then a 3525 hole, it might want to return two adjacent holes to the client. 3527 If the client specifies a rpa_count value of zero, the READ_PLUS 3528 succeeds and returns zero bytes of data. In all situations, the 3529 server may choose to return fewer bytes than specified by the client. 3530 The client needs to check for this condition and handle the condition 3531 appropriately. 3533 If the client specifies an rpa_offset and rpa_count value that is 3534 entirely contained within a hole of the file, then the di_offset and 3535 di_length returned must be for the entire hole. This result is 3536 considered valid until the file is changed (detected via the change 3537 attribute). The server MUST provide the same semantics for the hole 3538 as if the client read the region and received zeroes; the implied 3539 holes contents lifetime MUST be exactly the same as any other read 3540 data. 3542 If the client specifies an rpa_offset and rpa_count value that begins 3543 in a non-hole of the file but extends into hole the server should 3544 return an array comprised of both data and a hole. The client MUST 3545 be prepared for the server to return a short read describing just the 3546 data. The client will then issue another READ_PLUS for the remaining 3547 bytes, which the server will respond with information about the hole 3548 in the file. 3550 Except when special stateids are used, the stateid value for a 3551 READ_PLUS request represents a value returned from a previous byte- 3552 range lock or share reservation request or the stateid associated 3553 with a delegation. The stateid identifies the associated owners if 3554 any and is used by the server to verify that the associated locks are 3555 still valid (e.g., have not been revoked). 3557 If the read ended at the end-of-file (formally, in a correctly formed 3558 READ_PLUS operation, if rpa_offset + rpa_count is equal to the size 3559 of the file), or the READ_PLUS operation extends beyond the size of 3560 the file (if rpa_offset + rpa_count is greater than the size of the 3561 file), eof is returned as TRUE; otherwise, it is FALSE. A successful 3562 READ_PLUS of an empty file will always return eof as TRUE. 3564 If the current filehandle is not an ordinary file, an error will be 3565 returned to the client. In the case that the current filehandle 3566 represents an object of type NF4DIR, NFS4ERR_ISDIR is returned. If 3567 the current filehandle designates a symbolic link, NFS4ERR_SYMLINK is 3568 returned. In all other cases, NFS4ERR_WRONG_TYPE is returned. 3570 For a READ_PLUS with a stateid value of all bits equal to zero, the 3571 server MAY allow the READ_PLUS to be serviced subject to mandatory 3572 byte-range locks or the current share deny modes for the file. For a 3573 READ_PLUS with a stateid value of all bits equal to one, the server 3574 MAY allow READ_PLUS operations to bypass locking checks at the 3575 server. 3577 On success, the current filehandle retains its value. 3579 13.10.4. IMPLEMENTATION 3581 In general, the IMPLEMENTATION notes for READ in Section 18.22.4 of 3582 [2] also apply to READ_PLUS. One delta is that when the owner has a 3583 locked byte range, the server MUST return an array of rpr_contents 3584 with values inside that range. 3586 13.10.4.1. Additional pNFS Implementation Information 3588 With pNFS, the semantics of using READ_PLUS remains the same. Any 3589 data server MAY return a hole or ADH result for a READ_PLUS request 3590 that it receives. When a data server chooses to return such a 3591 result, it has the option of returning information for the data 3592 stored on that data server (as defined by the data layout), but it 3593 MUST not return results for a byte range that includes data managed 3594 by another data server. 3596 A data server should do its best to return as much information about 3597 a ADH as is feasible without having to contact the metadata server. 3598 If communication with the metadata server is required, then every 3599 attempt should be taken to minimize the number of requests. 3601 If mandatory locking is enforced, then the data server must also 3602 ensure that to return only information that is within the owner's 3603 locked byte range. 3605 13.10.5. READ_PLUS with Sparse Files Example 3607 The following table describes a sparse file. For each byte range, 3608 the file contains either non-zero data or a hole. In addition, the 3609 server in this example uses a Hole Threshold of 32K. 3611 +-------------+----------+ 3612 | Byte-Range | Contents | 3613 +-------------+----------+ 3614 | 0-15999 | Hole | 3615 | 16K-31999 | Non-Zero | 3616 | 32K-255999 | Hole | 3617 | 256K-287999 | Non-Zero | 3618 | 288K-353999 | Hole | 3619 | 354K-417999 | Non-Zero | 3620 +-------------+----------+ 3622 Table 5 3624 Under the given circumstances, if a client was to read from the file 3625 with a max read size of 64K, the following will be the results for 3626 the given READ_PLUS calls. This assumes the client has already 3627 opened the file, acquired a valid stateid ('s' in the example), and 3628 just needs to issue READ_PLUS requests. 3630 1. READ_PLUS(s, 0, 64K) --> NFS_OK, eof = false, . Since the first hole is less than the server's 3632 Hole Threshhold, the first 32K of the file is returned as data 3633 and the remaining 32K is returned as a hole which actually 3634 extends to 256K. 3636 2. READ_PLUS(s, 32K, 64K) --> NFS_OK, eof = false, 3637 The requested range was all zeros, and the current hole begins at 3638 offset 32K and is 224K in length. Note that the client should 3639 not have followed up the previous READ_PLUS request with this one 3640 as the hole information from the previous call extended past what 3641 the client was requesting. 3643 3. READ_PLUS(s, 256K, 64K) --> NFS_OK, eof = false, . Returns an array of the 32K data and 3645 the hole which extends to 354K. 3647 4. READ_PLUS(s, 354K, 64K) --> NFS_OK, eof = true, . Returns the final 64K of data and informs the client 3649 there is no more data in the file. 3651 13.11. Operation 66: SEEK 3653 SEEK is an operation that allows a client to determine the location 3654 of the next data_content4 in a file. It allows an implementation of 3655 the emerging extension to lseek(2) to allow clients to determine 3656 SEEK_HOLE and SEEK_DATA. 3658 13.11.1. ARGUMENT 3660 struct SEEK4args { 3661 /* CURRENT_FH: file */ 3662 stateid4 sa_stateid; 3663 offset4 sa_offset; 3664 data_content4 sa_what; 3665 }; 3667 13.11.2. RESULT 3669 union seek_content switch (data_content4 content) { 3670 case NFS4_CONTENT_DATA: 3671 data_info4 sc_data; 3672 case NFS4_CONTENT_APP_DATA_HOLE: 3673 app_data_hole4 sc_adh; 3674 case NFS4_CONTENT_HOLE: 3675 data_info4 sc_hole; 3676 default: 3677 void; 3678 }; 3680 struct seek_res4 { 3681 bool sr_eof; 3682 seek_content sr_contents; 3683 }; 3685 union SEEK4res switch (nfsstat4 status) { 3686 case NFS4_OK: 3687 seek_res4 resok4; 3688 default: 3689 void; 3690 }; 3692 13.11.3. DESCRIPTION 3694 From the given sa_offset, find the next data_content4 of type sa_what 3695 in the file. For either a hole or ADH, this must return the 3696 data_content4 in its entirety. For data, it must not return the 3697 actual data. 3699 SEEK must follow the same rules for stateids as READ_PLUS 3700 (Section 13.10.3). 3702 If the server could not find a corresponding sa_what, then the status 3703 would still be NFS4_OK, but sr_eof would be TRUE. The sr_contents 3704 would contain a zero-ed out content of the appropriate type. 3706 14. NFSv4.2 Callback Operations 3708 14.1. Operation 15: CB_COPY - Report results of a server-side copy 3710 14.1.1. ARGUMENT 3712 union copy_info4 switch (nfsstat4 cca_status) { 3713 case NFS4_OK: 3714 void; 3715 default: 3716 length4 cca_bytes_copied; 3717 }; 3719 struct CB_COPY4args { 3720 nfs_fh4 cca_fh; 3721 stateid4 cca_stateid; 3722 copy_info4 cca_copy_info; 3723 }; 3725 14.1.2. RESULT 3727 struct CB_COPY4res { 3728 nfsstat4 ccr_status; 3729 }; 3731 14.1.3. DESCRIPTION 3733 CB_COPY is used for both intra- and inter-server asynchronous copies. 3734 The CB_COPY callback informs the client of the result of an 3735 asynchronous server-side copy. This operation is sent by the 3736 destination server to the client in a CB_COMPOUND request. The copy 3737 is identified by the filehandle and stateid arguments. The result is 3738 indicated by the status field. If the copy failed, cca_bytes_copied 3739 contains the number of bytes copied before the failure occurred. The 3740 cca_bytes_copied value indicates the number of bytes copied but not 3741 which specific bytes have been copied. 3743 If the client supports the COPY operation, the client is REQUIRED to 3744 support the CB_COPY operation. 3746 There is a potential race between the reply to the original COPY on 3747 the forechannel and the CB_COPY callback on the backchannel. 3748 Sections 2.10.6.3 and 20.9.3 in [2] describes how to handle this type 3749 of issue. 3751 The CB_COPY operation may fail for the following reasons (this is a 3752 partial list): 3754 NFS4ERR_NOTSUPP: The copy offload operation is not supported by the 3755 NFS client receiving this request. 3757 15. IANA Considerations 3759 This section uses terms that are defined in [24]. 3761 16. References 3763 16.1. Normative References 3765 [1] Bradner, S., "Key words for use in RFCs to Indicate Requirement 3766 Levels", March 1997. 3768 [2] Shepler, S., Eisler, M., and D. Noveck, "Network File System 3769 (NFS) Version 4 Minor Version 1 Protocol", RFC 5661, 3770 January 2010. 3772 [3] Haynes, T., "Network File System (NFS) Version 4 Minor Version 3773 2 External Data Representation Standard (XDR) Description", 3774 March 2011. 3776 [4] Berners-Lee, T., Fielding, R., and L. Masinter, "Uniform 3777 Resource Identifier (URI): Generic Syntax", STD 66, RFC 3986, 3778 January 2005. 3780 [5] Haynes, T. and N. Williams, "Remote Procedure Call (RPC) 3781 Security Version 3", draft-williams-rpcsecgssv3 (work in 3782 progress), 2011. 3784 [6] The Open Group, "Section 'posix_fadvise()' of System Interfaces 3785 of The Open Group Base Specifications Issue 6, IEEE Std 1003.1, 3786 2004 Edition", 2004. 3788 [7] Haynes, T., "Requirements for Labeled NFS", 3789 draft-ietf-nfsv4-labreqs-00 (work in progress). 3791 [8] Eisler, M., Chiu, A., and L. Ling, "RPCSEC_GSS Protocol 3792 Specification", RFC 2203, September 1997. 3794 [9] Halevy, B., Welch, B., and J. Zelenka, "Object-Based Parallel 3795 NFS (pNFS) Operations", RFC 5664, January 2010. 3797 16.2. Informative References 3799 [10] Haynes, T. and D. Noveck, "Network File System (NFS) version 4 3800 Protocol", draft-ietf-nfsv4-rfc3530bis-09 (Work In Progress), 3801 March 2011. 3803 [11] Lentini, J., Everhart, C., Ellard, D., Tewari, R., and M. Naik, 3804 "NSDB Protocol for Federated Filesystems", 3805 draft-ietf-nfsv4-federated-fs-protocol (Work In Progress), 3806 2010. 3808 [12] Lentini, J., Everhart, C., Ellard, D., Tewari, R., and M. Naik, 3809 "Administration Protocol for Federated Filesystems", 3810 draft-ietf-nfsv4-federated-fs-admin (Work In Progress), 2010. 3812 [13] Fielding, R., Gettys, J., Mogul, J., Frystyk, H., Masinter, L., 3813 Leach, P., and T. Berners-Lee, "Hypertext Transfer Protocol -- 3814 HTTP/1.1", RFC 2616, June 1999. 3816 [14] Postel, J. and J. Reynolds, "File Transfer Protocol", STD 9, 3817 RFC 959, October 1985. 3819 [15] Simpson, W., "PPP Challenge Handshake Authentication Protocol 3820 (CHAP)", RFC 1994, August 1996. 3822 [16] Strohm, R., "Chapter 2, Data Blocks, Extents, and Segments, of 3823 Oracle Database Concepts 11g Release 1 (11.1)", January 2011. 3825 [17] Ashdown, L., "Chapter 15, Validating Database Files and 3826 Backups, of Oracle Database Backup and Recovery User's Guide 3827 11g Release 1 (11.1)", August 2008. 3829 [18] McDougall, R. and J. Mauro, "Section 11.4.3, Detecting Memory 3830 Corruption of Solaris Internals", 2007. 3832 [19] Bairavasundaram, L., Goodson, G., Schroeder, B., Arpaci- 3833 Dusseau, A., and R. Arpaci-Dusseau, "An Analysis of Data 3834 Corruption in the Storage Stack", Proceedings of the 6th USENIX 3835 Symposium on File and Storage Technologies (FAST '08) , 2008. 3837 [20] "Section 46.6. Multi-Level Security (MLS) of Deployment Guide: 3838 Deployment, configuration and administration of Red Hat 3839 Enterprise Linux 5, Edition 6", 2011. 3841 [21] Quigley, D. and J. Lu, "Registry Specification for MAC Security 3842 Label Formats", draft-quigley-label-format-registry (work in 3843 progress), 2011. 3845 [22] ISEG, "IESG Processing of RFC Errata for the IETF Stream", 3846 2008. 3848 [23] Eisler, M., "XDR: External Data Representation Standard", 3849 RFC 4506, May 2006. 3851 [24] Narten, T. and H. Alvestrand, "Guidelines for Writing an IANA 3852 Considerations Section in RFCs", BCP 26, RFC 5226, May 2008. 3854 [25] VanDeBogart, S., Frost, C., and E. Kohler, "Reducing Seek 3855 Overhead with Application-Directed Prefetching", Proceedings of 3856 USENIX Annual Technical Conference , June 2009. 3858 Appendix A. Acknowledgments 3860 For the pNFS Access Permissions Check, the original draft was by 3861 Sorin Faibish, David Black, Mike Eisler, and Jason Glasgow. The work 3862 was influenced by discussions with Benny Halevy and Bruce Fields. A 3863 review was done by Tom Haynes. 3865 For the Sharing change attribute implementation details with NFSv4 3866 clients, the original draft was by Trond Myklebust. 3868 For the NFS Server-side Copy, the original draft was by James 3869 Lentini, Mike Eisler, Deepak Kenchammana, Anshul Madan, and Rahul 3870 Iyer. Tom Talpey co-authored an unpublished version of that 3871 document. It was also was reviewed by a number of individuals: 3872 Pranoop Erasani, Tom Haynes, Arthur Lent, Trond Myklebust, Dave 3873 Noveck, Theresa Lingutla-Raj, Manjunath Shankararao, Satyam Vaghani, 3874 and Nico Williams. 3876 For the NFS space reservation operations, the original draft was by 3877 Mike Eisler, James Lentini, Manjunath Shankararao, and Rahul Iyer. 3879 For the sparse file support, the original draft was by Dean 3880 Hildebrand and Marc Eshel. Valuable input and advice was received 3881 from Sorin Faibish, Bruce Fields, Benny Halevy, Trond Myklebust, and 3882 Richard Scheffenegger. 3884 For the Application IO Hints, the original draft was by Dean 3885 Hildebrand, Mike Eisler, Trond Myklebust, and Sam Falkner. Some 3886 early reviwers included Benny Halevy and Pranoop Erasani. 3888 For Labeled NFS, the original draft was by David Quigley, James 3889 Morris, Jarret Lu, and Tom Haynes. Peter Staubach, Trond Myklebust, 3890 Stephen Smalley, Sorrin Faibish, Nico Williams, and David Black also 3891 contributed in the final push to get this accepted. 3893 During the review process, Talia Reyes-Ortiz helped the sessions run 3894 smoothly. While many people contributed here and there, the core 3895 reviewers were Andy Adamson, Pranoop Erasani, Bruce Fields, Chuck 3896 Lever, Trond Myklebust, David Noveck, and Peter Staubach. 3898 Appendix B. RFC Editor Notes 3900 [RFC Editor: please remove this section prior to publishing this 3901 document as an RFC] 3903 [RFC Editor: prior to publishing this document as an RFC, please 3904 replace all occurrences of RFCTBD10 with RFCxxxx where xxxx is the 3905 RFC number of this document] 3907 Author's Address 3909 Thomas Haynes 3910 NetApp 3911 9110 E 66th St 3912 Tulsa, OK 74133 3913 USA 3915 Phone: +1 918 307 1415 3916 Email: thomas@netapp.com 3917 URI: http://www.tulsalabs.com