idnits 2.17.1 draft-ietf-nfsv4-layout-types-08.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 : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year (Using the creation date from RFC5661, updated by this document, for RFC5378 checks: 2005-10-21) -- The document seems to lack a disclaimer for pre-RFC5378 work, but may have content which was first submitted before 10 November 2008. If you have contacted all the original authors and they are all willing to grant the BCP78 rights to the IETF Trust, then this is fine, and you can ignore this comment. If not, you may need to add the pre-RFC5378 disclaimer. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (August 31, 2017) is 2428 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) ** Obsolete normative reference: RFC 5661 (Obsoleted by RFC 8881) == Outdated reference: A later version (-19) exists of draft-ietf-nfsv4-flex-files-13 Summary: 1 error (**), 0 flaws (~~), 2 warnings (==), 2 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 NFSv4 T. Haynes 3 Internet-Draft Primary Data 4 Updates: 5661 (if approved) August 31, 2017 5 Intended status: Standards Track 6 Expires: March 4, 2018 8 Requirements for pNFS Layout Types 9 draft-ietf-nfsv4-layout-types-08.txt 11 Abstract 13 This document defines the requirements which individual pNFS layout 14 types need to meet in order to work within the parallel NFS (pNFS) 15 framework as defined in RFC5661. In so doing, it aims to clearly 16 distinguish between requirements for pNFS as a whole and those 17 specifically directed to the pNFS File Layout. The lack of a clear 18 separation between the two set of requirements has been troublesome 19 for those specifying and evaluating new Layout Types. In this 20 regard, this document effectively updates RFC5661. 22 Status of This Memo 24 This Internet-Draft is submitted in full conformance with the 25 provisions of BCP 78 and BCP 79. 27 Internet-Drafts are working documents of the Internet Engineering 28 Task Force (IETF). Note that other groups may also distribute 29 working documents as Internet-Drafts. The list of current Internet- 30 Drafts is at http://datatracker.ietf.org/drafts/current/. 32 Internet-Drafts are draft documents valid for a maximum of six months 33 and may be updated, replaced, or obsoleted by other documents at any 34 time. It is inappropriate to use Internet-Drafts as reference 35 material or to cite them other than as "work in progress." 37 This Internet-Draft will expire on March 4, 2018. 39 Copyright Notice 41 Copyright (c) 2017 IETF Trust and the persons identified as the 42 document authors. All rights reserved. 44 This document is subject to BCP 78 and the IETF Trust's Legal 45 Provisions Relating to IETF Documents 46 (http://trustee.ietf.org/license-info) in effect on the date of 47 publication of this document. Please review these documents 48 carefully, as they describe your rights and restrictions with respect 49 to this document. Code Components extracted from this document must 50 include Simplified BSD License text as described in Section 4.e of 51 the Trust Legal Provisions and are provided without warranty as 52 described in the Simplified BSD License. 54 Table of Contents 56 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 57 2. Definitions . . . . . . . . . . . . . . . . . . . . . . . . . 3 58 2.1. Use of the Terms "Data Server" and "Storage Device" . . . 5 59 2.2. Requirements Language . . . . . . . . . . . . . . . . . . 6 60 3. The Control Protocol . . . . . . . . . . . . . . . . . . . . 6 61 3.1. Control Protocol Requirements . . . . . . . . . . . . . . 8 62 3.2. Previously Undocumented Protocol Requirements . . . . . . 9 63 3.3. Editorial Requirements . . . . . . . . . . . . . . . . . 10 64 4. Specifications of Original Layout Types . . . . . . . . . . . 11 65 4.1. File Layout Type . . . . . . . . . . . . . . . . . . . . 11 66 4.2. Block Layout Type . . . . . . . . . . . . . . . . . . . . 12 67 4.3. Object Layout Type . . . . . . . . . . . . . . . . . . . 13 68 5. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 69 6. Security Considerations . . . . . . . . . . . . . . . . . . . 15 70 7. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 15 71 8. References . . . . . . . . . . . . . . . . . . . . . . . . . 15 72 8.1. Normative References . . . . . . . . . . . . . . . . . . 15 73 8.2. Informative References . . . . . . . . . . . . . . . . . 16 74 Appendix A. Acknowledgments . . . . . . . . . . . . . . . . . . 16 75 Appendix B. RFC Editor Notes . . . . . . . . . . . . . . . . . . 16 76 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 16 78 1. Introduction 80 The concept of layout type has a central role in the definition and 81 implementation of Parallel Network File System (pNFS). Clients and 82 servers implementing different layout types behave differently in 83 many ways while conforming to the overall pNFS framework defined in 84 [RFC5661] and this document. Layout types may differ as to: 86 o The method used to do I/O operations directed to data storage 87 devices. 89 o The requirements for communication between the metadata server 90 (MDS) and the storage devices. 92 o The means used to ensure that I/O requests are only processed when 93 the client holds an appropriate layout. 95 o The format and interpretation of nominally opaque data fields in 96 pNFS-related NFSv4.x data structures. 98 Such matters are defined in a standards-track layout type 99 specification. Except for the files layout type, which was defined 100 in Section 13 of [RFC5661], existing layout types are defined in 101 their own standards-track documents and it is anticipated that new 102 layout types will be defined in similar documents. 104 The file layout type was defined in the Network File System (NFS) 105 version 4.1 protocol specification [RFC5661]. The block layout type 106 was defined in [RFC5663] while the object layout type was defined in 107 [RFC5664]. Subsequently, the SCSI layout type was defined in 108 [RFC8154]. 110 Some implementers have interpreted the text in Sections 12 ("Parallel 111 NFS (pNFS)") and 13 ("NFSv4.1 as a Storage Protocol in pNFS: the File 112 Layout Type") of [RFC5661] as both being applying only to the file 113 layout type. Because Section 13 was not covered in a separate 114 standards-track document such as those for both the block and object 115 layout types, there had been some confusion as to the 116 responsibilities of both the metadata server and the data servers 117 (DS) which were laid out in Section 12. 119 As a consequence, new internet drafts (see [FlexFiles] and [Lustre]) 120 may struggle to meet the requirements to be a pNFS layout type. This 121 document gathers the requirements from all of the original layout 122 type standard documents and then specifies the requirements placed on 123 all layout types independent of the particular type chosen. 125 2. Definitions 127 control communication requirements: are for a layout type the 128 details regarding information on layouts, stateids, file metadata, 129 and file data which must be communicated between the metadata 130 server and the storage devices. 132 control protocol: is the particular mechanism that an implementation 133 of a layout type would use to meet the control communication 134 requirement for that layout type. This need not be a protocol as 135 normally understood. In some cases the same protocol may be used 136 as a control protocol and storage protocol. 138 (file) data: is that part of the file system object which contains 139 the data to read or written. It is the contents of the object 140 rather than the attributes of the object. 142 data server (DS): is a pNFS server which provides the file's data 143 when the file system object is accessed over a file-based 144 protocol. Note that this usage differs from that in [RFC5661] 145 which applies the term in some cases even when other sorts of 146 protocols are being used. Depending on the layout, there might be 147 one or more data servers over which the data is striped. While 148 the metadata server is strictly accessed over the NFSv4.1 149 protocol, the data server could be accessed via any file access 150 protocol that meets the pNFS requirements. 152 See Section 2.1 for a comparison of this term and "data storage 153 device". 155 fencing: is the process by which the metadata server prevents the 156 storage devices from processing I/O from a specific client to a 157 specific file. 159 layout: is the information a client uses to access file data on a 160 storage device. This information will include specification of 161 the protocol (layout type) and the identity of the storage devices 162 to be used. 164 The bulk of the contents of the layout are defined in [RFC5661] 165 as nominally opaque, but individual layout types are responsible 166 for specifying the format of the layout data. 168 layout iomode: is a grant of either read or read/write I/O to the 169 client. 171 layout stateid: is a 128-bit quantity returned by a server that 172 uniquely defines the layout state provided by the server for a 173 specific layout that describes a layout type and file (see 174 Section 12.5.2 of [RFC5661]). Further, Section 12.5.3 describes 175 differences in handling between layout stateids and other stateid 176 types. 178 layout type: is a specification of both the storage protocol used to 179 access the data and the aggregation scheme used to lay out the 180 file data on the underlying storage devices. 182 loose coupling: is when the control protocol is a storage protocol. 184 (file) metadata: is that part of the file system object that 185 contains various descriptive data relevant to the file object, as 186 opposed to the file data itself. This could include the time of 187 last modification, access time, end-of-file (EOF) position, etc. 189 metadata server (MDS): is the pNFS server which provides metadata 190 information for a file system object. It also is responsible for 191 generating, recalling, and revoking layouts for file system 192 objects, for performing directory operations, and for performing I 193 /O operations to regular files when the clients direct these to 194 the metadata server itself. 196 recalling a layout: is a graceful recall, via a callback, of a 197 specific layout by the metadata server to the client. Graceful 198 here means that the client would have the opportunity to flush any 199 writes, etc., before returning the layout to the metadata server. 201 revoking a layout: is an invalidation of a specific layout by the 202 metadata server. Once revocation occurs, the metadata server will 203 not accept as valid any reference to the revoked layout and a 204 storage device will not accept any client access based on the 205 layout. 207 stateid: is a 128-bit quantity returned by a server that uniquely 208 defines the set of locking-related state provided by the server. 209 Stateids may designate state related to open files, to byte-range 210 locks, to delegations, or to layouts. 212 storage device: is the target to which clients may direct I/O 213 requests when they hold an appropriate layout. Note that each 214 data server is a storage device but that some storage device are 215 not data servers. See Section 2.1 for further discussion. 217 storage protocol: is the protocol used by clients to do I/O 218 operations to the storage device. Each layout type specifies the 219 set of storage protocols. 221 tight coupling: is an arrangement in which the control protocol is 222 one designed specifically for that purpose. It may be either a 223 proprietary protocol, adapted specifically to a a particular 224 metadata server, or one based on a standards-track document. 226 2.1. Use of the Terms "Data Server" and "Storage Device" 228 In [RFC5661], these two terms of "Data Server" and "Storage Device" 229 are used somewhat inconsistently: 231 o In chapter 12, where pNFS in general is discussed, the term 232 "storage device" is used. 234 o In chapter 13, where the file layout type is discussed, the term 235 "data server" is used. 237 o In other chapters, the term "data server" is used, even in 238 contexts where the storage access type is not NFSv4.1 or any other 239 file access protocol. 241 As this document deals with pNFS in general, it uses the more generic 242 term "storage device" in preference to "data server". The term "data 243 server" is used only in contexts in which a file server is used as a 244 storage device. Note that every data server is a storage device but 245 storage devices which use protocols which are not file access 246 protocols (such as NFS) are not data servers. 248 Since a given storage device may support multiple layout types, a 249 given device can potentially act as a data server for some set of 250 storage protocols while simultaneously acting as a storage device for 251 others. 253 2.2. Requirements Language 255 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 256 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 257 document are to be interpreted as described in [RFC2119]. 259 This document differs from most standards-track documents in that it 260 specifies requirements for those defining future layout types rather 261 than defining the requirements for implementations directly. This 262 document makes clear whether: 264 (1) any particular requirement applies to implementations. 266 (2) any particular requirement applies to those defining layout 267 types. 269 (3) the requirement is a general requirement which implementations 270 need to conform to, with the specific means left to layout type 271 definitions type to specify. 273 3. The Control Protocol 275 A layout type has to meet the requirements that apply to the 276 interaction between the metadata server and the storage device such 277 that they present to the client a consistent view of stored data and 278 lock state (Section 12.2.6 of [RFC5661]). Particular implementations 279 may satisfy these requirements in any manner they choose and the 280 mechanism chosen need not be described as a protocol. Specifications 281 defining layout types need to clearly show how implementations can 282 meet the requirements discussed below, especially with respect to 283 those that have security implications. In addition, such 284 specifications may find it necessary to impose requirements on 285 implementations of the layout type to ensure appropriate 286 interoperability. 288 In some cases, there may be no control protocol other than the 289 storage protocol. This is often described as using a "loose 290 coupling" model. In such cases, the assumption is that the metadata 291 server, storage devices, and client may be changed independently and 292 that the implementation requirements in the layout type specification 293 need to ensure this degree of interoperability. This model is used 294 in the block and object layout type specification. 296 In other cases, it is assumed that there will be a purpose-built 297 control protocol which may be different for different implementations 298 of the metadata server and data server. The assumption here is that 299 the metadata server and data servers are designed and implemented as 300 a unit and interoperability needs to be assured between clients and 301 metadata-data server pairs, developed independently. This is the 302 model used for the files layout. 304 Another possibility is for the definition of a control protocol to be 305 specified in a standards-track document. There are two subcases to 306 consider: 308 o A new layout type includes a definition of a particular control 309 protocol whose use is obligatory for metadata servers and storage 310 devices implementing the layout type. In this case the 311 interoperability model is similar to the first case above and the 312 defining document should assure interoperability among metadata 313 servers, storage devices, and clients developed independently. 315 o A control protocol is defined in a standards-track document which 316 meets the control protocol requirements for one of the existing 317 layout types. In this case, the new document's job is to assure 318 interoperability between metadata servers and storage devices 319 developed separately. The existing definition document for the 320 selected layout type retains the function of assuring 321 interoperability between clients and a given collection of 322 metadata servers and storage devices. In this context, 323 implementations that implement the new protocol are treated in the 324 same way as those that use an internal control protocol or a 325 functional equivalent. 327 An example of this last case is the SCSI layout type [RFC8154], which 328 extends the block layout type. The block layout type had a 329 requirement for fencing of clients, but did not present a way for the 330 control protocol (in this case the SCSI storage protocol) to fence 331 the client. The SCSI layout type remedies that in [RFC8154] and in 332 effect has a tightly coupled model. 334 3.1. Control Protocol Requirements 336 The requirements of interactions between the metadata server and the 337 storage devices are: 339 (1) The metadata server MUST be able to service the client's I/O 340 requests if the client decides to make such requests to the 341 metadata server instead of to the storage device. The metadata 342 server must be able to retrieve the data from the constituent 343 storage devices and present it back to the client. A corollary 344 to this is that even though the metadata server has successfully 345 given the client a layout, the client MAY still send I/O 346 requests to the metadata server. 348 (2) The metadata server MUST be able to restrict access to a file on 349 the storage devices when it revokes a layout. The metadata 350 server typically would revoke a layout whenever a client fails 351 to respond to a recall or a client's lease is expired due to 352 non-renewal. It might also revoke the layout as a means of 353 enforcing a change in locking state or access permissions that 354 the storage device cannot directly enforce. 356 Effective revocation may require client co-operation in using a 357 particular stateid (files layout) or principal (e,g., flexible 358 files layout) when performing I/O. 360 In contrast, there is no requirement to restrict access to a 361 file on the storage devices when a layout is recalled. It is 362 only after the metadata server determines that the client is not 363 gracefully returning the layout and starts the revocation that 364 this requirement is enforced. 366 (3) A pNFS implementation MUST NOT allow the violation of NFSv4.1's 367 access controls: ACLs and file open modes. Section 12.9 of 368 [RFC5661] specifically lays this burden on the combination of 369 clients, storage devices, and the metadata server. However the 370 specification of the individual layout type might create 371 requirements as to how this is to be done. This may include a 372 possible requirement for the metadata server to update the 373 storage device so that it can enforce security. 375 The file layout requires the storage device to enforce access 376 whereas the flex file layout requires both the storage device 377 and the client to enforce security. 379 (4) Interactions between locking and I/O operations MUST obey 380 existing semantic restrictions. In particular, if an I/O 381 operation would be invalid when directed at the metadata server, 382 it is not to be allowed when performed on the storage device. 384 For the block and SCSI layout, as the storage device is not able 385 to reject the I/O operation, the client is responsible for 386 enforcing this requirement. 388 (5) Any disagreement between the metadata server and the data server 389 as to the value of attributes such as modify time, the change 390 attribute, and the EOF position MUST be of limited duration with 391 clear means of resolution of any discrepancies being provided. 392 Note that 394 (a) Discrepancies need not be resolved unless any client has 395 accessed the file in question via the metadata server, 396 typically by performing a GETATTR. 398 (b) A particular storage device might be striped such it has no 399 information regarding the EOF position. 401 (c) Both clock skew and network delay can lead to the metadata 402 server and the storage device having different values of 403 the time attributes. As long as those differences can be 404 accounted for in what is presented to the client in a 405 GETATTR, then no violation results. 407 (d) A LAYOUTCOMMIT requires that changes in attributes 408 resulting from operations on the storage device need to be 409 reflected in the metadata server by the completion of the 410 operation. 412 These requirements may be satisfied in different ways by different 413 layout types. As an example, while the file layout type uses the 414 stateid to fence off the client, there is no requirement that other 415 layout types use this stateid approach. 417 Each new standards-track document for a layout types MUST address how 418 the client, metadata server, and storage devices are to interact to 419 meet these requirements. 421 3.2. Previously Undocumented Protocol Requirements 423 While not explicitly stated as requirements in Section 12 of 424 [RFC5661], the existing layout types do have more requirements that 425 they need to enforce. 427 The client has these obligations when making I/O requests to the 428 storage devices: 430 (1) Clients MUST NOT perform I/O to the storage device if they do 431 not have layouts for the files in question. 433 (2) Clients MUST NOT perform I/O operations outside of the specified 434 ranges in the layout segment. 436 (3) Clients MUST NOT perform I/O operations which would be 437 inconsistent with the iomode specified in the layout segments it 438 holds. 440 Under the file layout type, the storage devices are able to reject 441 any request made not conforming to these requirements. This may not 442 be possible for other known layout types, which puts the burden of 443 enforcing such violations solely on the client. For these layout 444 types: 446 (1) The metadata server MAY use fencing operations to the storage 447 devices to enforce layout revocation against the client. 449 (2) The metadata server MUST allow the clients to perform data I/O 450 against it, even if it has already granted the client a layout. 451 A layout type might discourage such I/O, but it can not forbid 452 it. 454 (3) The metadata server MUST be able to do storage allocation, 455 whether that is to create, delete, extend, or truncate files. 457 The means to address these requirements will vary with the layout 458 type. A control protocol will be used to effect these, whether a 459 purpose-built one, one identical to the storage protocol, or a new 460 standards-track control protocol. 462 3.3. Editorial Requirements 464 This section discusses how the protocol requirements discussed above 465 need to be addressed in documents specifying a new layout type. 466 Depending on the interoperability model for the layout type in 467 question, this may involve the imposition of layout-type-specific 468 requirements that ensure appropriate interoperability of pNFS 469 components which are developed separately. 471 The specification of the layout type needs to make clear how the 472 client, metadata server, and storage device act together to meet the 473 protocol requirements discussed previously. If the document does not 474 impose implementation requirements sufficient to ensure that these 475 semantic requirements are met, it is not appropriate for the working 476 group to allow the document to move forward. 478 Some examples include: 480 o If the metadata server does not have a means to invalidate a 481 stateid issued to the storage device to keep a particular client 482 from accessing a specific file, then the layout type specification 483 has to document how the metadata server is going to fence the 484 client from access to the file on that storage device. 486 o If the metadata server implements mandatory byte-range locking 487 when accessed directly by the client, then the layout type 488 specification must require that this also be done when data is 489 read or written using the designated storage protocol. 491 4. Specifications of Original Layout Types 493 This section updates Section 12 of [RFC5661], which enumerates the 494 requirements of pNFS layout type specifications. It is not normative 495 with regards to the file layout type presented in Section 13 of 496 [RFC5661], the block layout type [RFC5663], and the object layout 497 type [RFC5664]. These are discussed here only to illuminate the 498 updates made to Section 12 of [RFC5661]. 500 4.1. File Layout Type 502 Because the storage protocol is a subset of NFSv4.1, the semantics of 503 the file layout type comes closest to the semantics of NFSv4.1 in the 504 absence of pNFS. In particular, the stateid and principal used for I 505 /O MUST have the same effect and be subject to the same validation on 506 a data server as it would have if the I/O were being performed on the 507 metadata server itself. The same set of validations are applied 508 whether pNFS is in effect or not. 510 And while for most implementations the storage devices can do the 511 following validations: 513 (1) client holds a valid layout, 515 (2) client I/O matches the layout iomode, and, 517 (3) client does not go out of the byte ranges, 519 these are each presented as a "SHOULD" and not a "MUST". Actually, 520 the first point is presented in [RFC5661] as both: 522 "MUST": in Section 13.6 523 "As described in Section 12.5.1, a client MUST NOT send an I/O to 524 a data server for which it does not hold a valid layout; the data 525 server MUST reject such an I/O." 527 "SHOULD": in Section 13.8 529 "The iomode need not be checked by the data servers when clients 530 perform I/O. However, the data servers SHOULD still validate that 531 the client holds a valid layout and return an error if the client 532 does not." 534 It should be noted that it is just these layout specific checks that 535 are optional, not the normal file access semantics. The storage 536 devices MUST make all of the required access checks on each READ or 537 WRITE I/O as determined by the NFSv4.1 protocol. If the metadata 538 server would deny a READ or WRITE operation on a file due to its ACL, 539 mode attribute, open access mode, open deny mode, mandatory byte- 540 range lock state, or any other attributes and state, the storage 541 device MUST also deny the READ or WRITE operation. Also while the 542 NFSv4.1 protocol does not mandate export access checks based on the 543 client's IP address, if the metadata server implements such a policy, 544 then that counts as such state as outlined above. 546 The data filehandle provided by the PUTFH operation to the data 547 server provides sufficient context to enable the data server to 548 ensure that for the subsequent READ or WRITE operation in the 549 compound, that the client has a valid layout for the I/O being 550 performed. 552 Finally, the data server can check the stateid presented in the READ 553 or WRITE operation to see if that stateid has been rejected by the 554 metadata server in order to cause the I/O to be fenced. Whilst it 555 might just be the open owner or lock owner on that client being 556 fenced, the client should take the NFS4ERR_BAD_STATEID error code to 557 mean it has been fenced from the file and contact the metadata 558 server. 560 4.2. Block Layout Type 562 With the block layout type, the storage devices are generally not 563 able to enforce file-based security. Typically, storage area network 564 (SAN) disk arrays and SAN protocols provide coarse-grained access 565 control mechanisms (e.g., Logical Unit Number (LUN) mapping and/or 566 masking), with a target granularity of disks rather than individual 567 blocks and a source granularity of individual hosts rather than of 568 users or owners. Access to block storage is logically at a lower 569 layer of the I/O stack than NFSv4. Since NFSv4 security is not 570 directly applicable to protocols that access such storage directly, 571 Section 2.1 [RFC5663] specifies that: 573 "in environments where pNFS clients cannot be trusted to enforce 574 such policies, pNFS block layout types SHOULD NOT be used." 576 Due to these granularity issues, the security burden has been shifted 577 from the storage devices to the client. Those deploying 578 implementations of this layout type need to be sure that the client 579 implementation can be trusted This is not a new sort of requirement 580 in the context of SAN protocols. In such environments, the client is 581 expected to provide block-based protection. 583 This shift of the burden also extends to locks and layouts. The 584 storage devices are not able to enforce any of these and the burden 585 is pushed to the client to make the appropriate checks before sending 586 I/O to the storage devices. For example, the server may use a layout 587 iomode only allowing reading to enforce a mandatory read-only lock, 588 In such cases, the client has to support that use by not sending 589 WRITEs to the storage devices. The fundamental issue here is that 590 the storage device is treated by this layout type in the same fashion 591 as a local disk device. Once the client has access to the storage 592 device, it is able to perform both READ and WRITE I/O to the entire 593 storage device. The byte ranges in the layout, any locks, the layout 594 iomode, etc, can only be enforced by the client. Therefore, the 595 client is required to provide that enforcement. 597 In the context of fencing off of the client upon revocation of a 598 layout, these limitations come into play again, i.e., the granularity 599 of the fencing can only be at the host/logical-unit level. Thus, if 600 one of a client's layouts is revoked by the server, it will 601 effectively revoke all of the client's layouts for files located on 602 the storage units comprising the logical volume. This may extend to 603 the client's layouts for files in other file systems. Clients need 604 to be prepared for such revocations and reacquire layouts as needed. 606 4.3. Object Layout Type 608 With the object layout type, security checks occur during the 609 allocation of the layout. The client will typically ask for layouts 610 covering all of the file and may do so for either READ or READ/WRITE. 611 This enables it to do subsequent I/O operations without the need to 612 obtain layouts for specific byte ranges. At that time, the metadata 613 server should verify permissions against the layout iomode, the file 614 mode bits or ACLs, etc. As the client may be acting for multiple 615 local users, it MUST authenticate and authorize the user by issuing 616 respective OPEN and ACCESS calls to the metadata server, similar to 617 having NFSv4 data delegations. 619 Upon successful authorization, the client receives within the layout 620 a set of object capabilities allowing it I/O access to the specified 621 objects corresponding to the requested iomode. These capabilities 622 are used to enforce access control and locking semantics at the 623 storage devices. Whenever one of the following occur on the metadata 624 server: 626 o the permissions on the object change, 628 o a conflicting mandatory byte-range lock is granted, or 630 o a layout is revoked and reassigned to another client, 632 then the metadata server MUST change the capability version attribute 633 on all objects comprising the file to in order to invalidate any 634 outstanding capabilities before committing to one of these changes. 636 When the metadata server wishes to fence off a client to a particular 637 object, then it can use the above approach to invalidate the 638 capability attribute on the given object. The client can be informed 639 via the storage device that the capability has been rejected and is 640 allowed to fetch a refreshed set of capabilities, i.e., re-acquire 641 the layout. 643 5. Summary 645 In the three original layout types, the burden of enforcing the 646 security of NFSv4.1 can fall to either the storage devices (files), 647 the client (blocks), or the metadata server (objects). Such choices 648 are conditioned by the native capabilities of the storage devices - 649 if a control protocol can be implemented, then the burden can be 650 shifted primarily to the storage devices. 652 In the context of this document, we treat the control protocol as a 653 set of requirements. And as new layout types are published, the 654 defining documents MUST address: 656 (1) The fencing of clients after a layout is revoked. 658 (2) The security implications of the native capabilities of the 659 storage devices with respect to the requirements of the NFSv4.1 660 security model. 662 In addition, these defining documents need to make clear how other 663 semantic requirements of NFSv4.1 (e.g., locking) are met in the 664 context of the proposed layout type. 666 6. Security Considerations 668 This section does not deal directly with security considerations for 669 existing or new layout types. Instead, it provides a general 670 framework for understating security-related issues within the pNFS 671 framework. Specific security considerations will be addressed in the 672 Security Considerations sections of documents specifying layout 673 types. 675 The layout type specification must ensure that only data accesses 676 consistent with the NFSV4.1 security model are allowed. It may do 677 this directly, by providing that appropriate checks be performed at 678 the time each access is performed. It may do it indirectly by 679 allowing the client or the storage device to be responsible for 680 making the appropriate checks. In the latter case, I/O access writes 681 are reflected in layouts and the layout type must provide a way to 682 prevent inappropriate access due to permissions changes between the 683 time a layout is granted and the time the access is performed. 685 The metadata server MUST be able to fence off a client's access to 686 the data file on a storage device. When it revokes the layout, the 687 client's access MUST be terminated at the storage devices. The 688 client has a subsequent opportunity to re-acquire the layout and 689 perform the security check in the context of the newly current access 690 permissions. 692 7. IANA Considerations 694 This document has no actions for IANA. 696 8. References 698 8.1. Normative References 700 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 701 Requirement Levels", March 1997. 703 [RFC5661] Shepler, S., Eisler, M., and D. Noveck, "Network File 704 System (NFS) Version 4 Minor Version 1 Protocol", RFC 705 5661, January 2010. 707 [RFC5663] Black, D., Fridella, S., and J. Glasgow, "pNFS Block/ 708 Volume Layout", RFC 5663, January 2010. 710 [RFC5664] Halevy, B., Welch, B., and J. Zelenka, "Object-Based 711 Parallel NFS (pNFS) Operations", RFC 5664, January 2010. 713 [RFC8154] Hellwig, C., "Parallel NFS (pNFS) Small Computer System 714 Interface (SCSI) Layout", RFC 8154, DOI 10.17487/RFC8154, 715 May 2017, . 717 8.2. Informative References 719 [FlexFiles] 720 Halevy, B. and T. Haynes, "Parallel NFS (pNFS) Flexible 721 File Layout", draft-ietf-nfsv4-flex-files-13 (Work In 722 Progress), August 2017. 724 [Lustre] Faibish, S. and P. Tao, "Parallel NFS (pNFS) Lustre Layout 725 Operations", draft-faibish-nfsv4-pnfs-lustre-layout-07 726 (Work In Progress), April 2014. 728 Appendix A. Acknowledgments 730 Dave Noveck provided an early review that sharpened the clarity of 731 the definitions. He also provided a more comprehensive review of the 732 document. 734 Both Chuck Lever and Christoph Helwig provided insightful comments 735 during the WGLC. 737 Appendix B. RFC Editor Notes 739 [RFC Editor: please remove this section prior to publishing this 740 document as an RFC] 742 [RFC Editor: prior to publishing this document as an RFC, please 743 replace all occurrences of RFCTBD10 with RFCxxxx where xxxx is the 744 RFC number of this document] 746 Author's Address 748 Thomas Haynes 749 Primary Data, Inc. 750 4300 El Camino Real Ste 100 751 Los Altos, CA 94022 752 USA 754 Phone: +1 408 215 1519 755 Email: thomas.haynes@primarydata.com