1NFS(5) File Formats Manual NFS(5)
2
6 nfs - fstab format and options for the nfs file systems
7
9 /etc/fstab
10
12 NFS is an Internet Standard protocol created by Sun Microsystems in
13 1984. NFS was developed to allow file sharing between systems residing
14 on a local area network. Depending on kernel configuration, the Linux
15 NFS client may support NFS versions 2, 3, 4.0, 4.1, or 4.2.
16 The mount(8) command attaches a file system to the system's name space
18 hierarchy at a given mount point. The /etc/fstab file describes how
19 mount(8) should assemble a system's file name hierarchy from various
20 independent file systems (including file systems exported by NFS
21 servers). Each line in the /etc/fstab file describes a single file
22 system, its mount point, and a set of default mount options for that
23 mount point.
24 For NFS file system mounts, a line in the /etc/fstab file specifies the
26 server name, the path name of the exported server directory to mount,
27 the local directory that is the mount point, the type of file system
28 that is being mounted, and a list of mount options that control the way
29 the filesystem is mounted and how the NFS client behaves when accessing
30 files on this mount point. The fifth and sixth fields on each line are
31 not used by NFS, thus conventionally each contain the digit zero. For
32 example:
33 server:path /mountpoint fstype option,option,... 0 0
35 The server's hostname and export pathname are separated by a colon,
37 while the mount options are separated by commas. The remaining fields
38 are separated by blanks or tabs.
39 The server's hostname can be an unqualified hostname, a fully qualified
41 domain name, a dotted quad IPv4 address, or an IPv6 address enclosed in
42 square brackets. Link-local and site-local IPv6 addresses must be
43 accompanied by an interface identifier. See ipv6(7) for details on
44 specifying raw IPv6 addresses.
45 The fstype field contains "nfs". Use of the "nfs4" fstype in
47 /etc/fstab is deprecated.
48
50 Refer to mount(8) for a description of generic mount options available
51 for all file systems. If you do not need to specify any mount options,
52 use the generic option defaults in /etc/fstab.
53 Options supported by all versions
55 These options are valid to use with any NFS version.
56 nfsvers=n The NFS protocol version number used to contact the
58 server's NFS service. If the server does not support
59 the requested version, the mount request fails. If this
60 option is not specified, the client tries version 4.2
61 first, then negotiates down until it finds a version
62 supported by the server.
63 vers=n This option is an alternative to the nfsvers option. It
65 is included for compatibility with other operating sys‐
66 tems
67 soft / hard Determines the recovery behavior of the NFS client after
69 an NFS request times out. If neither option is speci‐
70 fied (or if the hard option is specified), NFS requests
71 are retried indefinitely. If the soft option is speci‐
72 fied, then the NFS client fails an NFS request after
73 retrans retransmissions have been sent, causing the NFS
74 client to return an error to the calling application.
75 NB: A so-called "soft" timeout can cause silent data
77 corruption in certain cases. As such, use the soft
78 option only when client responsiveness is more important
79 than data integrity. Using NFS over TCP or increasing
80 the value of the retrans option may mitigate some of the
81 risks of using the soft option.
82 softreval / nosoftreval
84 In cases where the NFS server is down, it may be useful
85 to allow the NFS client to continue to serve up paths
86 and attributes from cache after retrans attempts to
87 revalidate that cache have timed out. This may, for
88 instance, be helpful when trying to unmount a filesystem
89 tree from a server that is permanently down.
90 It is possible to combine softreval with the soft mount
92 option, in which case operations that cannot be served
93 up from cache will time out and return an error after
94 retrans attempts. The combination with the default hard
95 mount option implies those uncached operations will con‐
96 tinue to retry until a response is received from the
97 server.
98 Note: the default mount option is nosoftreval which dis‐
100 allows fallback to cache when revalidation fails, and
101 instead follows the behavior dictated by the hard or
102 soft mount option.
103 intr / nointr This option is provided for backward compatibility. It
105 is ignored after kernel 2.6.25.
106 timeo=n The time in deciseconds (tenths of a second) the NFS
108 client waits for a response before it retries an NFS
109 request.
110 For NFS over TCP the default timeo value is 600 (60 sec‐
112 onds). The NFS client performs linear backoff: After
113 each retransmission the timeout is increased by timeo up
114 to the maximum of 600 seconds.
115 However, for NFS over UDP, the client uses an adaptive
117 algorithm to estimate an appropriate timeout value for
118 frequently used request types (such as READ and WRITE
119 requests), but uses the timeo setting for infrequently
120 used request types (such as FSINFO requests). If the
121 timeo option is not specified, infrequently used request
122 types are retried after 1.1 seconds. After each
123 retransmission, the NFS client doubles the timeout for
124 that request, up to a maximum timeout length of 60 sec‐
125 onds.
126 retrans=n The number of times the NFS client retries a request
128 before it attempts further recovery action. If the
129 retrans option is not specified, the NFS client tries
130 each UDP request three times and each TCP request twice.
131 The NFS client generates a "server not responding" mes‐
133 sage after retrans retries, then attempts further recov‐
134 ery (depending on whether the hard mount option is in
135 effect).
136 rsize=n The maximum number of bytes in each network READ request
138 that the NFS client can receive when reading data from a
139 file on an NFS server. The actual data payload size of
140 each NFS READ request is equal to or smaller than the
141 rsize setting. The largest read payload supported by the
142 Linux NFS client is 1,048,576 bytes (one megabyte).
143 The rsize value is a positive integral multiple of 1024.
145 Specified rsize values lower than 1024 are replaced with
146 4096; values larger than 1048576 are replaced with
147 1048576. If a specified value is within the supported
148 range but not a multiple of 1024, it is rounded down to
149 the nearest multiple of 1024.
150 If an rsize value is not specified, or if the specified
152 rsize value is larger than the maximum that either
153 client or server can support, the client and server
154 negotiate the largest rsize value that they can both
155 support.
156 The rsize mount option as specified on the mount(8) com‐
158 mand line appears in the /etc/mtab file. However, the
159 effective rsize value negotiated by the client and
160 server is reported in the /proc/mounts file.
161 wsize=n The maximum number of bytes per network WRITE request
163 that the NFS client can send when writing data to a file
164 on an NFS server. The actual data payload size of each
165 NFS WRITE request is equal to or smaller than the wsize
166 setting. The largest write payload supported by the
167 Linux NFS client is 1,048,576 bytes (one megabyte).
168 Similar to rsize , the wsize value is a positive inte‐
170 gral multiple of 1024. Specified wsize values lower
171 than 1024 are replaced with 4096; values larger than
172 1048576 are replaced with 1048576. If a specified value
173 is within the supported range but not a multiple of
174 1024, it is rounded down to the nearest multiple of
175 1024.
176 If a wsize value is not specified, or if the specified
178 wsize value is larger than the maximum that either
179 client or server can support, the client and server
180 negotiate the largest wsize value that they can both
181 support.
182 The wsize mount option as specified on the mount(8) com‐
184 mand line appears in the /etc/mtab file. However, the
185 effective wsize value negotiated by the client and
186 server is reported in the /proc/mounts file.
187 ac / noac Selects whether the client may cache file attributes. If
189 neither option is specified (or if ac is specified), the
190 client caches file attributes.
191 To improve performance, NFS clients cache file
193 attributes. Every few seconds, an NFS client checks the
194 server's version of each file's attributes for updates.
195 Changes that occur on the server in those small inter‐
196 vals remain undetected until the client checks the
197 server again. The noac option prevents clients from
198 caching file attributes so that applications can more
199 quickly detect file changes on the server.
200 In addition to preventing the client from caching file
202 attributes, the noac option forces application writes to
203 become synchronous so that local changes to a file
204 become visible on the server immediately. That way,
205 other clients can quickly detect recent writes when they
206 check the file's attributes.
207 Using the noac option provides greater cache coherence
209 among NFS clients accessing the same files, but it
210 extracts a significant performance penalty. As such,
211 judicious use of file locking is encouraged instead.
212 The DATA AND METADATA COHERENCE section contains a
213 detailed discussion of these trade-offs.
214 acregmin=n The minimum time (in seconds) that the NFS client caches
216 attributes of a regular file before it requests fresh
217 attribute information from a server. If this option is
218 not specified, the NFS client uses a 3-second minimum.
219 See the DATA AND METADATA COHERENCE section for a full
220 discussion of attribute caching.
221 acregmax=n The maximum time (in seconds) that the NFS client caches
223 attributes of a regular file before it requests fresh
224 attribute information from a server. If this option is
225 not specified, the NFS client uses a 60-second maximum.
226 See the DATA AND METADATA COHERENCE section for a full
227 discussion of attribute caching.
228 acdirmin=n The minimum time (in seconds) that the NFS client caches
230 attributes of a directory before it requests fresh
231 attribute information from a server. If this option is
232 not specified, the NFS client uses a 30-second minimum.
233 See the DATA AND METADATA COHERENCE section for a full
234 discussion of attribute caching.
235 acdirmax=n The maximum time (in seconds) that the NFS client caches
237 attributes of a directory before it requests fresh
238 attribute information from a server. If this option is
239 not specified, the NFS client uses a 60-second maximum.
240 See the DATA AND METADATA COHERENCE section for a full
241 discussion of attribute caching.
242 actimeo=n Using actimeo sets all of acregmin, acregmax, acdirmin,
244 and acdirmax to the same value. If this option is not
245 specified, the NFS client uses the defaults for each of
246 these options listed above.
247 bg / fg Determines how the mount(8) command behaves if an
249 attempt to mount an export fails. The fg option causes
250 mount(8) to exit with an error status if any part of the
251 mount request times out or fails outright. This is
252 called a "foreground" mount, and is the default behavior
253 if neither the fg nor bg mount option is specified.
254 If the bg option is specified, a timeout or failure
256 causes the mount(8) command to fork a child which con‐
257 tinues to attempt to mount the export. The parent imme‐
258 diately returns with a zero exit code. This is known as
259 a "background" mount.
260 If the local mount point directory is missing, the
262 mount(8) command acts as if the mount request timed out.
263 This permits nested NFS mounts specified in /etc/fstab
264 to proceed in any order during system initialization,
265 even if some NFS servers are not yet available. Alter‐
266 natively these issues can be addressed using an auto‐
267 mounter (refer to automount(8) for details).
268 nconnect=n When using a connection oriented protocol such as TCP,
270 it may sometimes be advantageous to set up multiple con‐
271 nections between the client and server. For instance, if
272 your clients and/or servers are equipped with multiple
273 network interface cards (NICs), using multiple connec‐
274 tions to spread the load may improve overall perfor‐
275 mance. In such cases, the nconnect option allows the
276 user to specify the number of connections that should be
277 established between the client and server up to a limit
278 of 16.
279 Note that the nconnect option may also be used by some
281 pNFS drivers to decide how many connections to set up to
282 the data servers.
283 rdirplus / nordirplus
285 Selects whether to use NFS v3 or v4 READDIRPLUS
286 requests. If this option is not specified, the NFS
287 client uses READDIRPLUS requests on NFS v3 or v4 mounts
288 to read small directories. Some applications perform
289 better if the client uses only READDIR requests for all
290 directories.
291 retry=n The number of minutes that the mount(8) command retries
293 an NFS mount operation in the foreground or background
294 before giving up. If this option is not specified, the
295 default value for foreground mounts is 2 minutes, and
296 the default value for background mounts is 10000 minutes
297 (80 minutes shy of one week). If a value of zero is
298 specified, the mount(8) command exits immediately after
299 the first failure.
300 Note that this only affects how many retries are made
302 and doesn't affect the delay caused by each retry. For
303 UDP each retry takes the time determined by the timeo
304 and retrans options, which by default will be about 7
305 seconds. For TCP the default is 3 minutes, but system
306 TCP connection timeouts will sometimes limit the timeout
307 of each retransmission to around 2 minutes.
308 sec=flavors A colon-separated list of one or more security flavors
310 to use for accessing files on the mounted export. If the
311 server does not support any of these flavors, the mount
312 operation fails. If sec= is not specified, the client
313 attempts to find a security flavor that both the client
314 and the server supports. Valid flavors are none, sys,
315 krb5, krb5i, and krb5p. Refer to the SECURITY CONSIDER‐
316 ATIONS section for details.
317 sharecache / nosharecache
319 Determines how the client's data cache and attribute
320 cache are shared when mounting the same export more than
321 once concurrently. Using the same cache reduces memory
322 requirements on the client and presents identical file
323 contents to applications when the same remote file is
324 accessed via different mount points.
325 If neither option is specified, or if the sharecache
327 option is specified, then a single cache is used for all
328 mount points that access the same export. If the
329 nosharecache option is specified, then that mount point
330 gets a unique cache. Note that when data and attribute
331 caches are shared, the mount options from the first
332 mount point take effect for subsequent concurrent mounts
333 of the same export.
334 As of kernel 2.6.18, the behavior specified by noshare‐
336 cache is legacy caching behavior. This is considered a
337 data risk since multiple cached copies of the same file
338 on the same client can become out of sync following a
339 local update of one of the copies.
340 resvport / noresvport
342 Specifies whether the NFS client should use a privileged
343 source port when communicating with an NFS server for
344 this mount point. If this option is not specified, or
345 the resvport option is specified, the NFS client uses a
346 privileged source port. If the noresvport option is
347 specified, the NFS client uses a non-privileged source
348 port. This option is supported in kernels 2.6.28 and
349 later.
350 Using non-privileged source ports helps increase the
352 maximum number of NFS mount points allowed on a client,
353 but NFS servers must be configured to allow clients to
354 connect via non-privileged source ports.
355 Refer to the SECURITY CONSIDERATIONS section for impor‐
357 tant details.
358 lookupcache=mode
360 Specifies how the kernel manages its cache of directory
361 entries for a given mount point. mode can be one of
362 all, none, pos, or positive. This option is supported
363 in kernels 2.6.28 and later.
364 The Linux NFS client caches the result of all NFS LOOKUP
366 requests. If the requested directory entry exists on
367 the server, the result is referred to as positive. If
368 the requested directory entry does not exist on the
369 server, the result is referred to as negative.
370 If this option is not specified, or if all is specified,
372 the client assumes both types of directory cache entries
373 are valid until their parent directory's cached
374 attributes expire.
375 If pos or positive is specified, the client assumes pos‐
377 itive entries are valid until their parent directory's
378 cached attributes expire, but always revalidates nega‐
379 tive entires before an application can use them.
380 If none is specified, the client revalidates both types
382 of directory cache entries before an application can use
383 them. This permits quick detection of files that were
384 created or removed by other clients, but can impact
385 application and server performance.
386 The DATA AND METADATA COHERENCE section contains a
388 detailed discussion of these trade-offs.
389 fsc / nofsc Enable/Disables the cache of (read-only) data pages to
391 the local disk using the FS-Cache facility. See
392 cachefilesd(8) and <kernel_soruce>/Documenta‐
393 tion/filesystems/caching for detail on how to configure
394 the FS-Cache facility. Default value is nofsc.
395 Options for NFS versions 2 and 3 only
397 Use these options, along with the options in the above subsection, for
398 NFS versions 2 and 3 only.
399 proto=netid The netid determines the transport that is used to com‐
401 municate with the NFS server. Available options are
402 udp, udp6, tcp, tcp6, and rdma. Those which end in 6
403 use IPv6 addresses and are only available if support for
404 TI-RPC is built in. Others use IPv4 addresses.
405 Each transport protocol uses different default retrans
407 and timeo settings. Refer to the description of these
408 two mount options for details.
409 In addition to controlling how the NFS client transmits
411 requests to the server, this mount option also controls
412 how the mount(8) command communicates with the server's
413 rpcbind and mountd services. Specifying a netid that
414 uses TCP forces all traffic from the mount(8) command
415 and the NFS client to use TCP. Specifying a netid that
416 uses UDP forces all traffic types to use UDP.
417 Before using NFS over UDP, refer to the TRANSPORT METH‐
419 ODS section.
420 If the proto mount option is not specified, the mount(8)
422 command discovers which protocols the server supports
423 and chooses an appropriate transport for each service.
424 Refer to the TRANSPORT METHODS section for more details.
425 udp The udp option is an alternative to specifying
427 proto=udp. It is included for compatibility with other
428 operating systems.
429 Before using NFS over UDP, refer to the TRANSPORT METH‐
431 ODS section.
432 tcp The tcp option is an alternative to specifying
434 proto=tcp. It is included for compatibility with other
435 operating systems.
436 rdma The rdma option is an alternative to specifying
438 proto=rdma.
439 port=n The numeric value of the server's NFS service port. If
441 the server's NFS service is not available on the speci‐
442 fied port, the mount request fails.
443 If this option is not specified, or if the specified
445 port value is 0, then the NFS client uses the NFS ser‐
446 vice port number advertised by the server's rpcbind ser‐
447 vice. The mount request fails if the server's rpcbind
448 service is not available, the server's NFS service is
449 not registered with its rpcbind service, or the server's
450 NFS service is not available on the advertised port.
451 mountport=n The numeric value of the server's mountd port. If the
453 server's mountd service is not available on the speci‐
454 fied port, the mount request fails.
455 If this option is not specified, or if the specified
457 port value is 0, then the mount(8) command uses the
458 mountd service port number advertised by the server's
459 rpcbind service. The mount request fails if the
460 server's rpcbind service is not available, the server's
461 mountd service is not registered with its rpcbind ser‐
462 vice, or the server's mountd service is not available on
463 the advertised port.
464 This option can be used when mounting an NFS server
466 through a firewall that blocks the rpcbind protocol.
467 mountproto=netid
469 The transport the NFS client uses to transmit requests
470 to the NFS server's mountd service when performing this
471 mount request, and when later unmounting this mount
472 point.
473 netid may be one of udp, and tcp which use IPv4 address
475 or, if TI-RPC is built into the mount.nfs command, udp6,
476 and tcp6 which use IPv6 addresses.
477 This option can be used when mounting an NFS server
479 through a firewall that blocks a particular transport.
480 When used in combination with the proto option, differ‐
481 ent transports for mountd requests and NFS requests can
482 be specified. If the server's mountd service is not
483 available via the specified transport, the mount request
484 fails.
485 Refer to the TRANSPORT METHODS section for more on how
487 the mountproto mount option interacts with the proto
488 mount option.
489 mounthost=name The hostname of the host running mountd. If this option
491 is not specified, the mount(8) command assumes that the
492 mountd service runs on the same host as the NFS service.
493 mountvers=n The RPC version number used to contact the server's
495 mountd. If this option is not specified, the client
496 uses a version number appropriate to the requested NFS
497 version. This option is useful when multiple NFS ser‐
498 vices are running on the same remote server host.
499 namlen=n The maximum length of a pathname component on this
501 mount. If this option is not specified, the maximum
502 length is negotiated with the server. In most cases,
503 this maximum length is 255 characters.
504 Some early versions of NFS did not support this negotia‐
506 tion. Using this option ensures that pathconf(3)
507 reports the proper maximum component length to applica‐
508 tions in such cases.
509 lock / nolock Selects whether to use the NLM sideband protocol to lock
511 files on the server. If neither option is specified (or
512 if lock is specified), NLM locking is used for this
513 mount point. When using the nolock option, applications
514 can lock files, but such locks provide exclusion only
515 against other applications running on the same client.
516 Remote applications are not affected by these locks.
517 NLM locking must be disabled with the nolock option when
519 using NFS to mount /var because /var contains files used
520 by the NLM implementation on Linux. Using the nolock
521 option is also required when mounting exports on NFS
522 servers that do not support the NLM protocol.
523 cto / nocto Selects whether to use close-to-open cache coherence
525 semantics. If neither option is specified (or if cto is
526 specified), the client uses close-to-open cache coher‐
527 ence semantics. If the nocto option is specified, the
528 client uses a non-standard heuristic to determine when
529 files on the server have changed.
530 Using the nocto option may improve performance for read-
532 only mounts, but should be used only if the data on the
533 server changes only occasionally. The DATA AND METADATA
534 COHERENCE section discusses the behavior of this option
535 in more detail.
536 acl / noacl Selects whether to use the NFSACL sideband protocol on
538 this mount point. The NFSACL sideband protocol is a
539 proprietary protocol implemented in Solaris that manages
540 Access Control Lists. NFSACL was never made a standard
541 part of the NFS protocol specification.
542 If neither acl nor noacl option is specified, the NFS
544 client negotiates with the server to see if the NFSACL
545 protocol is supported, and uses it if the server sup‐
546 ports it. Disabling the NFSACL sideband protocol may be
547 necessary if the negotiation causes problems on the
548 client or server. Refer to the SECURITY CONSIDERATIONS
549 section for more details.
550 local_lock=mechanism
552 Specifies whether to use local locking for any or both
553 of the flock and the POSIX locking mechanisms. mecha‐
554 nism can be one of all, flock, posix, or none. This
555 option is supported in kernels 2.6.37 and later.
556 The Linux NFS client provides a way to make locks local.
558 This means, the applications can lock files, but such
559 locks provide exclusion only against other applications
560 running on the same client. Remote applications are not
561 affected by these locks.
562 If this option is not specified, or if none is speci‐
564 fied, the client assumes that the locks are not local.
565 If all is specified, the client assumes that both flock
567 and POSIX locks are local.
568 If flock is specified, the client assumes that only
570 flock locks are local and uses NLM sideband protocol to
571 lock files when POSIX locks are used.
572 If posix is specified, the client assumes that POSIX
574 locks are local and uses NLM sideband protocol to lock
575 files when flock locks are used.
576 To support legacy flock behavior similar to that of NFS
578 clients < 2.6.12, use 'local_lock=flock'. This option is
579 required when exporting NFS mounts via Samba as Samba
580 maps Windows share mode locks as flock. Since NFS
581 clients > 2.6.12 implement flock by emulating POSIX
582 locks, this will result in conflicting locks.
583 NOTE: When used together, the 'local_lock' mount option
585 will be overridden by 'nolock'/'lock' mount option.
586 Options for NFS version 4 only
588 Use these options, along with the options in the first subsection
589 above, for NFS version 4.0 and newer.
590 proto=netid The netid determines the transport that is used to com‐
592 municate with the NFS server. Supported options are
593 tcp, tcp6, and rdma. tcp6 use IPv6 addresses and is
594 only available if support for TI-RPC is built in. Both
595 others use IPv4 addresses.
596 All NFS version 4 servers are required to support TCP,
598 so if this mount option is not specified, the NFS ver‐
599 sion 4 client uses the TCP protocol. Refer to the
600 TRANSPORT METHODS section for more details.
601 minorversion=n Specifies the protocol minor version number. NFSv4
603 introduces "minor versioning," where NFS protocol
604 enhancements can be introduced without bumping the NFS
605 protocol version number. Before kernel 2.6.38, the
606 minor version is always zero, and this option is not
607 recognized. After this kernel, specifying "minorver‐
608 sion=1" enables a number of advanced features, such as
609 NFSv4 sessions.
610 Recent kernels allow the minor version to be specified
612 using the vers= option. For example, specifying
613 vers=4.1 is the same as specifying vers=4,minorver‐
614 sion=1.
615 port=n The numeric value of the server's NFS service port. If
617 the server's NFS service is not available on the speci‐
618 fied port, the mount request fails.
619 If this mount option is not specified, the NFS client
621 uses the standard NFS port number of 2049 without first
622 checking the server's rpcbind service. This allows an
623 NFS version 4 client to contact an NFS version 4 server
624 through a firewall that may block rpcbind requests.
625 If the specified port value is 0, then the NFS client
627 uses the NFS service port number advertised by the
628 server's rpcbind service. The mount request fails if
629 the server's rpcbind service is not available, the
630 server's NFS service is not registered with its rpcbind
631 service, or the server's NFS service is not available on
632 the advertised port.
633 cto / nocto Selects whether to use close-to-open cache coherence
635 semantics for NFS directories on this mount point. If
636 neither cto nor nocto is specified, the default is to
637 use close-to-open cache coherence semantics for directo‐
638 ries.
639 File data caching behavior is not affected by this
641 option. The DATA AND METADATA COHERENCE section dis‐
642 cusses the behavior of this option in more detail.
643 clientaddr=n.n.n.n
645 clientaddr=n:n:...:n
647 Specifies a single IPv4 address (in dotted-quad form),
648 or a non-link-local IPv6 address, that the NFS client
649 advertises to allow servers to perform NFS version 4.0
650 callback requests against files on this mount point. If
651 the server is unable to establish callback connections
652 to clients, performance may degrade, or accesses to
653 files may temporarily hang. Can specify a value of
654 IPv4_ANY (0.0.0.0) or equivalent IPv6 any address which
655 will signal to the NFS server that this NFS client does
656 not want delegations.
657 If this option is not specified, the mount(8) command
659 attempts to discover an appropriate callback address
660 automatically. The automatic discovery process is not
661 perfect, however. In the presence of multiple client
662 network interfaces, special routing policies, or atypi‐
663 cal network topologies, the exact address to use for
664 callbacks may be nontrivial to determine.
665 NFS protocol versions 4.1 and 4.2 use the client-estab‐
667 lished TCP connection for callback requests, so do not
668 require the server to connect to the client. This
669 option is therefore only affect NFS version 4.0 mounts.
670 migration / nomigration
672 Selects whether the client uses an identification string
673 that is compatible with NFSv4 Transparent State Migra‐
674 tion (TSM). If the mounted server supports NFSv4 migra‐
675 tion with TSM, specify the migration option.
676 Some server features misbehave in the face of a migra‐
678 tion-compatible identification string. The nomigration
679 option retains the use of a traditional client indenti‐
680 fication string which is compatible with legacy NFS
681 servers. This is also the behavior if neither option is
682 specified. A client's open and lock state cannot be
683 migrated transparently when it identifies itself via a
684 traditional identification string.
685 This mount option has no effect with NFSv4 minor ver‐
687 sions newer than zero, which always use TSM-compatible
688 client identification strings.
689
691 The nfs4 file system type is an old syntax for specifying NFSv4 usage.
692 It can still be used with all NFSv4-specific and common options,
693 excepted the nfsvers mount option.
694
696 If the mount command is configured to do so, all of the mount options
697 described in the previous section can also be configured in the
698 /etc/nfsmount.conf file. See nfsmount.conf(5) for details.
699
701 To mount an export using NFS version 2, use the nfs file system type
702 and specify the nfsvers=2 mount option. To mount using NFS version 3,
703 use the nfs file system type and specify the nfsvers=3 mount option.
704 To mount using NFS version 4, use either the nfs file system type, with
705 the nfsvers=4 mount option, or the nfs4 file system type.
706 The following example from an /etc/fstab file causes the mount command
708 to negotiate reasonable defaults for NFS behavior.
709 server:/export /mnt nfs defaults 0 0
711 Here is an example from an /etc/fstab file for an NFS version 2 mount
713 over UDP.
714 server:/export /mnt nfs nfsvers=2,proto=udp 0 0
716 This example shows how to mount using NFS version 4 over TCP with Ker‐
718 beros 5 mutual authentication.
719 server:/export /mnt nfs4 sec=krb5 0 0
721 This example shows how to mount using NFS version 4 over TCP with Ker‐
723 beros 5 privacy or data integrity mode.
724 server:/export /mnt nfs4 sec=krb5p:krb5i 0 0
726 This example can be used to mount /usr over NFS.
728 server:/export /usr nfs ro,nolock,nocto,actimeo=3600 0 0
730 This example shows how to mount an NFS server using a raw IPv6 link-
732 local address.
733 [fe80::215:c5ff:fb3e:e2b1%eth0]:/export /mnt nfs defaults 0 0
735
737 NFS clients send requests to NFS servers via Remote Procedure Calls, or
738 RPCs. The RPC client discovers remote service endpoints automatically,
739 handles per-request authentication, adjusts request parameters for dif‐
740 ferent byte endianness on client and server, and retransmits requests
741 that may have been lost by the network or server. RPC requests and
742 replies flow over a network transport.
743 In most cases, the mount(8) command, NFS client, and NFS server can
745 automatically negotiate proper transport and data transfer size set‐
746 tings for a mount point. In some cases, however, it pays to specify
747 these settings explicitly using mount options.
748 Traditionally, NFS clients used the UDP transport exclusively for
750 transmitting requests to servers. Though its implementation is simple,
751 NFS over UDP has many limitations that prevent smooth operation and
752 good performance in some common deployment environments. Even an
753 insignificant packet loss rate results in the loss of whole NFS
754 requests; as such, retransmit timeouts are usually in the subsecond
755 range to allow clients to recover quickly from dropped requests, but
756 this can result in extraneous network traffic and server load.
757 However, UDP can be quite effective in specialized settings where the
759 networks MTU is large relative to NFSs data transfer size (such as net‐
760 work environments that enable jumbo Ethernet frames). In such environ‐
761 ments, trimming the rsize and wsize settings so that each NFS read or
762 write request fits in just a few network frames (or even in a single
763 frame) is advised. This reduces the probability that the loss of a
764 single MTU-sized network frame results in the loss of an entire large
765 read or write request.
766 TCP is the default transport protocol used for all modern NFS implemen‐
768 tations. It performs well in almost every conceivable network environ‐
769 ment and provides excellent guarantees against data corruption caused
770 by network unreliability. TCP is often a requirement for mounting a
771 server through a network firewall.
772 Under normal circumstances, networks drop packets much more frequently
774 than NFS servers drop requests. As such, an aggressive retransmit
775 timeout setting for NFS over TCP is unnecessary. Typical timeout set‐
776 tings for NFS over TCP are between one and ten minutes. After the
777 client exhausts its retransmits (the value of the retrans mount
778 option), it assumes a network partition has occurred, and attempts to
779 reconnect to the server on a fresh socket. Since TCP itself makes net‐
780 work data transfer reliable, rsize and wsize can safely be allowed to
781 default to the largest values supported by both client and server,
782 independent of the network's MTU size.
783 Using the mountproto mount option
785 This section applies only to NFS version 2 and version 3 mounts since
786 NFS version 4 does not use a separate protocol for mount requests.
787 The Linux NFS client can use a different transport for contacting an
789 NFS server's rpcbind service, its mountd service, its Network Lock Man‐
790 ager (NLM) service, and its NFS service. The exact transports employed
791 by the Linux NFS client for each mount point depends on the settings of
792 the transport mount options, which include proto, mountproto, udp, and
793 tcp.
794 The client sends Network Status Manager (NSM) notifications via UDP no
796 matter what transport options are specified, but listens for server NSM
797 notifications on both UDP and TCP. The NFS Access Control List
798 (NFSACL) protocol shares the same transport as the main NFS service.
799 If no transport options are specified, the Linux NFS client uses UDP to
801 contact the server's mountd service, and TCP to contact its NLM and NFS
802 services by default.
803 If the server does not support these transports for these services, the
805 mount(8) command attempts to discover what the server supports, and
806 then retries the mount request once using the discovered transports.
807 If the server does not advertise any transport supported by the client
808 or is misconfigured, the mount request fails. If the bg option is in
809 effect, the mount command backgrounds itself and continues to attempt
810 the specified mount request.
811 When the proto option, the udp option, or the tcp option is specified
813 but the mountproto option is not, the specified transport is used to
814 contact both the server's mountd service and for the NLM and NFS ser‐
815 vices.
816 If the mountproto option is specified but none of the proto, udp or tcp
818 options are specified, then the specified transport is used for the
819 initial mountd request, but the mount command attempts to discover what
820 the server supports for the NFS protocol, preferring TCP if both trans‐
821 ports are supported.
822 If both the mountproto and proto (or udp or tcp) options are specified,
824 then the transport specified by the mountproto option is used for the
825 initial mountd request, and the transport specified by the proto option
826 (or the udp or tcp options) is used for NFS, no matter what order these
827 options appear. No automatic service discovery is performed if these
828 options are specified.
829 If any of the proto, udp, tcp, or mountproto options are specified more
831 than once on the same mount command line, then the value of the right‐
832 most instance of each of these options takes effect.
833 Using NFS over UDP on high-speed links
835 Using NFS over UDP on high-speed links such as Gigabit can cause silent
836 data corruption.
837 The problem can be triggered at high loads, and is caused by problems
839 in IP fragment reassembly. NFS read and writes typically transmit UDP
840 packets of 4 Kilobytes or more, which have to be broken up into several
841 fragments in order to be sent over the Ethernet link, which limits
842 packets to 1500 bytes by default. This process happens at the IP net‐
843 work layer and is called fragmentation.
844 In order to identify fragments that belong together, IP assigns a 16bit
846 IP ID value to each packet; fragments generated from the same UDP
847 packet will have the same IP ID. The receiving system will collect
848 these fragments and combine them to form the original UDP packet. This
849 process is called reassembly. The default timeout for packet reassembly
850 is 30 seconds; if the network stack does not receive all fragments of a
851 given packet within this interval, it assumes the missing fragment(s)
852 got lost and discards those it already received.
853 The problem this creates over high-speed links is that it is possible
855 to send more than 65536 packets within 30 seconds. In fact, with heavy
856 NFS traffic one can observe that the IP IDs repeat after about 5 sec‐
857 onds.
858 This has serious effects on reassembly: if one fragment gets lost,
860 another fragment from a different packet but with the same IP ID will
861 arrive within the 30 second timeout, and the network stack will combine
862 these fragments to form a new packet. Most of the time, network layers
863 above IP will detect this mismatched reassembly - in the case of UDP,
864 the UDP checksum, which is a 16 bit checksum over the entire packet
865 payload, will usually not match, and UDP will discard the bad packet.
866 However, the UDP checksum is 16 bit only, so there is a chance of 1 in
868 65536 that it will match even if the packet payload is completely ran‐
869 dom (which very often isn't the case). If that is the case, silent data
870 corruption will occur.
871 This potential should be taken seriously, at least on Gigabit Ethernet.
873 Network speeds of 100Mbit/s should be considered less problematic,
874 because with most traffic patterns IP ID wrap around will take much
875 longer than 30 seconds.
876 It is therefore strongly recommended to use NFS over TCP where possi‐
878 ble, since TCP does not perform fragmentation.
879 If you absolutely have to use NFS over UDP over Gigabit Ethernet, some
881 steps can be taken to mitigate the problem and reduce the probability
882 of corruption:
883 Jumbo frames: Many Gigabit network cards are capable of transmitting
885 frames bigger than the 1500 byte limit of traditional
886 Ethernet, typically 9000 bytes. Using jumbo frames of
887 9000 bytes will allow you to run NFS over UDP at a page
888 size of 8K without fragmentation. Of course, this is
889 only feasible if all involved stations support jumbo
890 frames.
891 To enable a machine to send jumbo frames on cards that
893 support it, it is sufficient to configure the interface
894 for a MTU value of 9000.
895 Lower reassembly timeout:
897 By lowering this timeout below the time it takes the IP
898 ID counter to wrap around, incorrect reassembly of frag‐
899 ments can be prevented as well. To do so, simply write
900 the new timeout value (in seconds) to the file
901 /proc/sys/net/ipv4/ipfrag_time.
902 A value of 2 seconds will greatly reduce the probability
904 of IPID clashes on a single Gigabit link, while still
905 allowing for a reasonable timeout when receiving frag‐
906 mented traffic from distant peers.
907
909 Some modern cluster file systems provide perfect cache coherence among
910 their clients. Perfect cache coherence among disparate NFS clients is
911 expensive to achieve, especially on wide area networks. As such, NFS
912 settles for weaker cache coherence that satisfies the requirements of
913 most file sharing types.
914 Close-to-open cache consistency
916 Typically file sharing is completely sequential. First client A opens
917 a file, writes something to it, then closes it. Then client B opens
918 the same file, and reads the changes.
919 When an application opens a file stored on an NFS version 3 server, the
921 NFS client checks that the file exists on the server and is permitted
922 to the opener by sending a GETATTR or ACCESS request. The NFS client
923 sends these requests regardless of the freshness of the file's cached
924 attributes.
925 When the application closes the file, the NFS client writes back any
927 pending changes to the file so that the next opener can view the
928 changes. This also gives the NFS client an opportunity to report write
929 errors to the application via the return code from close(2).
930 The behavior of checking at open time and flushing at close time is
932 referred to as close-to-open cache consistency, or CTO. It can be dis‐
933 abled for an entire mount point using the nocto mount option.
934 Weak cache consistency
936 There are still opportunities for a client's data cache to contain
937 stale data. The NFS version 3 protocol introduced "weak cache consis‐
938 tency" (also known as WCC) which provides a way of efficiently checking
939 a file's attributes before and after a single request. This allows a
940 client to help identify changes that could have been made by other
941 clients.
942 When a client is using many concurrent operations that update the same
944 file at the same time (for example, during asynchronous write behind),
945 it is still difficult to tell whether it was that client's updates or
946 some other client's updates that altered the file.
947 Attribute caching
949 Use the noac mount option to achieve attribute cache coherence among
950 multiple clients. Almost every file system operation checks file
951 attribute information. The client keeps this information cached for a
952 period of time to reduce network and server load. When noac is in
953 effect, a client's file attribute cache is disabled, so each operation
954 that needs to check a file's attributes is forced to go back to the
955 server. This permits a client to see changes to a file very quickly,
956 at the cost of many extra network operations.
957 Be careful not to confuse the noac option with "no data caching." The
959 noac mount option prevents the client from caching file metadata, but
960 there are still races that may result in data cache incoherence between
961 client and server.
962 The NFS protocol is not designed to support true cluster file system
964 cache coherence without some type of application serialization. If
965 absolute cache coherence among clients is required, applications should
966 use file locking. Alternatively, applications can also open their files
967 with the O_DIRECT flag to disable data caching entirely.
968 File timestamp maintenance
970 NFS servers are responsible for managing file and directory timestamps
971 (atime, ctime, and mtime). When a file is accessed or updated on an
972 NFS server, the file's timestamps are updated just like they would be
973 on a filesystem local to an application.
974 NFS clients cache file attributes, including timestamps. A file's
976 timestamps are updated on NFS clients when its attributes are retrieved
977 from the NFS server. Thus there may be some delay before timestamp
978 updates on an NFS server appear to applications on NFS clients.
979 To comply with the POSIX filesystem standard, the Linux NFS client
981 relies on NFS servers to keep a file's mtime and ctime timestamps prop‐
982 erly up to date. It does this by flushing local data changes to the
983 server before reporting mtime to applications via system calls such as
984 stat(2).
985 The Linux client handles atime updates more loosely, however. NFS
987 clients maintain good performance by caching data, but that means that
988 application reads, which normally update atime, are not reflected to
989 the server where a file's atime is actually maintained.
990 Because of this caching behavior, the Linux NFS client does not support
992 generic atime-related mount options. See mount(8) for details on these
993 options.
994 In particular, the atime/noatime, diratime/nodiratime, relatime/norela‐
996 time, and strictatime/nostrictatime mount options have no effect on NFS
997 mounts.
998 /proc/mounts may report that the relatime mount option is set on NFS
1000 mounts, but in fact the atime semantics are always as described here,
1001 and are not like relatime semantics.
1002 Directory entry caching
1004 The Linux NFS client caches the result of all NFS LOOKUP requests. If
1005 the requested directory entry exists on the server, the result is
1006 referred to as a positive lookup result. If the requested directory
1007 entry does not exist on the server (that is, the server returned
1008 ENOENT), the result is referred to as negative lookup result.
1009 To detect when directory entries have been added or removed on the
1011 server, the Linux NFS client watches a directory's mtime. If the
1012 client detects a change in a directory's mtime, the client drops all
1013 cached LOOKUP results for that directory. Since the directory's mtime
1014 is a cached attribute, it may take some time before a client notices it
1015 has changed. See the descriptions of the acdirmin, acdirmax, and noac
1016 mount options for more information about how long a directory's mtime
1017 is cached.
1018 Caching directory entries improves the performance of applications that
1020 do not share files with applications on other clients. Using cached
1021 information about directories can interfere with applications that run
1022 concurrently on multiple clients and need to detect the creation or
1023 removal of files quickly, however. The lookupcache mount option allows
1024 some tuning of directory entry caching behavior.
1025 Before kernel release 2.6.28, the Linux NFS client tracked only posi‐
1027 tive lookup results. This permitted applications to detect new direc‐
1028 tory entries created by other clients quickly while still providing
1029 some of the performance benefits of caching. If an application depends
1030 on the previous lookup caching behavior of the Linux NFS client, you
1031 can use lookupcache=positive.
1032 If the client ignores its cache and validates every application lookup
1034 request with the server, that client can immediately detect when a new
1035 directory entry has been either created or removed by another client.
1036 You can specify this behavior using lookupcache=none. The extra NFS
1037 requests needed if the client does not cache directory entries can
1038 exact a performance penalty. Disabling lookup caching should result in
1039 less of a performance penalty than using noac, and has no effect on how
1040 the NFS client caches the attributes of files.
1041 The sync mount option
1043 The NFS client treats the sync mount option differently than some other
1044 file systems (refer to mount(8) for a description of the generic sync
1045 and async mount options). If neither sync nor async is specified (or
1046 if the async option is specified), the NFS client delays sending appli‐
1047 cation writes to the server until any of these events occur:
1048 Memory pressure forces reclamation of system memory resources.
1050 An application flushes file data explicitly with sync(2),
1052 msync(2), or fsync(3).
1053 An application closes a file with close(2).
1055 The file is locked/unlocked via fcntl(2).
1057 In other words, under normal circumstances, data written by an applica‐
1059 tion may not immediately appear on the server that hosts the file.
1060 If the sync option is specified on a mount point, any system call that
1062 writes data to files on that mount point causes that data to be flushed
1063 to the server before the system call returns control to user space.
1064 This provides greater data cache coherence among clients, but at a sig‐
1065 nificant performance cost.
1066 Applications can use the O_SYNC open flag to force application writes
1068 to individual files to go to the server immediately without the use of
1069 the sync mount option.
1070 Using file locks with NFS
1072 The Network Lock Manager protocol is a separate sideband protocol used
1073 to manage file locks in NFS version 2 and version 3. To support lock
1074 recovery after a client or server reboot, a second sideband protocol --
1075 known as the Network Status Manager protocol -- is also required. In
1076 NFS version 4, file locking is supported directly in the main NFS pro‐
1077 tocol, and the NLM and NSM sideband protocols are not used.
1078 In most cases, NLM and NSM services are started automatically, and no
1080 extra configuration is required. Configure all NFS clients with fully-
1081 qualified domain names to ensure that NFS servers can find clients to
1082 notify them of server reboots.
1083 NLM supports advisory file locks only. To lock NFS files, use fcntl(2)
1085 with the F_GETLK and F_SETLK commands. The NFS client converts file
1086 locks obtained via flock(2) to advisory locks.
1087 When mounting servers that do not support the NLM protocol, or when
1089 mounting an NFS server through a firewall that blocks the NLM service
1090 port, specify the nolock mount option. NLM locking must be disabled
1091 with the nolock option when using NFS to mount /var because /var con‐
1092 tains files used by the NLM implementation on Linux.
1093 Specifying the nolock option may also be advised to improve the perfor‐
1095 mance of a proprietary application which runs on a single client and
1096 uses file locks extensively.
1097 NFS version 4 caching features
1099 The data and metadata caching behavior of NFS version 4 clients is sim‐
1100 ilar to that of earlier versions. However, NFS version 4 adds two fea‐
1101 tures that improve cache behavior: change attributes and file delega‐
1102 tion.
1103 The change attribute is a new part of NFS file and directory metadata
1105 which tracks data changes. It replaces the use of a file's modifica‐
1106 tion and change time stamps as a way for clients to validate the con‐
1107 tent of their caches. Change attributes are independent of the time
1108 stamp resolution on either the server or client, however.
1109 A file delegation is a contract between an NFS version 4 client and
1111 server that allows the client to treat a file temporarily as if no
1112 other client is accessing it. The server promises to notify the client
1113 (via a callback request) if another client attempts to access that
1114 file. Once a file has been delegated to a client, the client can cache
1115 that file's data and metadata aggressively without contacting the
1116 server.
1117 File delegations come in two flavors: read and write. A read delega‐
1119 tion means that the server notifies the client about any other clients
1120 that want to write to the file. A write delegation means that the
1121 client gets notified about either read or write accessors.
1122 Servers grant file delegations when a file is opened, and can recall
1124 delegations at any time when another client wants access to the file
1125 that conflicts with any delegations already granted. Delegations on
1126 directories are not supported.
1127 In order to support delegation callback, the server checks the network
1129 return path to the client during the client's initial contact with the
1130 server. If contact with the client cannot be established, the server
1131 simply does not grant any delegations to that client.
1132
1134 NFS servers control access to file data, but they depend on their RPC
1135 implementation to provide authentication of NFS requests. Traditional
1136 NFS access control mimics the standard mode bit access control provided
1137 in local file systems. Traditional RPC authentication uses a number to
1138 represent each user (usually the user's own uid), a number to represent
1139 the user's group (the user's gid), and a set of up to 16 auxiliary
1140 group numbers to represent other groups of which the user may be a mem‐
1141 ber.
1142 Typically, file data and user ID values appear unencrypted (i.e. "in
1144 the clear") on the network. Moreover, NFS versions 2 and 3 use sepa‐
1145 rate sideband protocols for mounting, locking and unlocking files, and
1146 reporting system status of clients and servers. These auxiliary proto‐
1147 cols use no authentication.
1148 In addition to combining these sideband protocols with the main NFS
1150 protocol, NFS version 4 introduces more advanced forms of access con‐
1151 trol, authentication, and in-transit data protection. The NFS version
1152 4 specification mandates support for strong authentication and security
1153 flavors that provide per-RPC integrity checking and encryption.
1154 Because NFS version 4 combines the function of the sideband protocols
1155 into the main NFS protocol, the new security features apply to all NFS
1156 version 4 operations including mounting, file locking, and so on.
1157 RPCGSS authentication can also be used with NFS versions 2 and 3, but
1158 it does not protect their sideband protocols.
1159 The sec mount option specifies the security flavor used for operations
1161 on behalf of users on that NFS mount point. Specifying sec=krb5 pro‐
1162 vides cryptographic proof of a user's identity in each RPC request.
1163 This provides strong verification of the identity of users accessing
1164 data on the server. Note that additional configuration besides adding
1165 this mount option is required in order to enable Kerberos security.
1166 Refer to the rpc.gssd(8) man page for details.
1167 Two additional flavors of Kerberos security are supported: krb5i and
1169 krb5p. The krb5i security flavor provides a cryptographically strong
1170 guarantee that the data in each RPC request has not been tampered with.
1171 The krb5p security flavor encrypts every RPC request to prevent data
1172 exposure during network transit; however, expect some performance
1173 impact when using integrity checking or encryption. Similar support
1174 for other forms of cryptographic security is also available.
1175 NFS version 4 filesystem crossing
1177 The NFS version 4 protocol allows a client to renegotiate the security
1178 flavor when the client crosses into a new filesystem on the server.
1179 The newly negotiated flavor effects only accesses of the new filesys‐
1180 tem.
1181 Such negotiation typically occurs when a client crosses from a server's
1183 pseudo-fs into one of the server's exported physical filesystems, which
1184 often have more restrictive security settings than the pseudo-fs.
1185 NFS version 4 Leases
1187 In NFS version 4, a lease is a period of time during which a server
1188 irrevocably grants a file lock to a client. If the lease expires, the
1189 server is allowed to revoke that lock. Clients periodically renew
1190 their leases to prevent lock revocation.
1191 After an NFS version 4 server reboots, each client tells the server
1193 about all file open and lock state under its lease before operation can
1194 continue. If the client reboots, the server frees all open and lock
1195 state associated with that client's lease.
1196 As part of establishing a lease, therefore, a client must identify
1198 itself to a server. A fixed string is used to distinguish that client
1199 from others, and a changeable verifier is used to indicate when the
1200 client has rebooted.
1201 A client uses a particular security flavor and principal when perform‐
1203 ing the operations to establish a lease. If two clients happen to
1204 present the same identity string, a server can use their principals to
1205 detect that they are different clients, and prevent one client from
1206 interfering with the other's lease.
1207 The Linux NFS client establishes one lease for each server. Lease man‐
1209 agement operations, such as lease renewal, are not done on behalf of a
1210 particular file, lock, user, or mount point, but on behalf of the whole
1211 client that owns that lease. These operations must use the same secu‐
1212 rity flavor and principal that was used when the lease was established,
1213 even across client reboots.
1214 When Kerberos is configured on a Linux NFS client (i.e., there is a
1216 /etc/krb5.keytab on that client), the client attempts to use a Kerberos
1217 security flavor for its lease management operations. This provides
1218 strong authentication of the client to each server it contacts. By
1219 default, the client uses the host/ or nfs/ service principal in its
1220 /etc/krb5.keytab for this purpose.
1221 If the client has Kerberos configured, but the server does not, or if
1223 the client does not have a keytab or the requisite service principals,
1224 the client uses AUTH_SYS and UID 0 for lease management.
1225 Using non-privileged source ports
1227 NFS clients usually communicate with NFS servers via network sockets.
1228 Each end of a socket is assigned a port value, which is simply a number
1229 between 1 and 65535 that distinguishes socket endpoints at the same IP
1230 address. A socket is uniquely defined by a tuple that includes the
1231 transport protocol (TCP or UDP) and the port values and IP addresses of
1232 both endpoints.
1233 The NFS client can choose any source port value for its sockets, but
1235 usually chooses a privileged port. A privileged port is a port value
1236 less than 1024. Only a process with root privileges may create a
1237 socket with a privileged source port.
1238 The exact range of privileged source ports that can be chosen is set by
1240 a pair of sysctls to avoid choosing a well-known port, such as the port
1241 used by ssh. This means the number of source ports available for the
1242 NFS client, and therefore the number of socket connections that can be
1243 used at the same time, is practically limited to only a few hundred.
1244 As described above, the traditional default NFS authentication scheme,
1246 known as AUTH_SYS, relies on sending local UID and GID numbers to iden‐
1247 tify users making NFS requests. An NFS server assumes that if a con‐
1248 nection comes from a privileged port, the UID and GID numbers in the
1249 NFS requests on this connection have been verified by the client's ker‐
1250 nel or some other local authority. This is an easy system to spoof,
1251 but on a trusted physical network between trusted hosts, it is entirely
1252 adequate.
1253 Roughly speaking, one socket is used for each NFS mount point. If a
1255 client could use non-privileged source ports as well, the number of
1256 sockets allowed, and thus the maximum number of concurrent mount
1257 points, would be much larger.
1258 Using non-privileged source ports may compromise server security some‐
1260 what, since any user on AUTH_SYS mount points can now pretend to be any
1261 other when making NFS requests. Thus NFS servers do not support this
1262 by default. They explicitly allow it usually via an export option.
1263 To retain good security while allowing as many mount points as possi‐
1265 ble, it is best to allow non-privileged client connections only if the
1266 server and client both require strong authentication, such as Kerberos.
1267 Mounting through a firewall
1269 A firewall may reside between an NFS client and server, or the client
1270 or server may block some of its own ports via IP filter rules. It is
1271 still possible to mount an NFS server through a firewall, though some
1272 of the mount(8) command's automatic service endpoint discovery mecha‐
1273 nisms may not work; this requires you to provide specific endpoint
1274 details via NFS mount options.
1275 NFS servers normally run a portmapper or rpcbind daemon to advertise
1277 their service endpoints to clients. Clients use the rpcbind daemon to
1278 determine:
1279 What network port each RPC-based service is using
1281 What transport protocols each RPC-based service supports
1283 The rpcbind daemon uses a well-known port number (111) to help clients
1285 find a service endpoint. Although NFS often uses a standard port num‐
1286 ber (2049), auxiliary services such as the NLM service can choose any
1287 unused port number at random.
1288 Common firewall configurations block the well-known rpcbind port. In
1290 the absense of an rpcbind service, the server administrator fixes the
1291 port number of NFS-related services so that the firewall can allow
1292 access to specific NFS service ports. Client administrators then spec‐
1293 ify the port number for the mountd service via the mount(8) command's
1294 mountport option. It may also be necessary to enforce the use of TCP
1295 or UDP if the firewall blocks one of those transports.
1296 NFS Access Control Lists
1298 Solaris allows NFS version 3 clients direct access to POSIX Access Con‐
1299 trol Lists stored in its local file systems. This proprietary sideband
1300 protocol, known as NFSACL, provides richer access control than mode
1301 bits. Linux implements this protocol for compatibility with the
1302 Solaris NFS implementation. The NFSACL protocol never became a stan‐
1303 dard part of the NFS version 3 specification, however.
1304 The NFS version 4 specification mandates a new version of Access Con‐
1306 trol Lists that are semantically richer than POSIX ACLs. NFS version 4
1307 ACLs are not fully compatible with POSIX ACLs; as such, some transla‐
1308 tion between the two is required in an environment that mixes POSIX
1309 ACLs and NFS version 4.
1310
1312 Generic mount options such as rw and sync can be modified on NFS mount
1313 points using the remount option. See mount(8) for more information on
1314 generic mount options.
1315 With few exceptions, NFS-specific options are not able to be modified
1317 during a remount. The underlying transport or NFS version cannot be
1318 changed by a remount, for example.
1319 Performing a remount on an NFS file system mounted with the noac option
1321 may have unintended consequences. The noac option is a combination of
1322 the generic option sync, and the NFS-specific option actimeo=0.
1323 Unmounting after a remount
1325 For mount points that use NFS versions 2 or 3, the NFS umount subcom‐
1326 mand depends on knowing the original set of mount options used to per‐
1327 form the MNT operation. These options are stored on disk by the NFS
1328 mount subcommand, and can be erased by a remount.
1329 To ensure that the saved mount options are not erased during a remount,
1331 specify either the local mount directory, or the server hostname and
1332 export pathname, but not both, during a remount. For example,
1333 mount -o remount,ro /mnt
1335 merges the mount option ro with the mount options already saved on disk
1337 for the NFS server mounted at /mnt.
1338
1340 /etc/fstab file system table
1341 /etc/nfsmount.conf
1343 Configuration file for NFS mounts
1344
1346 Before 2.4.7, the Linux NFS client did not support NFS over TCP.
1347 Before 2.4.20, the Linux NFS client used a heuristic to determine
1349 whether cached file data was still valid rather than using the standard
1350 close-to-open cache coherency method described above.
1351 Starting with 2.4.22, the Linux NFS client employs a Van Jacobsen-based
1353 RTT estimator to determine retransmit timeout values when using NFS
1354 over UDP.
1355 Before 2.6.0, the Linux NFS client did not support NFS version 4.
1357 Before 2.6.8, the Linux NFS client used only synchronous reads and
1359 writes when the rsize and wsize settings were smaller than the system's
1360 page size.
1361 The Linux client's support for protocol versions depend on whether the
1363 kernel was built with options CONFIG_NFS_V2, CONFIG_NFS_V3, CON‐
1364 FIG_NFS_V4, CONFIG_NFS_V4_1, and CONFIG_NFS_V4_2.
1365
1367 fstab(5), mount(8), umount(8), mount.nfs(5), umount.nfs(5), exports(5),
1368 nfsmount.conf(5), netconfig(5), ipv6(7), nfsd(8), sm-notify(8),
1369 rpc.statd(8), rpc.idmapd(8), rpc.gssd(8), rpc.svcgssd(8), kerberos(1)
1370 RFC 768 for the UDP specification.
1372 RFC 793 for the TCP specification.
1373 RFC 1094 for the NFS version 2 specification.
1374 RFC 1813 for the NFS version 3 specification.
1375 RFC 1832 for the XDR specification.
1376 RFC 1833 for the RPC bind specification.
1377 RFC 2203 for the RPCSEC GSS API protocol specification.
1378 RFC 7530 for the NFS version 4.0 specification.
1379 RFC 5661 for the NFS version 4.1 specification.
1380 RFC 7862 for the NFS version 4.2 specification.
1381 9 October 2012 NFS(5)