1 .\" nfs.5 "Rick Sladkey" <jrs@world.std.com>
2 .\" Wed Feb 8 12:52:42 1995, faith@cs.unc.edu: updates for Ross Biro's
4 .TH NFS 5 "20 November 1993" "Linux 0.99" "Linux Programmer's Manual"
6 nfs \- nfs and nfs4 fstab format and options
12 file contains information about which filesystems
13 to mount where and with what options.
14 For NFS mounts, it contains the server name and
15 exported server directory to mount from,
16 the local directory that is the mount point,
17 and the NFS specific options that control
18 the way the filesystem is mounted.
20 Three different versions of the NFS protocol are
21 supported by the Linux NFS client:
22 NFS version 2, NFS version 3, and NFS version 4.
23 To mount via NFS version 2, use the
25 file system type and specify
27 To mount via NFS version 3, use the
29 file system type and specify
31 Version 3 is the default protocol version for the
35 is not specified on the mount command and both client and server
37 To mount via NFS version 4, use the
42 keyword is not supported for the
46 These file system types share similar mount options;
47 the differences are listed below.
49 Here is an example from an \fI/etc/fstab\fP file for an NFSv3 mount
53 .ta 2.5i +0.75i +0.75i +1.0i
54 server:/usr/local/pub /pub nfs rsize=32768,wsize=32768,timeo=14,intr
57 Here is an example for an NFSv4 mount over TCP using Kerberos
58 5 mutual authentication.
61 .ta 2.5i +0.75i +0.75i +1.0i
62 server:/usr/local/pub /pub nfs4 proto=tcp,sec=krb5,hard,intr
65 .SS Options for the nfs file system type
68 The number of bytes NFS uses when reading files from an NFS server.
69 The rsize is negotiated between the server and client to determine
70 the largest block size that both can support.
71 The value specified by this option is the maximum size that could
72 be used; however, the actual size used may be smaller.
73 Note: Setting this size to a value less than the largest supported
74 block size will adversely affect performance.
77 The number of bytes NFS uses when writing files to an NFS server.
78 The wsize is negotiated between the server and client to determine
79 the largest block size that both can support.
80 The value specified by this option is the maximum size that could
81 be used; however, the actual size used may be smaller.
82 Note: Setting this size to a value less than the largest supported
83 block size will adversely affect performance.
86 The value in tenths of a second before sending the
87 first retransmission after an RPC timeout.
88 The default value is 7 tenths of a second. After the first timeout,
89 the timeout is doubled after each successive timeout until a maximum
90 timeout of 60 seconds is reached or the enough retransmissions
91 have occured to cause a major timeout. Then, if the filesystem
92 is hard mounted, each new timeout cascade restarts at twice the
93 initial value of the previous cascade, again doubling at each
94 retransmission. The maximum timeout is always 60 seconds.
95 Better overall performance may be achieved by increasing the
96 timeout when mounting on a busy network, to a slow server, or through
97 several routers or gateways.
100 The number of minor timeouts and retransmissions that must occur before
101 a major timeout occurs. The default is 3 timeouts. When a major timeout
102 occurs, the file operation is either aborted or a "server not responding"
103 message is printed on the console.
106 The minimum time in seconds that attributes of a regular file should
107 be cached before requesting fresh information from a server.
108 The default is 3 seconds.
111 The maximum time in seconds that attributes of a regular file can
112 be cached before requesting fresh information from a server.
113 The default is 60 seconds.
116 The minimum time in seconds that attributes of a directory should
117 be cached before requesting fresh information from a server.
118 The default is 30 seconds.
121 The maximum time in seconds that attributes of a directory can
122 be cached before requesting fresh information from a server.
123 The default is 60 seconds.
126 Using actimeo sets all of
133 There is no default value.
136 The number of minutes to retry an NFS mount operation
137 in the foreground or background before giving up.
138 The default value for forground mounts is 2 minutes.
139 The default value for background mounts is 10000 minutes,
140 which is roughly one week.
143 When an NFS server does not support version two of the
144 RPC mount protocol, this option can be used to specify
145 the maximum length of a filename that is supported on
146 the remote filesystem. This is used to support the
147 POSIX pathconf functions. The default is 255 characters.
150 The numeric value of the port to connect to the NFS server on.
151 If the port number is 0 (the default) then query the
152 remote host's portmapper for the port number to use.
153 If the remote host's NFS daemon is not registered with
154 its portmapper, the standard NFS port number 2049 is
158 The numeric value of the
163 The name of the host running
167 Use an alternate RPC program number to contact the
168 mount daemon on the remote host. This option is useful
169 for hosts that can run multiple NFS servers.
170 The default value is 100005 which is the standard RPC
171 mount daemon program number.
174 Use an alternate RPC version number to contact the
175 mount daemon on the remote host. This option is useful
176 for hosts that can run multiple NFS servers.
177 The default value depends on which kernel you are using.
180 Use an alternate RPC program number to contact the
181 NFS daemon on the remote host. This option is useful
182 for hosts that can run multiple NFS servers.
183 The default value is 100003 which is the standard RPC
184 NFS daemon program number.
187 Use an alternate RPC version number to contact the
188 NFS daemon on the remote host. This option is useful
189 for hosts that can run multiple NFS servers.
190 The default value depends on which kernel you are using.
193 vers is an alternative to nfsvers and is compatible with
194 many other operating systems.
197 Disable NFS locking. Do not start lockd.
198 This is appropriate for mounting the root filesystem or
202 These filesystems are typically either read-only or not shared, and in
203 those cases, remote locking is not needed.
204 This also needs to be used with some old NFS servers
205 that don't support locking.
207 Note that applications can still get locks on files, but the locks
208 only provide exclusion locally. Other clients mounting the same
209 filesystem will not be able to detect the locks.
212 If the first NFS mount attempt times out, retry the mount
214 After a mount operation is backgrounded, all subsequent mounts
215 on the same NFS server will be backgrounded immediately, without
216 first attempting the mount.
217 A missing mount point is treated as a timeout,
218 to allow for nested NFS mounts.
221 If the first NFS mount attempt times out, retry the mount
223 This is the complement of the
225 option, and also the default behavior.
228 If an NFS file operation has a major timeout then report an I/O error to
230 The default is to continue retrying NFS file operations indefinitely.
233 If an NFS file operation has a major timeout then report
234 "server not responding" on the console and continue retrying indefinitely.
238 If an NFS file operation has a major timeout and it is hard mounted,
239 then allow signals to interupt the file operation and cause it to
240 return EINTR to the calling program. The default is to not
241 allow file operations to be interrupted.
244 Mount the NFS filesystem using POSIX semantics. This allows
245 an NFS filesystem to properly support the POSIX pathconf
246 command by querying the mount server for the maximum length
247 of a filename. To do this, the remote host must support version
248 two of the RPC mount protocol. Many NFS servers support only
252 Suppress the retrieval of new attributes when creating a file.
255 Disable all forms of attribute caching entirely. This extracts a
256 significant performance penalty but it allows two different NFS clients
257 to get reasonable results when both clients are actively
258 writing to a common export on the server.
261 Disables Access Control List (ACL) processing.
264 Set the security flavor for this mount to "mode".
265 The default setting is \f3sec=sys\f1, which uses local
266 unix uids and gids to authenticate NFS operations (AUTH_SYS).
267 Other currently supported settings are:
268 \f3sec=krb5\f1, which uses Kerberos V5 instead of local unix uids
269 and gids to authenticate users;
270 \f3sec=krb5i\f1, which uses Kerberos V5 for user authentication
271 and performs integrity checking of NFS operations using secure
272 checksums to prevent data tampering; and
273 \f3sec=krb5p\f1, which uses Kerberos V5 for user authentication
274 and integrity checking, and encrypts NFS traffic to prevent
275 traffic sniffing (this is the most secure setting).
276 Note that there is a performance penalty when using integrity
280 Mount the NFS filesystem using the TCP protocol. This is the default
281 if it is supported by both client and server. Many NFS servers only
285 Mount the NFS filesystem using the UDP protocol.
288 Disables NFSv3 READDIRPLUS RPCs. Use this option when
289 mounting servers that don't support or have broken
290 READDIRPLUS implementations.
293 As of kernel 2.6.18, it is no longer possible to mount the same
294 same filesystem with different mount options to a new mountpoint.
295 It was deemed unsafe to do so, since cached data cannot be shared
296 between the two mountpoints. In consequence, files or directories
297 that were common to both mountpoint subtrees could often be seen to
298 be out of sync following an update.
300 This option allows administrators to select the pre-2.6.18 behaviour,
301 permitting the same filesystem to be mounted with different mount
305 Use of this option is not recommended unless you are certain that there
306 are no hard links or subtrees of this mountpoint that are mounted
309 All of the non-value options have corresponding nooption forms.
310 For example, nointr means don't allow file operations to be
312 .SS Options for the nfs4 file system type
315 The number of bytes nfs4 uses when reading files from the server.
316 The rsize is negotiated between the server and client to determine
317 the largest block size that both can support.
318 The value specified by this option is the maximum size that could
319 be used; however, the actual size used may be smaller.
320 Note: Setting this size to a value less than the largest supported
321 block size will adversely affect performance.
324 The number of bytes nfs4 uses when writing files to the server.
325 The wsize is negotiated between the server and client to determine
326 the largest block size that both can support.
327 The value specified by this option is the maximum size that could
328 be used; however, the actual size used may be smaller.
329 Note: Setting this size to a value less than the largest supported
330 block size will adversely affect performance.
333 The value in tenths of a second before sending the
334 first retransmission after an RPC timeout.
335 The default value depends on whether
339 is in effect (see below).
340 The default value for UDP is 7 tenths of a second.
341 The default value for TCP is 60 seconds.
342 After the first timeout,
343 the timeout is doubled after each successive timeout until a maximum
344 timeout of 60 seconds is reached or the enough retransmissions
345 have occured to cause a major timeout. Then, if the filesystem
346 is hard mounted, each new timeout cascade restarts at twice the
347 initial value of the previous cascade, again doubling at each
348 retransmission. The maximum timeout is always 60 seconds.
351 The number of minor timeouts and retransmissions that must occur before
352 a major timeout occurs. The default is 5 timeouts for
357 occurs, the file operation is either aborted or a "server not responding"
358 message is printed on the console.
361 The minimum time in seconds that attributes of a regular file should
362 be cached before requesting fresh information from a server.
363 The default is 3 seconds.
366 The maximum time in seconds that attributes of a regular file can
367 be cached before requesting fresh information from a server.
368 The default is 60 seconds.
371 The minimum time in seconds that attributes of a directory should
372 be cached before requesting fresh information from a server.
373 The default is 30 seconds.
376 The maximum time in seconds that attributes of a directory can
377 be cached before requesting fresh information from a server.
378 The default is 60 seconds.
381 Using actimeo sets all of
388 There is no default value.
391 The number of minutes to retry an NFS mount operation
392 in the foreground or background before giving up.
393 The default value for forground mounts is 2 minutes.
394 The default value for background mounts is 10000 minutes,
395 which is roughly one week.
398 The numeric value of the port to connect to the NFS server on.
399 If the port number is 0 (the default) then query the
400 remote host's portmapper for the port number to use.
401 If the remote host's NFS daemon is not registered with
402 its portmapper, the standard NFS port number 2049 is
406 Mount the NFS filesystem using a specific network protocol
407 instead of the default UDP protocol.
408 Many NFS version 4 servers only support TCP.
409 Valid protocol types are
414 .I clientaddr=n.n.n.n
415 Specifies a single IPv4 address in dotted-quad form that
416 the NFS client advertises to allow servers to perform
417 NFSv4 callback requests against files on this mount point.
418 If the server is not able to establish callback connections
419 to clients, performance may degrade, or accesses to
420 files may temporarily hang.
422 If this option is not specified, the
424 command attempts to discover an appropriate callback
425 address automatically.
426 The automatic discovery process is not perfect, however.
427 In the presence of multiple client network interfaces, special
428 routing policies, or atypical network topologies, the exact
429 address to use for callbacks may be nontrivial to determine,
430 and should be explicitly set using this mount option.
433 Same as \f3sec=mode\f1 for the nfs filesystem type (see above).
436 If an NFS mount attempt times out, retry the mount
438 After a mount operation is backgrounded, all subsequent mounts
439 on the same NFS server will be backgrounded immediately, without
440 first attempting the mount.
441 A missing mount point is treated as a timeout,
442 to allow for nested NFS mounts.
445 If the first NFS mount attempt times out, retry the mount
447 This is the complement of the
449 option, and also the default behavior.
452 If an NFS file operation has a major timeout then report an I/O error to
454 The default is to continue retrying NFS file operations indefinitely.
457 If an NFS file operation has a major timeout then report
458 "server not responding" on the console and continue retrying indefinitely.
462 If an NFS file operation has a major timeout and it is hard mounted,
463 then allow signals to interupt the file operation and cause it to
464 return EINTR to the calling program. The default is to not
465 allow file operations to be interrupted.
468 Suppress the retrieval of new attributes when creating a file.
471 Disable attribute caching, and force synchronous writes.
473 server performance penalty but it allows two different NFS clients
474 to get reasonable good results when both clients are actively
475 writing to common filesystem on the server.
478 As of kernel 2.6.18, it is no longer possible to mount the same
479 same filesystem with different mount options to a new mountpoint.
480 It was deemed unsafe to do so, since cached data cannot be shared
481 between the two mountpoints. In consequence, files or directories
482 that were common to both mountpoint subtrees could often be seen to
483 be out of sync following an update.
485 This option allows administrators to select the pre-2.6.18 behaviour,
486 permitting the same filesystem to be mounted with different mount
490 Use of this option is not recommended unless you are certain that there
491 are no hard links or subtrees of this mountpoint that are mounted
494 All of the non-value options have corresponding nooption forms.
495 For example, nointr means don't allow file operations to be
500 .BR fstab "(5), " mount "(8), " umount "(8), " exports (5)
502 "Rick Sladkey" <jrs@world.std.com>
505 Checking files on NFS filesystem referenced by file descriptors (i.e. the
509 families of functions) may lead to inconsistent result due to the lack of
510 consistency check in kernel even if noac is used.