1 .TH EXPORTS 5 "4 March 2005" "Linux" "Linux File Formats Manual"
3 exports \- NFS file systems being exported (for Kernel based NFS)
9 serves as the access control list for file systems which may be
10 exported to NFS clients. It is used by
12 to give information to
14 and to the kernel based NFS file server daemon
17 The file format is similar to the SunOS
19 file. Each line contains an export point and a whitespace-separated list
20 of clients allowed to mount the file system at that point. Each listed
21 client may be immediately followed by a parenthesized, comma-separated
22 list of export options for that client. No whitespace is permitted
23 between a client and its option list.
25 Also, each line may have one or more specifications for default options
26 after the path name, in the form of a dash ("\-") followed by an option
27 list. The option list is used for all subsequent exports on that line
30 Blank lines are ignored. A pound sign ("#") introduces a comment to the
31 end of the line. Entries may be continued across newlines using a
32 backslash. If an export name contains spaces it should be quoted using
33 double quotes. You can also specify spaces or other unusual character in
34 the export name using a backslash followed by the character code as three
37 .SS Machine Name Formats
38 NFS clients may be specified in a number of ways:
40 This is the most common format. You may specify a host either by an
41 abbreviated name recognized be the resolver, the fully qualified domain
42 name, or an IP address.
44 NIS netgroups may be given as
46 Only the host part of each
47 netgroup members is consider in checking for membership. Empty host
48 parts or those containing a single dash (\-) are ignored.
50 Machine names may contain the wildcard characters \fI*\fR and \fI?\fR.
51 This can be used to make the \fIexports\fR file more compact; for instance,
52 \fI*.cs.foo.edu\fR matches all hosts in the domain
53 \fIcs.foo.edu\fR. As these characters also match the dots in a domain
54 name, the given pattern will also match all hosts within any subdomain
57 You can also export directories to all hosts on an IP (sub-) network
58 simultaneously. This is done by specifying an IP address and netmask pair
61 where the netmask can be specified in dotted-decimal format, or as a
62 contiguous mask length (for example, either `/255.255.252.0' or `/22' appended
63 to the network base address result in identical subnetworks with 10 bits of
64 host). Wildcard characters generally do not work on IP addresses, though they
65 may work by accident when reverse DNS lookups fail.
68 '''This is a special ``hostname'' that identifies the given directory name
69 '''as the public root directory (see the section on WebNFS in
71 '''for a discussion of WebNFS and the public root handle). When using this
74 '''must be the only entry on this line, and must have no export options
75 '''associated with it. Note that this does
77 '''actually export the named directory; you still have to set the exports
78 '''options in a separate entry.
80 '''The public root path can also be specified by invoking
83 '''.B \-\-public\-root
84 '''option. Multiple specifications of a public root will be ignored.
86 .SS RPCSEC_GSS security
87 You may use the special strings "gss/krb5", "gss/krb5i", or "gss/krb5p"
88 to restrict access to clients using rpcsec_gss security. However, this
89 syntax is deprecated; on linux kernels since 2.6.23, you should instead
90 use the "sec=" export option:
93 The sec= option, followed by a colon-delimited list of security flavors,
94 restricts the export to clients using those flavors. Available security
95 flavors include sys (the default--no cryptographic security), krb5
96 (authentication only), krb5i (integrity protection), and krb5p (privacy
97 protection). For the purposes of security flavor negotiation, order
98 counts: preferred flavors should be listed first. The order of the sec=
99 option with respect to the other options does not matter, unless you
100 want some options to be enforced differently depending on flavor.
101 In that case you may include multiple sec= options, and following options
102 will be enforced only for access using flavors listed in the immediately
103 preceding sec= option. The only options that are permitted to vary in
104 this way are ro, rw, no_root_squash, root_squash, and all_squash.
108 understands the following export options:
111 This option requires that requests originate on an Internet port less
112 than IPPORT_RESERVED (1024). This option is on by default. To turn it
117 Allow both read and write requests on this NFS volume. The
118 default is to disallow any request which changes the filesystem.
119 This can also be made explicit by using
124 This option allows the NFS server to violate the NFS protocol and
125 reply to requests before any changes made by that request have been
126 committed to stable storage (e.g. disc drive).
128 Using this option usually improves performance, but at the cost that
129 an unclean server restart (i.e. a crash) can cause data to be lost or
134 Reply to requests only after the changes have been committed to stable
139 In releases of nfs-utils up to and including 1.0.0, this option was the
140 default. In all subsequence releases,
144 must be explicitly requested if needed.
145 To help make system administrators aware of this change, 'exportfs'
146 will issue a warning if neither
153 This option has no effect if
155 is also set. The NFS server will normally delay committing a write request
156 to disc slightly if it suspects that another related write request may be in
157 progress or may arrive soon. This allows multiple write requests to
158 be committed to disc with the one operation which can improve
159 performance. If an NFS server received mainly small unrelated
160 requests, this behaviour could actually reduce performance, so
162 is available to turn it off.
163 The default can be explicitly requested with the
167 This option is based on the option of the same name provided in IRIX
168 NFS. Normally, if a server exports two filesystems one of which is
169 mounted on the other, then the client will have to mount both
170 filesystems explicitly to get access to them. If it just mounts the
171 parent, it will see an empty directory at the place where the other
172 filesystem is mounted. That filesystem is "hidden".
176 option on a filesystem causes it not to be hidden, and an
177 appropriately authorised client will be able to move from the parent to
178 that filesystem without noticing the change.
180 However, some NFS clients do not cope well with this situation as, for
181 instance, it is then possible for two files in the one apparent
182 filesystem to have the same inode number.
186 option is currently only effective on
188 exports. It does not work reliably with netgroup, subnet, or wildcard
191 This option can be very useful in some situations, but it should be
192 used with due care, and only after confirming that the client system
193 copes with the situation effectively.
195 The option can be explicitly disabled with
199 This option is similar to
201 but it makes it possible for clients to move from the filesystem marked
202 with crossmnt to exported filesystems mounted on it. Thus when a child
203 filesystem "B" is mounted on a parent "A", setting crossmnt on "A" has
204 the same effect as setting "nohide" on B.
207 This option disables subtree checking, which has mild security
208 implications, but can improve reliability in some circumstances.
210 If a subdirectory of a filesystem is exported, but the whole
211 filesystem isn't then whenever a NFS request arrives, the server must
212 check not only that the accessed file is in the appropriate filesystem
213 (which is easy) but also that it is in the exported tree (which is
214 harder). This check is called the
217 In order to perform this check, the server must include some
218 information about the location of the file in the "filehandle" that is
219 given to the client. This can cause problems with accessing files that
220 are renamed while a client has them open (though in many simple cases
223 subtree checking is also used to make sure that files inside
224 directories to which only root has access can only be accessed if the
225 filesystem is exported with
227 (see below), even if the file itself allows more general access.
229 As a general guide, a home directory filesystem, which is normally
230 exported at the root and may see lots of file renames, should be
231 exported with subtree checking disabled. A filesystem which is mostly
232 readonly, and at least doesn't see many file renames (e.g. /usr or
233 /var) and for which subdirectories may be exported, should probably be
234 exported with subtree checks enabled.
236 The default of having subtree checks enabled, can be explicitly
240 From release 1.1.0 of nfs-utils onwards, the default will be
242 as subtree_checking tends to cause more problems than it is worth.
243 If you genuinely require subtree checking, you should explicitly put
246 file. If you put neither option,
248 will warn you that the change is pending.
254 This option (the two names are synonymous) tells the NFS server not to require authentication of
255 locking requests (i.e. requests which use the NLM protocol). Normally
256 the NFS server will require a lock request to hold a credential for a
257 user who has read access to the file. With this flag no access checks
260 Early NFS client implementations did not send credentials with lock
261 requests, and many current NFS clients still exist which are based on
262 the old implementations. Use this flag if you find that you can only
263 lock files which are world readable.
265 The default behaviour of requiring authentication for NLM requests can
266 be explicitly requested with either of the synonymous
272 On some specially patched kernels, and when exporting filesystems that
273 support ACLs, this option tells nfsd not to reveal ACLs to clients, so
274 they will see only a subset of actual permissions on the given file
275 system. This option is safe for filesystems used by NFSv2 clients and
276 old NFSv3 clients that perform access decisions locally. Current
277 NFSv3 clients use the ACCESS RPC to perform all access decisions on
278 the server. Note that the
280 option only has effect on kernels specially patched to support it, and
281 when exporting filesystems with ACL support. The default is to export
282 with ACL support (i.e. by default,
288 '''This makes everything below the directory inaccessible for the named
289 '''client. This is useful when you want to export a directory hierarchy to
290 '''a client, but exclude certain subdirectories. The client's view of a
291 '''directory flagged with noaccess is very limited; it is allowed to read
292 '''its attributes, and lookup `.' and `..'. These are also the only entries
293 '''returned by a readdir.
296 '''Convert absolute symbolic links (where the link contents start with a
297 '''slash) into relative links by prepending the necessary number of ../'s
298 '''to get from the directory containing the link to the root on the
299 '''server. This has subtle, perhaps questionable, semantics when the file
300 '''hierarchy is not mounted at its root.
303 '''Leave all symbolic link as they are. This is the default operation.
309 This option makes it possible to only export a directory if it has
310 successfully been mounted.
311 If no path is given (e.g.
312 .IR mountpoint " or " mp )
313 then the export point must also be a mount point. If it isn't then
314 the export point is not exported. This allows you to be sure that the
315 directory underneath a mountpoint will never be exported by accident
316 if, for example, the filesystem failed to mount due to a disc error.
318 If a path is given (e.g.
319 .IR mountpoint= "/path or " mp= /path)
320 then the nominated path must be a mountpoint for the exportpoint to be
324 .IR fsid= num|root|uuid
325 NFS needs to be able to identify each filesystem that it exports.
326 Normally it will use a UUID for the filesystem (if the filesystem has
327 such a thing) or the device number of the device holding the
328 filesystem (if the filesystem is stored on the device).
330 As not all filesystems are stored on devices, and not all filesystems
331 have UUIDs, it is sometimes necessary to explicitly tell NFS how to
332 identify a filesystem. This is done with the
336 For NFSv4, there is a distinguished filesystem which is the root of
337 all exported filesystem. This is specified with
341 both of which mean exactly the same thing.
343 Other filesystems can be identified with a small integer, or a UUID
344 which should contain 32 hex digits and arbitrary punctuation.
346 Linux kernels version 2.6.20 and earlier do not understand the UUID
347 setting so a small integer must be used if an fsid option needs to be
348 set for such kernels. Setting both a small number and a UUID is
349 supported so the same configuration can be made to work on old and new
353 .IR refer= path@host[+host][:path@host[+host]]
354 A client referencing the export point will be directed to choose from
355 the given list an alternative location for the filesystem.
356 (Note that the server must have a mountpoint here, though a different
357 filesystem is not required; so, for example,
358 .IR "mount --bind" " /path /path"
361 .IR replicas= path@host[+host][:path@host[+host]]
362 If the client asks for alternative locations for the export point, it
363 will be given this list of alternatives. (Note that actual replication
364 of the filesystem must be handled elsewhere.)
367 .IR refer= path@host[+host][:path@host[+host]]
368 A client referencing the export point will be directed to choose from
369 the given list an alternative location for the filesystem.
370 (Note that the server must have a mountpoint here, though a different
371 filesystem is not required; so, for example,
372 .IR "mount --bind" " /path /path"
375 .IR replicas= path@host[+host][:path@host[+host]]
376 If the client asks for alternative locations for the export point, it
377 will be given this list of alternatives. (Note that actual replication
378 of the filesystem must be handled elsewhere.)
383 bases its access control to files on the server machine on the uid and
384 gid provided in each NFS RPC request. The normal behavior a user would
385 expect is that she can access her files on the server just as she would
386 on a normal file system. This requires that the same uids and gids are
387 used on the client and the server machine. This is not always true, nor
388 is it always desirable.
390 Very often, it is not desirable that the root user on a client machine
391 is also treated as root when accessing files on the NFS server. To this
392 end, uid 0 is normally mapped to a different id: the so-called
395 uid. This mode of operation (called `root squashing') is the default,
396 and can be turned off with
401 '''tries to obtain the anonymous uid and gid by looking up user
403 '''in the password file at startup time. If it isn't found, a uid and gid
405 chooses a uid and gid
406 of 65534 for squashed access. These values can also be overridden by
408 .IR anonuid " and " anongid
411 '''In addition to this,
413 '''lets you specify arbitrary uids and gids that should be mapped to user
415 Finally, you can map all user requests to the
416 anonymous uid by specifying the
417 .IR all_squash " option.
419 Here's the complete list of mapping options:
422 Map requests from uid/gid 0 to the anonymous uid/gid. Note that this does
423 not apply to any other uids or gids that might be equally sensitive, such as
430 Turn off root squashing. This option is mainly useful for diskless clients.
433 Map all uids and gids to the anonymous user. Useful for NFS-exported
434 public FTP directories, news spool directories, etc. The opposite option
437 which is the default setting.
439 .IR anonuid " and " anongid
440 These options explicitly set the uid and gid of the anonymous account.
441 This option is primarily useful for PC/NFS clients, where you might want
442 all requests appear to be from one user. As an example, consider the
445 in the example section below, which maps all requests to uid 150 (which
446 is supposedly that of user joe).
452 # sample /etc/exports file
453 / master(rw) trusty(rw,no_root_squash)
454 /projects proj*.local.domain(rw)
455 /usr *.local.domain(ro) @trusted(rw)
456 /home/joe pc001(rw,all_squash,anonuid=150,anongid=100)
457 /pub (ro,insecure,all_squash)
458 /srv/www \-sync,rw server @trusted @external(ro)
459 '''/pub/private (noaccess)
462 The first line exports the entire filesystem to machines master and trusty.
463 In addition to write access, all uid squashing is turned off for host
464 trusty. The second and third entry show examples for wildcard hostnames
465 and netgroups (this is the entry `@trusted'). The fourth line shows the
466 entry for the PC/NFS client discussed above. Line 5 exports the
467 public FTP directory to every host in the world, executing all requests
468 under the nobody account. The
470 option in this entry also allows clients with NFS implementations that
471 don't use a reserved port for NFS.
472 The sixth line exports a directory read-write to the machine 'server'
473 as well as the `@trusted' netgroup, and read-only to netgroup `@external',
474 all three mounts with the `sync' option enabled.
475 ''' The last line denies all NFS clients
476 '''access to the private directory.
478 '''Unlike other NFS server implementations, this
480 '''allows you to export both a directory and a subdirectory thereof to
481 '''the same host, for instance
482 '''.IR /usr " and " /usr/X11R6 .
483 '''In this case, the mount options of the most specific entry apply. For
484 '''instance, when a user on the client host accesses a file in
486 '''the mount options given in the
488 '''entry apply. This is also true when the latter is a wildcard or netgroup
499 '''An error parsing the file is reported using syslogd(8) as level NOTICE from
500 '''a DAEMON whenever nfsd(8) or mountd(8) is started up. Any unknown
501 '''host is reported at that time, but often not all hosts are not yet known
502 '''to named(8) at boot time, thus as hosts are found they are reported
503 '''with the same syslogd(8) parameters.