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 Blank lines are ignored. A pound sign ("#") introduces a comment to the
26 end of the line. Entries may be continued across newlines using a
27 backslash. If an export name contains spaces it should be quoted using
28 double quotes. You can also specify spaces or other unusual character in
29 the export name using a backslash followed by the character code as three
32 .SS Machine Name Formats
33 NFS clients may be specified in a number of ways:
35 This is the most common format. You may specify a host either by an
36 abbreviated name recognized be the resolver, the fully qualified domain
37 name, or an IP address.
39 NIS netgroups may be given as
41 Only the host part of each
42 netgroup members is consider in checking for membership. Empty host
43 parts or those containing a single dash (\-) are ignored.
45 Machine names may contain the wildcard characters \fI*\fR and \fI?\fR.
46 This can be used to make the \fIexports\fR file more compact; for instance,
47 \fI*.cs.foo.edu\fR matches all hosts in the domain
48 \fIcs.foo.edu\fR. As these characters also match the dots in a domain
49 name, the given pattern will also match all hosts within any subdomain
52 You can also export directories to all hosts on an IP (sub-) network
53 simultaneously. This is done by specifying an IP address and netmask pair
56 where the netmask can be specified in dotted-decimal format, or as a
57 contiguous mask length (for example, either `/255.255.252.0' or `/22' appended
58 to the network base address result in identical subnetworks with 10 bits of
59 host). Wildcard characters generally do not work on IP addresses, though they
60 may work by accident when reverse DNS lookups fail.
63 '''This is a special ``hostname'' that identifies the given directory name
64 '''as the public root directory (see the section on WebNFS in
66 '''for a discussion of WebNFS and the public root handle). When using this
69 '''must be the only entry on this line, and must have no export options
70 '''associated with it. Note that this does
72 '''actually export the named directory; you still have to set the exports
73 '''options in a separate entry.
75 '''The public root path can also be specified by invoking
78 '''.B \-\-public\-root
79 '''option. Multiple specifications of a public root will be ignored.
81 .SS RPCSEC_GSS security
82 To restrict access to an export using rpcsec_gss security, use the special
83 string "gss/krb5" as the client. It is not possible to simultaneously require
84 rpcsec_gss and to make requirements on the IP address of the client.
88 understands the following export options:
91 This option requires that requests originate on an internet port less
92 than IPPORT_RESERVED (1024). This option is on by default. To turn it
97 Allow both read and write requests on this NFS volume. The
98 default is to disallow any request which changes the filesystem.
99 This can also be made explicit by using
104 This option allows the NFS server to violate the NFS protocol and
105 reply to requests before any changes made by that request have been
106 committed to stable storage (e.g. disc drive).
108 Using this option usually improves performance, but at the cost that
109 an unclean server restart (i.e. a crash) can cause data to be lost or
114 Reply to requests only after the changes have been committed to stable
119 In releases of nfs-utils upto and including 1.0.0, this option was the
120 default. In this and future releases,
124 must be explicit requested if needed.
125 To help make system adminstrators aware of this change, 'exportfs'
126 will issue a warning if neither
133 This option has no effect if
135 is also set. The NFS server will normally delay committing a write request
136 to disc slightly if it suspects that another related write request may be in
137 progress or may arrive soon. This allows multiple write requests to
138 be committed to disc with the one operation which can improve
139 performance. If an NFS server received mainly small unrelated
140 requests, this behaviour could actually reduce performance, so
142 is available to turn it off.
143 The default can be explicitly requested with the
147 This option is based on the option of the same name provided in IRIX
148 NFS. Normally, if a server exports two filesystems one of which is
149 mounted on the other, then the client will have to mount both
150 filesystems explicitly to get access to them. If it just mounts the
151 parent, it will see an empty directory at the place where the other
152 filesystem is mounted. That filesystem is "hidden".
156 option on a filesystem causes it not to be hidden, and an
157 appropriately authorised client will be able to move from the parent to
158 that filesystem without noticing the change.
160 However, some NFS clients do not cope well with this situation as, for
161 instance, it is then possible for two files in the one apparent
162 filesystem to have the same inode number.
166 option is currently only effective on
168 exports. It does not work reliably with netgroup, subnet, or wildcard
171 This option can be very useful in some situations, but it should be
172 used with due care, and only after confirming that the client system
173 copes with the situation effectively.
175 The option can be explicitly disabled with
179 This option is similar to
181 but it makes it possible for clients to move from the filesystem marked
182 with crossmnt to exported filesystems mounted on it. Thus when a child
183 filesystem "B" is mounted on a parent "A", setting crossmnt on "A" has
184 the same effect as setting "nohide" on B.
187 This option disables subtree checking, which has mild security
188 implications, but can improve reliability in some circumstances.
190 If a subdirectory of a filesystem is exported, but the whole
191 filesystem isn't then whenever a NFS request arrives, the server must
192 check not only that the accessed file is in the appropriate filesystem
193 (which is easy) but also that it is in the exported tree (which is
194 harder). This check is called the
197 In order to perform this check, the server must include some
198 information about the location of the file in the "filehandle" that is
199 given to the client. This can cause problems with accessing files that
200 are renamed while a client has them open (though in many simple cases
203 subtree checking is also used to make sure that files inside
204 directories to which only root has access can only be accessed if the
205 filesystem is exported with
207 (see below), even if the file itself allows more general access.
209 As a general guide, a home directory filesystem, which is normally
210 exported at the root and may see lots of file renames, should be
211 exported with subtree checking disabled. A filesystem which is mostly
212 readonly, and at least doesn't see many file renames (e.g. /usr or
213 /var) and for which subdirectories may be exported, should probably be
214 exported with subtree checks enabled.
216 The default of having subtree checks enabled, can be explicitly
224 This option (the two names are synonymous) tells the NFS server not to require authentication of
225 locking requests (i.e. requests which use the NLM protocol). Normally
226 the NFS server will require a lock request to hold a credential for a
227 user who has read access to the file. With this flag no access checks
230 Early NFS client implementations did not send credentials with lock
231 requests, and many current NFS clients still exist which are based on
232 the old implementations. Use this flag if you find that you can only
233 lock files which are world readable.
235 The default behaviour of requiring authentication for NLM requests can
236 be explicitly requested with either of the synonymous
242 On some specially patched kernels, and when exporting filesystems that
243 support ACLs, this option tells nfsd not to reveal ACLs to clients, so
244 they will see only a subset of actual permissions on the given file
245 system. This option is safe for filesystems used by NFSv2 clients and
246 old NFSv3 clients that perform access decisions locally. Current
247 NFSv3 clients use the ACCESS RPC to perform all access decisions on
248 the server. Note that the
250 option only has effect on kernels specially patched to support it, and
251 when exporting filesystems with ACL support. The default is to export
252 with ACL support (i.e. by default,
258 '''This makes everything below the directory inaccessible for the named
259 '''client. This is useful when you want to export a directory hierarchy to
260 '''a client, but exclude certain subdirectories. The client's view of a
261 '''directory flagged with noaccess is very limited; it is allowed to read
262 '''its attributes, and lookup `.' and `..'. These are also the only entries
263 '''returned by a readdir.
266 '''Convert absolute symbolic links (where the link contents start with a
267 '''slash) into relative links by prepending the necessary number of ../'s
268 '''to get from the directory containing the link to the root on the
269 '''server. This has subtle, perhaps questionable, semantics when the file
270 '''hierarchy is not mounted at its root.
273 '''Leave all symbolic link as they are. This is the default operation.
279 This option makes it possible to only export a directory if it has
280 successfully been mounted.
281 If no path is given (e.g.
282 .IR mountpoint " or " mp )
283 then the export point must also be a mount point. If it isn't then
284 the export point is not exported. This allows you to be sure that the
285 directory underneath a mountpoint will never be exported by accident
286 if, for example, the filesystem failed to mount due to a disc error.
288 If a path is given (e.g.
289 .IR mountpoint= "/path or " mp= /path)
290 then the nominted path must be a mountpoint for the exportpoint to be
295 This option forces the filesystem identification portion of the file
296 handle and file attributes used on the wire to be
298 instead of a number derived from the major and minor number of the
299 block device on which the filesystem is mounted. Any 32 bit number
300 can be used, but it must be unique amongst all the exported filesystems.
302 This can be useful for NFS failover, to ensure that both servers of
303 the failover pair use the same NFS file handles for the shared filesystem
304 thus avoiding stale file handles after failover.
306 Some Linux filesystems are not mounted on a block device; exporting
307 these via NFS requires the use of the
309 option (although that may still not be enough).
311 The value 0 has a special meaning when use with NFSv4. NFSv4 has a
312 concept of a root of the overall exported filesystem. The export point
313 exported with fsid=0 will be used as this root.
318 bases its access control to files on the server machine on the uid and
319 gid provided in each NFS RPC request. The normal behavior a user would
320 expect is that she can access her files on the server just as she would
321 on a normal file system. This requires that the same uids and gids are
322 used on the client and the server machine. This is not always true, nor
323 is it always desirable.
325 Very often, it is not desirable that the root user on a client machine
326 is also treated as root when accessing files on the NFS server. To this
327 end, uid 0 is normally mapped to a different id: the so-called
330 uid. This mode of operation (called `root squashing') is the default,
331 and can be turned off with
336 '''tries to obtain the anonymous uid and gid by looking up user
338 '''in the password file at startup time. If it isn't found, a uid and gid
340 chooses a uid and gid
341 of 65534 for squashed access. These values can also be overridden by
343 .IR anonuid " and " anongid
346 '''In addition to this,
348 '''lets you specify arbitrary uids and gids that should be mapped to user
350 Finally, you can map all user requests to the
351 anonymous uid by specifying the
352 .IR all_squash " option.
354 '''For the benefit of installations where uids differ between different
357 '''provides several mechanism to dynamically map server uids to client
358 '''uids and vice versa: static mapping files, NIS-based mapping, and
363 '''mapping is enabled with the
365 '''option, and uses the UGID RPC protocol. For this to work, you have to run
368 '''mapping daemon on the client host. It is the least secure of the three methods,
369 '''because by running
371 '''everybody can query the client host for a list of valid user names. You
372 '''can protect yourself by restricting access to
374 '''to valid hosts only. This can be done by entering the list of valid
379 '''file. The service name is
381 '''For a description of the file's syntax, please read
382 '''.IR hosts_access (5).
384 '''Static mapping is enabled by using the
386 '''option, which takes a file name as an argument that describes the mapping.
387 '''NIS-based mapping queries the client's NIS server to obtain a mapping from
388 '''user and group names on the server host to user and group names on the
391 Here's the complete list of mapping options:
394 Map requests from uid/gid 0 to the anonymous uid/gid. Note that this does
395 not apply to any other uids that might be equally sensitive, such as user
399 Turn off root squashing. This option is mainly useful for diskless clients.
401 '''.IR squash_uids " and " squash_gids
402 '''This option specifies a list of uids or gids that should be subject to
403 '''anonymous mapping. A valid list of ids looks like this:
405 '''.IR squash_uids=0-15,20,25-50
407 '''Usually, your squash lists will look a lot simpler.
410 Map all uids and gids to the anonymous user. Useful for NFS-exported
411 public FTP directories, news spool directories, etc. The opposite option
414 which is the default setting.
417 '''This option turns on dynamic uid/gid mapping. Each uid in an NFS request
418 '''will be translated to the equivalent server uid, and each uid in an
419 '''NFS reply will be mapped the other way round. This option requires that
421 '''runs on the client host. The default setting is
422 '''.IR map_identity ,
423 '''which leaves all uids untouched. The normal squash options apply regardless
424 '''of whether dynamic mapping is requested or not.
427 '''This option enables static mapping. It specifies the name of the file
428 '''that describes the uid/gid mapping, e.g.
430 '''.IR map_static=/etc/nfs/foobar.map
432 '''The file's format looks like this
436 '''# Mapping for client foobar:
438 '''uid 0-99 - # squash these
439 '''uid 100-500 1000 # map 100-500 to 1000-1400
440 '''gid 0-49 - # squash these
441 '''gid 50-100 700 # map 50-100 to 700-750
445 '''This option enables NIS-based uid/gid mapping. For instance, when
446 '''the server encounters the uid 123 on the server, it will obtain the
447 '''login name associated with it, and contact the NFS client's NIS server
448 '''to obtain the uid the client associates with the name.
450 '''In order to do this, the NFS server must know the client's NIS domain.
451 '''This is specified as an argument to the
455 '''.I map_nis=foo.com
457 '''Note that it may not be sufficient to simply specify the NIS domain
458 '''here; you may have to take additional actions before
460 '''is actually able to contact the server. If your distribution uses
461 '''the NYS library, you can specify one or more NIS servers for the
462 '''client's domain in
463 '''.IR /etc/yp.conf .
464 '''If you are using a different NIS library, you may have to obtain a
467 '''daemon that can be configured via
470 .IR anonuid " and " anongid
471 These options explicitly set the uid and gid of the anonymous account.
472 This option is primarily useful for PC/NFS clients, where you might want
473 all requests appear to be from one user. As an example, consider the
476 in the example section below, which maps all requests to uid 150 (which
477 is supposedly that of user joe).
483 # sample /etc/exports file
484 / master(rw) trusty(rw,no_root_squash)
485 /projects proj*.local.domain(rw)
486 /usr *.local.domain(ro) @trusted(rw)
487 /home/joe pc001(rw,all_squash,anonuid=150,anongid=100)
488 /pub (ro,insecure,all_squash)
489 '''/pub/private (noaccess)
492 The first line exports the entire filesystem to machines master and trusty.
493 In addition to write access, all uid squashing is turned off for host
494 trusty. The second and third entry show examples for wildcard hostnames
495 and netgroups (this is the entry `@trusted'). The fourth line shows the
496 entry for the PC/NFS client discussed above. Line 5 exports the
497 public FTP directory to every host in the world, executing all requests
498 under the nobody account. The
500 option in this entry also allows clients with NFS implementations that
501 don't use a reserved port for NFS.
502 ''' The last line denies all NFS clients
503 '''access to the private directory.
505 '''Unlike other NFS server implementations, this
507 '''allows you to export both a directory and a subdirectory thereof to
508 '''the same host, for instance
509 '''.IR /usr " and " /usr/X11R6 .
510 '''In this case, the mount options of the most specific entry apply. For
511 '''instance, when a user on the client host accesses a file in
513 '''the mount options given in the
515 '''entry apply. This is also true when the latter is a wildcard or netgroup
526 '''An error parsing the file is reported using syslogd(8) as level NOTICE from
527 '''a DAEMON whenever nfsd(8) or mountd(8) is started up. Any unknown
528 '''host is reported at that time, but often not all hosts are not yet known
529 '''to named(8) at boot time, thus as hosts are found they are reported
530 '''with the same syslogd(8) parameters.