1 .TH EXPORTS 5 "28 October 1999"
4 exports \- NFS file systems being exported (for Kernel based NFS)
10 serves as the access control list for file systems which may be
11 exported to NFS clients. It is used by
13 to give information to
15 and to the kernel based NFS file server daemon
18 The file format is similar to the SunOS
20 file. Each line contains an export point and a whitespace-separated list
21 of clients allowed to mount the file system at that point. Each listed
22 client may be immediately followed by a parenthesized, comma-separated
23 list of export options for that client. No whitespace is permitted
24 between a client and its option list.
26 Blank lines are ignored. A pound sign ("#") introduces a comment to the
27 end of the line. Entries may be continued across newlines using a
28 backslash. If an export name contains spaces it should be quoted using
29 double quotes. You can also specify spaces or other unusual character in
30 the export name using a backslash followed by the character code as three
33 .SS Machine Name Formats
34 NFS clients may be specified in a number of ways:
36 This is the most common format. You may specify a host either by an
37 abbreviated name recognized be the resolver, the fully qualified domain
38 name, or an IP address.
40 NIS netgroups may be given as
42 Only the host part of each
43 netgroup members is consider in checking for membership. Empty host
44 parts or those containing a single dash (\-) are ignored.
46 Machine names may contain the wildcard characters \fI*\fR and \fI?\fR.
47 This can be used to make the \fIexports\fR file more compact; for instance,
48 \fI*.cs.foo.edu\fR matches all hosts in the domain
49 \fIcs.foo.edu\fR. As these characters also match the dots in a domain
50 name, the given pattern will also match all hosts within any subdomain
53 You can also export directories to all hosts on an IP (sub-) network
54 simultaneously. This is done by specifying an IP address and netmask pair
57 where the netmask can be specified in dotted-decimal format, or as a
58 contiguous mask length (for example, either `/255.255.252.0' or `/22' appended
59 to the network base address result in identical subnetworks with 10 bits of
60 host). Wildcard characters generally do not work on IP addresses, though they
61 may work by accident when reverse DNS lookups fail.
64 '''This is a special ``hostname'' that identifies the given directory name
65 '''as the public root directory (see the section on WebNFS in
67 '''for a discussion of WebNFS and the public root handle). When using this
70 '''must be the only entry on this line, and must have no export options
71 '''associated with it. Note that this does
73 '''actually export the named directory; you still have to set the exports
74 '''options in a separate entry.
76 '''The public root path can also be specified by invoking
79 '''.B \-\-public\-root
80 '''option. Multiple specifications of a public root will be ignored.
84 understands the following export options:
87 This option requires that requests originate on an internet port less
88 than IPPORT_RESERVED (1024). This option is on by default. To turn it
93 Allow both read and write requests on this NFS volume. The
94 default is to disallow any request which changes the filesystem.
95 This can also be made explicit by using
100 This option allows the NFS server to violate the NFS protocol and
101 reply to requests before any changes made by that request have been
102 committed to stable storage (e.g. disc drive).
104 Using this option usually improves performance, but at the cost that
105 an unclean server restart (i.e. a crash) can cause data to be lost or
108 In releases of nfs-utils upto and including 1.0.0, this option was the
109 default. In this and future releases,
113 must be explicit requested if needed.
114 To help make system adminstrators aware of this change, 'exportfs'
115 will issue a warning if neither
122 This option has no effect if
124 is also set. The NFS server will normally delay committing a write request
125 to disc slightly if it suspects that another related write request may be in
126 progress or may arrive soon. This allows multiple write requests to
127 be committed to disc with the one operation which can improve
128 performance. If an NFS server received mainly small unrelated
129 requests, this behaviour could actually reduce performance, so
131 is available to turn it off.
132 The default can be explicitly requested with the
136 This option is based on the option of the same name provided in IRIX
137 NFS. Normally, if a server exports two filesystems one of which is
138 mounted on the other, then the client will have to mount both
139 filesystems explicitly to get access to them. If it just mounts the
140 parent, it will see an empty directory at the place where the other
141 filesystem is mounted. That filesystem is "hidden".
145 option on a filesystem causes it not to be hidden, and an
146 appropriately authorised client will be able to move from the parent to
147 that filesystem without noticing the change.
149 However, some NFS clients do not cope well with this situation as, for
150 instance, it is then possible for two files in the one apparent
151 filesystem to have the same inode number.
155 option is currently only effective on
157 exports. It does not work reliably with netgroup, subnet, or wildcard
160 This option can be very useful in some situations, but it should be
161 used with due care, and only after confirming that the client system
162 copes with the situation effectively.
164 The option can be explicitly disabled with
168 This option disables subtree checking, which has mild security
169 implications, but can improve reliability is some circumstances.
171 If a subdirectory of a filesystem is exported, but the whole
172 filesystem isn't then whenever a NFS request arrives, the server must
173 check not only that the accessed file is in the appropriate filesystem
174 (which is easy) but also that it is in the exported tree (which is
175 harder). This check is called the
178 In order to perform this check, the server must include some
179 information about the location of the file in the "filehandle" that is
180 given to the client. This can cause problems with accessing files that
181 are renamed while a client has them open (though in many simple cases
184 subtree checking is also used to make sure that files inside
185 directories to which only root has access can only be accessed if the
186 filesystem is exported with
188 (see below), even the file itself allows more general access.
190 As a general guide, a home directory filesystem, which is normally
191 exported at the root and may see lots of file renames, should be
192 exported with subtree checking disabled. A filesystem which is mostly
193 readonly, and at least doesn't see many file renames (e.g. /usr or
194 /var) and for which subdirectories may be exported, should probably be
195 exported with subtree checks enabled.
197 The default of having subtree checks enabled, can be explicitly
205 This option (the two names are synonymous) tells the NFS server not to require authentication of
206 locking requests (i.e. requests which use the NLM protocol). Normally
207 the NFS server will require a lock request to hold a credential for a
208 user who has read access to the file. With this flag no access checks
211 Early NFS client implementations did not send credentials with lock
212 requests, and many current NFS clients still exist which are based on
213 the old implementations. Use this flag if you find that you can only
214 lock files which are world readable.
216 The default behaviour of requiring authentication for NLM requests can
217 be explicitly requested with either of the synonymous
224 '''This makes everything below the directory inaccessible for the named
225 '''client. This is useful when you want to export a directory hierarchy to
226 '''a client, but exclude certain subdirectories. The client's view of a
227 '''directory flagged with noaccess is very limited; it is allowed to read
228 '''its attributes, and lookup `.' and `..'. These are also the only entries
229 '''returned by a readdir.
232 '''Convert absolute symbolic links (where the link contents start with a
233 '''slash) into relative links by prepending the necessary number of ../'s
234 '''to get from the directory containing the link to the root on the
235 '''server. This has subtle, perhaps questionable, semantics when the file
236 '''hierarchy is not mounted at its root.
239 '''Leave all symbolic link as they are. This is the default operation.
245 This option makes it possible to only export a directory if it has
246 successfully been mounted.
247 If no path is given (e.g.
248 .IR mountpoint " or " mp )
249 then the export point must also be a mount point. If it isn't then
250 the export point is not exported. This allows you to be sure that the
251 directory underneath a mountpoint will never be exported by accident
252 if, for example, the filesystem failed to mount due to a disc error.
254 If a path is given (e.g.
255 .IR mountpoint= "/path or " mp= /path)
256 then the nominted path must be a mountpoint for the exportpoint to be
262 bases its access control to files on the server machine on the uid and
263 gid provided in each NFS RPC request. The normal behavior a user would
264 expect is that she can access her files on the server just as she would
265 on a normal file system. This requires that the same uids and gids are
266 used on the client and the server machine. This is not always true, nor
267 is it always desirable.
269 Very often, it is not desirable that the root user on a client machine
270 is also treated as root when accessing files on the NFS server. To this
271 end, uid 0 is normally mapped to a different id: the so-called
274 uid. This mode of operation (called `root squashing') is the default,
275 and can be turned off with
280 '''tries to obtain the anonymous uid and gid by looking up user
282 '''in the password file at startup time. If it isn't found, a uid and gid
284 chooses a uid and gid
285 of -2 (i.e. 65534) for squashed access. These values can also be overridden by
287 .IR anonuid " and " anongid
290 '''In addition to this,
292 '''lets you specify arbitrary uids and gids that should be mapped to user
294 Finally, you can map all user requests to the
295 anonymous uid by specifying the
296 .IR all_squash " option.
298 '''For the benefit of installations where uids differ between different
301 '''provides several mechanism to dynamically map server uids to client
302 '''uids and vice versa: static mapping files, NIS-based mapping, and
307 '''mapping is enabled with the
309 '''option, and uses the UGID RPC protocol. For this to work, you have to run
312 '''mapping daemon on the client host. It is the least secure of the three methods,
313 '''because by running
315 '''everybody can query the client host for a list of valid user names. You
316 '''can protect yourself by restricting access to
318 '''to valid hosts only. This can be done by entering the list of valid
323 '''file. The service name is
325 '''For a description of the file's syntax, please read
326 '''.IR hosts_access (5).
328 '''Static mapping is enabled by using the
330 '''option, which takes a file name as an argument that describes the mapping.
331 '''NIS-based mapping queries the client's NIS server to obtain a mapping from
332 '''user and group names on the server host to user and group names on the
335 Here's the complete list of mapping options:
338 Map requests from uid/gid 0 to the anonymous uid/gid. Note that this does
339 not apply to any other uids that might be equally sensitive, such as user
343 Turn off root squashing. This option is mainly useful for diskless clients.
345 '''.IR squash_uids " and " squash_gids
346 '''This option specifies a list of uids or gids that should be subject to
347 '''anonymous mapping. A valid list of ids looks like this:
349 '''.IR squash_uids=0-15,20,25-50
351 '''Usually, your squash lists will look a lot simpler.
354 Map all uids and gids to the anonymous user. Useful for NFS-exported
355 public FTP directories, news spool directories, etc. The opposite option
358 which is the default setting.
361 '''This option turns on dynamic uid/gid mapping. Each uid in an NFS request
362 '''will be translated to the equivalent server uid, and each uid in an
363 '''NFS reply will be mapped the other way round. This option requires that
365 '''runs on the client host. The default setting is
366 '''.IR map_identity ,
367 '''which leaves all uids untouched. The normal squash options apply regardless
368 '''of whether dynamic mapping is requested or not.
371 '''This option enables static mapping. It specifies the name of the file
372 '''that describes the uid/gid mapping, e.g.
374 '''.IR map_static=/etc/nfs/foobar.map
376 '''The file's format looks like this
380 '''# Mapping for client foobar:
382 '''uid 0-99 - # squash these
383 '''uid 100-500 1000 # map 100-500 to 1000-1400
384 '''gid 0-49 - # squash these
385 '''gid 50-100 700 # map 50-100 to 700-750
389 '''This option enables NIS-based uid/gid mapping. For instance, when
390 '''the server encounters the uid 123 on the server, it will obtain the
391 '''login name associated with it, and contact the NFS client's NIS server
392 '''to obtain the uid the client associates with the name.
394 '''In order to do this, the NFS server must know the client's NIS domain.
395 '''This is specified as an argument to the
399 '''.I map_nis=foo.com
401 '''Note that it may not be sufficient to simply specify the NIS domain
402 '''here; you may have to take additional actions before
404 '''is actually able to contact the server. If your distribution uses
405 '''the NYS library, you can specify one or more NIS servers for the
406 '''client's domain in
407 '''.IR /etc/yp.conf .
408 '''If you are using a different NIS library, you may have to obtain a
411 '''daemon that can be configured via
414 .IR anonuid " and " anongid
415 These options explicitly set the uid and gid of the anonymous account.
416 This option is primarily useful for PC/NFS clients, where you might want
417 all requests appear to be from one user. As an example, consider the
420 in the example section below, which maps all requests to uid 150 (which
421 is supposedly that of user joe).
427 # sample /etc/exports file
428 / master(rw) trusty(rw,no_root_squash)
429 /projects proj*.local.domain(rw)
430 /usr *.local.domain(ro) @trusted(rw)
431 /home/joe pc001(rw,all_squash,anonuid=150,anongid=100)
432 /pub (ro,insecure,all_squash)
433 '''/pub/private (noaccess)
436 The first line exports the entire filesystem to machines master and trusty.
437 In addition to write access, all uid squashing is turned off for host
438 trusty. The second and third entry show examples for wildcard hostnames
439 and netgroups (this is the entry `@trusted'). The fourth line shows the
440 entry for the PC/NFS client discussed above. Line 5 exports the
441 public FTP directory to every host in the world, executing all requests
442 under the nobody account. The
444 option in this entry also allows clients with NFS implementations that
445 don't use a reserved port for NFS.
446 ''' The last line denies all NFS clients
447 '''access to the private directory.
449 '''Unlike other NFS server implementations, this
451 '''allows you to export both a directory and a subdirectory thereof to
452 '''the same host, for instance
453 '''.IR /usr " and " /usr/X11R6 .
454 '''In this case, the mount options of the most specific entry apply. For
455 '''instance, when a user on the client host accesses a file in
457 '''the mount options given in the
459 '''entry apply. This is also true when the latter is a wildcard or netgroup
464 '''An error parsing the file is reported using syslogd(8) as level NOTICE from
465 '''a DAEMON whenever nfsd(8) or mountd(8) is started up. Any unknown
466 '''host is reported at that time, but often not all hosts are not yet known
467 '''to named(8) at boot time, thus as hosts are found they are reported
468 '''with the same syslogd(8) parameters.