1exports(5) File Formats Manual exports(5)
2
3
4
6 exports - NFS server export table
7
9 The file /etc/exports contains a table of local physical file systems
10 on an NFS server that are accessible to NFS clients. The contents of
11 the file are maintained by the server's system administrator.
12
13 Each file system in this table has a list of options and an access con‐
14 trol list. The table is used by exportfs(8) to give information to
15 mountd(8).
16
17 The file format is similar to the SunOS exports file. Each line con‐
18 tains an export point and a whitespace-separated list of clients
19 allowed to mount the file system at that point. Each listed client may
20 be immediately followed by a parenthesized, comma-separated list of
21 export options for that client. No whitespace is permitted between a
22 client and its option list.
23
24 Also, each line may have one or more specifications for default options
25 after the path name, in the form of a dash ("-") followed by an option
26 list. The option list is used for all subsequent exports on that line
27 only.
28
29 Blank lines are ignored. A pound sign ("#") introduces a comment to
30 the end of the line. Entries may be continued across newlines using a
31 backslash. If an export name contains spaces it should be quoted using
32 double quotes. You can also specify spaces or other unusual character
33 in the export name using a backslash followed by the character code as
34 three octal digits.
35
36 To apply changes to this file, run exportfs-ra or restart the NFS
37 server.
38
39 Machine Name Formats
40 NFS clients may be specified in a number of ways:
41
42 single host
43 You may specify a host either by an abbreviated name recognized
44 be the resolver, the fully qualified domain name, an IPv4
45 address, or an IPv6 address. IPv6 addresses must not be inside
46 square brackets in /etc/exports lest they be confused with char‐
47 acter-class wildcard matches.
48
49 netgroups
50 NIS netgroups may be given as @group. Only the host part of
51 each netgroup members is consider in checking for membership.
52 Empty host parts or those containing a single dash (-) are
53 ignored.
54
55 wildcards
56 Machine names may contain the wildcard characters * and ?, or
57 may contain character class lists within [square brackets].
58 This can be used to make the exports file more compact; for
59 instance, *.cs.foo.edu matches all hosts in the domain
60 cs.foo.edu. As these characters also match the dots in a domain
61 name, the given pattern will also match all hosts within any
62 subdomain of cs.foo.edu.
63
64 IP networks
65 You can also export directories to all hosts on an IP (sub-)
66 network simultaneously. This is done by specifying an IP address
67 and netmask pair as address/netmask where the netmask can be
68 specified in dotted-decimal format, or as a contiguous mask
69 length. For example, either `/255.255.252.0' or `/22' appended
70 to the network base IPv4 address results in identical subnet‐
71 works with 10 bits of host. IPv6 addresses must use a contiguous
72 mask length and must not be inside square brackets to avoid con‐
73 fusion with character-class wildcards. Wildcard characters gen‐
74 erally do not work on IP addresses, though they may work by
75 accident when reverse DNS lookups fail.
76
77 RPCSEC_GSS security
78 You may use the special strings "gss/krb5", "gss/krb5i", or "gss/krb5p"
79 to restrict access to clients using rpcsec_gss security. However, this
80 syntax is deprecated; on linux kernels since 2.6.23, you should instead
81 use the "sec=" export option:
82
83 sec= The sec= option, followed by a colon-delimited list of security
84 flavors, restricts the export to clients using those flavors.
85 Available security flavors include sys (the default--no crypto‐
86 graphic security), krb5 (authentication only), krb5i (integrity
87 protection), and krb5p (privacy protection). For the purposes
88 of security flavor negotiation, order counts: preferred flavors
89 should be listed first. The order of the sec= option with
90 respect to the other options does not matter, unless you want
91 some options to be enforced differently depending on flavor. In
92 that case you may include multiple sec= options, and following
93 options will be enforced only for access using flavors listed in
94 the immediately preceding sec= option. The only options that
95 are permitted to vary in this way are ro, rw, no_root_squash,
96 root_squash, and all_squash.
97
98 General Options
99 exportfs understands the following export options:
100
101 secure This option requires that requests originate on an Internet port
102 less than IPPORT_RESERVED (1024). This option is on by default.
103 To turn it off, specify insecure.
104
105 rw Allow both read and write requests on this NFS volume. The
106 default is to disallow any request which changes the filesystem.
107 This can also be made explicit by using the ro option.
108
109 async This option allows the NFS server to violate the NFS protocol
110 and reply to requests before any changes made by that request
111 have been committed to stable storage (e.g. disc drive).
112
113 Using this option usually improves performance, but at the cost
114 that an unclean server restart (i.e. a crash) can cause data to
115 be lost or corrupted.
116
117
118 sync Reply to requests only after the changes have been committed to
119 stable storage (see async above).
120
121 In releases of nfs-utils up to and including 1.0.0, the async
122 option was the default. In all releases after 1.0.0, sync is
123 the default, and async must be explicitly requested if needed.
124 To help make system administrators aware of this change,
125 exportfs will issue a warning if neither sync nor async is spec‐
126 ified.
127
128 no_wdelay
129 This option has no effect if async is also set. The NFS server
130 will normally delay committing a write request to disc slightly
131 if it suspects that another related write request may be in
132 progress or may arrive soon. This allows multiple write
133 requests to be committed to disc with the one operation which
134 can improve performance. If an NFS server received mainly small
135 unrelated requests, this behaviour could actually reduce perfor‐
136 mance, so no_wdelay is available to turn it off. The default
137 can be explicitly requested with the wdelay option.
138
139 nohide This option is based on the option of the same name provided in
140 IRIX NFS. Normally, if a server exports two filesystems one of
141 which is mounted on the other, then the client will have to
142 mount both filesystems explicitly to get access to them. If it
143 just mounts the parent, it will see an empty directory at the
144 place where the other filesystem is mounted. That filesystem is
145 "hidden".
146
147 Setting the nohide option on a filesystem causes it not to be
148 hidden, and an appropriately authorised client will be able to
149 move from the parent to that filesystem without noticing the
150 change.
151
152 However, some NFS clients do not cope well with this situation
153 as, for instance, it is then possible for two files in the one
154 apparent filesystem to have the same inode number.
155
156 The nohide option is currently only effective on single host
157 exports. It does not work reliably with netgroup, subnet, or
158 wildcard exports.
159
160 This option can be very useful in some situations, but it should
161 be used with due care, and only after confirming that the client
162 system copes with the situation effectively.
163
164 The option can be explicitly disabled for NFSv2 and NFSv3 with
165 hide.
166
167 This option is not relevant when NFSv4 is use. NFSv4 never
168 hides subordinate filesystems. Any filesystem that is exported
169 will be visible where expected when using NFSv4.
170
171 crossmnt
172 This option is similar to nohide but it makes it possible for
173 clients to access all filesystems mounted on a filesystem marked
174 with crossmnt. Thus when a child filesystem "B" is mounted on a
175 parent "A", setting crossmnt on "A" has a similar effect to set‐
176 ting "nohide" on B.
177
178 With nohide the child filesystem needs to be explicitly
179 exported. With crossmnt it need not. If a child of a crossmnt
180 file is not explicitly exported, then it will be implicitly
181 exported with the same export options as the parent, except for
182 fsid=. This makes it impossible to not export a child of a
183 crossmnt filesystem. If some but not all subordinate filesys‐
184 tems of a parent are to be exported, then they must be explic‐
185 itly exported and the parent should not have crossmnt set.
186
187 The nocrossmnt option can explictly disable crossmnt if it was
188 previously set. This is rarely useful.
189
190 no_subtree_check
191 This option disables subtree checking, which has mild security
192 implications, but can improve reliability in some circumstances.
193
194 If a subdirectory of a filesystem is exported, but the whole
195 filesystem isn't then whenever a NFS request arrives, the server
196 must check not only that the accessed file is in the appropriate
197 filesystem (which is easy) but also that it is in the exported
198 tree (which is harder). This check is called the subtree_check.
199
200 In order to perform this check, the server must include some
201 information about the location of the file in the "filehandle"
202 that is given to the client. This can cause problems with
203 accessing files that are renamed while a client has them open
204 (though in many simple cases it will still work).
205
206 subtree checking is also used to make sure that files inside
207 directories to which only root has access can only be accessed
208 if the filesystem is exported with no_root_squash (see below),
209 even if the file itself allows more general access.
210
211 As a general guide, a home directory filesystem, which is nor‐
212 mally exported at the root and may see lots of file renames,
213 should be exported with subtree checking disabled. A filesystem
214 which is mostly readonly, and at least doesn't see many file
215 renames (e.g. /usr or /var) and for which subdirectories may be
216 exported, should probably be exported with subtree checks
217 enabled.
218
219 The default of having subtree checks enabled, can be explicitly
220 requested with subtree_check.
221
222 From release 1.1.0 of nfs-utils onwards, the default will be
223 no_subtree_check as subtree_checking tends to cause more prob‐
224 lems than it is worth. If you genuinely require subtree check‐
225 ing, you should explicitly put that option in the exports file.
226 If you put neither option, exportfs will warn you that the
227 change is pending.
228
229
230 insecure_locks
231
232 no_auth_nlm
233 This option (the two names are synonymous) tells the NFS server
234 not to require authentication of locking requests (i.e. requests
235 which use the NLM protocol). Normally the NFS server will
236 require a lock request to hold a credential for a user who has
237 read access to the file. With this flag no access checks will
238 be performed.
239
240 Early NFS client implementations did not send credentials with
241 lock requests, and many current NFS clients still exist which
242 are based on the old implementations. Use this flag if you find
243 that you can only lock files which are world readable.
244
245 The default behaviour of requiring authentication for NLM
246 requests can be explicitly requested with either of the synony‐
247 mous auth_nlm, or secure_locks.
248
249
250 mountpoint=path
251
252 mp This option makes it possible to only export a directory if it
253 has successfully been mounted. If no path is given (e.g.
254 mountpoint or mp) then the export point must also be a mount
255 point. If it isn't then the export point is not exported. This
256 allows you to be sure that the directory underneath a mountpoint
257 will never be exported by accident if, for example, the filesys‐
258 tem failed to mount due to a disc error.
259
260 If a path is given (e.g. mountpoint=/path or mp=/path) then the
261 nominated path must be a mountpoint for the exportpoint to be
262 exported.
263
264
265 fsid=num|root|uuid
266 NFS needs to be able to identify each filesystem that it
267 exports. Normally it will use a UUID for the filesystem (if the
268 filesystem has such a thing) or the device number of the device
269 holding the filesystem (if the filesystem is stored on the
270 device).
271
272 As not all filesystems are stored on devices, and not all
273 filesystems have UUIDs, it is sometimes necessary to explicitly
274 tell NFS how to identify a filesystem. This is done with the
275 fsid= option.
276
277 For NFSv4, there is a distinguished filesystem which is the root
278 of all exported filesystem. This is specified with fsid=root or
279 fsid=0 both of which mean exactly the same thing.
280
281 Other filesystems can be identified with a small integer, or a
282 UUID which should contain 32 hex digits and arbitrary punctua‐
283 tion.
284
285 Linux kernels version 2.6.20 and earlier do not understand the
286 UUID setting so a small integer must be used if an fsid option
287 needs to be set for such kernels. Setting both a small number
288 and a UUID is supported so the same configuration can be made to
289 work on old and new kernels alike.
290
291
292 nordirplus
293 This option will disable READDIRPLUS request handling. When
294 set, READDIRPLUS requests from NFS clients return NFS3ERR_NOT‐
295 SUPP, and clients fall back on READDIR. This option affects
296 only NFSv3 clients.
297
298 refer=path@host[+host][:path@host[+host]]
299 A client referencing the export point will be directed to choose
300 from the given list an alternative location for the filesystem.
301 (Note that the server must have a mountpoint here, though a dif‐
302 ferent filesystem is not required; so, for example, mount --bind
303 /path /path is sufficient.)
304
305 replicas=path@host[+host][:path@host[+host]]
306 If the client asks for alternative locations for the export
307 point, it will be given this list of alternatives. (Note that
308 actual replication of the filesystem must be handled elsewhere.)
309
310
311 User ID Mapping
312 nfsd bases its access control to files on the server machine on the uid
313 and gid provided in each NFS RPC request. The normal behavior a user
314 would expect is that she can access her files on the server just as she
315 would on a normal file system. This requires that the same uids and
316 gids are used on the client and the server machine. This is not always
317 true, nor is it always desirable.
318
319 Very often, it is not desirable that the root user on a client machine
320 is also treated as root when accessing files on the NFS server. To this
321 end, uid 0 is normally mapped to a different id: the so-called anony‐
322 mous or nobody uid. This mode of operation (called `root squashing') is
323 the default, and can be turned off with no_root_squash.
324
325 By default, exportfs chooses a uid and gid of 65534 for squashed
326 access. These values can also be overridden by the anonuid and anongid
327 options. Finally, you can map all user requests to the anonymous uid
328 by specifying the all_squash option.
329
330 Here's the complete list of mapping options:
331
332 root_squash
333 Map requests from uid/gid 0 to the anonymous uid/gid. Note that
334 this does not apply to any other uids or gids that might be
335 equally sensitive, such as user bin or group staff.
336
337 no_root_squash
338 Turn off root squashing. This option is mainly useful for disk‐
339 less clients.
340
341 all_squash
342 Map all uids and gids to the anonymous user. Useful for NFS-
343 exported public FTP directories, news spool directories, etc.
344 The opposite option is no_all_squash, which is the default set‐
345 ting.
346
347 anonuid and anongid
348 These options explicitly set the uid and gid of the anonymous
349 account. This option is primarily useful for PC/NFS clients,
350 where you might want all requests appear to be from one user. As
351 an example, consider the export entry for /home/joe in the exam‐
352 ple section below, which maps all requests to uid 150 (which is
353 supposedly that of user joe).
354
355 Extra Export Tables
356 After reading /etc/exports exportfs reads files under /etc/exports.d.
357 directory as extra export tables. exportfs regards only a file which
358 name is ended with .exports and not started with . as an extra export
359 file. A file which name is not met this condition is just ignored. The
360 format for extra export tables is the same as /etc/exports
361
363 # sample /etc/exports file
364 / master(rw) trusty(rw,no_root_squash)
365 /projects proj*.local.domain(rw)
366 /usr *.local.domain(ro) @trusted(rw)
367 /home/joe pc001(rw,all_squash,anonuid=150,anongid=100)
368 /pub *(ro,insecure,all_squash)
369 /srv/www -sync,rw server @trusted @external(ro)
370 /foo 2001:db8:9:e54::/64(rw) 192.0.2.0/24(rw)
371 /build buildhost[0-9].local.domain(rw)
372
373 The first line exports the entire filesystem to machines master and
374 trusty. In addition to write access, all uid squashing is turned off
375 for host trusty. The second and third entry show examples for wildcard
376 hostnames and netgroups (this is the entry `@trusted'). The fourth line
377 shows the entry for the PC/NFS client discussed above. Line 5 exports
378 the public FTP directory to every host in the world, executing all
379 requests under the nobody account. The insecure option in this entry
380 also allows clients with NFS implementations that don't use a reserved
381 port for NFS. The sixth line exports a directory read-write to the
382 machine 'server' as well as the `@trusted' netgroup, and read-only to
383 netgroup `@external', all three mounts with the `sync' option enabled.
384 The seventh line exports a directory to both an IPv6 and an IPv4 sub‐
385 net. The eighth line demonstrates a character class wildcard match.
386
388 /etc/exports /etc/exports.d
389
391 exportfs(8), netgroup(5), mountd(8), nfsd(8), showmount(8).
392
393
394
395 31 December 2009 exports(5)