Debian Codegraph @

upstream/debian-unstable/man8/exportfs.8

raw
1 .\"@(#)exportfs.8"
2 .\"
3 .\" Copyright (C) 1995 Olaf Kirch <[email protected]>
4 .\" Modifications 1999-2003 Neil Brown <[email protected]>
5 .\"
6 .TH exportfs 8 "30 September 2013"
7 .SH NAME
8 exportfs \- maintain table of exported NFS file systems
9 .SH SYNOPSIS
10 .BI "/usr/sbin/exportfs [-avi] [-o " "options,.." "] [" "client:/path" " ..]
11 .br
12 .BI "/usr/sbin/exportfs -r [-v]"
13 .br
14 .BI "/usr/sbin/exportfs [-av] -u [" "client:/path" " ..]
15 .br
16 .BI "/usr/sbin/exportfs [-v]
17 .br
18 .BI "/usr/sbin/exportfs -f"
19 .br
20 .BI "/usr/sbin/exportfs -s"
21 .br
22 .SH DESCRIPTION
23 An NFS server maintains a table of local physical file systems
24 that are accessible to NFS clients.
25 Each file system in this table is referred to as an
26 .IR "exported file system" ,
27 or
28 .IR export ,
29 for short.
30 .PP
31 The
32 .B exportfs
33 command maintains the current table of exports for the NFS server.
34 The master export table is kept in a file named
35 .IR /var/lib/nfs/etab .
36 This file is read by
37 .B rpc.mountd
38 when a client sends an NFS MOUNT request.
39 .PP
40 Normally the master export table is initialized with the contents of
41 .I /etc/exports
42 and files under
43 .I /etc/exports.d
44 by invoking
45 .BR "exportfs -a" .
46 However, a system administrator can choose to add or delete
47 exports without modifying
48 .I /etc/exports
49 or files under
50 .I /etc/exports.d
51 by using the
52 .B exportfs
53 command.
54 .PP
55 .B exportfs
56 does not communicate with the kernel directly.
57 It writes export information to
58 .I /var/lib/nfs/etab
59 and relies on its partner programs
60 .B rpc.mountd
61 and
62 .B nfsv4.exportd
63 to manage kernel communication.
64 These daemons work in one of two modes: a netlink mode and a
65 .I /proc
66 mode.
67 .PP
68 In the netlink mode, available on sufficiently recent kernels,
69 .B rpc.mountd
70 (or
71 .BR nfsv4.exportd )
72 communicates with the kernel via generic netlink sockets.
73 The kernel sends multicast notifications when cache entries need to be
74 resolved, and the daemon responds with the appropriate export
75 information.
76 Cache flushing (via
77 .BR "exportfs \-f" )
78 is also performed over netlink.
79 This mode can be disabled with the
80 .B \-L
81 option.
82 .PP
83 In the
84 .I /proc
85 mode, used when netlink is unavailable,
86 .B rpc.mountd
87 manages kernel requests for information about exports
88 via the
89 .I /proc/net/rpc
90 channel files.
91 .SH OPTIONS
92 .TP
93 .B \-d kind " or " \-\-debug kind
94 Turn on debugging. Valid kinds are: all, auth, call, general and parse.
95 Debugging can also be turned on by setting
96 .B debug=
97 in the
98 .B [exportfs]
99 section of
100 .IR /etc/nfs.conf .
101
102 .TP
103 .B -a
104 Export or unexport all directories.
105 .TP
106 .BI "-o " options,...
107 Specify a list of export options in the same manner as in
108 .BR exports (5).
109 .TP
110 .B -i
111 Ignore the
112 .I /etc/exports
113 file and files under
114 .I /etc/exports.d
115 directory. Only default options and options given on the command line are used.
116 .TP
117 .B -r
118 Reexport all directories, synchronizing
119 .I /var/lib/nfs/etab
120 with
121 .IR /etc/exports
122 and files under
123 .IR /etc/exports.d .
124 This option removes entries in
125 .I /var/lib/nfs/etab
126 which have been deleted from
127 .I /etc/exports
128 or files under
129 .IR /etc/exports.d ,
130 and removes any entries from the
131 kernel export table which are no longer valid.
132 .TP
133 .B -u
134 Unexport one or more directories.
135 .TP
136 .B -f
137 If
138 .I /proc/fs/nfsd
139 or
140 .I /proc/fs/nfs
141 is mounted, flush everything out of the kernel's export table.
142 Fresh entries for active clients are added to the kernel's export table by
143 .B rpc.mountd
144 when they make their next NFS mount request.
145 .TP
146 .B -v
147 Be verbose. When exporting or unexporting, show what's going on. When
148 displaying the current export list, also display the list of export
149 options.
150 .TP
151 .B -s
152 Display the current export list suitable for /etc/exports.
153 .TP
154 .B -L
155 Disable the use of netlink for kernel communication and force the use
156 of the legacy
157 .I /proc
158 interfaces for cache flushing and export validation.
159
160 .SH CONFIGURATION FILE
161 The
162 .B [exportfs]
163 section of the
164 .I /etc/nfs.conf
165 configuration file can contain a
166 .B debug
167 value, which can be one or more from the list
168 .BR general ,
169 .BR call ,
170 .BR auth ,
171 .BR parse ,
172 .BR all .
173 When a list is given, the members should be comma-separated.
174
175 .B exportfs
176 will also recognize the
177 .B state-directory-path
178 and
179 .B no\-netlink
180 values from both the
181 .B [mountd]
182 section and the
183 .B [exportd]
184 section.
185 When
186 .B no\-netlink
187 is set,
188 .B exportfs
189 will skip the netlink probe and use the legacy
190 .I /proc
191 interfaces for cache flushing and export validation
192
193 .SH DISCUSSION
194 .SS Exporting Directories
195 The first synopsis shows how to invoke
196 .B exportfs
197 when adding new entries to the export table. When using
198 .BR "exportfs -a" ,
199 all exports listed in
200 .I /etc/exports
201 and files under
202 .I /etc/exports.d
203 are added to
204 .IR /var/lib/nfs/etab .
205 The kernel's export table is also updated as needed.
206 .PP
207 The
208 .I host:/path
209 argument specifies a local directory to export,
210 along with the client or clients who are permitted to access it.
211 See
212 .B exports(5)
213 for a description of supported options and access list formats.
214 .PP
215 IPv6 presentation addresses contain colons, which are already used
216 to separate the "host" and "path" command line arguments.
217 When specifying a client using a raw IPv6 address,
218 enclose the address in square brackets.
219 For IPv6 network addresses, place the prefix just after the closing
220 bracket.
221 .PP
222 To export a directory to the world, simply specify
223 .IR :/path .
224 .PP
225 The export options for a particular host/directory pair derive from
226 several sources.
227 The default export options are
228 .BR sync,ro,root_squash,wdelay .
229 These can be overridden by entries in
230 .IR /etc/exports
231 or files under
232 .IR /etc/exports.d .
233 .PP
234 A system administrator may override options from these sources using the
235 .B -o
236 command-line option on
237 .BR exportfs .
238 This option takes a comma-separated list of options in the same fashion
239 as one would specify them in
240 .IR /etc/exports .
241 In this way
242 .B exportfs
243 can be used to modify the export options of an already exported directory.
244 .SS Unexporting Directories
245 The third synopsis shows how to unexport a currently exported directory.
246 When using
247 .BR "exportfs -ua" ,
248 all entries listed in
249 .I /var/lib/nfs/etab
250 are removed from the kernel export tables, and the file is cleared. This
251 effectively shuts down all NFS activity.
252 .PP
253 To remove an export, specify a
254 .I host:/path
255 pair. This deletes the specified entry from
256 .I /var/lib/nfs/etab
257 and removes the corresponding kernel entry (if any).
258 .PP
259 .SS Dumping the Export Table
260 Invoking
261 .B exportfs
262 without options shows the current list of exported file systems.
263 Adding the
264 .B -v
265 option causes
266 .B exportfs
267 to display the export options for each export.
268 .SH EXAMPLES
269 The following adds all directories listed in
270 .I /etc/exports
271 and files under
272 .I /etc/exports.d
273 to
274 .I /var/lib/nfs/etab
275 and pushes the resulting export entries into the kernel:
276 .PP
277 .nf
278 .B "# exportfs -a
279 .fi
280 .PP
281 To export the
282 .I /usr/tmp
283 directory to host
284 .BR django ,
285 allowing insecure file locking requests from clients:
286 .PP
287 .nf
288 .B "# exportfs -o insecure_locks django:/usr/tmp
289 .fi
290 .PP
291 To unexport the
292 .I /usr/tmp
293 directory:
294 .PP
295 .nf
296 .B "# exportfs -u django:/usr/tmp
297 .fi
298 .PP
299 To unexport all exports listed in
300 .IR /etc/exports
301 and files under
302 .IR /etc/exports.d :
303 .PP
304 .nf
305 .B "# exportfs -au
306 .fi
307 .PP
308 To export the
309 .I /usr/tmp
310 directory to IPv6 link-local clients:
311 .PP
312 .nf
313 .B "# exportfs [fe80::]/64:/usr/tmp
314 .fi
315 .SH USAGE NOTES
316 Exporting to IP networks or DNS and NIS domains does not enable clients
317 from these groups to access NFS immediately.
318 Rather, these sorts of exports are hints to
319 .BR rpc.mountd (8)
320 to grant any mount requests from these clients.
321 This is usually not a problem, because any existing mounts are preserved in
322 .I rmtab
323 across reboots.
324 .PP
325 When unexporting a network or domain entry, any current exports to members
326 of this group will be checked against the remaining valid exports and
327 if they themselves are no longer valid they will be removed.
328 .SH FILES
329 .TP 2.5i
330 .I /etc/exports
331 input file listing exports, export options, and access control lists
332 .TP 2.5i
333 .I /etc/exports.d
334 directory where extra input files are stored.
335 .B Note:
336 only files that end with
337 .I .exports
338 are used.
339 .TP 2.5i
340 .I /var/lib/nfs/etab
341 master table of exports
342 .TP 2.5i
343 .I /var/lib/nfs/rmtab
344 table of clients accessing server's exports
345 .SH SEE ALSO
346 .BR exports (5),
347 .BR nfs.conf (5),
348 .BR rpc.mountd (8),
349 .BR exportd (8),
350 .BR netgroup (5)
351 .SH AUTHORS
352 Olaf Kirch <[email protected]>
353 .br
354 Neil Brown <[email protected]>