1DUMP(8) System management commands DUMP(8)
2
3
4
6 dump - ext2/3 filesystem backup
7
9 dump [-level#] [-ackMnqSuv] [-A file] [-B records] [-b blocksize] [-d
10 density] [-D file] [-e inode numbers] [-E file] [-f file] [-F script]
11 [-h level] [-I nr errors] [-jcompression level] [-L label] [-Q file]
12 [-s feet] [-T date] [-y] [-zcompression level] files-to-dump
13
14 dump [-W | -w]
15
17 Dump examines files on an ext2/3 filesystem and determines which files
18 need to be backed up. These files are copied to the given disk, tape or
19 other storage medium for safe keeping (see the -f option below for
20 doing remote backups). A dump that is larger than the output medium is
21 broken into multiple volumes. On most media the size is determined by
22 writing until an end-of-media indication is returned.
23
24 On media that cannot reliably return an end-of-media indication (such
25 as some cartridge tape drives), each volume is of a fixed size; the
26 actual size is determined by specifying cartridge media, or via the
27 tape size, density and/or block count options below. By default, the
28 same output file name is used for each volume after prompting the oper‐
29 ator to change media.
30
31 files-to-dump is either a mountpoint of a filesystem or a list of files
32 and directories to be backed up as a subset of a filesystem. In the
33 former case, either the path to a mounted filesystem or the device of
34 an unmounted filesystem can be used. In the latter case, certain
35 restrictions are placed on the backup: -u is not allowed, the only dump
36 level that is supported is 0 and all the files and directories must
37 reside on the same filesystem.
38
40 The following options are supported by dump:
41
42 -level#
43 The dump level (any integer). A level 0, full backup, guarantees
44 the entire file system is copied (but see also the -h option
45 below). A level number above 0, incremental backup, tells dump
46 to copy all files new or modified since the last dump of a lower
47 level. The default level is 9. Historically only levels 0 to 9
48 were usable in dump, this version is able to understand any
49 integer as a dump level.
50
51 -a “auto-size”. Bypass all tape length calculations, and write
52 until an end-of-media indication is returned. This works best
53 for most modern tape drives, and is the default. Use of this
54 option is particularly recommended when appending to an existing
55 tape, or using a tape drive with hardware compression (where you
56 can never be sure about the compression ratio).
57
58 -A archive_file
59 Archive a dump table-of-contents in the specified archive_file
60 to be used by restore(8) to determine whether a file is in the
61 dump file that is being restored.
62
63 -b blocksize
64 The number of kilobytes per dump record. The default blocksize
65 is 10, unless the -d option has been used to specify a tape den‐
66 sity of 6250BPI or more, in which case the default blocksize is
67 32. Th maximal value is 1024. Note however that, since the IO
68 system slices all requests into chunks of MAXBSIZE (which can be
69 as low as 64kB), you can experience problems with dump(8) and
70 restore(8) when using a higher value, depending on your kernel
71 and/or libC versions.
72
73 -B records
74 The number of 1 kB blocks per volume. Not normally required, as
75 dump can detect end-of-media. When the specified size is
76 reached, dump waits for you to change the volume. This option
77 overrides the calculation of tape size based on length and den‐
78 sity. If compression is on this limits the size of the com‐
79 pressed output per volume. Multiple values may be given as a
80 single argument separated by commas. Each value will be used
81 for one dump volume in the order listed; if dump creates more
82 volumes than the number of values given, the last value will be
83 used for the remaining volumes. This is useful for filling up
84 already partially filled media (and then continuing with full
85 size volumes on empty media) or mixing media of different sizes.
86
87 -c Change the defaults for use with a cartridge tape drive, with a
88 density of 8000 bpi, and a length of 1700 feet. Specifying a
89 cartridge drive overrides the end-of-media detection.
90
91 -d density
92 Set tape density to density. The default is 1600BPI. Specifying
93 a tape density overrides the end-of-media detection.
94
95 -D file
96 Set the path name of the file storing the information about the
97 previous full and incremental dumps. The default location is
98 /etc/dumpdates.
99
100 -e inodes
101 Exclude inodes from the dump. The inodes parameter is a comma
102 separated list of inode numbers (you can use stat(1) to find the
103 inode number for a file or directory).
104
105 -E file
106 Read list of inodes to be excluded from the dump from the text
107 file file. The file file should be an ordinary file containing
108 inode numbers separated by newlines.
109
110 -f file
111 Write the backup to file; file may be a special device file like
112 /dev/st0 (a tape drive), /dev/rsd1c (a floppy disk drive), an
113 ordinary file, or - (the standard output). Multiple file names
114 may be given as a single argument separated by commas. Each file
115 will be used for one dump volume in the order listed; if the
116 dump requires more volumes than the number of names given, the
117 last file name will used for all remaining volumes after prompt‐
118 ing for media changes. If the name of the file is of the form
119 host:file or user@host:file dump writes to the named file on the
120 remote host (which should already exist, dump doesn't create a
121 new remote file) using rmt(8). The default path name of the
122 remote rmt(8) program is /etc/rmt; this can be overridden by the
123 environment variable RMT.
124
125 -F script
126 Run script at the end of each tape (except for the last one).
127 The device name and the current volume number are passed on the
128 command line. The script must return 0 if dump should continue
129 without asking the user to change the tape, 1 if dump should
130 continue but ask the user to change the tape. Any other exit
131 code will cause dump to abort. For security reasons, dump
132 reverts back to the real user ID and the real group ID before
133 running the script.
134
135 -h level
136 Honor the user nodump flag UF_NODUMP only for dumps at or above
137 the given level. The default honor level is 1, so that incre‐
138 mental backups omit such files but full backups retain them.
139
140 -I nr errors
141 By default, dump will ignore the first 32 read errors on the
142 file system before asking for operator intervention. You can
143 change this using this flag to any value. This is useful when
144 running dump on an active filesystem where read errors simply
145 indicate an inconsistency between the mapping and dumping
146 passes.
147
148 A value of 0 means that all read errors will be ignored.
149
150 -jcompression level
151 Compress every block to be written on the tape using bzlib
152 library. This option will work only when dumping to a file or
153 pipe or, when dumping to a tape drive, if the tape drive is
154 capable of writing variable length blocks. You will need at
155 least the 0.4b24 version of restore in order to extract com‐
156 pressed tapes. Tapes written using compression will not be com‐
157 patible with the BSD tape format. The (optional) parameter spec‐
158 ifies the compression level bzlib will use. The default compres‐
159 sion level is 2. If the optional parameter is specified, there
160 should be no white space between the option letter and the
161 parameter.
162
163 -k Use Kerberos authentication to talk to remote tape servers.
164 (Only available if this option was enabled when dump was com‐
165 piled.)
166
167 -L label
168 The user-supplied text string label is placed into the dump
169 header, where tools like restore(8) and file(8) can access it.
170 Note that this label is limited to be at most LBLSIZE (currently
171 16) characters, which must include the terminating \0.
172
173 -m If this flag is specified, dump will optimise the output for
174 inodes having been changed but not modified since the last dump
175 ('changed' and 'modified' have the meaning defined in stat(2) ).
176 For those inodes, dump will save only the metadata, instead of
177 saving the entire inode contents. Inodes which are either
178 directories or have been modified since the last dump are saved
179 in a regular way. Uses of this flag must be consistent, meaning
180 that either every dump in an incremental dump set have the flag,
181 or no one has it.
182
183 If you use this option, be aware that many programs that unpack
184 files from archives (e.g. tar, rpm, unzip, dpkg) may set files'
185 mtimes to dates in the past. Files installed in this way may
186 not be dumped correctly using "dump -m" if the modified mtime is
187 earlier than the previous level dump.
188
189 Tapes written using such 'metadata only' inodes will not be com‐
190 patible with the BSD tape format or older versions of restore.
191
192 -M Enable the multi-volume feature. The name specified with f is
193 treated as a prefix and dump writes in sequence to <prefix>001,
194 <prefix>002 etc. This can be useful when dumping to files on an
195 ext2 partition, in order to bypass the 2GB file size limitation.
196
197 -n Whenever dump requires operator attention, notify all operators
198 in the group operator by means similar to a wall(1).
199
200 -q Make dump abort immediately whenever operator attention is
201 required, without prompting in case of write errors, tape
202 changes etc.
203
204 -Q file
205 Enable the Quick File Access support. Tape positions for each
206 inode are stored into the file file which is used by restore (if
207 called with parameter -Q and the filename) to directly position
208 the tape at the file restore is currently working on. This saves
209 hours when restoring single files from large backups, saves the
210 tapes and the drive's head.
211
212 It is recommended to set up the st driver to return logical tape
213 positions rather than physical before calling dump/restore with
214 parameter -Q. Since not all tape devices support physical tape
215 positions those tape devices return an error during dump/restore
216 when the st driver is set to the default physical setting.
217 Please see the st(4) man page, option MTSETDRVBUFFER , or the
218 mt(1) man page, on how to set the driver to return logical tape
219 positions.
220
221 Before calling restore with parameter -Q, always make sure the
222 st driver is set to return the same type of tape position used
223 during the call to dump. Otherwise restore may be confused.
224
225 This option can be used when dumping to local tapes (see above)
226 or to local files.
227
228 -s feet
229 Attempt to calculate the amount of tape needed at a particular
230 density. If this amount is exceeded, dump prompts for a new
231 tape. It is recommended to be a bit conservative on this option.
232 The default tape length is 2300 feet. Specifying the tape size
233 overrides end-of-media detection.
234
235 -S Size estimate. Determine the amount of space that is needed to
236 perform the dump without actually doing it, and display the
237 estimated number of bytes it will take. This is useful with
238 incremental dumps to determine how many volumes of media will be
239 needed.
240
241 -T date
242 Use the specified date as the starting time for the dump instead
243 of the time determined from looking in /etc/dumpdates . The
244 format of date is the same as that of ctime(3) followed by an
245 rfc822 timezone specification: either a plus or minus sign fol‐
246 lowed by two digits for the number of hours and two digits for
247 the minutes. For example, -0800 for eight hours west of Green‐
248 wich or +0230 for two hours and a half east of Greenwich. This
249 timezone offset takes into account daylight savings time (if
250 applicable to the timezone): UTC offsets when daylight savings
251 time is in effect will be different than offsets when daylight
252 savings time is not in effect. For backward compatibility, if no
253 timezone is specified, a local time is assumed. This option is
254 useful for automated dump scripts that wish to dump over a spe‐
255 cific period of time. The -T option is mutually exclusive from
256 the -u option.
257
258 -u Update the file /etc/dumpdates after a successful dump. The for‐
259 mat of /etc/dumpdates is readable by people, consisting of one
260 free format record per line: filesystem name, increment level
261 and ctime(3) format dump date followed by a rfc822 timezone
262 specification (see the -u option for details). If no timezone
263 offset is specified, times are interpreted as local. Whenever
264 the file is written, all dates in the file are converted to the
265 local time zone, without changing the UTC times. There may be
266 only one entry per filesystem at each level. The file /etc/dump‐
267 dates may be edited to change any of the fields, if necessary.
268
269 -v The -v (verbose) makes dump to print extra information which
270 could be helpful in debug sessions.
271
272 -W Dump tells the operator what file systems need to be dumped.
273 This information is gleaned from the files /etc/dumpdates and
274 /etc/fstab. The -W option causes dump to print out, for all
275 file systems in /etc/dumpdates , and regognized file systems in
276 /etc/mtab and /etc/fstab. the most recent dump date and level,
277 and highlights those that should be dumped. If the -W option is
278 set, all other options are ignored, and dump exits immediately.
279
280 -w Is like -W, but prints only recognized filesystems in /etc/mtab
281 and /etc/fstab which need to be dumped.
282
283 -y Compress every block to be written to the tape using the lzo
284 library. This doesn't compress as well as the zlib library but
285 it's much faster. This option will work only when dumping to a
286 file or pipe or, when dumping to a tape drive, if the tape drive
287 is capable of writing variable length blocks. You will need at
288 least the 0.4b34 version of restore in order to extract com‐
289 pressed tapes. Tapes written using compression will not be com‐
290 patible with the BSD tape format.
291
292 -zcompression level
293 Compress every block to be written on the tape using zlib
294 library. This option will work only when dumping to a file or
295 pipe or, when dumping to a tape drive, if the tape drive is
296 capable of writing variable length blocks. You will need at
297 least the 0.4b22 version of restore in order to extract com‐
298 pressed tapes. Tapes written using compression will not be com‐
299 patible with the BSD tape format. The (optional) parameter spec‐
300 ifies the compression level zlib will use. The default compres‐
301 sion level is 2. If the optional parameter is specified, there
302 should be no white space between the option letter and the
303 parameter.
304
305 Dump requires operator intervention on these conditions: end of tape,
306 end of dump, tape write error, tape open error or disk read error (if
307 there is more than a threshold of nr errors). In addition to alerting
308 all operators implied by the -n key, dump interacts with the operator
309 on dump's control terminal at times when dump can no longer proceed, or
310 if something is grossly wrong. All questions dump poses must be
311 answered by typing “yes” or “no”, appropriately.
312
313 Since making a dump involves a lot of time and effort for full dumps,
314 dump checkpoints itself at the start of each tape volume. If writing
315 that volume fails for some reason, dump will, with operator permission,
316 restart itself from the checkpoint after the old tape has been rewound
317 and removed, and a new tape has been mounted.
318
319 Dump tells the operator what is going on at periodic intervals, includ‐
320 ing usually low estimates of the number of blocks to write, the number
321 of tapes it will take, the time to completion, and the time to the tape
322 change. The output is verbose, so that others know that the terminal
323 controlling dump is busy, and will be for some time.
324
325 In the event of a catastrophic disk event, the time required to restore
326 all the necessary backup tapes or files to disk can be kept to a mini‐
327 mum by staggering the incremental dumps. An efficient method of stag‐
328 gering incremental dumps to minimize the number of tapes follows:
329
330 — Always start with a level 0 backup, for example:
331 /sbin/dump -0u -f /dev/st0 /usr/src
332
333 This should be done at set intervals, say once a month or once
334 every two months, and on a set of fresh tapes that is saved for‐
335 ever.
336
337 — After a level 0, dumps of active file systems are taken on a
338 daily basis, using a modified Tower of Hanoi algorithm, with
339 this sequence of dump levels:
340 3 2 5 4 7 6 9 8 9 9 ...
341
342 For the daily dumps, it should be possible to use a fixed number
343 of tapes for each day, used on a weekly basis. Each week, a
344 level 1 dump is taken, and the daily Hanoi sequence repeats
345 beginning with 3. For weekly dumps, another fixed set of tapes
346 per dumped file system is used, also on a cyclical basis.
347
348 After several months or so, the daily and weekly tapes should get
349 rotated out of the dump cycle and fresh tapes brought in.
350
351 (The 4.3BSD option syntax is implemented for backward compatibility but
352 is not documented here.)
353
355 TAPE If no -f option was specified, dump will use the device speci‐
356 fied via TAPE as the dump device. TAPE may be of the form tape‐
357 name, host:tapename, or user@host:tapename.
358
359 RMT The environment variable RMT will be used to determine the path‐
360 name of the remote rmt(8) program.
361
362 RSH Dump uses the contents of this variable to determine the name of
363 the remote shell command to use when doing remote backups (rsh,
364 ssh etc.). If this variable is not set, rcmd(3) will be used,
365 but only root will be able to do remote backups.
366
368 /dev/st0
369 default tape unit to dump to
370
371 /etc/dumpdates
372 dump date records
373
374 /etc/fstab
375 dump table: file systems and frequency
376
377 /etc/mtab
378 dump table: mounted file systems
379
380 /etc/group
381 to find group operator
382
384 fstab(5), restore(8), rmt(8)
385
387 Many, and verbose.
388
390 The format of the /etc/dumpdates file has changed in release 0.4b34,
391 however, the file will be read correctly with either pre-0.4b34 or
392 0.4b34 and later versions of dump provided that the machine on which
393 dump is run did not change timezones (which should be a fairly rare
394 occurence).
395
397 Dump exits with zero status on success. Startup errors are indicated
398 with an exit code of 1; abnormal termination is indicated with an exit
399 code of 3.
400
402 It might be considered a bug that this version of dump can only handle
403 ext2/3 filesystems. Specifically, it does not work with FAT filesys‐
404 tems.
405
406 Fewer than 32 read errors (change this with -I) on the filesystem are
407 ignored. If noticing read errors is important, the output from dump can
408 be parsed to look for lines that contain the text 'read error'.
409
410 When a read error occurs, dump prints out the corresponding physical
411 disk block and sector number and the ext2/3 logical block number. It
412 doesn't print out the corresponing file name or even the inode number.
413 The user has to use debugfs(8), commands ncheck and icheck to translate
414 the ext2blk number printed out by dump into an inode number, then into
415 a file name.
416
417 Each reel requires a new process, so parent processes for reels already
418 written just hang around until the entire tape is written.
419
420 The estimated number of tapes is not correct if compression is on.
421
422 It would be nice if dump knew about the dump sequence, kept track of
423 the tapes scribbled on, told the operator which tape to mount when, and
424 provided more assistance for the operator running restore.
425
426 Dump cannot do remote backups without being run as root, due to its
427 security history. Presently, it works if you set it setuid (like it
428 used to be), but this might constitute a security risk. Note that you
429 can set RSH to use a remote shell program instead.
430
432 The dump/restore backup suite was ported to Linux's Second Extended
433 File System by Remy Card <card@Linux.EU.Org>. He maintained the initial
434 versions of dump (up and including 0.4b4, released in january 1997).
435
436 Starting with 0.4b5, the new maintainer is Stelian Pop
437 <stelian@popies.net>.
438
440 The dump/restore backup suite is available from <http://dump.source‐
441 forge.net>
442
444 A dump command appeared in Version 6 AT&T UNIX.
445
446
447
448BSD version 0.4b41 of January 2, 2006 DUMP(8)