1xfs_repair(8) System Manager's Manual xfs_repair(8)
2
3
4
6 xfs_repair - repair an XFS filesystem
7
9 xfs_repair [ -dfLnPv ] [ -m maxmem ] [ -c subopt=value ] [ -o sub‐
10 opt[=value] ] [ -t interval ] [ -l logdev ] [ -r rtdev ] device
11 xfs_repair -V
12
14 xfs_repair repairs corrupt or damaged XFS filesystems (see xfs(5)).
15 The filesystem is specified using the device argument which should be
16 the device name of the disk partition or volume containing the filesys‐
17 tem. If given the name of a block device, xfs_repair will attempt to
18 find the raw device associated with the specified block device and will
19 use the raw device instead.
20
21 Regardless, the filesystem to be repaired must be unmounted, otherwise,
22 the resulting filesystem may be inconsistent or corrupt.
23
25 -f Specifies that the filesystem image to be processed is stored in
26 a regular file at device (see the mkfs.xfs -d file option). This
27 might happen if an image copy of a filesystem has been copied or
28 written into an ordinary file. This option implies that any
29 external log or realtime section is also in an ordinary file.
30
31 -L Force Log Zeroing. Forces xfs_repair to zero the log even if it
32 is dirty (contains metadata changes). When using this option
33 the filesystem will likely appear to be corrupt, and can cause
34 the loss of user files and/or data.
35
36 -l logdev
37 Specifies the device special file where the filesystem's exter‐
38 nal log resides. Only for those filesystems which use an exter‐
39 nal log. See the mkfs.xfs -l option, and refer to xfs(5) for a
40 detailed description of the XFS log.
41
42 -r rtdev
43 Specifies the device special file where the filesystem's real‐
44 time section resides. Only for those filesystems which use a
45 realtime section. See the mkfs.xfs -r option, and refer to
46 xfs(5) for a detailed description of the XFS realtime section.
47
48 -n No modify mode. Specifies that xfs_repair should not modify the
49 filesystem but should only scan the filesystem and indicate what
50 repairs would have been made.
51
52 -P Disable prefetching of inode and directory blocks. Use this
53 option if you find xfs_repair gets stuck and stops proceeding.
54 Interrupting a stuck xfs_repair is safe.
55
56 -m maxmem
57 Specifies the approximate maximum amount of memory, in
58 megabytes, to use for xfs_repair. xfs_repair has its own inter‐
59 nal block cache which will scale out up to the lesser of the
60 process's virtual address limit or about 75% of the system's
61 physical RAM. This option overrides these limits.
62
63 NOTE: These memory limits are only approximate and may use more
64 than the specified limit.
65
66 -c subopt=value
67 Change filesystem parameters. Refer to xfs_admin(8) for informa‐
68 tion on changing filesystem parameters.
69
70 -o subopt[=value]
71 Override what the program might conclude about the filesystem if
72 left to its own devices.
73
74 The suboptions supported are:
75
76 bhash=bhashsize
77 overrides the default buffer cache hash size. The
78 total number of buffer cache entries are limited to 8
79 times this amount. The default size is set to use up
80 the remainder of 75% of the system's physical RAM
81 size.
82
83 ag_stride=ags_per_concat_unit
84 This creates additional processing threads to parallel
85 process AGs that span multiple concat units. This can
86 significantly reduce repair times on concat based
87 filesystems.
88
89 force_geometry
90 Check the filesystem even if geometry information
91 could not be validated. Geometry information can not
92 be validated if only a single allocation group exists
93 and thus we do not have a backup superblock available,
94 or if there are two allocation groups and the two
95 superblocks do not agree on the filesystem geometry.
96 Only use this option if you validated the geometry
97 yourself and know what you are doing. If In doubt run
98 in no modify mode first.
99
100 -t interval
101 Modify reporting interval, specified in seconds. During long
102 runs xfs_repair outputs its progress every 15 minutes. Reporting
103 is only activated when ag_stride is enabled.
104
105 -v Verbose output. May be specified multiple times to increase
106 verbosity.
107
108 -d Repair dangerously. Allow xfs_repair to repair an XFS filesystem
109 mounted read only. This is typically done on a root filesystem
110 from single user mode, immediately followed by a reboot.
111
112 -V Prints the version number and exits.
113
114 Checks Performed
115 Inconsistencies corrected include the following:
116
117 1. Inode and inode blockmap (addressing) checks: bad magic number
118 in inode, bad magic numbers in inode blockmap blocks, extents
119 out of order, incorrect number of records in inode blockmap
120 blocks, blocks claimed that are not in a legal data area of the
121 filesystem, blocks that are claimed by more than one inode.
122
123 2. Inode allocation map checks: bad magic number in inode map
124 blocks, inode state as indicated by map (free or in-use) incon‐
125 sistent with state indicated by the inode, inodes referenced by
126 the filesystem that do not appear in the inode allocation map,
127 inode allocation map referencing blocks that do not appear to
128 contain inodes.
129
130 3. Size checks: number of blocks claimed by inode inconsistent with
131 inode size, directory size not block aligned, inode size not
132 consistent with inode format.
133
134 4. Directory checks: bad magic numbers in directory blocks, incor‐
135 rect number of entries in a directory block, bad freespace
136 information in a directory leaf block, entry pointing to an
137 unallocated (free) or out of range inode, overlapping entries,
138 missing or incorrect dot and dotdot entries, entries out of
139 hashvalue order, incorrect internal directory pointers, direc‐
140 tory type not consistent with inode format and size.
141
142 5. Pathname checks: files or directories not referenced by a path‐
143 name starting from the filesystem root, illegal pathname compo‐
144 nents.
145
146 6. Link count checks: link counts that do not agree with the number
147 of directory references to the inode.
148
149 7. Freemap checks: blocks claimed free by the freemap but also
150 claimed by an inode, blocks unclaimed by any inode but not
151 appearing in the freemap.
152
153 8. Super Block checks: total free block and/or free i-node count
154 incorrect, filesystem geometry inconsistent, secondary and pri‐
155 mary superblocks contradictory.
156
157 Orphaned files and directories (allocated, in-use but unreferenced) are
158 reconnected by placing them in the lost+found directory. The name
159 assigned is the inode number.
160
161 Disk Errors
162 xfs_repair aborts on most disk I/O errors. Therefore, if you are trying
163 to repair a filesystem that was damaged due to a disk drive failure,
164 steps should be taken to ensure that all blocks in the filesystem are
165 readable and writable before attempting to use xfs_repair to repair the
166 filesystem. A possible method is using dd(8) to copy the data onto a
167 good disk.
168
169 lost+found
170 The directory lost+found does not have to already exist in the filesys‐
171 tem being repaired. If the directory does not exist, it is automati‐
172 cally created if required. If it already exists, it will be checked
173 for consistency and if valid will be used for additional orphaned
174 files. Invalid lost+found directories are removed and recreated. Exist‐
175 ing files in a valid lost+found are not removed or renamed.
176
177 Corrupted Superblocks
178 XFS has both primary and secondary superblocks. xfs_repair uses infor‐
179 mation in the primary superblock to automatically find and validate the
180 primary superblock against the secondary superblocks before proceeding.
181 Should the primary be too corrupted to be useful in locating the sec‐
182 ondary superblocks, the program scans the filesystem until it finds and
183 validates some secondary superblocks. At that point, it generates a
184 primary superblock.
185
186 Quotas
187 If quotas are in use, it is possible that xfs_repair will clear some or
188 all of the filesystem quota information. If so, the program issues a
189 warning just before it terminates. If all quota information is lost,
190 quotas are disabled and the program issues a warning to that effect.
191
192 Note that xfs_repair does not check the validity of quota limits. It is
193 recommended that you check the quota limit information manually after
194 xfs_repair. Also, space usage information is automatically regenerated
195 the next time the filesystem is mounted with quotas turned on, so the
196 next quota mount of the filesystem may take some time.
197
199 xfs_repair issues informative messages as it proceeds indicating what
200 it has found that is abnormal or any corrective action that it has
201 taken. Most of the messages are completely understandable only to
202 those who are knowledgeable about the structure of the filesystem.
203 Some of the more common messages are explained here. Note that the
204 language of the messages is slightly different if xfs_repair is run in
205 no-modify mode because the program is not changing anything on disk.
206 No-modify mode indicates what it would do to repair the filesystem if
207 run without the no-modify flag.
208
209 disconnected inode ino, moving to lost+found
210
211 An inode numbered ino was not connected to the filesystem direc‐
212 tory tree and was reconnected to the lost+found directory. The
213 inode is assigned the name of its inode number (ino). If a
214 lost+found directory does not exist, it is automatically cre‐
215 ated.
216
217 disconnected dir inode ino, moving to lost+found
218
219 As above only the inode is a directory inode. If a directory
220 inode is attached to lost+found, all of its children (if any)
221 stay attached to the directory and therefore get automatically
222 reconnected when the directory is reconnected.
223
224 imap claims in-use inode ino is free, correcting imap
225
226 The inode allocation map thinks that inode ino is free whereas
227 examination of the inode indicates that the inode may be in use
228 (although it may be disconnected). The program updates the
229 inode allocation map.
230
231 imap claims free inode ino is in use, correcting imap
232
233 The inode allocation map thinks that inode ino is in use whereas
234 examination of the inode indicates that the inode is not in use
235 and therefore is free. The program updates the inode allocation
236 map.
237
238 resetting inode ino nlinks from x to y
239
240 The program detected a mismatch between the number of valid
241 directory entries referencing inode ino and the number of refer‐
242 ences recorded in the inode and corrected the the number in the
243 inode.
244
245 fork-type fork in ino ino claims used block bno
246
247 Inode ino claims a block bno that is used (claimed) by either
248 another inode or the filesystem itself for metadata storage. The
249 fork-type is either data or attr indicating whether the problem
250 lies in the portion of the inode that tracks regular data or the
251 portion of the inode that stores XFS attributes. If the inode
252 is a real-time (rt) inode, the message says so. Any inode that
253 claims blocks used by the filesystem is deleted. If two or more
254 inodes claim the same block, they are both deleted.
255
256 fork-type fork in ino ino claims dup extent ...
257
258 Inode ino claims a block in an extent known to be claimed more
259 than once. The offset in the inode, start and length of the
260 extent is given. The message is slightly different if the inode
261 is a real-time (rt) inode and the extent is therefore a real-
262 time (rt) extent.
263
264 inode ino - bad extent ...
265
266 An extent record in the blockmap of inode ino claims blocks that
267 are out of the legal range of the filesystem. The message sup‐
268 plies the start, end, and file offset of the extent. The mes‐
269 sage is slightly different if the extent is a real-time (rt)
270 extent.
271
272 bad fork-type fork in inode ino
273
274 There was something structurally wrong or inconsistent with the
275 data structures that map offsets to filesystem blocks.
276
277 cleared inode ino
278
279 There was something wrong with the inode that was uncorrectable
280 so the program freed the inode. This usually happens because
281 the inode claims blocks that are used by something else or the
282 inode itself is badly corrupted. Typically, this message is pre‐
283 ceded by one or more messages indicating why the inode needed to
284 be cleared.
285
286 bad attribute fork in inode ino, clearing attr fork
287
288 There was something wrong with the portion of the inode that
289 stores XFS attributes (the attribute fork) so the program reset
290 the attribute fork. As a result of this, all attributes on that
291 inode are lost.
292
293 correcting nextents for inode ino, was x - counted y
294
295 The program found that the number of extents used to store the
296 data in the inode is wrong and corrected the number. The mes‐
297 sage refers to nextents if the count is wrong on the number of
298 extents used to store attribute information.
299
300 entry name in dir dir_ino not consistent with .. value (xxxx) in dir
301 ino ino, junking entry name in directory inode dir_ino
302
303 The entry name in directory inode dir_ino references a directory
304 inode ino. However, the .. entry in directory ino does not
305 point back to directory dir_ino, so the program deletes the
306 entry name in directory inode dir_ino. If the directory inode
307 ino winds up becoming a disconnected inode as a result of this,
308 it is moved to lost+found later.
309
310 entry name in dir dir_ino references already connected dir ino ino,
311 junking entry name in directory inode dir_ino
312
313 The entry name in directory inode dir_ino points to a directory
314 inode ino that is known to be a child of another directory.
315 Therefore, the entry is invalid and is deleted. This message
316 refers to an entry in a small directory. If this were a large
317 directory, the last phrase would read "will clear entry".
318
319 entry references free inode ino in directory dir_ino, will clear entry
320
321 An entry in directory inode dir_ino references an inode ino that
322 is known to be free. The entry is therefore invalid and is
323 deleted. This message refers to a large directory. If the
324 directory were small, the message would read "junking entry
325 ...".
326
328 xfs_repair -n (no modify node) will return a status of 1 if filesystem
329 corruption was detected and 0 if no filesystem corruption was detected.
330 xfs_repair run without the -n option will always return a status code
331 of 0.
332
334 The filesystem to be checked and repaired must have been unmounted
335 cleanly using normal system administration procedures (the umount(8)
336 command or system shutdown), not as a result of a crash or system
337 reset. If the filesystem has not been unmounted cleanly, mount it and
338 unmount it cleanly before running xfs_repair.
339
340 xfs_repair does not do a thorough job on XFS extended attributes. The
341 structure of the attribute fork will be consistent, but only the con‐
342 tents of attribute forks that will fit into an inode are checked. This
343 limitation will be fixed in the future.
344
345 The no-modify mode (-n option) is not completely accurate. It does not
346 catch inconsistencies in the freespace and inode maps, particularly
347 lost blocks or subtly corrupted maps (trees).
348
349 The no-modify mode can generate repeated warnings about the same prob‐
350 lems because it cannot fix the problems as they are encountered.
351
352 If a filesystem fails to be repaired, a metadump image can be generated
353 with xfs_metadump(8) and be sent to an XFS maintainer to be analysed
354 and xfs_repair fixed and/or improved.
355
357 dd(1), mkfs.xfs(8), umount(8), xfs_admin(8), xfs_metadump(8), xfs(5).
358
359
360
361 xfs_repair(8)