1E2FSCK(8) System Manager's Manual E2FSCK(8)
2
6 e2fsck - check a Linux ext2/ext3/ext4 file system
7
9 e2fsck [ -pacnyrdfkvtDFV ] [ -b superblock ] [ -B blocksize ] [ -l|-L
10 bad_blocks_file ] [ -C fd ] [ -j external-journal ] [ -E extended_op‐
11 tions ] [ -z undo_file ] device
12
14 e2fsck is used to check the ext2/ext3/ext4 family of file systems. For
15 ext3 and ext4 filesystems that use a journal, if the system has been
16 shut down uncleanly without any errors, normally, after replaying the
17 committed transactions in the journal, the file system should be
18 marked as clean. Hence, for filesystems that use journalling, e2fsck
19 will normally replay the journal and exit, unless its superblock indi‐
20 cates that further checking is required.
21 device is a block device (e.g., /dev/sdc1) or file containing the file
23 system.
24 Note that in general it is not safe to run e2fsck on mounted filesys‐
26 tems. The only exception is if the -n option is specified, and -c, -l,
27 or -L options are not specified. However, even if it is safe to do
28 so, the results printed by e2fsck are not valid if the filesystem is
29 mounted. If e2fsck asks whether or not you should check a filesystem
30 which is mounted, the only correct answer is ``no''. Only experts who
31 really know what they are doing should consider answering this question
32 in any other way.
33 If e2fsck is run in interactive mode (meaning that none of -y, -n, or
35 -p are specified), the program will ask the user to fix each problem
36 found in the filesystem. A response of 'y' will fix the error; 'n'
37 will leave the error unfixed; and 'a' will fix the problem and all sub‐
38 sequent problems; pressing Enter will proceed with the default re‐
39 sponse, which is printed before the question mark. Pressing Control-C
40 terminates e2fsck immediately.
41
43 -a This option does the same thing as the -p option. It is pro‐
44 vided for backwards compatibility only; it is suggested that
45 people use -p option whenever possible.
46 -b superblock
48 Instead of using the normal superblock, use an alternative su‐
49 perblock specified by superblock. This option is normally used
50 when the primary superblock has been corrupted. The location of
51 backup superblocks is dependent on the filesystem's blocksize,
52 the number of blocks per group, and features such as sparse_su‐
53 per.
54 Additional backup superblocks can be determined by using the
56 mke2fs program using the -n option to print out where the su‐
57 perblocks exist, supposing mke2fs is supplied with arguments
58 that are consistent with the filesystem's layout (e.g. block‐
59 size, blocks per group, sparse_super, etc.).
60 If an alternative superblock is specified and the filesystem is
62 not opened read-only, e2fsck will make sure that the primary su‐
63 perblock is updated appropriately upon completion of the
64 filesystem check.
65 -B blocksize
67 Normally, e2fsck will search for the superblock at various dif‐
68 ferent block sizes in an attempt to find the appropriate block
69 size. This search can be fooled in some cases. This option
70 forces e2fsck to only try locating the superblock at a particu‐
71 lar blocksize. If the superblock is not found, e2fsck will ter‐
72 minate with a fatal error.
73 -c This option causes e2fsck to use badblocks(8) program to do a
75 read-only scan of the device in order to find any bad blocks.
76 If any bad blocks are found, they are added to the bad block in‐
77 ode to prevent them from being allocated to a file or directory.
78 If this option is specified twice, then the bad block scan will
79 be done using a non-destructive read-write test.
80 -C fd This option causes e2fsck to write completion information to the
82 specified file descriptor so that the progress of the filesystem
83 check can be monitored. This option is typically used by pro‐
84 grams which are running e2fsck. If the file descriptor number
85 is negative, then absolute value of the file descriptor will be
86 used, and the progress information will be suppressed initially.
87 It can later be enabled by sending the e2fsck process a SIGUSR1
88 signal. If the file descriptor specified is 0, e2fsck will
89 print a completion bar as it goes about its business. This re‐
90 quires that e2fsck is running on a video console or terminal.
91 -d Print debugging output (useless unless you are debugging
93 e2fsck).
94 -D Optimize directories in filesystem. This option causes e2fsck
96 to try to optimize all directories, either by reindexing them if
97 the filesystem supports directory indexing, or by sorting and
98 compressing directories for smaller directories, or for filesys‐
99 tems using traditional linear directories.
100 Even without the -D option, e2fsck may sometimes optimize a few
102 directories --- for example, if directory indexing is enabled
103 and a directory is not indexed and would benefit from being in‐
104 dexed, or if the index structures are corrupted and need to be
105 rebuilt. The -D option forces all directories in the filesystem
106 to be optimized. This can sometimes make them a little smaller
107 and slightly faster to search, but in practice, you should
108 rarely need to use this option.
109 The -D option will detect directory entries with duplicate names
111 in a single directory, which e2fsck normally does not enforce
112 for performance reasons.
113 -E extended_options
115 Set e2fsck extended options. Extended options are comma sepa‐
116 rated, and may take an argument using the equals ('=') sign.
117 The following options are supported:
118 ea_ver=extended_attribute_version
120 Set the version of the extended attribute blocks
121 which e2fsck will require while checking the
122 filesystem. The version number may be 1 or 2. The
123 default extended attribute version format is 2.
124 journal_only
126 Only replay the journal if required, but do not per‐
127 form any further checks or repairs.
128 fragcheck
130 During pass 1, print a detailed report of any dis‐
131 contiguous blocks for files in the filesystem.
132 discard
134 Attempt to discard free blocks and unused inode
135 blocks after the full filesystem check (discarding
136 blocks is useful on solid state devices and sparse /
137 thin-provisioned storage). Note that discard is done
138 in pass 5 AFTER the filesystem has been fully
139 checked and only if it does not contain recognizable
140 errors. However there might be cases where e2fsck
141 does not fully recognize a problem and hence in this
142 case this option may prevent you from further manual
143 data recovery.
144 nodiscard
146 Do not attempt to discard free blocks and unused in‐
147 ode blocks. This option is exactly the opposite of
148 discard option. This is set as default.
149 no_optimize_extents
151 Do not offer to optimize the extent tree by elimi‐
152 nating unnecessary width or depth. This can also be
153 enabled in the options section of /etc/e2fsck.conf.
154 optimize_extents
156 Offer to optimize the extent tree by eliminating un‐
157 necessary width or depth. This is the default un‐
158 less otherwise specified in /etc/e2fsck.conf.
159 inode_count_fullmap
161 Trade off using memory for speed when checking a
162 file system with a large number of hard-linked
163 files. The amount of memory required is propor‐
164 tional to the number of inodes in the file system.
165 For large file systems, this can be gigabytes of
166 memory. (For example, a 40TB file system with 2.8
167 billion inodes will consume an additional 5.7 GB
168 memory if this optimization is enabled.) This opti‐
169 mization can also be enabled in the options section
170 of /etc/e2fsck.conf.
171 no_inode_count_fullmap
173 Disable the inode_count_fullmap optimization. This
174 is the default unless otherwise specified in
175 /etc/e2fsck.conf.
176 readahead_kb
178 Use this many KiB of memory to pre-fetch metadata in
179 the hopes of reducing e2fsck runtime. By default,
180 this is set to the size of two block groups' inode
181 tables (typically 4MiB on a regular ext4 filesys‐
182 tem); if this amount is more than 1/50th of total
183 physical memory, readahead is disabled. Set this to
184 zero to disable readahead entirely.
185 bmap2extent
187 Convert block-mapped files to extent-mapped files.
188 fixes_only
190 Only fix damaged metadata; do not optimize htree di‐
191 rectories or compress extent trees. This option is
192 incompatible with the -D and -E bmap2extent options.
193 check_encoding
195 Force verification of encoded filenames in case-in‐
196 sensitive directories. This is the default mode if
197 the filesystem has the strict flag enabled.
198 unshare_blocks
200 If the filesystem has shared blocks, with the shared
201 blocks read-only feature enabled, then this will un‐
202 share all shared blocks and unset the read-only fea‐
203 ture bit. If there is not enough free space then the
204 operation will fail. If the filesystem does not
205 have the read-only feature bit, but has shared
206 blocks anyway, then this option will have no effect.
207 Note when using this option, if there is no free
208 space to clone blocks, there is no prompt to delete
209 files and instead the operation will fail.
210 Note that unshare_blocks implies the "-f" option to
212 ensure that all passes are run. Additionally, if
213 "-n" is also specified, e2fsck will simulate trying
214 to allocate enough space to deduplicate. If this
215 fails, the exit code will be non-zero.
216 -f Force checking even if the file system seems clean.
218 -F Flush the filesystem device's buffer caches before beginning.
220 Only really useful for doing e2fsck time trials.
221 -j external-journal
223 Set the pathname where the external-journal for this filesystem
224 can be found.
225 -k When combined with the -c option, any existing bad blocks in the
227 bad blocks list are preserved, and any new bad blocks found by
228 running badblocks(8) will be added to the existing bad blocks
229 list.
230 -l filename
232 Add the block numbers listed in the file specified by filename
233 to the list of bad blocks. The format of this file is the same
234 as the one generated by the badblocks(8) program. Note that the
235 block numbers are based on the blocksize of the filesystem.
236 Hence, badblocks(8) must be given the blocksize of the filesys‐
237 tem in order to obtain correct results. As a result, it is much
238 simpler and safer to use the -c option to e2fsck, since it will
239 assure that the correct parameters are passed to the badblocks
240 program.
241 -L filename
243 Set the bad blocks list to be the list of blocks specified by
244 filename. (This option is the same as the -l option, except the
245 bad blocks list is cleared before the blocks listed in the file
246 are added to the bad blocks list.)
247 -n Open the filesystem read-only, and assume an answer of `no' to
249 all questions. Allows e2fsck to be used non-interactively.
250 This option may not be specified at the same time as the -p or
251 -y options.
252 -p Automatically repair ("preen") the file system. This option
254 will cause e2fsck to automatically fix any filesystem problems
255 that can be safely fixed without human intervention. If e2fsck
256 discovers a problem which may require the system administrator
257 to take additional corrective action, e2fsck will print a de‐
258 scription of the problem and then exit with the value 4 logi‐
259 cally or'ed into the exit code. (See the EXIT CODE section.)
260 This option is normally used by the system's boot scripts. It
261 may not be specified at the same time as the -n or -y options.
262 -r This option does nothing at all; it is provided only for back‐
264 wards compatibility.
265 -t Print timing statistics for e2fsck. If this option is used
267 twice, additional timing statistics are printed on a pass by
268 pass basis.
269 -v Verbose mode.
271 -V Print version information and exit.
273 -y Assume an answer of `yes' to all questions; allows e2fsck to be
275 used non-interactively. This option may not be specified at the
276 same time as the -n or -p options.
277 -z undo_file
279 Before overwriting a file system block, write the old contents
280 of the block to an undo file. This undo file can be used with
281 e2undo(8) to restore the old contents of the file system should
282 something go wrong. If the empty string is passed as the
283 undo_file argument, the undo file will be written to a file
284 named e2fsck-device.e2undo in the directory specified via the
285 E2FSPROGS_UNDO_DIR environment variable.
286 WARNING: The undo file cannot be used to recover from a power or
288 system crash.
289
291 The exit code returned by e2fsck is the sum of the following condi‐
292 tions:
293 0 - No errors
294 1 - File system errors corrected
295 2 - File system errors corrected, system should
296 be rebooted
297 4 - File system errors left uncorrected
298 8 - Operational error
299 16 - Usage or syntax error
300 32 - E2fsck canceled by user request
301 128 - Shared library error
302
304 The following signals have the following effect when sent to e2fsck.
305 SIGUSR1
307 This signal causes e2fsck to start displaying a completion bar
308 or emitting progress information. (See discussion of the -C op‐
309 tion.)
310 SIGUSR2
312 This signal causes e2fsck to stop displaying a completion bar or
313 emitting progress information.
314
316 Almost any piece of software will have bugs. If you manage to find a
317 filesystem which causes e2fsck to crash, or which e2fsck is unable to
318 repair, please report it to the author.
319 Please include as much information as possible in your bug report.
321 Ideally, include a complete transcript of the e2fsck run, so I can see
322 exactly what error messages are displayed. (Make sure the messages
323 printed by e2fsck are in English; if your system has been configured so
324 that e2fsck's messages have been translated into another language,
325 please set the the LC_ALL environment variable to C so that the tran‐
326 script of e2fsck's output will be useful to me.) If you have a
327 writable filesystem where the transcript can be stored, the script(1)
328 program is a handy way to save the output of e2fsck to a file.
329 It is also useful to send the output of dumpe2fs(8). If a specific in‐
331 ode or inodes seems to be giving e2fsck trouble, try running the de‐
332 bugfs(8) command and send the output of the stat(1u) command run on the
333 relevant inode(s). If the inode is a directory, the debugfs dump com‐
334 mand will allow you to extract the contents of the directory inode,
335 which can sent to me after being first run through uuencode(1). The
336 most useful data you can send to help reproduce the bug is a compressed
337 raw image dump of the filesystem, generated using e2image(8). See the
338 e2image(8) man page for more details.
339 Always include the full version string which e2fsck displays when it is
341 run, so I know which version you are running.
342
344 E2FSCK_CONFIG
345 Determines the location of the configuration file (see
346 e2fsck.conf(5)).
347
349 This version of e2fsck was written by Theodore Ts'o <tytso@mit.edu>.
350
352 e2fsck.conf(5), badblocks(8), dumpe2fs(8), debugfs(8), e2image(8),
353 mke2fs(8), tune2fs(8)
354E2fsprogs version 1.46.3 July 2021 E2FSCK(8)