1MOUNT(8) System Administration MOUNT(8)
2
3
4
6 mount - mount a filesystem
7
9 mount [-lhV]
10
11 mount -a [-fFnrsvw] [-t vfstype] [-O optlist]
12
13 mount [-fnrsvw] [-o option[,option]...] device|dir
14
15 mount [-fnrsvw] [-t vfstype] [-o options] device dir
16
18 All files accessible in a Unix system are arranged in one big tree, the
19 file hierarchy, rooted at /. These files can be spread out over sev‐
20 eral devices. The mount command serves to attach the filesystem found
21 on some device to the big file tree. Conversely, the umount(8) command
22 will detach it again.
23
24 The standard form of the mount command, is
25
26 mount -t type device dir
27
28 This tells the kernel to attach the filesystem found on device (which
29 is of type type) at the directory dir. The previous contents (if any)
30 and owner and mode of dir become invisible, and as long as this
31 filesystem remains mounted, the pathname dir refers to the root of the
32 filesystem on device.
33
34 If only directory or device is given, for example:
35
36 mount /dir
37
38 then mount looks for a mountpoint and if not found then for a device in
39 the /etc/fstab file. It's possible to use --target or --source options
40 to avoid ambivalent interpretation of the given argument. For example
41
42 mount --target /mountpoint
43
44
45
46 The listing and help.
47 The listing mode is maintained for backward compatibility only.
48
49 For more robust and definable output use findmnt(8), especially
50 in your scripts. Note that control characters in the mountpoint
51 name are replaced with '?'.
52
53
54 mount [-l] [-t type]
55 lists all mounted filesystems (of type type). The option
56 -l adds the labels in this listing. See below.
57
58 The device indication.
59 Most devices are indicated by a file name (of a block special
60 device), like /dev/sda1, but there are other possibilities. For
61 example, in the case of an NFS mount, device may look like
62 knuth.cwi.nl:/dir. It is possible to indicate a block special
63 device using its filesystem LABEL or UUID (see the -L and -U
64 options below) and partition PARTUUID or PARTLABEL (partition
65 identifiers are supported for GUID Partition Table (GPT) and MAC
66 partition tables only).
67
68 The recommended setup is to use tags (e.g. LABEL=<label>) rather
69 than /dev/disk/by-{label,uuid,partuuid,partlabel} udev symlinks
70 in the /etc/fstab file. The tags are more readable, robust and
71 portable. The mount(8) command internally uses udev symlinks, so
72 use the symlinks in /etc/fstab has no advantage over the tags.
73 For more details see libblkid(3).
74
75 Note that mount(8) uses UUIDs as strings. The UUIDs from command
76 line or fstab(5) are not converted to internal binary represen‐
77 tation. The string representation of the UUID should be based on
78 lower case characters.
79
80 The proc filesystem is not associated with a special device, and
81 when mounting it, an arbitrary keyword, such as proc can be used
82 instead of a device specification. (The customary choice none
83 is less fortunate: the error message `none busy' from umount can
84 be confusing.)
85
86 The /etc/fstab, /etc/mtab and /proc/mounts files.
87 The file /etc/fstab (see fstab(5)), may contain lines describing
88 what devices are usually mounted where, using which options. The
89 default location of the fstab(5) file could be overridden by
90 --fstab <path> command line option (see below for more details).
91
92 The command
93
94 mount -a [-t type] [-O optlist]
95
96 (usually given in a bootscript) causes all filesystems mentioned
97 in fstab (of the proper type and/or having or not having the
98 proper options) to be mounted as indicated, except for those
99 whose line contains the noauto keyword. Adding the -F option
100 will make mount fork, so that the filesystems are mounted simul‐
101 taneously.
102
103 When mounting a filesystem mentioned in fstab or mtab, it suf‐
104 fices to give only the device, or only the mount point.
105
106
107 The programs mount and umount maintain a list of currently
108 mounted filesystems in the file /etc/mtab. If no arguments are
109 given to mount, this list is printed.
110
111 The mount program does not read the /etc/fstab file if device
112 (or LABEL, UUID, PARTUUID or PARTLABEL) and dir are specified.
113 For example:
114
115 mount /dev/foo /dir
116
117 If you want to override mount options from /etc/fstab you have
118 to use:
119
120 mount device|dir -o <options>
121
122 and then the mount options from command line will be appended to
123 the list of options from /etc/fstab. The usual behaviour is
124 that the last option wins if there is more duplicated options.
125
126 When the proc filesystem is mounted (say at /proc), the files
127 /etc/mtab and /proc/mounts have very similar contents. The for‐
128 mer has somewhat more information, such as the mount options
129 used, but is not necessarily up-to-date (cf. the -n option
130 below). It is possible to replace /etc/mtab by a symbolic link
131 to /proc/mounts, and especially when you have very large numbers
132 of mounts things will be much faster with that symlink, but some
133 information is lost that way, and in particular using the "user"
134 option will fail.
135
136 The non-superuser mounts.
137 Normally, only the superuser can mount filesystems. However,
138 when fstab contains the user option on a line, anybody can mount
139 the corresponding system.
140
141 Thus, given a line
142
143 /dev/cdrom /cd iso9660 ro,user,noauto,unhide
144
145 any user can mount the iso9660 filesystem found on his CDROM
146 using the command
147
148 mount /dev/cdrom
149
150 or
151
152 mount /cd
153
154 For more details, see fstab(5). Only the user that mounted a
155 filesystem can unmount it again. If any user should be able to
156 unmount, then use users instead of user in the fstab line. The
157 owner option is similar to the user option, with the restriction
158 that the user must be the owner of the special file. This may be
159 useful e.g. for /dev/fd if a login script makes the console user
160 owner of this device. The group option is similar, with the
161 restriction that the user must be member of the group of the
162 special file.
163
164
165 The bind mounts.
166 Since Linux 2.4.0 it is possible to remount part of the file
167 hierarchy somewhere else. The call is
168 mount --bind olddir newdir
169 or shortoption
170 mount -B olddir newdir
171 or fstab entry is:
172 /olddir /newdir none bind
173
174 After this call the same contents is accessible in two places.
175 One can also remount a single file (on a single file). It's also
176 possible to use the bind mount to create a mountpoint from a
177 regular directory, for example:
178
179 mount --bind foo foo
180
181 The bind mount call attaches only (part of) a single filesystem,
182 not possible submounts. The entire file hierarchy including sub‐
183 mounts is attached a second place using
184
185 mount --rbind olddir newdir
186
187 or shortoption
188
189 mount -R olddir newdir
190
191 Note that the filesystem mount options will remain the same as
192 those on the original mount point.
193
194 mount(8) since v2.27 (backported to RHEL7.3) allow to change the
195 options by passing the -o option along with --bind for example:
196
197 mount --bind,ro foo foo
198
199 This feature is not supported by Linux kernel and it is imple‐
200 mented in userspace by additional remount mount(2) syscall. This
201 solution is not atomic.
202
203 The alternative (classic) way to create a read-only bind mount
204 is to use remount operation, for example:
205
206 mount --bind olddir newdir
207 mount -o remount,ro,bind olddir newdir
208
209 Note that read-only bind will create a read-only mountpoint (VFS
210 entry), but the original filesystem superblock will still be
211 writable, meaning that the olddir will be writable, but the
212 newdir will be read-only.
213
214 It's impossible to change mount options recursively (for example
215 with -o rbind,ro).
216
217 The move operation.
218 Since Linux 2.5.1 it is possible to atomically move a mounted
219 tree to another place. The call is
220 mount --move olddir newdir
221 or shortoption
222 mount -M olddir newdir
223 This will cause the contents which previously appeared under
224 olddir to be accessed under newdir. The physical location of
225 the files is not changed. Note that the olddir has to be a
226 mountpoint.
227
228 Note that moving a mount residing under a shared mount is
229 invalid and unsupported. Use findmnt -o TARGET,PROPAGATION to
230 see the current propagation flags.
231
232 The shared subtrees operations.
233 Since Linux 2.6.15 it is possible to mark a mount and its sub‐
234 mounts as shared, private, slave or unbindable. A shared mount
235 provides ability to create mirrors of that mount such that
236 mounts and umounts within any of the mirrors propagate to the
237 other mirror. A slave mount receives propagation from its mas‐
238 ter, but any not vice-versa. A private mount carries no propa‐
239 gation abilities. A unbindable mount is a private mount which
240 cannot be cloned through a bind operation. Detailed semantics is
241 documented in Documentation/filesystems/sharedsubtree.txt file
242 in the kernel source tree.
243
244 Supported operations:
245 mount --make-shared mountpoint
246 mount --make-slave mountpoint
247 mount --make-private mountpoint
248 mount --make-unbindable mountpoint
249
250 The following commands allows one to recursively change the type
251 of all the mounts under a given mountpoint.
252
253 mount --make-rshared mountpoint
254 mount --make-rslave mountpoint
255 mount --make-rprivate mountpoint
256 mount --make-runbindable mountpoint
257
258 mount(8) does not read fstab(5) when --make-* operation is
259 requested. All necessary information has to be specified on com‐
260 mand line.
261
262 Note that Linux kernel does not allow to change more propagation
263 flags by one mount(2) syscall and the flags cannot be mixed with
264 another mount options.
265
266 Since util-linux 2.23 mount command allows to use more propaga‐
267 tion flags together and with another mount operations. This fea‐
268 ture is EXPERIMENTAL. The propagation flags are applied by
269 additional mount(2) syscalls after previous successful mount
270 operation. Note that this use case is not atomic. The propaga‐
271 tion flags is possible to specify in fstab(5) as mount options
272 (private, slave, shared, unbindable, rprivate, rslave, rshared,
273 runbindable).
274
275 For example
276 mount --make-private --make-unbindable /dev/sda1 /A
277
278 is the same as
279 mount /dev/sda1 /A
280 mount --make-private /A
281 mount --make-unbindable /A
282
283
285 The full set of mount options used by an invocation of mount is deter‐
286 mined by first extracting the mount options for the filesystem from the
287 fstab table, then applying any options specified by the -o argument,
288 and finally applying a -r or -w option, when present.
289
290 Command line options available for the mount command:
291
292 -V, --version
293 Output version.
294
295 -h, --help
296 Print a help message.
297
298 -v, --verbose
299 Verbose mode.
300
301 -a, --all
302 Mount all filesystems (of the given types) mentioned in fstab.
303
304 -F, --fork
305 (Used in conjunction with -a.) Fork off a new incarnation of
306 mount for each device. This will do the mounts on different
307 devices or different NFS servers in parallel. This has the
308 advantage that it is faster; also NFS timeouts go in parallel. A
309 disadvantage is that the mounts are done in undefined order.
310 Thus, you cannot use this option if you want to mount both /usr
311 and /usr/spool.
312
313 -f, --fake
314 Causes everything to be done except for the actual system call;
315 if it's not obvious, this ``fakes'' mounting the filesystem.
316 This option is useful in conjunction with the -v flag to deter‐
317 mine what the mount command is trying to do. It can also be used
318 to add entries for devices that were mounted earlier with the -n
319 option. The -f option checks for existing record in /etc/mtab
320 and fails when the record already exists (with regular non-fake
321 mount, this check is done by kernel).
322
323 -i, --internal-only
324 Don't call the /sbin/mount.<filesystem> helper even if it
325 exists.
326
327 -l, --show-labels
328 Add the labels in the mount output. Mount must have permission
329 to read the disk device (e.g. be suid root) for this to work.
330 One can set such a label for ext2, ext3 or ext4 using the
331 e2label(8) utility, or for XFS using xfs_admin(8), or for reis‐
332 erfs using reiserfstune(8).
333
334 -n, --no-mtab
335 Mount without writing in /etc/mtab. This is necessary for exam‐
336 ple when /etc is on a read-only filesystem.
337
338 -c, --no-canonicalize
339 Don't canonicalize paths. The mount command canonicalizes all
340 paths (from command line or fstab) and stores canonicalized
341 paths to the /etc/mtab file. This option can be used together
342 with the -f flag for already canonicalized absolute paths.
343
344 -s Tolerate sloppy mount options rather than failing. This will
345 ignore mount options not supported by a filesystem type. Not all
346 filesystems support this option. This option exists for support
347 of the Linux autofs-based automounter.
348
349 --source src
350 If only one argument for the mount command is given then the
351 argument might be interpreted as target (mountpoint) or source
352 (device). This option allows to explicitly define that the argu‐
353 ment is mount source.
354
355 -r, --read-only
356 Mount the filesystem read-only. A synonym is -o ro.
357
358 Note that, depending on the filesystem type, state and kernel
359 behavior, the system may still write to the device. For example,
360 Ext3 or ext4 will replay its journal if the filesystem is dirty.
361 To prevent this kind of write access, you may want to mount ext3
362 or ext4 filesystem with "ro,noload" mount options or set the
363 block device to read-only mode, see command blockdev(8).
364
365 -w, --rw, --read-write
366 Mount the filesystem read/write. This is the default. A synonym
367 is -o rw.
368
369 -L, --label label
370 Mount the partition that has the specified label.
371
372 -U, --uuid uuid
373 Mount the partition that has the specified uuid. These two
374 options require the file /proc/partitions (present since Linux
375 2.1.116) to exist.
376
377 -T, --fstab path
378 Specifies alternative fstab file. If the path is directory then
379 the files in the directory are sorted by strverscmp(3), files
380 that starts with "." or without .fstab extension are ignored.
381 The option can be specified more than once. This option is
382 mostly designed for initramfs or chroot scripts where additional
383 configuration is specified outside standard system configura‐
384 tion.
385
386 Note that mount(8) does not pass the option --fstab to
387 /sbin/mount.<type> helpers, it means that the alternative fstab
388 files will be invisible for the helpers. This is no problem for
389 normal mounts, but user (non-root) mounts always require fstab
390 to verify user's rights.
391
392 -t, --types vfstype
393 The argument following the -t is used to indicate the filesystem
394 type. The filesystem types which are currently supported
395 include: adfs, affs, autofs, cifs, coda, coherent, cramfs,
396 debugfs, devpts, efs, ext, ext2, ext3, ext4, hfs, hfsplus, hpfs,
397 iso9660, jfs, minix, msdos, ncpfs, nfs, nfs4, ntfs, proc, qnx4,
398 ramfs, reiserfs, romfs, squashfs, smbfs, sysv, tmpfs, ubifs,
399 udf, ufs, umsdos, usbfs, vfat, xenix, xfs, xiafs. Note that
400 coherent, sysv and xenix are equivalent and that xenix and
401 coherent will be removed at some point in the future — use sysv
402 instead. Since kernel version 2.1.21 the types ext and xiafs do
403 not exist anymore. Earlier, usbfs was known as usbdevfs. Note,
404 the real list of all supported filesystems depends on your ker‐
405 nel.
406
407 The programs mount and umount support filesystem subtypes. The
408 subtype is defined by '.subtype' suffix. For example
409 'fuse.sshfs'. It's recommended to use subtype notation rather
410 than add any prefix to the mount source (for example
411 'sshfs#example.com' is depreacated).
412
413 For most types all the mount program has to do is issue a simple
414 mount(2) system call, and no detailed knowledge of the filesys‐
415 tem type is required. For a few types however (like nfs, nfs4,
416 cifs, smbfs, ncpfs) ad hoc code is necessary. The nfs, nfs4,
417 cifs, smbfs, and ncpfs filesystems have a separate mount pro‐
418 gram. In order to make it possible to treat all types in a uni‐
419 form way, mount will execute the program /sbin/mount.TYPE (if
420 that exists) when called with type TYPE. Since various versions
421 of the smbmount program have different calling conventions,
422 /sbin/mount.smbfs may have to be a shell script that sets up the
423 desired call.
424
425 If no -t option is given, or if the auto type is specified,
426 mount will try to guess the desired type. Mount uses the blkid
427 library for guessing the filesystem type; if that does not turn
428 up anything that looks familiar, mount will try to read the file
429 /etc/filesystems, or, if that does not exist, /proc/filesystems.
430 All of the filesystem types listed there will be tried, except
431 for those that are labeled "nodev" (e.g., devpts, proc and nfs).
432 If /etc/filesystems ends in a line with a single * only, mount
433 will read /proc/filesystems afterwards. All of the filesystem
434 types will be mounted with mount option "silent".
435
436 The auto type may be useful for user-mounted floppies. Creating
437 a file /etc/filesystems can be useful to change the probe order
438 (e.g., to try vfat before msdos or ext3 before ext2) or if you
439 use a kernel module autoloader.
440
441 More than one type may be specified in a comma separated list.
442 The list of filesystem types can be prefixed with no to specify
443 the filesystem types on which no action should be taken. (This
444 can be meaningful with the -a option.) For example, the command:
445
446 mount -a -t nomsdos,ext
447
448 mounts all filesystems except those of type msdos and ext.
449
450 --target dir
451 If only one argument for the mount command is given then the
452 argument might be interpreted as target (mountpoint) or source
453 (device). This option allows to explicitly define that the argu‐
454 ment is mount target.
455
456 -O, --test-opts opts
457 Used in conjunction with -a, to limit the set of filesystems to
458 which the -a is applied. Like -t in this regard except that it
459 is useless except in the context of -a. For example, the com‐
460 mand:
461
462 mount -a -O no_netdev
463
464 mounts all filesystems except those which have the option _net‐
465 dev specified in the options field in the /etc/fstab file.
466
467 It is different from -t in that each option is matched exactly;
468 a leading no at the beginning of one option does not negate the
469 rest.
470
471 The -t and -O options are cumulative in effect; that is, the
472 command
473
474 mount -a -t ext2 -O _netdev
475
476 mounts all ext2 filesystems with the _netdev option, not all
477 filesystems that are either ext2 or have the _netdev option
478 specified.
479
480 -o, --options opts
481 Options are specified with a -o flag followed by a comma sepa‐
482 rated string of options. For example:
483
484 mount LABEL=mydisk -o noatime,nouser
485
486
487 For more details, see FILESYSTEM INDEPENDENT MOUNT OPTIONS and
488 FILESYSTEM SPECIFIC MOUNT OPTIONS sections.
489
490 -B, --bind
491 Remount a subtree somewhere else (so that its contents are
492 available in both places). See above.
493
494 -R, --rbind
495 Remount a subtree and all possible submounts somewhere else (so
496 that its contents are available in both places). See above.
497
498 -M, --move
499 Move a subtree to some other place. See above.
500
501
503 Some of these options are only useful when they appear in the
504 /etc/fstab file.
505
506 Some of these options could be enabled or disabled by default in the
507 system kernel. To check the current setting see the options in
508 /proc/mounts. Note that filesystems also have per-filesystem specific
509 default mount options (see for example tune2fs -l output for extN
510 filesystems).
511
512 The following options apply to any filesystem that is being mounted
513 (but not every filesystem actually honors them - e.g., the sync option
514 today has effect only for ext2, ext3, ext4, fat, vfat and ufs):
515
516
517 async All I/O to the filesystem should be done asynchronously. (See
518 also the sync option.)
519
520 atime Do not use noatime feature, then the inode access time is con‐
521 trolled by kernel defaults. See also the description for stric‐
522 tatime and relatime mount options.
523
524 noatime
525 Do not update inode access times on this filesystem (e.g., for
526 faster access on the news spool to speed up news servers).
527
528 auto Can be mounted with the -a option.
529
530 noauto Can only be mounted explicitly (i.e., the -a option will not
531 cause the filesystem to be mounted).
532
533 context=context, fscontext=context, defcontext=context and rootcon‐
534 text=context
535 The context= option is useful when mounting filesystems that do
536 not support extended attributes, such as a floppy or hard disk
537 formatted with VFAT, or systems that are not normally running
538 under SELinux, such as an ext3 or ext4 formatted disk from a
539 non-SELinux workstation. You can also use context= on filesys‐
540 tems you do not trust, such as a floppy. It also helps in com‐
541 patibility with xattr-supporting filesystems on earlier 2.4.<x>
542 kernel versions. Even where xattrs are supported, you can save
543 time not having to label every file by assigning the entire disk
544 one security context.
545
546 A commonly used option for removable media is context="sys‐
547 tem_u:object_r:removable_t".
548
549 Two other options are fscontext= and defcontext=, both of which
550 are mutually exclusive of the context option. This means you can
551 use fscontext and defcontext with each other, but neither can be
552 used with context.
553
554 The fscontext= option works for all filesystems, regardless of
555 their xattr support. The fscontext option sets the overarching
556 filesystem label to a specific security context. This filesystem
557 label is separate from the individual labels on the files. It
558 represents the entire filesystem for certain kinds of permission
559 checks, such as during mount or file creation. Individual file
560 labels are still obtained from the xattrs on the files them‐
561 selves. The context option actually sets the aggregate context
562 that fscontext provides, in addition to supplying the same label
563 for individual files.
564
565 You can set the default security context for unlabeled files
566 using defcontext= option. This overrides the value set for unla‐
567 beled files in the policy and requires a filesystem that sup‐
568 ports xattr labeling.
569
570 The rootcontext= option allows you to explicitly label the root
571 inode of a FS being mounted before that FS or inode becomes vis‐
572 ible to userspace. This was found to be useful for things like
573 stateless linux.
574
575 Note that the kernel rejects any remount request that includes
576 the context option, even when unchanged from the current con‐
577 text.
578
579 Warning: the context value might contain commas, in which case
580 the value has to be properly quoted, otherwise mount(8) will
581 interpret the comma as a separator between mount options. Don't
582 forget that the shell strips off quotes and thus double quoting
583 is required. For example:
584
585 mount -t tmpfs none /mnt -o 'context="sys‐
586 tem_u:object_r:tmp_t:s0:c127,c456",noexec'
587
588 For more details, see selinux(8).
589
590
591 defaults
592 Use default options: rw, suid, dev, exec, auto, nouser, and
593 async.
594
595 Note that the real set of the all default mount options depends
596 on kernel and filesystem type. See the begin of this section for
597 more details.
598
599 dev Interpret character or block special devices on the filesystem.
600
601 nodev Do not interpret character or block special devices on the file
602 system.
603
604 diratime
605 Update directory inode access times on this filesystem. This is
606 the default.
607
608 nodiratime
609 Do not update directory inode access times on this filesystem.
610
611 dirsync
612 All directory updates within the filesystem should be done syn‐
613 chronously. This affects the following system calls: creat,
614 link, unlink, symlink, mkdir, rmdir, mknod and rename.
615
616 exec Permit execution of binaries.
617
618 noexec Do not allow direct execution of any binaries on the mounted
619 filesystem. (Until recently it was possible to run binaries
620 anyway using a command like /lib/ld*.so /mnt/binary. This trick
621 fails since Linux 2.4.25 / 2.6.0.)
622
623 group Allow an ordinary (i.e., non-root) user to mount the filesystem
624 if one of his groups matches the group of the device. This
625 option implies the options nosuid and nodev (unless overridden
626 by subsequent options, as in the option line group,dev,suid).
627
628 iversion
629 Every time the inode is modified, the i_version field will be
630 incremented.
631
632 noiversion
633 Do not increment the i_version inode field.
634
635 mand Allow mandatory locks on this filesystem. See fcntl(2).
636
637 nomand Do not allow mandatory locks on this filesystem.
638
639 _netdev
640 The filesystem resides on a device that requires network access
641 (used to prevent the system from attempting to mount these
642 filesystems until the network has been enabled on the system).
643
644 nofail Do not report errors for this device if it does not exist.
645
646 relatime
647 Update inode access times relative to modify or change time.
648 Access time is only updated if the previous access time was ear‐
649 lier than the current modify or change time. (Similar to noat‐
650 ime, but doesn't break mutt or other applications that need to
651 know if a file has been read since the last time it was modi‐
652 fied.)
653
654 Since Linux 2.6.30, the kernel defaults to the behavior provided
655 by this option (unless noatime was specified), and the stricta‐
656 time option is required to obtain traditional semantics. In
657 addition, since Linux 2.6.30, the file's last access time is
658 always updated if it is more than 1 day old.
659
660 norelatime
661 Do not use relatime feature. See also the strictatime mount
662 option.
663
664 strictatime
665 Allows to explicitly requesting full atime updates. This makes
666 it possible for kernel to defaults to relatime or noatime but
667 still allow userspace to override it. For more details about the
668 default system mount options see /proc/mounts.
669
670 nostrictatime
671 Use the kernel's default behaviour for inode access time
672 updates.
673
674 suid Allow set-user-identifier or set-group-identifier bits to take
675 effect.
676
677 nosuid Do not allow set-user-identifier or set-group-identifier bits to
678 take effect. (This seems safe, but is in fact rather unsafe if
679 you have suidperl(1) installed.)
680
681 silent Turn on the silent flag.
682
683 loud Turn off the silent flag.
684
685 owner Allow an ordinary (i.e., non-root) user to mount the filesystem
686 if he is the owner of the device. This option implies the
687 options nosuid and nodev (unless overridden by subsequent
688 options, as in the option line owner,dev,suid).
689
690 remount
691 Attempt to remount an already-mounted filesystem. This is com‐
692 monly used to change the mount flags for a filesystem, espe‐
693 cially to make a readonly filesystem writable. It does not
694 change device or mount point.
695
696 The remount functionality follows the standard way how the mount
697 command works with options from fstab. It means the mount com‐
698 mand doesn't read fstab (or mtab) only when a device and dir are
699 fully specified.
700
701 mount -o remount,rw /dev/foo /dir
702
703 After this call all old mount options are replaced and arbitrary
704 stuff from fstab is ignored, except the loop= option which is
705 internally generated and maintained by the mount command.
706
707 mount -o remount,rw /dir
708
709 After this call mount reads fstab (or mtab) and merges these
710 options with options from command line ( -o ).
711
712 ro Mount the filesystem read-only.
713
714 rw Mount the filesystem read-write.
715
716 sync All I/O to the filesystem should be done synchronously. In case
717 of media with limited number of write cycles (e.g. some flash
718 drives) "sync" may cause life-cycle shortening.
719
720 user Allow an ordinary user to mount the filesystem. The name of the
721 mounting user is written to mtab so that he can unmount the
722 filesystem again. This option implies the options noexec,
723 nosuid, and nodev (unless overridden by subsequent options, as
724 in the option line user,exec,dev,suid).
725
726 nouser Forbid an ordinary (i.e., non-root) user to mount the filesys‐
727 tem. This is the default.
728
729 users Allow every user to mount and unmount the filesystem. This
730 option implies the options noexec, nosuid, and nodev (unless
731 overridden by subsequent options, as in the option line
732 users,exec,dev,suid).
733
734 x-* All options prefixed with "x-" are interpreted as comments or
735 userspace applications specific options. These options are not
736 stored to mtab file, send to mount.<type> helpers or mount(2)
737 system call. The suggested format is x-<appname>.<option> (e.g.
738 x-systemd.automount).
739
740 x-mount.mkdir[=<mode>]
741 Allow to make a target directory (mountpoint). The optional
742 argument <mode> specifies the file system access mode used for
743 mkdir (2) in octal notation. The default mode is 0755. This
744 functionality is supported only for root users.
745
746
748 The following options apply only to certain filesystems. We sort them
749 by filesystem. They all follow the -o flag.
750
751 What options are supported depends a bit on the running kernel. More
752 info may be found in the kernel source subdirectory Documenta‐
753 tion/filesystems.
754
755
757 uid=value and gid=value
758 Set the owner and group of the files in the filesystem (default:
759 uid=gid=0).
760
761 ownmask=value and othmask=value
762 Set the permission mask for ADFS 'owner' permissions and 'other'
763 permissions, respectively (default: 0700 and 0077, respec‐
764 tively). See also /usr/src/linux/Documentation/filesys‐
765 tems/adfs.txt.
766
768 uid=value and gid=value
769 Set the owner and group of the root of the filesystem (default:
770 uid=gid=0, but with option uid or gid without specified value,
771 the uid and gid of the current process are taken).
772
773 setuid=value and setgid=value
774 Set the owner and group of all files.
775
776 mode=value
777 Set the mode of all files to value & 0777 disregarding the orig‐
778 inal permissions. Add search permission to directories that
779 have read permission. The value is given in octal.
780
781 protect
782 Do not allow any changes to the protection bits on the filesys‐
783 tem.
784
785 usemp Set uid and gid of the root of the filesystem to the uid and gid
786 of the mount point upon the first sync or umount, and then clear
787 this option. Strange...
788
789 verbose
790 Print an informational message for each successful mount.
791
792 prefix=string
793 Prefix used before volume name, when following a link.
794
795 volume=string
796 Prefix (of length at most 30) used before '/' when following a
797 symbolic link.
798
799 reserved=value
800 (Default: 2.) Number of unused blocks at the start of the
801 device.
802
803 root=value
804 Give explicitly the location of the root block.
805
806 bs=value
807 Give blocksize. Allowed values are 512, 1024, 2048, 4096.
808
809 grpquota|noquota|quota|usrquota
810 These options are accepted but ignored. (However, quota utili‐
811 ties may react to such strings in /etc/fstab.)
812
813
815 See the options section of the mount.cifs(8) man page (cifs-utils pack‐
816 age must be installed).
817
818
820 None.
821
822
824 The debugfs filesystem is a pseudo filesystem, traditionally mounted on
825 /sys/kernel/debug. As of kernel version 3.4, debugfs has the following
826 options:
827
828 uid=n, gid=n
829 Set the owner and group of the mountpoint.
830
831 mode=value
832 Sets the mode of the mountpoint.
833
835 The devpts filesystem is a pseudo filesystem, traditionally mounted on
836 /dev/pts. In order to acquire a pseudo terminal, a process opens
837 /dev/ptmx; the number of the pseudo terminal is then made available to
838 the process and the pseudo terminal slave can be accessed as
839 /dev/pts/<number>.
840
841 uid=value and gid=value
842 This sets the owner or the group of newly created PTYs to the
843 specified values. When nothing is specified, they will be set to
844 the UID and GID of the creating process. For example, if there
845 is a tty group with GID 5, then gid=5 will cause newly created
846 PTYs to belong to the tty group.
847
848 mode=value
849 Set the mode of newly created PTYs to the specified value. The
850 default is 0600. A value of mode=620 and gid=5 makes "mesg y"
851 the default on newly created PTYs.
852
853 newinstance
854 Create a private instance of devpts filesystem, such that
855 indices of ptys allocated in this new instance are independent
856 of indices created in other instances of devpts.
857
858 All mounts of devpts without this newinstance option share the
859 same set of pty indices (i.e legacy mode). Each mount of devpts
860 with the newinstance option has a private set of pty indices.
861
862 This option is mainly used to support containers in the linux
863 kernel. It is implemented in linux kernel versions starting with
864 2.6.29. Further, this mount option is valid only if CON‐
865 FIG_DEVPTS_MULTIPLE_INSTANCES is enabled in the kernel configu‐
866 ration.
867
868 To use this option effectively, /dev/ptmx must be a symbolic
869 link to pts/ptmx. See Documentation/filesystems/devpts.txt in
870 the linux kernel source tree for details.
871
872 ptmxmode=value
873
874 Set the mode for the new ptmx device node in the devpts filesys‐
875 tem.
876
877 With the support for multiple instances of devpts (see newin‐
878 stance option above), each instance has a private ptmx node in
879 the root of the devpts filesystem (typically /dev/pts/ptmx).
880
881 For compatibility with older versions of the kernel, the default
882 mode of the new ptmx node is 0000. ptmxmode=value specifies a
883 more useful mode for the ptmx node and is highly recommended
884 when the newinstance option is specified.
885
886 This option is only implemented in linux kernel versions start‐
887 ing with 2.6.29. Further this option is valid only if CON‐
888 FIG_DEVPTS_MULTIPLE_INSTANCES is enabled in the kernel configu‐
889 ration.
890
891
893 None. Note that the `ext' filesystem is obsolete. Don't use it. Since
894 Linux version 2.1.21 extfs is no longer part of the kernel source.
895
896
898 The `ext2' filesystem is the standard Linux filesystem. Since Linux
899 2.5.46, for most mount options the default is determined by the
900 filesystem superblock. Set them with tune2fs(8).
901
902 acl|noacl
903 Support POSIX Access Control Lists (or not).
904
905 bsddf|minixdf
906 Set the behaviour for the statfs system call. The minixdf behav‐
907 iour is to return in the f_blocks field the total number of
908 blocks of the filesystem, while the bsddf behaviour (which is
909 the default) is to subtract the overhead blocks used by the ext2
910 filesystem and not available for file storage. Thus
911
912 % mount /k -o minixdf; df /k; umount /k
913 Filesystem 1024-blocks Used Available Capacity Mounted on
914 /dev/sda6 2630655 86954 2412169 3% /k
915 % mount /k -o bsddf; df /k; umount /k
916 Filesystem 1024-blocks Used Available Capacity Mounted on
917 /dev/sda6 2543714 13 2412169 0% /k
918
919 (Note that this example shows that one can add command line
920 options to the options given in /etc/fstab.)
921
922
923 check=none or nocheck
924 No checking is done at mount time. This is the default. This is
925 fast. It is wise to invoke e2fsck(8) every now and then, e.g.
926 at boot time. The non-default behavior is unsupported
927 (check=normal and check=strict options have been removed). Note
928 that these mount options don't have to be supported if ext4 ker‐
929 nel driver is used for ext2 and ext3 filesystems.
930
931 debug Print debugging info upon each (re)mount.
932
933 errors={continue|remount-ro|panic}
934 Define the behaviour when an error is encountered. (Either
935 ignore errors and just mark the filesystem erroneous and con‐
936 tinue, or remount the filesystem read-only, or panic and halt
937 the system.) The default is set in the filesystem superblock,
938 and can be changed using tune2fs(8).
939
940 grpid|bsdgroups and nogrpid|sysvgroups
941 These options define what group id a newly created file gets.
942 When grpid is set, it takes the group id of the directory in
943 which it is created; otherwise (the default) it takes the fsgid
944 of the current process, unless the directory has the setgid bit
945 set, in which case it takes the gid from the parent directory,
946 and also gets the setgid bit set if it is a directory itself.
947
948 grpquota|noquota|quota|usrquota
949 The usrquota (same as quota) mount option enables user quota
950 support on the filesystem. grpquota enables group quotas sup‐
951 port. You need the quota utilities to actually enable and manage
952 the quota system.
953
954 nouid32
955 Disables 32-bit UIDs and GIDs. This is for interoperability
956 with older kernels which only store and expect 16-bit values.
957
958 oldalloc or orlov
959 Use old allocator or Orlov allocator for new inodes. Orlov is
960 default.
961
962 resgid=n and resuid=n
963 The ext2 filesystem reserves a certain percentage of the avail‐
964 able space (by default 5%, see mke2fs(8) and tune2fs(8)). These
965 options determine who can use the reserved blocks. (Roughly:
966 whoever has the specified uid, or belongs to the specified
967 group.)
968
969 sb=n Instead of block 1, use block n as superblock. This could be
970 useful when the filesystem has been damaged. (Earlier, copies
971 of the superblock would be made every 8192 blocks: in block 1,
972 8193, 16385, ... (and one got thousands of copies on a big
973 filesystem). Since version 1.08, mke2fs has a -s (sparse
974 superblock) option to reduce the number of backup superblocks,
975 and since version 1.15 this is the default. Note that this may
976 mean that ext2 filesystems created by a recent mke2fs cannot be
977 mounted r/w under Linux 2.0.*.) The block number here uses 1k
978 units. Thus, if you want to use logical block 32768 on a
979 filesystem with 4k blocks, use "sb=131072".
980
981 user_xattr|nouser_xattr
982 Support "user." extended attributes (or not).
983
984
985
987 The ext3 filesystem is a version of the ext2 filesystem which has been
988 enhanced with journaling. It supports the same options as ext2 as well
989 as the following additions:
990
991 journal=update
992 Update the ext3 filesystem's journal to the current format.
993
994 journal=inum
995 When a journal already exists, this option is ignored. Other‐
996 wise, it specifies the number of the inode which will represent
997 the ext3 filesystem's journal file; ext3 will create a new
998 journal, overwriting the old contents of the file whose inode
999 number is inum.
1000
1001 journal_dev=devnum
1002 When the external journal device's major/minor numbers have
1003 changed, this option allows the user to specify the new journal
1004 location. The journal device is identified through its new
1005 major/minor numbers encoded in devnum.
1006
1007 norecovery/noload
1008 Don't load the journal on mounting. Note that if the filesystem
1009 was not unmounted cleanly, skipping the journal replay will lead
1010 to the filesystem containing inconsistencies that can lead to
1011 any number of problems.
1012
1013 data={journal|ordered|writeback}
1014 Specifies the journaling mode for file data. Metadata is always
1015 journaled. To use modes other than ordered on the root filesys‐
1016 tem, pass the mode to the kernel as boot parameter, e.g. root‐
1017 flags=data=journal.
1018
1019 journal
1020 All data is committed into the journal prior to being
1021 written into the main filesystem.
1022
1023 ordered
1024 This is the default mode. All data is forced directly
1025 out to the main file system prior to its metadata being
1026 committed to the journal.
1027
1028 writeback
1029 Data ordering is not preserved - data may be written into
1030 the main filesystem after its metadata has been committed
1031 to the journal. This is rumoured to be the highest-
1032 throughput option. It guarantees internal filesystem
1033 integrity, however it can allow old data to appear in
1034 files after a crash and journal recovery.
1035
1036 barrier=0 / barrier=1
1037 This enables/disables barriers. barrier=0 disables it, bar‐
1038 rier=1 enables it. Write barriers enforce proper on-disk order‐
1039 ing of journal commits, making volatile disk write caches safe
1040 to use, at some performance penalty. The ext3 filesystem does
1041 not enable write barriers by default. Be sure to enable barri‐
1042 ers unless your disks are battery-backed one way or another.
1043 Otherwise you risk filesystem corruption in case of power fail‐
1044 ure.
1045
1046 commit=nrsec
1047 Sync all data and metadata every nrsec seconds. The default
1048 value is 5 seconds. Zero means default.
1049
1050 user_xattr
1051 Enable Extended User Attributes. See the attr(5) manual page.
1052
1053 acl Enable POSIX Access Control Lists. See the acl(5) manual page.
1054
1055 usrjquota=aquota.user|grpjquota=aquota.group|jqfmt=vfsv0
1056 Apart from the old quota system (as in ext2, jqfmt=vfsold aka
1057 version 1 quota) ext3 also supports journaled quotas (version 2
1058 quota). jqfmt=vfsv0 enables journaled quotas. For journaled quo‐
1059 tas the mount options usrjquota=aquota.user and
1060 grpjquota=aquota.group are required to tell the quota system
1061 which quota database files to use. Journaled quotas have the
1062 advantage that even after a crash no quota check is required.
1063
1064
1066 The ext4 filesystem is an advanced level of the ext3 filesystem which
1067 incorporates scalability and reliability enhancements for supporting
1068 large filesystem.
1069
1070 The options journal_dev, noload, data, commit, orlov, oldalloc,
1071 [no]user_xattr [no]acl, bsddf, minixdf, debug, errors, data_err, grpid,
1072 bsdgroups, nogrpid sysvgroups, resgid, resuid, sb, quota, noquota,
1073 grpquota, usrquota usrjquota, grpjquota and jqfmt are backwardly com‐
1074 patible with ext3 or ext2.
1075
1076 journal_checksum
1077 Enable checksumming of the journal transactions. This will
1078 allow the recovery code in e2fsck and the kernel to detect cor‐
1079 ruption in the kernel. It is a compatible change and will be
1080 ignored by older kernels.
1081
1082 journal_async_commit
1083 Commit block can be written to disk without waiting for descrip‐
1084 tor blocks. If enabled older kernels cannot mount the device.
1085 This will enable 'journal_checksum' internally.
1086
1087 barrier=0 / barrier=1 / barrier / nobarrier
1088 This enables/disables the use of write barriers in the jbd code.
1089 barrier=0 disables, barrier=1 enables. This also requires an IO
1090 stack which can support barriers, and if jbd gets an error on a
1091 barrier write, it will disable again with a warning. Write bar‐
1092 riers enforce proper on-disk ordering of journal commits, making
1093 volatile disk write caches safe to use, at some performance
1094 penalty. If your disks are battery-backed in one way or
1095 another, disabling barriers may safely improve performance. The
1096 mount options "barrier" and "nobarrier" can also be used to
1097 enable or disable barriers, for consistency with other ext4
1098 mount options.
1099
1100 The ext4 filesystem enables write barriers by default.
1101
1102 inode_readahead_blks=n
1103 This tuning parameter controls the maximum number of inode table
1104 blocks that ext4's inode table readahead algorithm will pre-read
1105 into the buffer cache. The value must be a power of 2. The
1106 default value is 32 blocks.
1107
1108 stripe=n
1109 Number of filesystem blocks that mballoc will try to use for
1110 allocation size and alignment. For RAID5/6 systems this should
1111 be the number of data disks * RAID chunk size in filesystem
1112 blocks.
1113
1114 delalloc
1115 Deferring block allocation until write-out time.
1116
1117 nodelalloc
1118 Disable delayed allocation. Blocks are allocated when data is
1119 copied from user to page cache.
1120
1121 max_batch_time=usec
1122 Maximum amount of time ext4 should wait for additional filesys‐
1123 tem operations to be batch together with a synchronous write
1124 operation. Since a synchronous write operation is going to force
1125 a commit and then a wait for the I/O complete, it doesn't cost
1126 much, and can be a huge throughput win, we wait for a small
1127 amount of time to see if any other transactions can piggyback on
1128 the synchronous write. The algorithm used is designed to auto‐
1129 matically tune for the speed of the disk, by measuring the
1130 amount of time (on average) that it takes to finish committing a
1131 transaction. Call this time the "commit time". If the time that
1132 the transaction has been running is less than the commit time,
1133 ext4 will try sleeping for the commit time to see if other oper‐
1134 ations will join the transaction. The commit time is capped by
1135 the max_batch_time, which defaults to 15000us (15ms). This opti‐
1136 mization can be turned off entirely by setting max_batch_time to
1137 0.
1138
1139 min_batch_time=usec
1140 This parameter sets the commit time (as described above) to be
1141 at least min_batch_time. It defaults to zero microseconds.
1142 Increasing this parameter may improve the throughput of multi-
1143 threaded, synchronous workloads on very fast disks, at the cost
1144 of increasing latency.
1145
1146 journal_ioprio=prio
1147 The I/O priority (from 0 to 7, where 0 is the highest priority)
1148 which should be used for I/O operations submitted by kjournald2
1149 during a commit operation. This defaults to 3, which is a
1150 slightly higher priority than the default I/O priority.
1151
1152 abort Simulate the effects of calling ext4_abort() for debugging pur‐
1153 poses. This is normally used while remounting a filesystem
1154 which is already mounted.
1155
1156 auto_da_alloc|noauto_da_alloc
1157 Many broken applications don't use fsync() when replacing exist‐
1158 ing files via patterns such as
1159
1160 fd = open("foo.new")/write(fd,..)/close(fd)/ rename("foo.new",
1161 "foo")
1162
1163 or worse yet
1164
1165 fd = open("foo", O_TRUNC)/write(fd,..)/close(fd).
1166
1167 If auto_da_alloc is enabled, ext4 will detect the replace-via-
1168 rename and replace-via-truncate patterns and force that any
1169 delayed allocation blocks are allocated such that at the next
1170 journal commit, in the default data=ordered mode, the data
1171 blocks of the new file are forced to disk before the rename()
1172 operation is committed. This provides roughly the same level of
1173 guarantees as ext3, and avoids the "zero-length" problem that
1174 can happen when a system crashes before the delayed allocation
1175 blocks are forced to disk.
1176
1177 discard/nodiscard
1178 Controls whether ext4 should issue discard/TRIM commands to the
1179 underlying block device when blocks are freed. This is useful
1180 for SSD devices and sparse/thinly-provisioned LUNs, but it is
1181 off by default until sufficient testing has been done.
1182
1183 nouid32
1184 Disables 32-bit UIDs and GIDs. This is for interoperability
1185 with older kernels which only store and expect 16-bit values.
1186
1187 resize Allows to resize filesystem to the end of the last existing
1188 block group, further resize has to be done with resize2fs either
1189 online, or offline. It can be used only with conjunction with
1190 remount.
1191
1192 block_validity/noblock_validity
1193 This options allows to enables/disables the in-kernel facility
1194 for tracking filesystem metadata blocks within internal data
1195 structures. This allows multi- block allocator and other rou‐
1196 tines to quickly locate extents which might overlap with
1197 filesystem metadata blocks. This option is intended for debug‐
1198 ging purposes and since it negatively affects the performance,
1199 it is off by default.
1200
1201 dioread_lock/dioread_nolock
1202 Controls whether or not ext4 should use the DIO read locking. If
1203 the dioread_nolock option is specified ext4 will allocate unini‐
1204 tialized extent before buffer write and convert the extent to
1205 initialized after IO completes. This approach allows ext4 code
1206 to avoid using inode mutex, which improves scalability on high
1207 speed storages. However this does not work with data journaling
1208 and dioread_nolock option will be ignored with kernel warning.
1209 Note that dioread_nolock code path is only used for extent-based
1210 files. Because of the restrictions this options comprises it is
1211 off by default (e.g. dioread_lock).
1212
1213 i_version
1214 Enable 64-bit inode version support. This option is off by
1215 default.
1216
1217
1219 (Note: fat is not a separate filesystem, but a common part of the
1220 msdos, umsdos and vfat filesystems.)
1221
1222 blocksize={512|1024|2048}
1223 Set blocksize (default 512). This option is obsolete.
1224
1225 uid=value and gid=value
1226 Set the owner and group of all files. (Default: the uid and gid
1227 of the current process.)
1228
1229 umask=value
1230 Set the umask (the bitmask of the permissions that are not
1231 present). The default is the umask of the current process. The
1232 value is given in octal.
1233
1234 dmask=value
1235 Set the umask applied to directories only. The default is the
1236 umask of the current process. The value is given in octal.
1237
1238 fmask=value
1239 Set the umask applied to regular files only. The default is the
1240 umask of the current process. The value is given in octal.
1241
1242 allow_utime=value
1243 This option controls the permission check of mtime/atime.
1244
1245 20 If current process is in group of file's group ID, you
1246 can change timestamp.
1247
1248 2 Other users can change timestamp.
1249
1250 The default is set from `dmask' option. (If the directory is
1251 writable, utime(2) is also allowed. I.e. ~dmask & 022)
1252
1253 Normally utime(2) checks current process is owner of the file,
1254 or it has CAP_FOWNER capability. But FAT filesystem doesn't
1255 have uid/gid on disk, so normal check is too inflexible. With
1256 this option you can relax it.
1257
1258 check=value
1259 Three different levels of pickyness can be chosen:
1260
1261 r[elaxed]
1262 Upper and lower case are accepted and equivalent, long
1263 name parts are truncated (e.g. verylongname.foobar
1264 becomes verylong.foo), leading and embedded spaces are
1265 accepted in each name part (name and extension).
1266
1267 n[ormal]
1268 Like "relaxed", but many special characters (*, ?, <,
1269 spaces, etc.) are rejected. This is the default.
1270
1271 s[trict]
1272 Like "normal", but names may not contain long parts and
1273 special characters that are sometimes used on Linux, but
1274 are not accepted by MS-DOS are rejected. (+, =, spaces,
1275 etc.)
1276
1277 codepage=value
1278 Sets the codepage for converting to shortname characters on FAT
1279 and VFAT filesystems. By default, codepage 437 is used.
1280
1281 conv={b[inary]|t[ext]|a[uto]}
1282 The fat filesystem can perform CRLF<-->NL (MS-DOS text format to
1283 UNIX text format) conversion in the kernel. The following con‐
1284 version modes are available:
1285
1286 binary no translation is performed. This is the default.
1287
1288 text CRLF<-->NL translation is performed on all files.
1289
1290 auto CRLF<-->NL translation is performed on all files that
1291 don't have a "well-known binary" extension. The list of
1292 known extensions can be found at the beginning of
1293 fs/fat/misc.c (as of 2.0, the list is: exe, com, bin,
1294 app, sys, drv, ovl, ovr, obj, lib, dll, pif, arc, zip,
1295 lha, lzh, zoo, tar, z, arj, tz, taz, tzp, tpz, gz, tgz,
1296 deb, gif, bmp, tif, gl, jpg, pcx, tfm, vf, gf, pk, pxl,
1297 dvi).
1298
1299 Programs that do computed lseeks won't like in-kernel text con‐
1300 version. Several people have had their data ruined by this
1301 translation. Beware!
1302
1303 For filesystems mounted in binary mode, a conversion tool (from‐
1304 dos/todos) is available. This option is obsolete.
1305
1306 cvf_format=module
1307 Forces the driver to use the CVF (Compressed Volume File) module
1308 cvf_module instead of auto-detection. If the kernel supports
1309 kmod, the cvf_format=xxx option also controls on-demand CVF mod‐
1310 ule loading. This option is obsolete.
1311
1312 cvf_option=option
1313 Option passed to the CVF module. This option is obsolete.
1314
1315 debug Turn on the debug flag. A version string and a list of filesys‐
1316 tem parameters will be printed (these data are also printed if
1317 the parameters appear to be inconsistent).
1318
1319 discard
1320 If set, causes discard/TRIM commands to be issued to the block
1321 device when blocks are freed. This is useful for SSD devices and
1322 sparse/thinly-provisioned LUNs.
1323
1324 fat={12|16|32}
1325 Specify a 12, 16 or 32 bit fat. This overrides the automatic
1326 FAT type detection routine. Use with caution!
1327
1328 iocharset=value
1329 Character set to use for converting between 8 bit characters and
1330 16 bit Unicode characters. The default is iso8859-1. Long file‐
1331 names are stored on disk in Unicode format.
1332
1333 nfs If set, enables in-memory indexing of directory inodes to reduce
1334 the frequency of ESTALE errors in NFS client operations. Useful
1335 only when the filesystem is exported via NFS.
1336
1337 tz=UTC This option disables the conversion of timestamps between local
1338 time (as used by Windows on FAT) and UTC (which Linux uses
1339 internally). This is particularly useful when mounting devices
1340 (like digital cameras) that are set to UTC in order to avoid the
1341 pitfalls of local time.
1342
1343 quiet Turn on the quiet flag. Attempts to chown or chmod files do not
1344 return errors, although they fail. Use with caution!
1345
1346 showexec
1347 If set, the execute permission bits of the file will be allowed
1348 only if the extension part of the name is .EXE, .COM, or .BAT.
1349 Not set by default.
1350
1351 sys_immutable
1352 If set, ATTR_SYS attribute on FAT is handled as IMMUTABLE flag
1353 on Linux. Not set by default.
1354
1355 flush If set, the filesystem will try to flush to disk more early than
1356 normal. Not set by default.
1357
1358 usefree
1359 Use the "free clusters" value stored on FSINFO. It'll be used to
1360 determine number of free clusters without scanning disk. But
1361 it's not used by default, because recent Windows don't update it
1362 correctly in some case. If you are sure the "free clusters" on
1363 FSINFO is correct, by this option you can avoid scanning disk.
1364
1365 dots, nodots, dotsOK=[yes|no]
1366 Various misguided attempts to force Unix or DOS conventions onto
1367 a FAT filesystem.
1368
1369
1371 creator=cccc, type=cccc
1372 Set the creator/type values as shown by the MacOS finder used
1373 for creating new files. Default values: '????'.
1374
1375 uid=n, gid=n
1376 Set the owner and group of all files. (Default: the uid and gid
1377 of the current process.)
1378
1379 dir_umask=n, file_umask=n, umask=n
1380 Set the umask used for all directories, all regular files, or
1381 all files and directories. Defaults to the umask of the current
1382 process.
1383
1384 session=n
1385 Select the CDROM session to mount. Defaults to leaving that
1386 decision to the CDROM driver. This option will fail with any‐
1387 thing but a CDROM as underlying device.
1388
1389 part=n Select partition number n from the device. Only makes sense for
1390 CDROMs. Defaults to not parsing the partition table at all.
1391
1392 quiet Don't complain about invalid mount options.
1393
1394
1396 uid=value and gid=value
1397 Set the owner and group of all files. (Default: the uid and gid
1398 of the current process.)
1399
1400 umask=value
1401 Set the umask (the bitmask of the permissions that are not
1402 present). The default is the umask of the current process. The
1403 value is given in octal.
1404
1405 case={lower|asis}
1406 Convert all files names to lower case, or leave them. (Default:
1407 case=lower.)
1408
1409 conv={binary|text|auto}
1410 For conv=text, delete some random CRs (in particular, all fol‐
1411 lowed by NL) when reading a file. For conv=auto, choose more or
1412 less at random between conv=binary and conv=text. For
1413 conv=binary, just read what is in the file. This is the default.
1414
1415 nocheck
1416 Do not abort mounting when certain consistency checks fail.
1417
1418
1420 ISO 9660 is a standard describing a filesystem structure to be used on
1421 CD-ROMs. (This filesystem type is also seen on some DVDs. See also the
1422 udf filesystem.)
1423
1424 Normal iso9660 filenames appear in a 8.3 format (i.e., DOS-like
1425 restrictions on filename length), and in addition all characters are in
1426 upper case. Also there is no field for file ownership, protection,
1427 number of links, provision for block/character devices, etc.
1428
1429 Rock Ridge is an extension to iso9660 that provides all of these UNIX-
1430 like features. Basically there are extensions to each directory record
1431 that supply all of the additional information, and when Rock Ridge is
1432 in use, the filesystem is indistinguishable from a normal UNIX filesys‐
1433 tem (except that it is read-only, of course).
1434
1435 norock Disable the use of Rock Ridge extensions, even if available. Cf.
1436 map.
1437
1438 nojoliet
1439 Disable the use of Microsoft Joliet extensions, even if avail‐
1440 able. Cf. map.
1441
1442 check={r[elaxed]|s[trict]}
1443 With check=relaxed, a filename is first converted to lower case
1444 before doing the lookup. This is probably only meaningful
1445 together with norock and map=normal. (Default: check=strict.)
1446
1447 uid=value and gid=value
1448 Give all files in the filesystem the indicated user or group id,
1449 possibly overriding the information found in the Rock Ridge
1450 extensions. (Default: uid=0,gid=0.)
1451
1452 map={n[ormal]|o[ff]|a[corn]}
1453 For non-Rock Ridge volumes, normal name translation maps upper
1454 to lower case ASCII, drops a trailing `;1', and converts `;' to
1455 `.'. With map=off no name translation is done. See norock.
1456 (Default: map=normal.) map=acorn is like map=normal but also
1457 apply Acorn extensions if present.
1458
1459 mode=value
1460 For non-Rock Ridge volumes, give all files the indicated mode.
1461 (Default: read permission for everybody.) Since Linux 2.1.37
1462 one no longer needs to specify the mode in decimal. (Octal is
1463 indicated by a leading 0.)
1464
1465 unhide Also show hidden and associated files. (If the ordinary files
1466 and the associated or hidden files have the same filenames, this
1467 may make the ordinary files inaccessible.)
1468
1469 block={512|1024|2048}
1470 Set the block size to the indicated value. (Default:
1471 block=1024.)
1472
1473 conv={a[uto]|b[inary]|m[text]|t[ext]}
1474 (Default: conv=binary.) Since Linux 1.3.54 this option has no
1475 effect anymore. (And non-binary settings used to be very dan‐
1476 gerous, possibly leading to silent data corruption.)
1477
1478 cruft If the high byte of the file length contains other garbage, set
1479 this mount option to ignore the high order bits of the file
1480 length. This implies that a file cannot be larger than 16MB.
1481
1482 session=x
1483 Select number of session on multisession CD. (Since 2.3.4.)
1484
1485 sbsector=xxx
1486 Session begins from sector xxx. (Since 2.3.4.)
1487
1488 The following options are the same as for vfat and specifying them only
1489 makes sense when using discs encoded using Microsoft's Joliet exten‐
1490 sions.
1491
1492 iocharset=value
1493 Character set to use for converting 16 bit Unicode characters on
1494 CD to 8 bit characters. The default is iso8859-1.
1495
1496 utf8 Convert 16 bit Unicode characters on CD to UTF-8.
1497
1498
1500 iocharset=name
1501 Character set to use for converting from Unicode to ASCII. The
1502 default is to do no conversion. Use iocharset=utf8 for UTF8
1503 translations. This requires CONFIG_NLS_UTF8 to be set in the
1504 kernel .config file.
1505
1506 resize=value
1507 Resize the volume to value blocks. JFS only supports growing a
1508 volume, not shrinking it. This option is only valid during a
1509 remount, when the volume is mounted read-write. The resize key‐
1510 word with no value will grow the volume to the full size of the
1511 partition.
1512
1513 nointegrity
1514 Do not write to the journal. The primary use of this option is
1515 to allow for higher performance when restoring a volume from
1516 backup media. The integrity of the volume is not guaranteed if
1517 the system abnormally ends.
1518
1519 integrity
1520 Default. Commit metadata changes to the journal. Use this
1521 option to remount a volume where the nointegrity option was pre‐
1522 viously specified in order to restore normal behavior.
1523
1524 errors={continue|remount-ro|panic}
1525 Define the behaviour when an error is encountered. (Either
1526 ignore errors and just mark the filesystem erroneous and con‐
1527 tinue, or remount the filesystem read-only, or panic and halt
1528 the system.)
1529
1530 noquota|quota|usrquota|grpquota
1531 These options are accepted but ignored.
1532
1533
1535 None.
1536
1537
1539 See mount options for fat. If the msdos filesystem detects an incon‐
1540 sistency, it reports an error and sets the file system read-only. The
1541 filesystem can be made writable again by remounting it.
1542
1543
1545 Just like nfs, the ncpfs implementation expects a binary argument (a
1546 struct ncp_mount_data) to the mount system call. This argument is con‐
1547 structed by ncpmount(8) and the current version of mount (2.12) does
1548 not know anything about ncpfs.
1549
1550
1552 See the options section of the nfs(5) man page (nfs-utils package must
1553 be installed).
1554
1555 The nfs and nfs4 implementation expects a binary argument (a struct
1556 nfs_mount_data) to the mount system call. This argument is constructed
1557 by mount.nfs(8) and the current version of mount (2.13) does not know
1558 anything about nfs and nfs4.
1559
1560
1562 iocharset=name
1563 Character set to use when returning file names. Unlike VFAT,
1564 NTFS suppresses names that contain nonconvertible characters.
1565 Deprecated.
1566
1567 nls=name
1568 New name for the option earlier called iocharset.
1569
1570 utf8 Use UTF-8 for converting file names.
1571
1572 uni_xlate={0|1|2}
1573 For 0 (or `no' or `false'), do not use escape sequences for
1574 unknown Unicode characters. For 1 (or `yes' or `true') or 2,
1575 use vfat-style 4-byte escape sequences starting with ":". Here 2
1576 give a little-endian encoding and 1 a byteswapped bigendian
1577 encoding.
1578
1579 posix=[0|1]
1580 If enabled (posix=1), the filesystem distinguishes between upper
1581 and lower case. The 8.3 alias names are presented as hard links
1582 instead of being suppressed. This option is obsolete.
1583
1584 uid=value, gid=value and umask=value
1585 Set the file permission on the filesystem. The umask value is
1586 given in octal. By default, the files are owned by root and not
1587 readable by somebody else.
1588
1589
1591 uid=value and gid=value
1592 These options are recognized, but have no effect as far as I can
1593 see.
1594
1595
1597 Ramfs is a memory based filesystem. Mount it and you have it. Unmount
1598 it and it is gone. Present since Linux 2.3.99pre4. There are no mount
1599 options.
1600
1601
1603 Reiserfs is a journaling filesystem.
1604
1605 conv Instructs version 3.6 reiserfs software to mount a version 3.5
1606 filesystem, using the 3.6 format for newly created objects. This
1607 filesystem will no longer be compatible with reiserfs 3.5 tools.
1608
1609 hash={rupasov|tea|r5|detect}
1610 Choose which hash function reiserfs will use to find files
1611 within directories.
1612
1613 rupasov
1614 A hash invented by Yury Yu. Rupasov. It is fast and pre‐
1615 serves locality, mapping lexicographically close file
1616 names to close hash values. This option should not be
1617 used, as it causes a high probability of hash collisions.
1618
1619 tea A Davis-Meyer function implemented by Jeremy
1620 Fitzhardinge. It uses hash permuting bits in the name.
1621 It gets high randomness and, therefore, low probability
1622 of hash collisions at some CPU cost. This may be used if
1623 EHASHCOLLISION errors are experienced with the r5 hash.
1624
1625 r5 A modified version of the rupasov hash. It is used by
1626 default and is the best choice unless the filesystem has
1627 huge directories and unusual file-name patterns.
1628
1629 detect Instructs mount to detect which hash function is in use
1630 by examining the filesystem being mounted, and to write
1631 this information into the reiserfs superblock. This is
1632 only useful on the first mount of an old format filesys‐
1633 tem.
1634
1635 hashed_relocation
1636 Tunes the block allocator. This may provide performance improve‐
1637 ments in some situations.
1638
1639 no_unhashed_relocation
1640 Tunes the block allocator. This may provide performance improve‐
1641 ments in some situations.
1642
1643 noborder
1644 Disable the border allocator algorithm invented by Yury Yu.
1645 Rupasov. This may provide performance improvements in some sit‐
1646 uations.
1647
1648 nolog Disable journaling. This will provide slight performance
1649 improvements in some situations at the cost of losing reiserfs's
1650 fast recovery from crashes. Even with this option turned on,
1651 reiserfs still performs all journaling operations, save for
1652 actual writes into its journaling area. Implementation of nolog
1653 is a work in progress.
1654
1655 notail By default, reiserfs stores small files and `file tails'
1656 directly into its tree. This confuses some utilities such as
1657 LILO(8). This option is used to disable packing of files into
1658 the tree.
1659
1660 replayonly
1661 Replay the transactions which are in the journal, but do not
1662 actually mount the filesystem. Mainly used by reiserfsck.
1663
1664 resize=number
1665 A remount option which permits online expansion of reiserfs par‐
1666 titions. Instructs reiserfs to assume that the device has num‐
1667 ber blocks. This option is designed for use with devices which
1668 are under logical volume management (LVM). There is a special
1669 resizer utility which can be obtained from
1670 ftp://ftp.namesys.com/pub/reiserfsprogs.
1671
1672 user_xattr
1673 Enable Extended User Attributes. See the attr(5) manual page.
1674
1675 acl Enable POSIX Access Control Lists. See the acl(5) manual page.
1676
1677 barrier=none / barrier=flush
1678 This enables/disables the use of write barriers in the journal‐
1679 ing code. barrier=none disables it, barrier=flush enables it.
1680 Write barriers enforce proper on-disk ordering of journal com‐
1681 mits, making volatile disk write caches safe to use, at some
1682 performance penalty. The reiserfs filesystem does not enable
1683 write barriers by default. Be sure to enable barriers unless
1684 your disks are battery-backed one way or another. Otherwise you
1685 risk filesystem corruption in case of power failure.
1686
1687
1689 None.
1690
1691
1693 None.
1694
1695
1697 Just like nfs, the smbfs implementation expects a binary argument (a
1698 struct smb_mount_data) to the mount system call. This argument is con‐
1699 structed by smbmount(8) and the current version of mount (2.12) does
1700 not know anything about smbfs.
1701
1702
1704 None.
1705
1706
1708 size=nbytes
1709 Override default maximum size of the filesystem. The size is
1710 given in bytes, and rounded up to entire pages. The default is
1711 half of the memory. The size parameter also accepts a suffix %
1712 to limit this tmpfs instance to that percentage of your physical
1713 RAM: the default, when neither size nor nr_blocks is specified,
1714 is size=50%
1715
1716 nr_blocks=
1717 The same as size, but in blocks of PAGE_CACHE_SIZE
1718
1719 nr_inodes=
1720 The maximum number of inodes for this instance. The default is
1721 half of the number of your physical RAM pages, or (on a machine
1722 with highmem) the number of lowmem RAM pages, whichever is the
1723 lower.
1724
1725 The tmpfs mount options for sizing ( size, nr_blocks, and nr_inodes)
1726 accept a suffix k, m or g for Ki, Mi, Gi (binary kilo, mega and giga)
1727 and can be changed on remount.
1728
1729
1730 mode= Set initial permissions of the root directory.
1731
1732 uid= The user id.
1733
1734 gid= The group id.
1735
1736 mpol=[default|prefer:Node|bind:NodeList|interleave|interleave:NodeList]
1737 Set the NUMA memory allocation policy for all files in that
1738 instance (if the kernel CONFIG_NUMA is enabled) - which can be
1739 adjusted on the fly via 'mount -o remount ...'
1740
1741 default
1742 prefers to allocate memory from the local node
1743
1744 prefer:Node
1745 prefers to allocate memory from the given Node
1746
1747 bind:NodeList
1748 allocates memory only from nodes in NodeList
1749
1750 interleave
1751 prefers to allocate from each node in turn
1752
1753 interleave:NodeList
1754 allocates from each node of NodeList in turn.
1755
1756 The NodeList format is a comma-separated list of decimal numbers
1757 and ranges, a range being two hyphen-separated decimal numbers,
1758 the smallest and largest node numbers in the range. For exam‐
1759 ple, mpol=bind:0-3,5,7,9-15
1760
1761 Note that trying to mount a tmpfs with an mpol option will fail
1762 if the running kernel does not support NUMA; and will fail if
1763 its nodelist specifies a node which is not online. If your sys‐
1764 tem relies on that tmpfs being mounted, but from time to time
1765 runs a kernel built without NUMA capability (perhaps a safe
1766 recovery kernel), or with fewer nodes online, then it is advis‐
1767 able to omit the mpol option from automatic mount options. It
1768 can be added later, when the tmpfs is already mounted on Mount‐
1769 Point, by 'mount -o remount,mpol=Policy:NodeList MountPoint'.
1770
1771
1773 UBIFS is a flash file system which works on top of UBI volumes. Note
1774 that atime is not supported and is always turned off.
1775
1776 The device name may be specified as
1777 ubiX_Y UBI device number X, volume number Y
1778
1779 ubiY UBI device number 0, volume number Y
1780
1781 ubiX:NAME
1782 UBI device number X, volume with name NAME
1783
1784 ubi:NAME
1785 UBI device number 0, volume with name NAME
1786 Alternative ! separator may be used instead of :.
1787
1788 The following mount options are available:
1789
1790 bulk_read
1791 Enable bulk-read. VFS read-ahead is disabled because it slows
1792 down the file system. Bulk-Read is an internal optimization.
1793 Some flashes may read faster if the data are read at one go,
1794 rather than at several read requests. For example, OneNAND can
1795 do "read-while-load" if it reads more than one NAND page.
1796
1797 no_bulk_read
1798 Do not bulk-read. This is the default.
1799
1800 chk_data_crc
1801 Check data CRC-32 checksums. This is the default.
1802
1803 no_chk_data_crc.
1804 Do not check data CRC-32 checksums. With this option, the
1805 filesystem does not check CRC-32 checksum for data, but it does
1806 check it for the internal indexing information. This option only
1807 affects reading, not writing. CRC-32 is always calculated when
1808 writing the data.
1809
1810 compr={none|lzo|zlib}
1811 Select the default compressor which is used when new files are
1812 written. It is still possible to read compressed files if
1813 mounted with the none option.
1814
1815
1817 udf is the "Universal Disk Format" filesystem defined by the Optical
1818 Storage Technology Association, and is often used for DVD-ROM. See
1819 also iso9660.
1820
1821 gid= Set the default group.
1822
1823 umask= Set the default umask. The value is given in octal.
1824
1825 uid= Set the default user.
1826
1827 unhide Show otherwise hidden files.
1828
1829 undelete
1830 Show deleted files in lists.
1831
1832 nostrict
1833 Unset strict conformance.
1834
1835 iocharset
1836 Set the NLS character set.
1837
1838 bs= Set the block size. (May not work unless 2048.)
1839
1840 novrs Skip volume sequence recognition.
1841
1842 session=
1843 Set the CDROM session counting from 0. Default: last session.
1844
1845 anchor=
1846 Override standard anchor location. Default: 256.
1847
1848 volume=
1849 Override the VolumeDesc location. (unused)
1850
1851 partition=
1852 Override the PartitionDesc location. (unused)
1853
1854 lastblock=
1855 Set the last block of the filesystem.
1856
1857 fileset=
1858 Override the fileset block location. (unused)
1859
1860 rootdir=
1861 Override the root directory location. (unused)
1862
1863
1865 ufstype=value
1866 UFS is a filesystem widely used in different operating systems.
1867 The problem are differences among implementations. Features of
1868 some implementations are undocumented, so its hard to recognize
1869 the type of ufs automatically. That's why the user must specify
1870 the type of ufs by mount option. Possible values are:
1871
1872 old Old format of ufs, this is the default, read only.
1873 (Don't forget to give the -r option.)
1874
1875 44bsd For filesystems created by a BSD-like system (Net‐
1876 BSD,FreeBSD,OpenBSD).
1877
1878 ufs2 Used in FreeBSD 5.x supported as read-write.
1879
1880 5xbsd Synonym for ufs2.
1881
1882 sun For filesystems created by SunOS or Solaris on Sparc.
1883
1884 sunx86 For filesystems created by Solaris on x86.
1885
1886 hp For filesystems created by HP-UX, read-only.
1887
1888 nextstep
1889 For filesystems created by NeXTStep (on NeXT station)
1890 (currently read only).
1891
1892 nextstep-cd
1893 For NextStep CDROMs (block_size == 2048), read-only.
1894
1895 openstep
1896 For filesystems created by OpenStep (currently read
1897 only). The same filesystem type is also used by Mac OS
1898 X.
1899
1900
1901 onerror=value
1902 Set behaviour on error:
1903
1904 panic If an error is encountered, cause a kernel panic.
1905
1906 [lock|umount|repair]
1907 These mount options don't do anything at present; when an
1908 error is encountered only a console message is printed.
1909
1910
1912 See mount options for msdos. The dotsOK option is explicitly killed by
1913 umsdos.
1914
1915
1917 First of all, the mount options for fat are recognized. The dotsOK
1918 option is explicitly killed by vfat. Furthermore, there are
1919
1920 uni_xlate
1921 Translate unhandled Unicode characters to special escaped
1922 sequences. This lets you backup and restore filenames that are
1923 created with any Unicode characters. Without this option, a '?'
1924 is used when no translation is possible. The escape character is
1925 ':' because it is otherwise illegal on the vfat filesystem. The
1926 escape sequence that gets used, where u is the unicode charac‐
1927 ter, is: ':', (u & 0x3f), ((u>>6) & 0x3f), (u>>12).
1928
1929 posix Allow two files with names that only differ in case. This
1930 option is obsolete.
1931
1932 nonumtail
1933 First try to make a short name without sequence number, before
1934 trying name~num.ext.
1935
1936 utf8 UTF8 is the filesystem safe 8-bit encoding of Unicode that is
1937 used by the console. It can be enabled for the filesystem with
1938 this option or disabled with utf8=0, utf8=no or utf8=false. If
1939 `uni_xlate' gets set, UTF8 gets disabled.
1940
1941 shortname={lower|win95|winnt|mixed}
1942
1943 Defines the behaviour for creation and display of filenames
1944 which fit into 8.3 characters. If a long name for a file exists,
1945 it will always be preferred display. There are four modes: :
1946
1947 lower Force the short name to lower case upon display; store a
1948 long name when the short name is not all upper case.
1949
1950 win95 Force the short name to upper case upon display; store a
1951 long name when the short name is not all upper case.
1952
1953 winnt Display the shortname as is; store a long name when the
1954 short name is not all lower case or all upper case.
1955
1956 mixed Display the short name as is; store a long name when the
1957 short name is not all upper case. This mode is the
1958 default since Linux 2.6.32.
1959
1960
1961
1963 devuid=uid and devgid=gid and devmode=mode
1964 Set the owner and group and mode of the device files in the
1965 usbfs filesystem (default: uid=gid=0, mode=0644). The mode is
1966 given in octal.
1967
1968 busuid=uid and busgid=gid and busmode=mode
1969 Set the owner and group and mode of the bus directories in the
1970 usbfs filesystem (default: uid=gid=0, mode=0555). The mode is
1971 given in octal.
1972
1973 listuid=uid and listgid=gid and listmode=mode
1974 Set the owner and group and mode of the file devices (default:
1975 uid=gid=0, mode=0444). The mode is given in octal.
1976
1977
1979 None.
1980
1981
1983 allocsize=size
1984 Sets the buffered I/O end-of-file preallocation size when doing
1985 delayed allocation writeout. Valid values for this option are
1986 page size (typically 4KiB) through to 1GiB, inclusive, in power-
1987 of-2 increments.
1988
1989 The default behaviour is for dynamic end-of-file preallocation
1990 size, which uses a set of heuristics to optimise the prealloca‐
1991 tion size based on the current allocation patterns within the
1992 file and the access patterns to the file. Specifying a fixed
1993 allocsize value turns off the dynamic behaviour.
1994
1995 attr2|noattr2
1996 The options enable/disable an "opportunistic" improvement to be
1997 made in the way inline extended attributes are stored on-disk.
1998 When the new form is used for the first time when attr2 is
1999 selected (either when setting or removing extended attributes)
2000 the on-disk superblock feature bit field will be updated to
2001 reflect this format being in use.
2002
2003 The default behaviour is determined by the on-disk feature bit
2004 indicating that attr2 behaviour is active. If either mount
2005 option it set, then that becomes the new default used by the
2006 filesystem.
2007
2008 CRC enabled filesystems always use the attr2 format, and so will
2009 reject the noattr2 mount option if it is set.
2010
2011 barrier|nobarrier
2012 Enables/disables the use of block layer write barriers for
2013 writes into the journal and for data integrity operations. This
2014 allows for drive level write caching to be enabled, for devices
2015 that support write barriers.
2016
2017 discard|nodiscard
2018 Enable/disable the issuing of commands to let the block device
2019 reclaim space freed by the filesystem. This is useful for SSD
2020 devices, thinly provisioned LUNs and virtual machine images, but
2021 may have a performance impact.
2022
2023 Note: It is currently recommended that you use the fstrim appli‐
2024 cation to discard unused blocks rather than the discard mount
2025 option because the performance impact of this option is quite
2026 severe.
2027
2028 grpid|bsdgroups|nogrpid|sysvgroups
2029 These options define what group ID a newly created file gets.
2030 When grpid is set, it takes the group ID of the directory in
2031 which it is created; otherwise it takes the fsgid of the current
2032 process, unless the directory has the setgid bit set, in which
2033 case it takes the gid from the parent directory, and also gets
2034 the setgid bit set if it is a directory itself.
2035
2036 filestreams
2037 Make the data allocator use the filestreams allocation mode
2038 across the entire filesystem rather than just on directories
2039 configured to use it.
2040
2041 When ikeep is specified, XFS does not delete empty inode
2042 clusters and keeps them around on disk. When noikeep is speci‐
2043 fied, empty inode clusters are returned to the free space pool.
2044
2045 inode32|inode64
2046 When inode32 is specified, it indicates that XFS limits inode
2047 creation to locations which will not result in inode numbers
2048 with more than 32 bits of significance.
2049
2050 When inode64 is specified, it indicates that XFS is allowed to
2051 create inodes at any location in the filesystem, including those
2052 which will result in inode numbers occupying more than 32 bits
2053 of significance.
2054
2055 inode32 is provided for backwards compatibility with older sys‐
2056 tems and applications, since 64 bits inode numbers might cause
2057 problems for some applications that cannot handle large inode
2058 numbers. If applications are in use which do not handle inode
2059 numbers bigger than 32 bits, the inode32 option should be speci‐
2060 fied.
2061
2062 largeio|nolargeio
2063 If "nolargeio" is specified, the optimal I/O reported in st_blk‐
2064 size by stat(2) will be as small as possible to allow user
2065 applications to avoid inefficient read/modify/write I/O. This
2066 is typically the page size of the machine, as this is the granu‐
2067 larity of the page cache.
2068
2069 If "largeio" specified, a filesystem that was created with a
2070 "swidth" specified will return the "swidth" value (in bytes) in
2071 st_blksize. If the filesystem does not have a "swidth" specified
2072 but does specify an "allocsize" then "allocsize" (in bytes) will
2073 be returned instead. Otherwise the behaviour is the same as if
2074 "nolargeio" was specified.
2075
2076 logbufs=value
2077 Set the number of in-memory log buffers. Valid numbers range
2078 from 2-8 inclusive.
2079
2080 The default value is 8 buffers.
2081
2082 If the memory cost of 8 log buffers is too high on small sys‐
2083 tems, then it may be reduced at some cost to performance on
2084 metadata intensive workloads. The logbsize option below controls
2085 the size of each buffer and so is also relevent to this case.
2086
2087 logbsize=value
2088 Set the size of each in-memory log buffer. The size may be
2089 specified in bytes, or in kilobytes with a "k" suffix. Valid
2090 sizes for version 1 and version 2 logs are 16384 (16k) and 32768
2091 (32k). Valid sizes for version 2 logs also include 65536 (64k),
2092 131072 (128k) and 262144 (256k). The logbsize must be an integer
2093 multiple of the log stripe unit configured at mkfs time.
2094
2095 The default value for version 1 logs is 32768, while the default
2096 value for version 2 logs is MAX(32768, log_sunit).
2097
2098 logdev=deviceandrtdev=device
2099 Use an external log (metadata journal) and/or real-time device.
2100 An XFS filesystem has up to three parts: a data section, a log
2101 section, and a real-time section. The real-time section is
2102 optional, and the log section can be separate from the data sec‐
2103 tion or contained within it.
2104
2105 noalign
2106 Data allocations will not be aligned at stripe unit boundaries.
2107 This is only relevant to filesystems created with non-zero data
2108 alignment parameters (sunit, swidth) by mkfs.
2109
2110 norecovery
2111 The filesystem will be mounted without running log recovery. If
2112 the filesystem was not cleanly unmounted, it is likely to be
2113 inconsistent when mounted in "norecovery" mode. Some files or
2114 directories may not be accessible because of this. Filesystems
2115 mounted "norecovery" must be mounted read-only or the mount will
2116 fail.
2117
2118 nouuid Don't check for double mounted file systems using the file sys‐
2119 tem uuid. This is useful to mount LVM snapshot volumes, and
2120 often used in combination with "norecovery" for mounting read-
2121 only snapshots.
2122
2123 noquota
2124 Forcibly turns off all quota accounting and enforcement within
2125 the filesystem.
2126
2127 uquota/usrquota/uqnoenforce/quota
2128 User disk quota accounting enabled, and limits (optionally)
2129 enforced. Refer to xfs_quota(8) for further details.
2130
2131 gquota/grpquota/gqnoenforce
2132 Group disk quota accounting enabled and limits (optionally)
2133 enforced. Refer to xfs_quota(8) for further details.
2134
2135 pquota/prjquota/pqnoenforce
2136 Project disk quota accounting enabled and limits (optionally)
2137 enforced. Refer to xfs_quota(8) for further details.
2138
2139 sunit=value and swidth=value
2140 Used to specify the stripe unit and width for a RAID device or a
2141 stripe volume. "value" must be specified in 512-byte block
2142 units. These options are only relevant to filesystems that were
2143 created with non-zero data alignment parameters.
2144
2145 The sunit and swidth parameters specified must be compatible
2146 with the existing filesystem alignment characteristics. In gen‐
2147 eral, that means the only valid changes to sunit are increasing
2148 it by a power-of-2 multiple. Valid swidth values are any integer
2149 multiple of a valid sunit value.
2150
2151 Typically the only time these mount options are necessary if
2152 after an underlying RAID device has had it's geometry modified,
2153 such as adding a new disk to a RAID5 lun and reshaping it.
2154
2155 swalloc
2156 Data allocations will be rounded up to stripe width boundaries
2157 when the current end of file is being extended and the file size
2158 is larger than the stripe width size.
2159
2160 wsync When specified, all filesystem namespace operations are executed
2161 synchronously. This ensures that when the namespace operation
2162 (create, unlink, etc) completes, the change to the namespace is
2163 on stable storage. This is useful in HA setups where failover
2164 must not result in clients seeing inconsistent namespace presen‐
2165 tation during or after a failover event.
2166
2167
2169 None. Although nothing is wrong with xiafs, it is not used much, and is
2170 not maintained. Probably one shouldn't use it. Since Linux version
2171 2.1.21 xiafs is no longer part of the kernel source.
2172
2173
2175 One further possible type is a mount via the loop device. For example,
2176 the command
2177
2178 mount /tmp/disk.img /mnt -t vfat -o loop=/dev/loop
2179
2180 will set up the loop device /dev/loop3 to correspond to the file
2181 /tmp/disk.img, and then mount this device on /mnt.
2182
2183 If no explicit loop device is mentioned (but just an option `-o loop'
2184 is given), then mount will try to find some unused loop device and use
2185 that, for example
2186
2187 mount /tmp/disk.img /mnt -o loop
2188
2189 The mount command automatically creates a loop device from a regular
2190 file if a filesystem type is not specified or the filesystem is known
2191 for libblkid, for example:
2192
2193 mount /tmp/disk.img /mnt
2194
2195 mount -t ext4 /tmp/disk.img /mnt
2196
2197 This type of mount knows about four options, namely loop, offset and
2198 sizelimit , that are really options to losetup(8). (These options can
2199 be used in addition to those specific to the filesystem type.)
2200
2201 Since Linux 2.6.25 is supported auto-destruction of loop devices and
2202 then any loop device allocated by mount will be freed by umount inde‐
2203 pendently on /etc/mtab.
2204
2205 You can also free a loop device by hand, using `losetup -d' or `umount
2206 -d`.
2207
2208
2210 mount has the following return codes (the bits can be ORed):
2211
2212 0 success
2213
2214 1 incorrect invocation or permissions
2215
2216 2 system error (out of memory, cannot fork, no more loop devices)
2217
2218 4 internal mount bug
2219
2220 8 user interrupt
2221
2222 16 problems writing or locking /etc/mtab
2223
2224 32 mount failure
2225
2226 64 some mount succeeded
2227
2228 The command mount -a returns 0 (all success), 32 (all failed) or 64
2229 (some failed, some success).
2230
2231
2233 The syntax of external mount helpers is:
2234
2235 /sbin/mount.<suffix> spec dir [-sfnv] [-o options] [-t type.sub‐
2236 type]
2237
2238 where the <type> is filesystem type and -sfnvo options have same mean‐
2239 ing like standard mount options. The -t option is used for filesystems
2240 with subtypes support (for example /sbin/mount.fuse -t fuse.sshfs).
2241
2242
2244 /etc/fstab filesystem table
2245
2246 /etc/mtab table of mounted filesystems
2247
2248 /etc/mtab~ lock file
2249
2250 /etc/mtab.tmp temporary file
2251
2252 /etc/filesystems a list of filesystem types to try
2253
2255 LIBMOUNT_FSTAB=<path>
2256 overrides the default location of the fstab file
2257
2258 LIBMOUNT_MTAB=<path>
2259 overrides the default location of the mtab file
2260
2261 LIBMOUNT_DEBUG=0xffff
2262 enables debug output
2263
2265 mount(2), umount(2), fstab(5), umount(8), swapon(8), findmnt(8),
2266 nfs(5), xfs(5), e2label(8), xfs_admin(8), mountd(8), nfsd(8),
2267 mke2fs(8), tune2fs(8), losetup(8)
2268
2270 It is possible for a corrupted filesystem to cause a crash.
2271
2272 Some Linux filesystems don't support -o sync and -o dirsync (the ext2,
2273 ext3, ext4, fat and vfat filesystems do support synchronous updates (a
2274 la BSD) when mounted with the sync option).
2275
2276 The -o remount may not be able to change mount parameters (all ext2fs-
2277 specific parameters, except sb, are changeable with a remount, for
2278 example, but you can't change gid or umask for the fatfs).
2279
2280 It is possible that files /etc/mtab and /proc/mounts don't match. The
2281 first file is based only on the mount command options, but the content
2282 of the second file also depends on the kernel and others settings (e.g.
2283 remote NFS server. In particular case the mount command may reports
2284 unreliable information about a NFS mount point and the /proc/mounts
2285 file usually contains more reliable information.)
2286
2287 Checking files on NFS filesystem referenced by file descriptors (i.e.
2288 the fcntl and ioctl families of functions) may lead to inconsistent
2289 result due to the lack of consistency check in kernel even if noac is
2290 used.
2291
2292 The loop option with the offset or sizelimit options used may fail when
2293 using older kernels if the mount command can't confirm that the size of
2294 the block device has been configured as requested. This situation can
2295 be worked around by using the losetup command manually before calling
2296 mount with the configured loop device.
2297
2299 A mount command existed in Version 5 AT&T UNIX.
2300
2302 Karel Zak <kzak@redhat.com>
2303
2305 The mount command is part of the util-linux package and is available
2306 from ftp://ftp.kernel.org/pub/linux/utils/util-linux/.
2307
2308
2309
2310
2311util-linux January 2012 MOUNT(8)