2 .\" Copyright 1993, 1994, 1995 by Theodore Ts'o. All Rights Reserved.
3 .\" This file may be copied under the terms of the GNU Public License.
5 .TH MKE2FS 8 "@E2FSPROGS_MONTH@ @E2FSPROGS_YEAR@" "E2fsprogs version @E2FSPROGS_VERSION@"
7 mke2fs \- create an ext2/ext3/ext4 filesystem
63 .I reserved-blocks-percentage
71 [^]\fIfeature\fR[,...]
96 .I last-mounted-directory
129 @JDEV@.B "mke2fs \-O journal_dev"
134 .\" No external-journal specific journal options yet (size is ignored)
137 .\" @JDEV@.I journal-options
141 @JDEV@.I volume-label
152 @JDEV@.I external-journal
158 is used to create an ext2, ext3, or ext4 filesystem, usually in a disk
159 partition (or file) named by
162 The file system size is specified by
166 does not have a suffix, it is interpreted as power-of-two kilobytes,
170 option is specified, in which case
172 is interpreted as the number of
174 blocks. If the fs-size is suffixed by 'k', 'm', 'g', 't'
175 (either upper-case or lower-case), then it is interpreted in
176 power-of-two kilobytes, megabytes, gigabytes, terabytes, etc.
181 will create the file system based on the device size.
197 will create a file system for use with ext3,
199 will create a file system for use with ext4, and so on.
201 The defaults of the parameters for the newly created filesystem, if not
202 overridden by the options listed below, are controlled by the
204 configuration file. See the
206 manual page for more details.
209 .BI \-b " block-size"
210 Specify the size of blocks in bytes. Valid block-size values are 1024,
211 2048 and 4096 bytes per block. If omitted,
212 block-size is heuristically determined by the filesystem size and
213 the expected usage of the filesystem (see the
217 is preceded by a negative sign ('-'), then
219 will use heuristics to determine the
220 appropriate block size, with the constraint that the block size will be
223 bytes. This is useful for certain hardware devices which require that
224 the blocksize be a multiple of 2k.
227 Check the device for bad blocks before creating the file system. If
228 this option is specified twice, then a slower read-write
229 test is used instead of a fast read-only test.
231 .B \-C " cluster-size"
232 Specify the size of cluster in bytes for filesystems using the bigalloc
233 feature. Valid cluster-size values are from 2048 to 256M bytes per
234 cluster. This can only be specified if the bigalloc feature is
237 man page for more details about bigalloc.) The default cluster size if
238 bigalloc is enabled is 16 times the block size.
240 .BI \-d " root-directory"
241 Copy the contents of the given directory into the root directory of the
245 Use direct I/O when writing to the disk. This avoids mke2fs dirtying a
246 lot of buffer cache memory, which may impact other applications running
247 on a busy server. This option will cause mke2fs to run much more
248 slowly, however, so there is a tradeoff to using direct I/O.
250 .BI \-e " error-behavior"
251 Change the behavior of the kernel code when errors are detected.
252 In all cases, a filesystem error will cause
254 to check the filesystem on the next boot.
256 can be one of the following:
260 Continue normal execution.
263 Remount filesystem read-only.
266 Cause a kernel panic.
269 .BI \-E " extended-options"
270 Set extended options for the filesystem. Extended options are comma
271 separated, and may take an argument using the equals ('=') sign. The
275 in earlier versions of
279 option is still accepted for backwards compatibility, but is deprecated.
280 The following extended options are supported:
283 .BI encoding= encoding-name
286 feature in the super block and set
288 as the encoding to be used. If
290 is not specified, the encoding defined in
294 .BI encoding_flags= encoding-flags
295 Define parameters for file name character encoding operations. If a
296 flag is not changed using this parameter, its default value is used.
298 should be a comma-separated lists of flags to be enabled. To disable a
299 flag, add it to the list with the prefix "no".
301 The only flag that can be set right now is
303 which means that invalid strings should be rejected by the file system.
304 In the default configuration, the
308 .BI mmp_update_interval= interval
309 Adjust the initial MMP update interval to
311 seconds. Specifying an
313 of 0 means to use the default interval. The specified interval must
314 be less than 300 seconds. Requires that the
318 .BI stride= stride-size
319 Configure the filesystem for a RAID array with
321 filesystem blocks. This is the number of blocks read or written to disk
322 before moving to the next disk, which is sometimes referred to as the
324 This mostly affects placement of filesystem metadata like bitmaps at
326 time to avoid placing them on a single disk, which can hurt performance.
327 It may also be used by the block allocator.
329 .BI stripe_width= stripe-width
330 Configure the filesystem for a RAID array with
332 filesystem blocks per stripe. This is typically stride-size * N, where
333 N is the number of data-bearing disks in the RAID (e.g. for RAID 5 there is one
334 parity disk, so N will be the number of disks in the array minus 1).
335 This allows the block allocator to prevent read-modify-write of the
336 parity in a RAID stripe if possible when the data is written.
339 Create the filesystem at an offset from the beginning of the device or
340 file. This can be useful when creating disk images for virtual machines.
342 .BI resize= max-online-resize
343 Reserve enough space so that the block group descriptor table can grow
344 to support a filesystem that has
348 .B lazy_itable_init\fR[\fB= \fI<0 to disable, 1 to enable>\fR]
349 If enabled and the uninit_bg feature is enabled, the inode table will
350 not be fully initialized by
352 This speeds up filesystem
353 initialization noticeably, but it requires the kernel to finish
354 initializing the filesystem in the background when the filesystem is
355 first mounted. If the option value is omitted, it defaults to 1 to
356 enable lazy inode table zeroing.
358 .B lazy_journal_init\fR[\fB= \fI<0 to disable, 1 to enable>\fR]
359 If enabled, the journal inode will not be fully zeroed out by
361 This speeds up filesystem initialization noticeably, but carries some
362 small risk if the system crashes before the journal has been overwritten
363 entirely one time. If the option value is omitted, it defaults to 1 to
364 enable lazy journal inode zeroing.
369 will copy the extended attributes of the files in the directory
370 hierarchy specified via the (optional)
372 option. This will disable the copy and leaves the files in the newly
373 created file system without any extended attributes.
375 .BI num_backup_sb= <0|1|2>
378 file system feature is enabled this option controls whether there will
379 be 0, 1, or 2 backup superblocks created in the file system.
381 .B packed_meta_blocks\fR[\fB= \fI<0 to disable, 1 to enable>\fR]
382 Place the allocation bitmaps and the inode table at the beginning of the
383 disk. This option requires that the flex_bg file system feature to be
384 enabled in order for it to have effect, and will also create the journal
385 at the beginning of the file system. This option is useful for flash
386 devices that use SLC flash at the beginning of the disk.
387 It also maximizes the range of contiguous data blocks, which
388 can be useful for certain specialized use cases, such as supported
391 .BI root_owner [=uid:gid]
392 Specify the numeric user and group ID of the root directory. If no UID:GID
393 is specified, use the user and group ID of the user running \fBmke2fs\fR.
394 In \fBmke2fs\fR 1.42 and earlier the UID and GID of the root directory were
395 set by default to the UID and GID of the user running the mke2fs command.
396 The \fBroot_owner=\fR option allows explicitly specifying these values,
397 and avoid side-effects for users that do not expect the contents of the
398 filesystem to change based on the user running \fBmke2fs\fR.
401 Set a flag in the filesystem superblock indicating that it may be
402 mounted using experimental kernel code, such as the ext4dev filesystem.
405 Attempt to discard blocks at mkfs time (discarding blocks initially is useful
406 on solid state devices and sparse / thin-provisioned storage). When the device
407 advertises that discard also zeroes data (any subsequent read after the discard
408 and before write returns zero), then mark all not-yet-zeroed inode tables as
409 zeroed. This significantly speeds up filesystem initialization. This is set
413 Do not attempt to discard blocks at mkfs time.
416 Specify the which quota types (usrquota, grpquota, prjquota) which
417 should be enabled in the created file system. The argument of this
418 extended option should be a colon separated list. This option has
421 feature is set. The default quota types to be initialized if this
422 option is not specified is both user and group quotas. If the project
423 feature is enabled that project quotas will be initialized as well.
429 to create a filesystem, even if the specified device is not a partition
430 on a block special device, or if other parameters do not make sense.
433 to create a filesystem even if the filesystem appears to be in use
434 or is mounted (a truly dangerous thing to do), this option must be
437 .BI \-g " blocks-per-group"
438 Specify the number of blocks in a block group. There is generally no
439 reason for the user to ever set this parameter, as the default is optimal
440 for the filesystem. (For administrators who are creating
441 filesystems on RAID arrays, it is preferable to use the
443 RAID parameter as part of the
445 option rather than manipulating the number of blocks per group.)
446 This option is generally used by developers who
447 are developing test cases.
449 If the bigalloc feature is enabled, the
451 option will specify the number of clusters in a block group.
453 .BI \-G " number-of-groups"
454 Specify the number of block groups that will be packed together to
455 create a larger virtual block group (or "flex_bg group") in an
456 ext4 filesystem. This improves meta-data locality and performance
457 on meta-data heavy workloads. The number of groups must be a power
458 of 2 and may only be specified if the
460 filesystem feature is enabled.
462 .BI \-i " bytes-per-inode"
463 Specify the bytes/inode ratio.
465 creates an inode for every
467 bytes of space on the disk. The larger the
469 ratio, the fewer inodes will be created. This value generally shouldn't
470 be smaller than the blocksize of the filesystem, since in that case more
471 inodes would be made than can ever be used. Be warned that it is not
472 possible to change this ratio on a filesystem after it is created, so be
473 careful deciding the correct value for this parameter. Note that resizing
474 a filesystem changes the number of inodes to maintain this ratio.
476 .BI \-I " inode-size"
477 Specify the size of each inode in bytes.
480 value must be a power of 2 larger or equal to 128. The larger the
482 the more space the inode table will consume, and this reduces the usable
483 space in the filesystem and can also negatively impact performance.
485 possible to change this value after the filesystem is created.
487 In kernels after 2.6.10 and some
488 earlier vendor kernels it is possible to utilize inodes larger than
490 extended attributes for improved performance.
492 stored in large inodes are not visible with older kernels, and such
493 filesystems will not be mountable with 2.4 kernels at all.
495 The default inode size is controlled by the
499 file shipped with e2fsprogs, the default inode size is 256 bytes for
500 most file systems, except for small file systems where the inode size
504 Create the filesystem with an ext3 journal. If the
506 option is not specified, the default journal parameters will be used to
507 create an appropriately sized journal (given the size of the filesystem)
508 stored within the filesystem. Note that you must be using a kernel
509 which has ext3 support in order to actually make use of the journal.
511 .BI \-J " journal-options"
512 Create the ext3 journal using options specified on the command-line.
513 Journal options are comma
514 separated, and may take an argument using the equals ('=') sign.
515 The following journal options are supported:
518 .BI size= journal-size
519 Create an internal journal (i.e., stored inside the filesystem) of size
522 The size of the journal must be at least 1024 filesystem blocks
523 (i.e., 1MB if using 1k blocks, 4MB if using 4k blocks, etc.)
524 and may be no more than 10,240,000 filesystem blocks or half the total
525 file system size (whichever is smaller)
527 .BI location =journal-location
528 Specify the location of the journal. The argument
530 can either be specified as a block number, or if the number has a units
531 suffix (e.g., 'M', 'G', etc.) interpret it as the offset from the
532 beginning of the file system.
534 @JDEV@.BI device= external-journal
535 @JDEV@Attach the filesystem to the journal block device located on
536 @JDEV@.IR external-journal .
538 @JDEV@journal must already have been created using the command
540 @JDEV@.B mke2fs -O journal_dev
541 @JDEV@.I external-journal
544 @JDEV@.I external-journal
545 @JDEV@must have been created with the
546 @JDEV@same block size as the new filesystem.
547 @JDEV@In addition, while there is support for attaching
548 @JDEV@multiple filesystems to a single external journal,
549 @JDEV@the Linux kernel and
551 @JDEV@do not currently support shared external journals yet.
553 @JDEV@Instead of specifying a device name directly,
554 @JDEV@.I external-journal
555 @JDEV@can also be specified by either
556 @JDEV@.BI LABEL= label
559 @JDEV@to locate the external journal by either the volume label or UUID
560 @JDEV@stored in the ext2 superblock at the start of the journal. Use
561 @JDEV@.BR dumpe2fs (8)
562 @JDEV@to display a journal device's volume label and UUID. See also the
565 @JDEV@.BR tune2fs (8).
568 @JDEV@Only one of the
569 @JDEV@.BR size " or " device
570 @JDEV@options can be given for a filesystem.
573 Read the bad blocks list from
575 Note that the block numbers in the bad block list must be generated
576 using the same block size as used by
582 is a much simpler and less error-prone method of checking a disk for bad
583 blocks before formatting it, as
585 will automatically pass the correct parameters to the
589 .BI \-L " new-volume-label"
590 Set the volume label for the filesystem to
591 .IR new-volume-label .
592 The maximum length of the
593 volume label is 16 bytes.
595 .BI \-m " reserved-blocks-percentage"
596 Specify the percentage of the filesystem blocks reserved for
597 the super-user. This avoids fragmentation, and allows root-owned
600 to continue to function correctly after non-privileged processes are
601 prevented from writing to the filesystem. The default percentage
604 .BI \-M " last-mounted-directory"
605 Set the last mounted directory for the filesystem. This might be useful
606 for the sake of utilities that key off of the last mounted directory to
607 determine where the filesystem should be mounted.
612 to not actually create a filesystem, but display what it
613 would do if it were to create a filesystem. This can be used to
614 determine the location of the backup superblocks for a particular
615 filesystem, so long as the
617 parameters that were passed when the
618 filesystem was originally created are used again. (With the
620 option added, of course!)
622 .BI \-N " number-of-inodes"
623 Overrides the default calculation of the number of inodes that should be
624 reserved for the filesystem (which is based on the number of blocks and
627 ratio). This allows the user to specify the number
628 of desired inodes directly.
630 .BI \-o " creator-os"
631 Overrides the default value of the "creator operating system" field of the
632 filesystem. The creator field is set by default to the name of the OS the
634 executable was compiled for.
636 .B "\-O \fR[^]\fIfeature\fR[,...]"
637 Create a filesystem with the given features (filesystem options),
638 overriding the default filesystem options. The features that are
639 enabled by default are specified by the
641 relation, either in the
648 subsections for the usage types as specified by the
650 option, further modified by the
652 relation found in the
654 subsections for the filesystem and usage types. See the
656 manual page for more details.
657 The filesystem type-specific configuration setting found in the
659 section will override the global default found in
662 The filesystem feature set will be further edited
663 using either the feature set specified by this option,
664 or if this option is not given, by the
666 relation for the filesystem type being created, or in the
668 section of the configuration file.
670 The filesystem feature set is comprised of a list of features, separated
671 by commas, that are to be enabled. To disable a feature, simply
672 prefix the feature name with a caret ('^') character.
673 Features with dependencies will not be removed successfully.
674 The pseudo-filesystem feature "none" will clear all filesystem features.
676 For more information about the features which can be set, please see
681 Quiet execution. Useful if
686 Set the filesystem revision for the new filesystem. Note that 1.2
687 kernels only support revision 0 filesystems. The default is to
688 create revision 1 filesystems.
691 Write superblock and group descriptors only. This is an extreme
692 measure to be taken only in the very unlikely case that all of
693 the superblock and backup superblocks are corrupted, and a last-ditch
694 recovery method is desired by experienced users. It causes
696 to reinitialize the superblock and group descriptors, while not
697 touching the inode table and the block and inode bitmaps. The
699 program should be run immediately after this option is used, and there
700 is no guarantee that any data will be salvageable. Due to the wide
701 variety of possible options to
703 that affect the on-disk layout, it is critical to specify exactly
704 the same format options, such as blocksize, fs-type, feature flags, and
705 other tunables when using this option, or the filesystem will be further
706 corrupted. In some cases, such as filesystems that have been resized,
707 or have had features enabled after format time, it is impossible to
708 overwrite all of the superblocks correctly, and at least some filesystem
709 corruption will occur. It is best to run this on a full copy of the
710 filesystem so other options can be tried if this doesn't work.
713 .\" Check the device for bad blocks before creating the file system
714 .\" using the specified test.
717 Specify the filesystem type (i.e., ext2, ext3, ext4, etc.) that is
719 If this option is not specified,
721 will pick a default either via how
722 the command was run (for example, using a name of the form mkfs.ext2,
723 mkfs.ext3, etc.) or via a default as defined by the
725 file. This option controls which filesystem options are used by
726 default, based on the
728 configuration stanza in
729 .BR /etc/mke2fs.conf .
733 option is used to explicitly add or remove filesystem options that
734 should be set in the newly created filesystem, the
735 resulting filesystem may not be supported by the requested
737 (e.g., "\fBmke2fs \-t ext3 \-O extent /dev/sdXX\fR" will create a
738 filesystem that is not supported by the ext3 implementation as found in
739 the Linux kernel; and "\fBmke2fs \-t ext3 \-O ^has_journal /dev/hdXX\fR"
740 will create a filesystem that does not have a journal and hence will not
741 be supported by the ext3 filesystem code in the Linux kernel.)
743 .BI \-T " usage-type[,...]"
744 Specify how the filesystem is going to be used, so that
746 can choose optimal filesystem parameters for that use. The usage
747 types that are supported are defined in the configuration file
748 .BR /etc/mke2fs.conf .
749 The user may specify one or more usage types
750 using a comma separated list.
752 If this option is is not specified,
754 will pick a single default usage type based on the size of the filesystem to
755 be created. If the filesystem size is less than 3 megabytes,
757 will use the filesystem type
759 If the filesystem size is greater than or equal to 3 but less than
762 will use the filesystem type
764 If the filesystem size is greater than or equal to 4 terabytes but less than
767 will use the filesystem type
769 If the filesystem size is greater than or equal to 16 terabytes,
771 will use the filesystem type
775 will use the default filesystem type
779 Set the universally unique identifier (UUID) of the filesystem to
781 The format of the UUID is a series of hex digits separated by hyphens,
783 "c1b9d5a2-f162-11cf-9ece-0020afc76f16".
786 parameter may also be one of the following:
790 clear the filesystem UUID
793 generate a new randomly-generated UUID
796 generate a new time-based UUID
803 Print the version number of
808 Before overwriting a file system block, write the old contents of the block to
809 an undo file. This undo file can be used with e2undo(8) to restore the old
810 contents of the file system should something go wrong. If the empty string is
811 passed as the undo_file argument, the undo file will be written to a file named
812 mke2fs-\fIdevice\fR.e2undo in the directory specified via the
813 \fIE2FSPROGS_UNDO_DIR\fR environment variable or the \fIundo_dir\fR directive
814 in the configuration file.
816 WARNING: The undo file cannot be used to recover from a power or system crash.
820 If set to non-zero integer value, its value is used to determine how often
822 is called during inode table initialization.
825 Determines the location of the configuration file (see
826 .BR mke2fs.conf (5)).
828 .B MKE2FS_FIRST_META_BG
829 If set to non-zero integer value, its value is used to determine first meta
830 block group. This is mostly for debugging purposes.
832 .B MKE2FS_DEVICE_SECTSIZE
833 If set to non-zero integer value, its value is used to determine logical
837 .B MKE2FS_DEVICE_PHYS_SECTSIZE
838 If set to non-zero integer value, its value is used to determine physical
842 .B MKE2FS_SKIP_CHECK_MSG
843 If set, do not show the message of filesystem automatic check caused by
844 mount count or check interval.
848 has been written by Theodore Ts'o <tytso@mit.edu>.
851 is part of the e2fsprogs package and is available from
852 http://e2fsprogs.sourceforge.net.