Section: File Formats (5)
Updated: August 2014Index
Return to Main Contents
ext2 - the second extended file system
ext2 - the third extended file system
ext4 - the fourth extended file system
The second, third, and fourth extended file systems, or ext2, ext3, and
ext4 as they are commonly known, are Linux file systems that have
historically been the default file system for many Linux distributions.
They are general purpose file systems that have been designed for
extensibility and backwards compatibility. In particular, file systems
previously intended for use with the ext2 and ext3 file systems can be
mounted using the ext4 file system driver, and indeed in many modern
Linux distributions, the ext4 file system driver has been configured
handle mount requests for ext2 and ext3 file systems.
FILE SYSTEM FEATURES
A file system formated for ext2, ext3, or ext4 can be have some
collection of the follow file system feature flags enabled. Some of
these features are not supported by all implementations of the ext2,
ext3, and ext4 file system drivers, depending on Linux kernel version in
use. On other operating systems, such as the GNU/HURD or FreeBSD, only
a very restrictive set of file system features may be supported in their
implementations of ext2.
Enables the file system to be larger than 2^32 blocks. This feature is set
automatically, as needed, but it can be useful to specify this feature
explicitly if the file system might need to be resized larger than 2^32
blocks, even if it was smaller than that threshold when it was
originally created. Note that some older kernels and older versions
of e2fsprogs will not support file systems with this ext4 feature enabled.
This ext4 feature enables clustered block allocation, so that the unit of
allocation is a power of two number of blocks. That is, each bit in the
what had traditionally been known as the block allocation bitmap now
indicates whether a cluster is in use or not, where a cluster is by
default composed of 16 blocks. This feature can decrease the time
spent on doing block allocation and brings smaller fragmentation, especially
for large files. The size can be specified using the
The bigalloc feature is still under development, and may not be fully
supported with your kernel or may have various bugs. Please see the web
page http://ext4.wiki.kernel.org/index.php/Bigalloc for details.
May clash with delayed allocation (see
This feature requires that the
features be enabled.
Use hashed b-trees to speed up name lookups in large directories. This
feature is supported by ext3 and ext4 file systems, and is ignored by
ext2 file systems.
This ext4 feature allows more than 65000 subdirectories per directory.
This ext4 feature allows the mapping of logical block numbers for a
particular inode to physical blocks on the storage device to be stored
using an extent tree, which is a more efficient data structure than the
traditional indirect block scheme used by the ext2 and ext3 file
systems. The use of the extent tree decreases metadata block overhead,
improves file system performance, and decreases the needed to run
on the file system.
are accepted as valid names for this feature for
historical/backwards compatibility reasons.)
This ext4 feature reserves a specific amount of space in each inode for
extended metadata such as nanosecond timestamps and file creation time,
even if the current kernel does not current need to reserve this much
space. Without this feature, the kernel will reserve the amount of
space for features currently it currently needs, and the rest may be
consumed by extended attributes.
For this feature to be useful the inode size must be 256 bytes in size
This feature enables the use of extended attributes. This feature is
supported by ext2, ext3, and ext4.
This feature enables the storage file type information in directory
entries. This feature is supported by ext2, ext3, and ext4.
This ext4 feature allows the per-block group metadata (allocation
and inode tables)
to be placed anywhere on the storage media. In addition,
will place the per-block group metadata together starting at the first
block group of each "flex_bg group". The size of the flex_bg group
can be specified using the
Create a journal to ensure filesystem consistency even across unclean
shutdowns. Setting the filesystem feature is equivalent to using the
option. This feature is supported by ext3 and ext4, and ignored by the
ext2 file system driver.
This ext4 feature allows files to be larger than 2 terabytes in size.
This feature is enabled on the superblock found on an external journal
device. The block size for the external journal must be the same as the
file system which uses it.
The external journal device can be used by a file system by specifying
This feature flag is set automatically by modern kernels when a file
larger than 2 gigabytes is created. Very old kernels could not
handle large files, so this feature flag was used to prohibit those
kernels from mounting file systems that they could not understand.
This feature indicates that there will only at most two backup
superblock and block group descriptors. The block groups used to store
the backup superblock and blockgroup descriptors are stored in the
superblock, but typically, one will be located at the beginning of block
group #1, and one in the last block group in the file system. This is
feature is essentially a more extreme version of sparse_super and is
designed to allow the a much larger percentage of the disk to have
contiguous blocks available for data files.
This ext4 feature allows file systems to be resized on-line without explicitly
needing to reserve space for growth in the size of the block group
descriptors. This scheme is also used to resize file systems which are
larger than 2^32 blocks. It is not recommended that this feature be set
when a file system is created, since this alternate method of storing
the block group descriptor will slow down the time needed to mount the
file system, and newer kernels can automatically set this feature as
necessary when doing an online resize and no more reserved space is
available in the resize inode.
This ext4 feature provides multiple mount protection (MMP). MMP helps to
protect the filesystem from being multiply mounted and is useful in
shared storage environments.
Create quota inodes (inode #3 for userquota and inode
#4 for group quota) and set them in the superblock.
With this feature, the quotas will be enabled
automatically when the filesystem is mounted.
Causes the quota files (i.e., user.quota and
group.quota which existed
in the older quota design) to be hidden inodes.
This file system feature indicates that space has been reserved so
the block group descriptor table can be extended by the file system is
resized while the file system is mounted. The online resize operation
is carried out by the kernel, triggered, by
will attempt to reserve enough space so that the
filesystem may grow to 1024 times its initial size. This can be changed
This feature requires that the
feature be enabled.
This file system feature is set on all modern ext2, ext3, and ext4 file
system. It indicates that backup copies of the superblock and block
group descriptors be present only on a few block groups, and not all of
This ext4 file system feature indicates that the block group descriptors
will be protected using checksums, making it safe for
to create a file system without initializing all of the block groups.
The kernel will keep a high watermark of unused inodes, and initialize
inode tables and block lazily. This feature speeds up the time to check
the file system using
and it also speeds up the time required for
to create the file system.
This section describes mount options which are specific to ext2, ext3,
and ext4. Other generic mount options may be used as well; see
Mount options for ext2
The `ext2' filesystem is the standard Linux filesystem.
Since Linux 2.5.46, for most mount options the default
is determined by the filesystem superblock. Set them with
Support POSIX Access Control Lists (or not).
Set the behavior for the
system call. The
behavior is to return in the
field the total number of blocks of the filesystem, while the
behavior (which is the default) is to subtract the overhead blocks
used by the ext2 filesystem and not available for file storage. Thus
% mount /k -o minixdf; df /k; umount /k
|Filesystem ||1024-blocks ||Used ||Available ||Capacity ||Mounted on|
% mount /k -o bsddf; df /k; umount /k
|Filesystem ||1024-blocks ||Used ||Available ||Capacity ||Mounted on|
(Note that this example shows that one can add command line options
to the options given in
- check=none or nocheck
No checking is done at mount time. This is the default. This is fast.
It is wise to invoke
every now and then, e.g. at boot time. The non-default behavior is unsupported
(check=normal and check=strict options have been removed). Note that these mount options
don't have to be supported if ext4 kernel driver is used for ext2 and ext3 filesystems.
Print debugging info upon each (re)mount.
Define the behavior when an error is encountered.
(Either ignore errors and just mark the filesystem erroneous and continue,
or remount the filesystem read-only, or panic and halt the system.)
The default is set in the filesystem superblock, and can be
- grpid|bsdgroups and nogrpid|sysvgroups
These options define what group id a newly created file gets.
is set, it takes the group id of the directory in which it is created;
otherwise (the default) it takes the fsgid of the current process, unless
the directory has the setgid bit set, in which case it takes the gid
from the parent directory, and also gets the setgid bit set
if it is a directory itself.
The usrquota (same as quota) mount option enables user quota support on the
filesystem. grpquota enables group quotas support. You need the quota utilities
to actually enable and manage the quota system.
Disables 32-bit UIDs and GIDs. This is for interoperability with older
kernels which only store and expect 16-bit values.
- oldalloc or orlov
Use old allocator or Orlov allocator for new inodes. Orlov is default.
- resgid=,n and resuid=,n
The ext2 filesystem reserves a certain percentage of the available
space (by default 5%, see
These options determine who can use the reserved blocks.
(Roughly: whoever has the specified uid, or belongs to the specified group.)
Instead of block 1, use block
as superblock. This could be useful when the filesystem has been damaged.
(Earlier, copies of the superblock would be made every 8192 blocks: in
block 1, 8193, 16385, ... (and one got thousands of copies on
a big filesystem). Since version 1.08,
has a -s (sparse superblock) option to reduce the number of backup
superblocks, and since version 1.15 this is the default. Note
that this may mean that ext2 filesystems created by a recent
cannot be mounted r/w under Linux 2.0.*.)
The block number here uses 1 k units. Thus, if you want to use logical
block 32768 on a filesystem with 4 k blocks, use "sb=131072".
Support "user." extended attributes (or not).
Mount options for ext3
The ext3 filesystem is a version of the ext2 filesystem which has been
enhanced with journaling. It supports the same options as ext2 as
well as the following additions:
Update the ext3 filesystem's journal to the current format.
When a journal already exists, this option is ignored. Otherwise, it
specifies the number of the inode which will represent the ext3 filesystem's
journal file; ext3 will create a new journal, overwriting the old contents
of the file whose inode number is
When the external journal device's major/minor numbers
have changed, these options allow the user to specify
the new journal location. The journal device is
identified either through its new major/minor numbers encoded
in devnum, or via a path to the device.
Don't load the journal on mounting. Note that
if the filesystem was not unmounted cleanly,
skipping the journal replay will lead to the
filesystem containing inconsistencies that can
lead to any number of problems.
Specifies the journaling mode for file data. Metadata is always journaled.
To use modes other than
on the root filesystem, pass the mode to the kernel as boot parameter, e.g.
All data is committed into the journal prior to being written into the
This is the default mode. All data is forced directly out to the main file
system prior to its metadata being committed to the journal.
Data ordering is not preserved - data may be written into the main
filesystem after its metadata has been committed to the journal.
This is rumoured to be the highest-throughput option. It guarantees
internal filesystem integrity, however it can allow old data to appear
in files after a crash and journal recovery.
Just print an error message if an error occurs in a file data buffer in
Abort the journal if an error occurs in a file data buffer in ordered mode.
- barrier=0 / barrier=1
This disables / enables the use of write barriers in the jbd code. barrier=0
disables, barrier=1 enables (default). This also requires an IO stack which can
support barriers, and if jbd gets an error on a barrier write, it will disable
barriers again with a warning. Write barriers enforce proper on-disk ordering
of journal commits, making volatile disk write caches safe to use, at some
performance penalty. If your disks are battery-backed in one way or another,
disabling barriers may safely improve performance.
Sync all data and metadata every
seconds. The default value is 5 seconds. Zero means default.
Enable Extended User Attributes. See the
Enable POSIX Access Control Lists. See the
Apart from the old quota system (as in ext2, jqfmt=vfsold aka version 1 quota)
ext3 also supports journaled quotas (version 2 quota). jqfmt=vfsv0
enables journaled quotas. For journaled quotas the mount options
usrjquota=aquota.user and grpjquota=aquota.group are required to tell the
quota system which quota database files to use. Journaled quotas have the
advantage that even after a crash no quota check is required.
Mount options for ext4
The ext4 filesystem is an advanced level of the ext3 filesystem which
incorporates scalability and reliability enhancements for supporting large
journal_dev, norecovery, noload, data, commit, orlov, oldalloc, [no]user_xattr
[no]acl, bsddf, minixdf, debug, errors, data_err, grpid, bsdgroups, nogrpid
sysvgroups, resgid, resuid, sb, quota, noquota, grpquota, usrquota
usrjquota, grpjquota and jqfmt
are backwardly compatible with ext3 or ext2.
Enable checksumming of the journal transactions. This will allow the recovery
code in e2fsck and the kernel to detect corruption in the kernel. It is a
compatible change and will be ignored by older kernels.
Commit block can be written to disk without waiting for descriptor blocks. If
enabled older kernels cannot mount the device.
This will enable 'journal_checksum' internally.
- barrier=0 / barrier=1 / barrier / nobarrier
These mount options have the same effect as in ext3. The mount options
"barrier" and "nobarrier" are added for consistency with other ext4 mount
The ext4 filesystem enables write barriers by default.
This tuning parameter controls the maximum number of inode table blocks that
ext4's inode table readahead algorithm will pre-read into the buffer cache.
The value must be a power of 2. The default value is 32 blocks.
Number of filesystem blocks that mballoc will try to use for allocation size
and alignment. For RAID5/6 systems this should be the number of data disks *
RAID chunk size in filesystem blocks.
Deferring block allocation until write-out time.
Disable delayed allocation. Blocks are allocated when data is copied from user
to page cache.
Maximum amount of time ext4 should wait for additional filesystem operations to
be batch together with a synchronous write operation. Since a synchronous
write operation is going to force a commit and then a wait for the I/O
complete, it doesn't cost much, and can be a huge throughput win, we wait for a
small amount of time to see if any other transactions can piggyback on the
synchronous write. The algorithm used is designed to automatically tune for
the speed of the disk, by measuring the amount of time (on average) that it
takes to finish committing a transaction. Call this time the "commit time".
If the time that the transaction has been running is less than the commit time,
ext4 will try sleeping for the commit time to see if other operations will join
the transaction. The commit time is capped by the max_batch_time, which
defaults to 15000 µs (15 ms). This optimization can be turned off entirely by
setting max_batch_time to 0.
This parameter sets the commit time (as described above) to be at least
min_batch_time. It defaults to zero microseconds. Increasing this parameter
may improve the throughput of multi-threaded, synchronous workloads on very
fast disks, at the cost of increasing latency.
The I/O priority (from 0 to 7, where 0 is the highest priority) which should be
used for I/O operations submitted by kjournald2 during a commit operation.
This defaults to 3, which is a slightly higher priority than the default I/O
Simulate the effects of calling ext4_abort() for
debugging purposes. This is normally used while
remounting a filesystem which is already mounted.
Many broken applications don't use fsync() when
replacing existing files via patterns such as
fd = open("foo.new")/write(fd,...)/close(fd)/ rename("foo.new", "foo")
or worse yet
fd = open("foo", O_TRUNC)/write(fd,...)/close(fd).
If auto_da_alloc is enabled, ext4 will detect the replace-via-rename and
replace-via-truncate patterns and force that any delayed allocation blocks are
allocated such that at the next journal commit, in the default data=ordered
mode, the data blocks of the new file are forced to disk before the rename()
operation is committed. This provides roughly the same level of guarantees as
ext3, and avoids the "zero-length" problem that can happen when a system
crashes before the delayed allocation blocks are forced to disk.
Do not initialize any uninitialized inode table blocks in the background. This
feature may be used by installation CD's so that the install process can
complete as quickly as possible; the inode table initialization process would
then be deferred until the next time the filesystem is mounted.
The lazy itable init code will wait n times the number of milliseconds it took
to zero out the previous block group's inode table. This minimizes the impact on
system performance while the filesystem's inode table is being initialized.
Controls whether ext4 should issue discard/TRIM commands to the underlying
block device when blocks are freed. This is useful for SSD devices and
sparse/thinly-provisioned LUNs, but it is off by default until sufficient
testing has been done.
Disables 32-bit UIDs and GIDs. This is for
interoperability with older kernels which only
store and expect 16-bit values.
This options allows to enables/disables the in-kernel facility for tracking
filesystem metadata blocks within internal data structures. This allows multi-block allocator and other routines to quickly locate extents which might
overlap with filesystem metadata blocks. This option is intended for debugging
purposes and since it negatively affects the performance, it is off by default.
Controls whether or not ext4 should use the DIO read locking. If the
dioread_nolock option is specified ext4 will allocate uninitialized extent
before buffer write and convert the extent to initialized after IO completes.
This approach allows ext4 code to avoid using inode mutex, which improves
scalability on high speed storages. However this does not work with data
journaling and dioread_nolock option will be ignored with kernel warning.
Note that dioread_nolock code path is only used for extent-based files.
Because of the restrictions this options comprises it is off by default
This limits the size of the directories so that any attempt to expand them
beyond the specified limit in kilobytes will cause an ENOSPC error. This is
useful in memory-constrained environments, where a very large directory can
cause severe performance problems or even provoke the Out Of Memory killer. (For
example, if there is only 512 MB memory available, a 176 MB directory may
seriously cramp the system's style.)
Enable 64-bit inode version support. This option is off by default.
The ext2, ext3, and ext4 filesystems support setting the following file
attributes on Linux systems using the
a - append only
A - no atime updates
d - no dump
D - synchronous directory updates
i - immutable
S - synchronous updates
u - undeletable
In addition, the ext3 and ext4 filesystems support the following flag:
j - data journaling
Finally, the ext4 filesystem also supports the following flag:
e - extents format
For descriptions of these attribute flags, please refer to the
- FILE SYSTEM FEATURES
- MOUNT OPTIONS
- Mount options for ext2
- Mount options for ext3
- Mount options for ext4
- FILE ATTRIBUTES
- SEE ALSO