5 This is a development version of the ext4 filesystem, an advanced level
6 of the ext3 filesystem which incorporates scalability and reliability
7 enhancements for supporting large filesystems (64 bit) in keeping with
8 increasing disk capacities and state-of-the-art feature requirements.
10 Mailing list: linux-ext4@vger.kernel.org
13 1. Quick usage instructions:
14 ===========================
16 - Compile and install the latest version of e2fsprogs (as of this
17 writing version 1.41) from:
19 http://sourceforge.net/project/showfiles.php?group_id=2406
23 ftp://ftp.kernel.org/pub/linux/kernel/people/tytso/e2fsprogs/
25 or grab the latest git repository from:
27 git://git.kernel.org/pub/scm/fs/ext2/e2fsprogs.git
29 - Note that it is highly important to install the mke2fs.conf file
30 that comes with the e2fsprogs 1.41.x sources in /etc/mke2fs.conf. If
31 you have edited the /etc/mke2fs.conf file installed on your system,
32 you will need to merge your changes with the version from e2fsprogs
35 - Create a new filesystem using the ext4 filesystem type:
37 # mke2fs -t ext4 /dev/hda1
39 Or configure an existing ext3 filesystem to support extents and set
40 the test_fs flag to indicate that it's ok for an in-development
41 filesystem to touch this filesystem:
43 # tune2fs -O extents -E test_fs /dev/hda1
45 If the filesystem was created with 128 byte inodes, it can be
46 converted to use 256 byte for greater efficiency via:
48 # tune2fs -I 256 /dev/hda1
50 (Note: we currently do not have tools to convert an ext4
51 filesystem back to ext3; so please do not do try this on production
56 # mount -t ext4 /dev/hda1 /wherever
58 - When comparing performance with other filesystems, remember that
59 ext3/4 by default offers higher data integrity guarantees than most.
60 So when comparing with a metadata-only journalling filesystem, such
61 as ext3, use `mount -o data=writeback'. And you might as well use
62 `mount -o nobh' too along with it. Making the journal larger than
63 the mke2fs default often helps performance with metadata-intensive
69 2.1 Currently available
71 * ability to use filesystems > 16TB (e2fsprogs support not available yet)
72 * extent format reduces metadata overhead (RAM, IO for access, transactions)
73 * extent format more robust in face of on-disk corruption due to magics,
74 * internal redunancy in tree
75 * improved file allocation (multi-block alloc)
76 * fix 32000 subdirectory limit
77 * nsec timestamps for mtime, atime, ctime, create time
78 * inode version field on disk (NFSv4, Lustre)
79 * reduced e2fsck time via uninit_bg feature
80 * journal checksumming for robustness, performance
81 * persistent file preallocation (e.g for streaming media, databases)
82 * ability to pack bitmaps and inode tables into larger virtual groups via the
85 * Inode allocation using large virtual block groups via flex_bg
87 * large block (up to pagesize) support
88 * efficent new ordered mode in JBD2 and ext4(avoid using buffer head to force
91 2.2 Candidate features for future inclusion
93 * Online defrag (patches available but not well tested)
94 * reduced mke2fs time via lazy itable initialization in conjuction with
95 the uninit_bg feature (capability to do this is available in e2fsprogs
96 but a kernel thread to do lazy zeroing of unused inode table blocks
97 after filesystem is first mounted is required for safety)
99 There are several others under discussion, whether they all make it in is
100 partly a function of how much time everyone has to work on them. Features like
101 metadata checksumming have been discussed and planned for a bit but no patches
102 exist yet so I'm not sure they're in the near-term roadmap.
104 The big performance win will come with mballoc, delalloc and flex_bg
105 grouping of bitmaps and inode tables. Some test results available here:
107 - http://www.bullopensource.org/ext4/20080530/ffsb-write-2.6.26-rc2.html
108 - http://www.bullopensource.org/ext4/20080530/ffsb-readwrite-2.6.26-rc2.html
113 When mounting an ext4 filesystem, the following option are accepted:
116 extents (*) ext4 will use extents to address file data. The
117 file system will no longer be mountable by ext3.
119 noextents ext4 will not use extents for newly created files
121 journal_checksum Enable checksumming of the journal transactions.
122 This will allow the recovery code in e2fsck and the
123 kernel to detect corruption in the kernel. It is a
124 compatible change and will be ignored by older kernels.
126 journal_async_commit Commit block can be written to disk without waiting
127 for descriptor blocks. If enabled older kernels cannot
128 mount the device. This will enable 'journal_checksum'
131 journal=update Update the ext4 file system's journal to the current
134 journal=inum When a journal already exists, this option is ignored.
135 Otherwise, it specifies the number of the inode which
136 will represent the ext4 file system's journal file.
138 journal_dev=devnum When the external journal device's major/minor numbers
139 have changed, this option allows the user to specify
140 the new journal location. The journal device is
141 identified through its new major/minor numbers encoded
144 noload Don't load the journal on mounting.
146 data=journal All data are committed into the journal prior to being
147 written into the main file system.
149 data=ordered (*) All data are forced directly out to the main file
150 system prior to its metadata being committed to the
153 data=writeback Data ordering is not preserved, data may be written
154 into the main file system after its metadata has been
155 committed to the journal.
157 commit=nrsec (*) Ext4 can be told to sync all its data and metadata
158 every 'nrsec' seconds. The default value is 5 seconds.
159 This means that if you lose your power, you will lose
160 as much as the latest 5 seconds of work (your
161 filesystem will not be damaged though, thanks to the
162 journaling). This default value (or any low value)
163 will hurt performance, but it's good for data-safety.
164 Setting it to 0 will have the same effect as leaving
165 it at the default (5 seconds).
166 Setting it to very large values will improve
169 barrier=<0|1(*)> This enables/disables the use of write barriers in
170 the jbd code. barrier=0 disables, barrier=1 enables.
171 This also requires an IO stack which can support
172 barriers, and if jbd gets an error on a barrier
173 write, it will disable again with a warning.
174 Write barriers enforce proper on-disk ordering
175 of journal commits, making volatile disk write caches
176 safe to use, at some performance penalty. If
177 your disks are battery-backed in one way or another,
178 disabling barriers may safely improve performance.
180 inode_readahead=n This tuning parameter controls the maximum
181 number of inode table blocks that ext4's inode
182 table readahead algorithm will pre-read into
183 the buffer cache. The default value is 32 blocks.
185 orlov (*) This enables the new Orlov block allocator. It is
188 oldalloc This disables the Orlov block allocator and enables
189 the old block allocator. Orlov should have better
190 performance - we'd like to get some feedback if it's
191 the contrary for you.
193 user_xattr Enables Extended User Attributes. Additionally, you
194 need to have extended attribute support enabled in the
195 kernel configuration (CONFIG_EXT4_FS_XATTR). See the
196 attr(5) manual page and http://acl.bestbits.at/ to
197 learn more about extended attributes.
199 nouser_xattr Disables Extended User Attributes.
201 acl Enables POSIX Access Control Lists support.
202 Additionally, you need to have ACL support enabled in
203 the kernel configuration (CONFIG_EXT4_FS_POSIX_ACL).
204 See the acl(5) manual page and http://acl.bestbits.at/
205 for more information.
207 noacl This option disables POSIX Access Control List
214 bsddf (*) Make 'df' act like BSD.
215 minixdf Make 'df' act like Minix.
217 check=none Don't do extra checking of bitmaps on mount.
220 debug Extra debugging information is sent to syslog.
222 errors=remount-ro(*) Remount the filesystem read-only on an error.
223 errors=continue Keep going on a filesystem error.
224 errors=panic Panic and halt the machine if an error occurs.
226 data_err=ignore(*) Just print an error message if an error occurs
227 in a file data buffer in ordered mode.
228 data_err=abort Abort the journal if an error occurs in a file
229 data buffer in ordered mode.
231 grpid Give objects the same group ID as their creator.
234 nogrpid (*) New objects have the group ID of their creator.
237 resgid=n The group ID which may use the reserved blocks.
239 resuid=n The user ID which may use the reserved blocks.
241 sb=n Use alternate superblock at this location.
248 bh (*) ext4 associates buffer heads to data pages to
249 nobh (a) cache disk block mapping information
250 (b) link pages into transaction to provide
252 "bh" option forces use of buffer heads.
253 "nobh" option tries to avoid associating buffer
254 heads (supported only for "writeback" mode).
256 mballoc (*) Use the multiple block allocator for block allocation
257 nomballoc disabled multiple block allocator for block allocation.
258 stripe=n Number of filesystem blocks that mballoc will try
259 to use for allocation size and alignment. For RAID5/6
260 systems this should be the number of data
261 disks * RAID chunk size in file system blocks.
262 delalloc (*) Deferring block allocation until write-out time.
263 nodelalloc Disable delayed allocation. Blocks are allocation
264 when data is copied from user to page cache.
268 There are 3 different data modes:
271 In data=writeback mode, ext4 does not journal data at all. This mode provides
272 a similar level of journaling as that of XFS, JFS, and ReiserFS in its default
273 mode - metadata journaling. A crash+recovery can cause incorrect data to
274 appear in files which were written shortly before the crash. This mode will
275 typically provide the best ext4 performance.
278 In data=ordered mode, ext4 only officially journals metadata, but it logically
279 groups metadata information related to data changes with the data blocks into a
280 single unit called a transaction. When it's time to write the new metadata
281 out to disk, the associated data blocks are written first. In general,
282 this mode performs slightly slower than writeback but significantly faster than journal mode.
285 data=journal mode provides full data and metadata journaling. All new data is
286 written to the journal first, and then to its final location.
287 In the event of a crash, the journal can be replayed, bringing both data and
288 metadata into a consistent state. This mode is the slowest except when data
289 needs to be read from and written to disk at the same time where it
290 outperforms all others modes. Curently ext4 does not have delayed
291 allocation support if this data journalling mode is selected.
296 kernel source: <file:fs/ext4/>
299 programs: http://e2fsprogs.sourceforge.net/
301 useful links: http://fedoraproject.org/wiki/ext3-devel
302 http://www.bullopensource.org/ext4/
303 http://ext4.wiki.kernel.org/index.php/Main_Page
304 http://fedoraproject.org/wiki/Features/Ext4