Documentation/filesystems/ext4.txt

   1
   2 Ext4 Filesystem
   3 ===============
   4
   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.
   9
  10 Mailing list: linux-ext4@vger.kernel.org
  11
  12
  13 1. Quick usage instructions:
  14 ===========================
  15
  16   - Compile and install the latest version of e2fsprogs (as of this
  17     writing version 1.41) from:
  18
  19     http://sourceforge.net/project/showfiles.php?group_id=2406
  20
  21         or
  22
  23     ftp://ftp.kernel.org/pub/linux/kernel/people/tytso/e2fsprogs/
  24
  25         or grab the latest git repository from:
  26
  27     git://git.kernel.org/pub/scm/fs/ext2/e2fsprogs.git
  28
  29   - Create a new filesystem using the ext4dev filesystem type:
  30
  31         # mke2fs -t ext4dev /dev/hda1
  32
  33     Or configure an existing ext3 filesystem to support extents and set
  34     the test_fs flag to indicate that it's ok for an in-development
  35     filesystem to touch this filesystem:
  36
  37         # tune2fs -O extents -E test_fs /dev/hda1
  38
  39     If the filesystem was created with 128 byte inodes, it can be
  40     converted to use 256 byte for greater efficiency via:
  41
  42         # tune2fs -I 256 /dev/hda1
  43
  44     (Note: we currently do not have tools to convert an ext4dev
  45     filesystem back to ext3; so please do not do try this on production
  46     filesystems.)
  47
  48   - Mounting:
  49
  50         # mount -t ext4dev /dev/hda1 /wherever
  51
  52   - When comparing performance with other filesystems, remember that
  53     ext3/4 by default offers higher data integrity guarantees than most.
  54     So when comparing with a metadata-only journalling filesystem, such
  55     as ext3, use `mount -o data=writeback'.  And you might as well use
  56     `mount -o nobh' too along with it.  Making the journal larger than
  57     the mke2fs default often helps performance with metadata-intensive
  58     workloads.
  59
  60 2. Features
  61 ===========
  62
  63 2.1 Currently available
  64
  65 * ability to use filesystems > 16TB (e2fsprogs support not available yet)
  66 * extent format reduces metadata overhead (RAM, IO for access, transactions)
  67 * extent format more robust in face of on-disk corruption due to magics,
  68 * internal redunancy in tree
  69 * improved file allocation (multi-block alloc)
  70 * fix 32000 subdirectory limit
  71 * nsec timestamps for mtime, atime, ctime, create time
  72 * inode version field on disk (NFSv4, Lustre)
  73 * reduced e2fsck time via uninit_bg feature
  74 * journal checksumming for robustness, performance
  75 * persistent file preallocation (e.g for streaming media, databases)
  76 * ability to pack bitmaps and inode tables into larger virtual groups via the
  77   flex_bg feature
  78 * large file support
  79 * Inode allocation using large virtual block groups via flex_bg
  80 * delayed allocation
  81 * large block (up to pagesize) support
  82 * efficent new ordered mode in JBD2 and ext4(avoid using buffer head to force
  83   the ordering)
  84
  85 2.2 Candidate features for future inclusion
  86
  87 * Online defrag (patches available but not well tested)
  88 * reduced mke2fs time via lazy itable initialization in conjuction with
  89   the uninit_bg feature (capability to do this is available in e2fsprogs
  90   but a kernel thread to do lazy zeroing of unused inode table blocks
  91   after filesystem is first mounted is required for safety)
  92
  93 There are several others under discussion, whether they all make it in is
  94 partly a function of how much time everyone has to work on them. Features like
  95 metadata checksumming have been discussed and planned for a bit but no patches
  96 exist yet so I'm not sure they're in the near-term roadmap.
  97
  98 The big performance win will come with mballoc, delalloc and flex_bg
  99 grouping of bitmaps and inode tables.  Some test results available here:
 100
 101  - http://www.bullopensource.org/ext4/20080530/ffsb-write-2.6.26-rc2.html
 102  - http://www.bullopensource.org/ext4/20080530/ffsb-readwrite-2.6.26-rc2.html
 103
 104 3. Options
 105 ==========
 106
 107 When mounting an ext4 filesystem, the following option are accepted:
 108 (*) == default
 109
 110 extents         (*)     ext4 will use extents to address file data.  The
 111                         file system will no longer be mountable by ext3.
 112
 113 noextents               ext4 will not use extents for newly created files
 114
 115 journal_checksum        Enable checksumming of the journal transactions.
 116                         This will allow the recovery code in e2fsck and the
 117                         kernel to detect corruption in the kernel.  It is a
 118                         compatible change and will be ignored by older kernels.
 119
 120 journal_async_commit    Commit block can be written to disk without waiting
 121                         for descriptor blocks. If enabled older kernels cannot
 122                         mount the device. This will enable 'journal_checksum'
 123                         internally.
 124
 125 journal=update          Update the ext4 file system's journal to the current
 126                         format.
 127
 128 journal=inum            When a journal already exists, this option is ignored.
 129                         Otherwise, it specifies the number of the inode which
 130                         will represent the ext4 file system's journal file.
 131
 132 journal_dev=devnum      When the external journal device's major/minor numbers
 133                         have changed, this option allows the user to specify
 134                         the new journal location.  The journal device is
 135                         identified through its new major/minor numbers encoded
 136                         in devnum.
 137
 138 noload                  Don't load the journal on mounting.
 139
 140 data=journal            All data are committed into the journal prior to being
 141                         written into the main file system.
 142
 143 data=ordered    (*)     All data are forced directly out to the main file
 144                         system prior to its metadata being committed to the
 145                         journal.
 146
 147 data=writeback          Data ordering is not preserved, data may be written
 148                         into the main file system after its metadata has been
 149                         committed to the journal.
 150
 151 commit=nrsec    (*)     Ext4 can be told to sync all its data and metadata
 152                         every 'nrsec' seconds. The default value is 5 seconds.
 153                         This means that if you lose your power, you will lose
 154                         as much as the latest 5 seconds of work (your
 155                         filesystem will not be damaged though, thanks to the
 156                         journaling).  This default value (or any low value)
 157                         will hurt performance, but it's good for data-safety.
 158                         Setting it to 0 will have the same effect as leaving
 159                         it at the default (5 seconds).
 160                         Setting it to very large values will improve
 161                         performance.
 162
 163 barrier=<0|1(*)>        This enables/disables the use of write barriers in
 164                         the jbd code.  barrier=0 disables, barrier=1 enables.
 165                         This also requires an IO stack which can support
 166                         barriers, and if jbd gets an error on a barrier
 167                         write, it will disable again with a warning.
 168                         Write barriers enforce proper on-disk ordering
 169                         of journal commits, making volatile disk write caches
 170                         safe to use, at some performance penalty.  If
 171                         your disks are battery-backed in one way or another,
 172                         disabling barriers may safely improve performance.
 173
 174 orlov           (*)     This enables the new Orlov block allocator. It is
 175                         enabled by default.
 176
 177 oldalloc                This disables the Orlov block allocator and enables
 178                         the old block allocator.  Orlov should have better
 179                         performance - we'd like to get some feedback if it's
 180                         the contrary for you.
 181
 182 user_xattr              Enables Extended User Attributes.  Additionally, you
 183                         need to have extended attribute support enabled in the
 184                         kernel configuration (CONFIG_EXT4_FS_XATTR).  See the
 185                         attr(5) manual page and http://acl.bestbits.at/ to
 186                         learn more about extended attributes.
 187
 188 nouser_xattr            Disables Extended User Attributes.
 189
 190 acl                     Enables POSIX Access Control Lists support.
 191                         Additionally, you need to have ACL support enabled in
 192                         the kernel configuration (CONFIG_EXT4_FS_POSIX_ACL).
 193                         See the acl(5) manual page and http://acl.bestbits.at/
 194                         for more information.
 195
 196 noacl                   This option disables POSIX Access Control List
 197                         support.
 198
 199 reservation
 200
 201 noreservation
 202
 203 bsddf           (*)     Make 'df' act like BSD.
 204 minixdf                 Make 'df' act like Minix.
 205
 206 check=none              Don't do extra checking of bitmaps on mount.
 207 nocheck
 208
 209 debug                   Extra debugging information is sent to syslog.
 210
 211 errors=remount-ro(*)    Remount the filesystem read-only on an error.
 212 errors=continue         Keep going on a filesystem error.
 213 errors=panic            Panic and halt the machine if an error occurs.
 214
 215 grpid                   Give objects the same group ID as their creator.
 216 bsdgroups
 217
 218 nogrpid         (*)     New objects have the group ID of their creator.
 219 sysvgroups
 220
 221 resgid=n                The group ID which may use the reserved blocks.
 222
 223 resuid=n                The user ID which may use the reserved blocks.
 224
 225 sb=n                    Use alternate superblock at this location.
 226
 227 quota
 228 noquota
 229 grpquota
 230 usrquota
 231
 232 bh              (*)     ext4 associates buffer heads to data pages to
 233 nobh                    (a) cache disk block mapping information
 234                         (b) link pages into transaction to provide
 235                             ordering guarantees.
 236                         "bh" option forces use of buffer heads.
 237                         "nobh" option tries to avoid associating buffer
 238                         heads (supported only for "writeback" mode).
 239
 240 mballoc         (*)     Use the multiple block allocator for block allocation
 241 nomballoc               disabled multiple block allocator for block allocation.
 242 stripe=n                Number of filesystem blocks that mballoc will try
 243                         to use for allocation size and alignment. For RAID5/6
 244                         systems this should be the number of data
 245                         disks *  RAID chunk size in file system blocks.
 246 delalloc        (*)     Deferring block allocation until write-out time.
 247 nodelalloc              Disable delayed allocation. Blocks are allocation
 248                         when data is copied from user to page cache.
 249 Data Mode
 250 =========
 251 There are 3 different data modes:
 252
 253 * writeback mode
 254 In data=writeback mode, ext4 does not journal data at all.  This mode provides
 255 a similar level of journaling as that of XFS, JFS, and ReiserFS in its default
 256 mode - metadata journaling.  A crash+recovery can cause incorrect data to
 257 appear in files which were written shortly before the crash.  This mode will
 258 typically provide the best ext4 performance.
 259
 260 * ordered mode
 261 In data=ordered mode, ext4 only officially journals metadata, but it logically
 262 groups metadata information related to data changes with the data blocks into a
 263 single unit called a transaction.  When it's time to write the new metadata
 264 out to disk, the associated data blocks are written first.  In general,
 265 this mode performs slightly slower than writeback but significantly faster than journal mode.
 266
 267 * journal mode
 268 data=journal mode provides full data and metadata journaling.  All new data is
 269 written to the journal first, and then to its final location.
 270 In the event of a crash, the journal can be replayed, bringing both data and
 271 metadata into a consistent state.  This mode is the slowest except when data
 272 needs to be read from and written to disk at the same time where it
 273 outperforms all others modes.  Curently ext4 does not have delayed
 274 allocation support if this data journalling mode is selected.
 275
 276 References
 277 ==========
 278
 279 kernel source:  <file:fs/ext4/>
 280                 <file:fs/jbd2/>
 281
 282 programs:       http://e2fsprogs.sourceforge.net/
 283
 284 useful links:   http://fedoraproject.org/wiki/ext3-devel
 285                 http://www.bullopensource.org/ext4/
 286                 http://ext4.wiki.kernel.org/index.php/Main_Page
 287                 http://fedoraproject.org/wiki/Features/Ext4