blob: 74aeb142ae5f2b14865cba7224a8d04020a73b6d [file] [log] [blame]
Linus Torvalds1da177e2005-04-16 15:20:36 -07001
2The SGI XFS Filesystem
3======================
4
5XFS is a high performance journaling filesystem which originated
6on the SGI IRIX platform. It is completely multi-threaded, can
7support large files and large filesystems, extended attributes,
8variable block sizes, is extent based, and makes extensive use of
9Btrees (directories, extents, free space) to aid both performance
10and scalability.
11
12Refer to the documentation at http://oss.sgi.com/projects/xfs/
13for further details. This implementation is on-disk compatible
14with the IRIX version of XFS.
15
16
17Mount Options
18=============
19
20When mounting an XFS filesystem, the following options are accepted.
21
Nathan Scottfc97bbf2005-11-03 13:46:43 +110022 allocsize=size
23 Sets the buffered I/O end-of-file preallocation size when
24 doing delayed allocation writeout (default size is 64KiB).
25 Valid values for this option are page size (typically 4KiB)
26 through to 1GiB, inclusive, in power-of-2 increments.
27
28 attr2/noattr2
29 The options enable/disable (default is disabled for backward
30 compatibility on-disk) an "opportunistic" improvement to be
31 made in the way inline extended attributes are stored on-disk.
32 When the new form is used for the first time (by setting or
33 removing extended attributes) the on-disk superblock feature
34 bit field will be updated to reflect this format being in use.
35
36 barrier
37 Enables the use of block layer write barriers for writes into
38 the journal and unwritten extent conversion. This allows for
39 drive level write caching to be enabled, for devices that
40 support write barriers.
41
42 dmapi
43 Enable the DMAPI (Data Management API) event callouts.
44 Use with the "mtpt" option.
45
46 grpid/bsdgroups and nogrpid/sysvgroups
47 These options define what group ID a newly created file gets.
48 When grpid is set, it takes the group ID of the directory in
49 which it is created; otherwise (the default) it takes the fsgid
50 of the current process, unless the directory has the setgid bit
51 set, in which case it takes the gid from the parent directory,
52 and also gets the setgid bit set if it is a directory itself.
53
54 ihashsize=value
55 Sets the number of hash buckets available for hashing the
56 in-memory inodes of the specified mount point. If a value
57 of zero is used, the value selected by the default algorithm
58 will be displayed in /proc/mounts.
Linus Torvalds1da177e2005-04-16 15:20:36 -070059
60 ikeep/noikeep
61 When inode clusters are emptied of inodes, keep them around
62 on the disk (ikeep) - this is the traditional XFS behaviour
63 and is still the default for now. Using the noikeep option,
64 inode clusters are returned to the free space pool.
65
Nathan Scottfc97bbf2005-11-03 13:46:43 +110066 inode64
67 Indicates that XFS is allowed to create inodes at any location
68 in the filesystem, including those which will result in inode
69 numbers occupying more than 32 bits of significance. This is
70 provided for backwards compatibility, but causes problems for
71 backup applications that cannot handle large inode numbers.
72
73 largeio/nolargeio
74 If "nolargeio" is specified, the optimal I/O reported in
75 st_blksize by stat(2) will be as small as possible to allow user
76 applications to avoid inefficient read/modify/write I/O.
77 If "largeio" specified, a filesystem that has a "swidth" specified
78 will return the "swidth" value (in bytes) in st_blksize. If the
79 filesystem does not have a "swidth" specified but does specify
80 an "allocsize" then "allocsize" (in bytes) will be returned
81 instead.
82 If neither of these two options are specified, then filesystem
83 will behave as if "nolargeio" was specified.
84
Linus Torvalds1da177e2005-04-16 15:20:36 -070085 logbufs=value
86 Set the number of in-memory log buffers. Valid numbers range
87 from 2-8 inclusive.
88 The default value is 8 buffers for filesystems with a
Nathan Scottfc97bbf2005-11-03 13:46:43 +110089 blocksize of 64KiB, 4 buffers for filesystems with a blocksize
90 of 32KiB, 3 buffers for filesystems with a blocksize of 16KiB
Linus Torvalds1da177e2005-04-16 15:20:36 -070091 and 2 buffers for all other configurations. Increasing the
92 number of buffers may increase performance on some workloads
93 at the cost of the memory used for the additional log buffers
94 and their associated control structures.
95
96 logbsize=value
97 Set the size of each in-memory log buffer.
98 Size may be specified in bytes, or in kilobytes with a "k" suffix.
Nathan Scottfc97bbf2005-11-03 13:46:43 +110099 Valid sizes for version 1 and version 2 logs are 16384 (16k) and
100 32768 (32k). Valid sizes for version 2 logs also include
Linus Torvalds1da177e2005-04-16 15:20:36 -0700101 65536 (64k), 131072 (128k) and 262144 (256k).
Nathan Scottfc97bbf2005-11-03 13:46:43 +1100102 The default value for machines with more than 32MiB of memory
Linus Torvalds1da177e2005-04-16 15:20:36 -0700103 is 32768, machines with less memory use 16384 by default.
104
105 logdev=device and rtdev=device
106 Use an external log (metadata journal) and/or real-time device.
107 An XFS filesystem has up to three parts: a data section, a log
108 section, and a real-time section. The real-time section is
109 optional, and the log section can be separate from the data
110 section or contained within it.
111
Nathan Scottfc97bbf2005-11-03 13:46:43 +1100112 mtpt=mountpoint
113 Use with the "dmapi" option. The value specified here will be
114 included in the DMAPI mount event, and should be the path of
115 the actual mountpoint that is used.
116
Linus Torvalds1da177e2005-04-16 15:20:36 -0700117 noalign
118 Data allocations will not be aligned at stripe unit boundaries.
119
120 noatime
121 Access timestamps are not updated when a file is read.
122
123 norecovery
124 The filesystem will be mounted without running log recovery.
125 If the filesystem was not cleanly unmounted, it is likely to
126 be inconsistent when mounted in "norecovery" mode.
127 Some files or directories may not be accessible because of this.
128 Filesystems mounted "norecovery" must be mounted read-only or
129 the mount will fail.
130
131 nouuid
132 Don't check for double mounted file systems using the file system uuid.
133 This is useful to mount LVM snapshot volumes.
134
135 osyncisosync
136 Make O_SYNC writes implement true O_SYNC. WITHOUT this option,
137 Linux XFS behaves as if an "osyncisdsync" option is used,
138 which will make writes to files opened with the O_SYNC flag set
139 behave as if the O_DSYNC flag had been used instead.
140 This can result in better performance without compromising
141 data safety.
142 However if this option is not in effect, timestamp updates from
143 O_SYNC writes can be lost if the system crashes.
144 If timestamp updates are critical, use the osyncisosync option.
145
Nathan Scottfc97bbf2005-11-03 13:46:43 +1100146 uquota/usrquota/uqnoenforce/quota
Linus Torvalds1da177e2005-04-16 15:20:36 -0700147 User disk quota accounting enabled, and limits (optionally)
Nathan Scottfc97bbf2005-11-03 13:46:43 +1100148 enforced. Refer to xfs_quota(8) for further details.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700149
Nathan Scottfc97bbf2005-11-03 13:46:43 +1100150 gquota/grpquota/gqnoenforce
Linus Torvalds1da177e2005-04-16 15:20:36 -0700151 Group disk quota accounting enabled and limits (optionally)
Nathan Scottfc97bbf2005-11-03 13:46:43 +1100152 enforced. Refer to xfs_quota(8) for further details.
153
154 pquota/prjquota/pqnoenforce
155 Project disk quota accounting enabled and limits (optionally)
156 enforced. Refer to xfs_quota(8) for further details.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700157
158 sunit=value and swidth=value
159 Used to specify the stripe unit and width for a RAID device or
160 a stripe volume. "value" must be specified in 512-byte block
161 units.
162 If this option is not specified and the filesystem was made on
163 a stripe volume or the stripe width or unit were specified for
164 the RAID device at mkfs time, then the mount system call will
165 restore the value from the superblock. For filesystems that
166 are made directly on RAID devices, these options can be used
167 to override the information in the superblock if the underlying
168 disk layout changes after the filesystem has been created.
169 The "swidth" option is required if the "sunit" option has been
170 specified, and must be a multiple of the "sunit" value.
171
Nathan Scottfc97bbf2005-11-03 13:46:43 +1100172 swalloc
173 Data allocations will be rounded up to stripe width boundaries
174 when the current end of file is being extended and the file
175 size is larger than the stripe width size.
176
177
Linus Torvalds1da177e2005-04-16 15:20:36 -0700178sysctls
179=======
180
181The following sysctls are available for the XFS filesystem:
182
183 fs.xfs.stats_clear (Min: 0 Default: 0 Max: 1)
Nathan Scottfc97bbf2005-11-03 13:46:43 +1100184 Setting this to "1" clears accumulated XFS statistics
Linus Torvalds1da177e2005-04-16 15:20:36 -0700185 in /proc/fs/xfs/stat. It then immediately resets to "0".
Nathan Scottfc97bbf2005-11-03 13:46:43 +1100186
Linus Torvalds1da177e2005-04-16 15:20:36 -0700187 fs.xfs.xfssyncd_centisecs (Min: 100 Default: 3000 Max: 720000)
188 The interval at which the xfssyncd thread flushes metadata
189 out to disk. This thread will flush log activity out, and
190 do some processing on unlinked inodes.
191
192 fs.xfs.xfsbufd_centisecs (Min: 50 Default: 100 Max: 3000)
193 The interval at which xfsbufd scans the dirty metadata buffers list.
194
195 fs.xfs.age_buffer_centisecs (Min: 100 Default: 1500 Max: 720000)
196 The age at which xfsbufd flushes dirty metadata buffers to disk.
197
198 fs.xfs.error_level (Min: 0 Default: 3 Max: 11)
199 A volume knob for error reporting when internal errors occur.
200 This will generate detailed messages & backtraces for filesystem
201 shutdowns, for example. Current threshold values are:
202
203 XFS_ERRLEVEL_OFF: 0
204 XFS_ERRLEVEL_LOW: 1
205 XFS_ERRLEVEL_HIGH: 5
206
207 fs.xfs.panic_mask (Min: 0 Default: 0 Max: 127)
Nathan Scottfc97bbf2005-11-03 13:46:43 +1100208 Causes certain error conditions to call BUG(). Value is a bitmask;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700209 AND together the tags which represent errors which should cause panics:
Nathan Scottfc97bbf2005-11-03 13:46:43 +1100210
Linus Torvalds1da177e2005-04-16 15:20:36 -0700211 XFS_NO_PTAG 0
212 XFS_PTAG_IFLUSH 0x00000001
213 XFS_PTAG_LOGRES 0x00000002
214 XFS_PTAG_AILDELETE 0x00000004
215 XFS_PTAG_ERROR_REPORT 0x00000008
216 XFS_PTAG_SHUTDOWN_CORRUPT 0x00000010
217 XFS_PTAG_SHUTDOWN_IOERROR 0x00000020
218 XFS_PTAG_SHUTDOWN_LOGERROR 0x00000040
219
Nathan Scottfc97bbf2005-11-03 13:46:43 +1100220 This option is intended for debugging only.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700221
222 fs.xfs.irix_symlink_mode (Min: 0 Default: 0 Max: 1)
223 Controls whether symlinks are created with mode 0777 (default)
224 or whether their mode is affected by the umask (irix mode).
225
226 fs.xfs.irix_sgid_inherit (Min: 0 Default: 0 Max: 1)
227 Controls files created in SGID directories.
228 If the group ID of the new file does not match the effective group
Nathan Scottfc97bbf2005-11-03 13:46:43 +1100229 ID or one of the supplementary group IDs of the parent dir, the
230 ISGID bit is cleared if the irix_sgid_inherit compatibility sysctl
Linus Torvalds1da177e2005-04-16 15:20:36 -0700231 is set.
232
233 fs.xfs.restrict_chown (Min: 0 Default: 1 Max: 1)
234 Controls whether unprivileged users can use chown to "give away"
235 a file to another user.
236
Nathan Scottfc97bbf2005-11-03 13:46:43 +1100237 fs.xfs.inherit_sync (Min: 0 Default: 1 Max: 1)
238 Setting this to "1" will cause the "sync" flag set
239 by the xfs_io(8) chattr command on a directory to be
Linus Torvalds1da177e2005-04-16 15:20:36 -0700240 inherited by files in that directory.
241
Nathan Scottfc97bbf2005-11-03 13:46:43 +1100242 fs.xfs.inherit_nodump (Min: 0 Default: 1 Max: 1)
243 Setting this to "1" will cause the "nodump" flag set
244 by the xfs_io(8) chattr command on a directory to be
Linus Torvalds1da177e2005-04-16 15:20:36 -0700245 inherited by files in that directory.
246
Nathan Scottfc97bbf2005-11-03 13:46:43 +1100247 fs.xfs.inherit_noatime (Min: 0 Default: 1 Max: 1)
248 Setting this to "1" will cause the "noatime" flag set
249 by the xfs_io(8) chattr command on a directory to be
Linus Torvalds1da177e2005-04-16 15:20:36 -0700250 inherited by files in that directory.
Nathan Scottfc97bbf2005-11-03 13:46:43 +1100251
252 fs.xfs.inherit_nosymlinks (Min: 0 Default: 1 Max: 1)
253 Setting this to "1" will cause the "nosymlinks" flag set
254 by the xfs_io(8) chattr command on a directory to be
255 inherited by files in that directory.
256
257 fs.xfs.rotorstep (Min: 1 Default: 1 Max: 256)
258 In "inode32" allocation mode, this option determines how many
259 files the allocator attempts to allocate in the same allocation
260 group before moving to the next allocation group. The intent
261 is to control the rate at which the allocator moves between
262 allocation groups when allocating extents for new files.