1 .. SPDX-License-Identifier: GPL-2.0 1 .. SPDX-License-Identifier: GPL-2.0
2 2
3 ====================== 3 ======================
4 The SGI XFS Filesystem 4 The SGI XFS Filesystem
5 ====================== 5 ======================
6 6
7 XFS is a high performance journaling filesyste 7 XFS is a high performance journaling filesystem which originated
8 on the SGI IRIX platform. It is completely mu 8 on the SGI IRIX platform. It is completely multi-threaded, can
9 support large files and large filesystems, ext 9 support large files and large filesystems, extended attributes,
10 variable block sizes, is extent based, and mak 10 variable block sizes, is extent based, and makes extensive use of
11 Btrees (directories, extents, free space) to a 11 Btrees (directories, extents, free space) to aid both performance
12 and scalability. 12 and scalability.
13 13
14 Refer to the documentation at https://xfs.wiki 14 Refer to the documentation at https://xfs.wiki.kernel.org/
15 for further details. This implementation is o 15 for further details. This implementation is on-disk compatible
16 with the IRIX version of XFS. 16 with the IRIX version of XFS.
17 17
18 18
19 Mount Options 19 Mount Options
20 ============= 20 =============
21 21
22 When mounting an XFS filesystem, the following 22 When mounting an XFS filesystem, the following options are accepted.
23 23
24 allocsize=size 24 allocsize=size
25 Sets the buffered I/O end-of-file prea 25 Sets the buffered I/O end-of-file preallocation size when
26 doing delayed allocation writeout (def 26 doing delayed allocation writeout (default size is 64KiB).
27 Valid values for this option are page 27 Valid values for this option are page size (typically 4KiB)
28 through to 1GiB, inclusive, in power-o 28 through to 1GiB, inclusive, in power-of-2 increments.
29 29
30 The default behaviour is for dynamic e 30 The default behaviour is for dynamic end-of-file
31 preallocation size, which uses a set o 31 preallocation size, which uses a set of heuristics to
32 optimise the preallocation size based 32 optimise the preallocation size based on the current
33 allocation patterns within the file an 33 allocation patterns within the file and the access patterns
34 to the file. Specifying a fixed ``allo 34 to the file. Specifying a fixed ``allocsize`` value turns off
35 the dynamic behaviour. 35 the dynamic behaviour.
36 36
37 attr2 or noattr2 37 attr2 or noattr2
38 The options enable/disable an "opportu 38 The options enable/disable an "opportunistic" improvement to
39 be made in the way inline extended att 39 be made in the way inline extended attributes are stored
40 on-disk. When the new form is used fo 40 on-disk. When the new form is used for the first time when
41 ``attr2`` is selected (either when set 41 ``attr2`` is selected (either when setting or removing extended
42 attributes) the on-disk superblock fea 42 attributes) the on-disk superblock feature bit field will be
43 updated to reflect this format being i 43 updated to reflect this format being in use.
44 44
45 The default behaviour is determined by 45 The default behaviour is determined by the on-disk feature
46 bit indicating that ``attr2`` behaviou 46 bit indicating that ``attr2`` behaviour is active. If either
47 mount option is set, then that becomes 47 mount option is set, then that becomes the new default used
48 by the filesystem. 48 by the filesystem.
49 49
50 CRC enabled filesystems always use the 50 CRC enabled filesystems always use the ``attr2`` format, and so
51 will reject the ``noattr2`` mount opti 51 will reject the ``noattr2`` mount option if it is set.
52 52
53 discard or nodiscard (default) 53 discard or nodiscard (default)
54 Enable/disable the issuing of commands 54 Enable/disable the issuing of commands to let the block
55 device reclaim space freed by the file 55 device reclaim space freed by the filesystem. This is
56 useful for SSD devices, thinly provisi 56 useful for SSD devices, thinly provisioned LUNs and virtual
57 machine images, but may have a perform 57 machine images, but may have a performance impact.
58 58
59 Note: It is currently recommended that 59 Note: It is currently recommended that you use the ``fstrim``
60 application to ``discard`` unused bloc 60 application to ``discard`` unused blocks rather than the ``discard``
61 mount option because the performance i 61 mount option because the performance impact of this option
62 is quite severe. 62 is quite severe.
63 63
64 grpid/bsdgroups or nogrpid/sysvgroups (defau 64 grpid/bsdgroups or nogrpid/sysvgroups (default)
65 These options define what group ID a n 65 These options define what group ID a newly created file
66 gets. When ``grpid`` is set, it takes 66 gets. When ``grpid`` is set, it takes the group ID of the
67 directory in which it is created; othe 67 directory in which it is created; otherwise it takes the
68 ``fsgid`` of the current process, unle 68 ``fsgid`` of the current process, unless the directory has the
69 ``setgid`` bit set, in which case it t 69 ``setgid`` bit set, in which case it takes the ``gid`` from the
70 parent directory, and also gets the `` 70 parent directory, and also gets the ``setgid`` bit set if it is
71 a directory itself. 71 a directory itself.
72 72
73 filestreams 73 filestreams
74 Make the data allocator use the filest 74 Make the data allocator use the filestreams allocation mode
75 across the entire filesystem rather th 75 across the entire filesystem rather than just on directories
76 configured to use it. 76 configured to use it.
77 77
78 ikeep or noikeep (default) 78 ikeep or noikeep (default)
79 When ``ikeep`` is specified, XFS does 79 When ``ikeep`` is specified, XFS does not delete empty inode
80 clusters and keeps them around on disk 80 clusters and keeps them around on disk. When ``noikeep`` is
81 specified, empty inode clusters are re 81 specified, empty inode clusters are returned to the free
82 space pool. 82 space pool.
83 83
84 inode32 or inode64 (default) 84 inode32 or inode64 (default)
85 When ``inode32`` is specified, it indi 85 When ``inode32`` is specified, it indicates that XFS limits
86 inode creation to locations which will 86 inode creation to locations which will not result in inode
87 numbers with more than 32 bits of sign 87 numbers with more than 32 bits of significance.
88 88
89 When ``inode64`` is specified, it indi 89 When ``inode64`` is specified, it indicates that XFS is allowed
90 to create inodes at any location in th 90 to create inodes at any location in the filesystem,
91 including those which will result in i 91 including those which will result in inode numbers occupying
92 more than 32 bits of significance. 92 more than 32 bits of significance.
93 93
94 ``inode32`` is provided for backwards 94 ``inode32`` is provided for backwards compatibility with older
95 systems and applications, since 64 bit 95 systems and applications, since 64 bits inode numbers might
96 cause problems for some applications t 96 cause problems for some applications that cannot handle
97 large inode numbers. If applications 97 large inode numbers. If applications are in use which do
98 not handle inode numbers bigger than 3 98 not handle inode numbers bigger than 32 bits, the ``inode32``
99 option should be specified. 99 option should be specified.
100 100
101 largeio or nolargeio (default) 101 largeio or nolargeio (default)
102 If ``nolargeio`` is specified, the opt 102 If ``nolargeio`` is specified, the optimal I/O reported in
103 ``st_blksize`` by **stat(2)** will be 103 ``st_blksize`` by **stat(2)** will be as small as possible to allow
104 user applications to avoid inefficient 104 user applications to avoid inefficient read/modify/write
105 I/O. This is typically the page size 105 I/O. This is typically the page size of the machine, as
106 this is the granularity of the page ca 106 this is the granularity of the page cache.
107 107
108 If ``largeio`` is specified, a filesys 108 If ``largeio`` is specified, a filesystem that was created with a
109 ``swidth`` specified will return the ` 109 ``swidth`` specified will return the ``swidth`` value (in bytes)
110 in ``st_blksize``. If the filesystem d 110 in ``st_blksize``. If the filesystem does not have a ``swidth``
111 specified but does specify an ``allocs 111 specified but does specify an ``allocsize`` then ``allocsize``
112 (in bytes) will be returned instead. O 112 (in bytes) will be returned instead. Otherwise the behaviour
113 is the same as if ``nolargeio`` was sp 113 is the same as if ``nolargeio`` was specified.
114 114
115 logbufs=value 115 logbufs=value
116 Set the number of in-memory log buffer 116 Set the number of in-memory log buffers. Valid numbers
117 range from 2-8 inclusive. 117 range from 2-8 inclusive.
118 118
119 The default value is 8 buffers. 119 The default value is 8 buffers.
120 120
121 If the memory cost of 8 log buffers is 121 If the memory cost of 8 log buffers is too high on small
122 systems, then it may be reduced at som 122 systems, then it may be reduced at some cost to performance
123 on metadata intensive workloads. The ` 123 on metadata intensive workloads. The ``logbsize`` option below
124 controls the size of each buffer and s 124 controls the size of each buffer and so is also relevant to
125 this case. 125 this case.
126 126
127 logbsize=value 127 logbsize=value
128 Set the size of each in-memory log buf 128 Set the size of each in-memory log buffer. The size may be
129 specified in bytes, or in kilobytes wi 129 specified in bytes, or in kilobytes with a "k" suffix.
130 Valid sizes for version 1 and version 130 Valid sizes for version 1 and version 2 logs are 16384 (16k)
131 and 32768 (32k). Valid sizes for vers 131 and 32768 (32k). Valid sizes for version 2 logs also
132 include 65536 (64k), 131072 (128k) and 132 include 65536 (64k), 131072 (128k) and 262144 (256k). The
133 logbsize must be an integer multiple o 133 logbsize must be an integer multiple of the log
134 stripe unit configured at **mkfs(8)** 134 stripe unit configured at **mkfs(8)** time.
135 135
136 The default value for version 1 logs i 136 The default value for version 1 logs is 32768, while the
137 default value for version 2 logs is MA 137 default value for version 2 logs is MAX(32768, log_sunit).
138 138
139 logdev=device and rtdev=device 139 logdev=device and rtdev=device
140 Use an external log (metadata journal) 140 Use an external log (metadata journal) and/or real-time device.
141 An XFS filesystem has up to three part 141 An XFS filesystem has up to three parts: a data section, a log
142 section, and a real-time section. The 142 section, and a real-time section. The real-time section is
143 optional, and the log section can be s 143 optional, and the log section can be separate from the data
144 section or contained within it. 144 section or contained within it.
145 145
146 noalign 146 noalign
147 Data allocations will not be aligned a 147 Data allocations will not be aligned at stripe unit
148 boundaries. This is only relevant to f 148 boundaries. This is only relevant to filesystems created
149 with non-zero data alignment parameter 149 with non-zero data alignment parameters (``sunit``, ``swidth``) by
150 **mkfs(8)**. 150 **mkfs(8)**.
151 151
152 norecovery 152 norecovery
153 The filesystem will be mounted without 153 The filesystem will be mounted without running log recovery.
154 If the filesystem was not cleanly unmo 154 If the filesystem was not cleanly unmounted, it is likely to
155 be inconsistent when mounted in ``nore 155 be inconsistent when mounted in ``norecovery`` mode.
156 Some files or directories may not be a 156 Some files or directories may not be accessible because of this.
157 Filesystems mounted ``norecovery`` mus 157 Filesystems mounted ``norecovery`` must be mounted read-only or
158 the mount will fail. 158 the mount will fail.
159 159
160 nouuid 160 nouuid
161 Don't check for double mounted file sy 161 Don't check for double mounted file systems using the file
162 system ``uuid``. This is useful to mo 162 system ``uuid``. This is useful to mount LVM snapshot volumes,
163 and often used in combination with ``n 163 and often used in combination with ``norecovery`` for mounting
164 read-only snapshots. 164 read-only snapshots.
165 165
166 noquota 166 noquota
167 Forcibly turns off all quota accountin 167 Forcibly turns off all quota accounting and enforcement
168 within the filesystem. 168 within the filesystem.
169 169
170 uquota/usrquota/uqnoenforce/quota 170 uquota/usrquota/uqnoenforce/quota
171 User disk quota accounting enabled, an 171 User disk quota accounting enabled, and limits (optionally)
172 enforced. Refer to **xfs_quota(8)** f 172 enforced. Refer to **xfs_quota(8)** for further details.
173 173
174 gquota/grpquota/gqnoenforce 174 gquota/grpquota/gqnoenforce
175 Group disk quota accounting enabled an 175 Group disk quota accounting enabled and limits (optionally)
176 enforced. Refer to **xfs_quota(8)** f 176 enforced. Refer to **xfs_quota(8)** for further details.
177 177
178 pquota/prjquota/pqnoenforce 178 pquota/prjquota/pqnoenforce
179 Project disk quota accounting enabled 179 Project disk quota accounting enabled and limits (optionally)
180 enforced. Refer to **xfs_quota(8)** f 180 enforced. Refer to **xfs_quota(8)** for further details.
181 181
182 sunit=value and swidth=value 182 sunit=value and swidth=value
183 Used to specify the stripe unit and wi 183 Used to specify the stripe unit and width for a RAID device
184 or a stripe volume. "value" must be s 184 or a stripe volume. "value" must be specified in 512-byte
185 block units. These options are only re 185 block units. These options are only relevant to filesystems
186 that were created with non-zero data a 186 that were created with non-zero data alignment parameters.
187 187
188 The ``sunit`` and ``swidth`` parameter 188 The ``sunit`` and ``swidth`` parameters specified must be compatible
189 with the existing filesystem alignment 189 with the existing filesystem alignment characteristics. In
190 general, that means the only valid cha 190 general, that means the only valid changes to ``sunit`` are
191 increasing it by a power-of-2 multiple 191 increasing it by a power-of-2 multiple. Valid ``swidth`` values
192 are any integer multiple of a valid `` 192 are any integer multiple of a valid ``sunit`` value.
193 193
194 Typically the only time these mount op 194 Typically the only time these mount options are necessary if
195 after an underlying RAID device has ha 195 after an underlying RAID device has had its geometry
196 modified, such as adding a new disk to 196 modified, such as adding a new disk to a RAID5 lun and
197 reshaping it. 197 reshaping it.
198 198
199 swalloc 199 swalloc
200 Data allocations will be rounded up to 200 Data allocations will be rounded up to stripe width boundaries
201 when the current end of file is being 201 when the current end of file is being extended and the file
202 size is larger than the stripe width s 202 size is larger than the stripe width size.
203 203
204 wsync 204 wsync
205 When specified, all filesystem namespa 205 When specified, all filesystem namespace operations are
206 executed synchronously. This ensures t 206 executed synchronously. This ensures that when the namespace
207 operation (create, unlink, etc) comple 207 operation (create, unlink, etc) completes, the change to the
208 namespace is on stable storage. This i 208 namespace is on stable storage. This is useful in HA setups
209 where failover must not result in clie 209 where failover must not result in clients seeing
210 inconsistent namespace presentation du 210 inconsistent namespace presentation during or after a
211 failover event. 211 failover event.
212 212
213 Deprecation of V4 Format 213 Deprecation of V4 Format
214 ======================== 214 ========================
215 215
216 The V4 filesystem format lacks certain feature 216 The V4 filesystem format lacks certain features that are supported by
217 the V5 format, such as metadata checksumming, 217 the V5 format, such as metadata checksumming, strengthened metadata
218 verification, and the ability to store timesta 218 verification, and the ability to store timestamps past the year 2038.
219 Because of this, the V4 format is deprecated. 219 Because of this, the V4 format is deprecated. All users should upgrade
220 by backing up their files, reformatting, and r 220 by backing up their files, reformatting, and restoring from the backup.
221 221
222 Administrators and users can detect a V4 files 222 Administrators and users can detect a V4 filesystem by running xfs_info
223 against a filesystem mountpoint and checking f 223 against a filesystem mountpoint and checking for a string containing
224 "crc=". If no such string is found, please up 224 "crc=". If no such string is found, please upgrade xfsprogs to the
225 latest version and try again. 225 latest version and try again.
226 226
227 The deprecation will take place in two parts. 227 The deprecation will take place in two parts. Support for mounting V4
228 filesystems can now be disabled at kernel buil 228 filesystems can now be disabled at kernel build time via Kconfig option.
229 The option will default to yes until September 229 The option will default to yes until September 2025, at which time it
230 will be changed to default to no. In Septembe 230 will be changed to default to no. In September 2030, support will be
231 removed from the codebase entirely. 231 removed from the codebase entirely.
232 232
233 Note: Distributors may choose to withdraw V4 f 233 Note: Distributors may choose to withdraw V4 format support earlier than
234 the dates listed above. 234 the dates listed above.
235 235
236 Deprecated Mount Options 236 Deprecated Mount Options
237 ======================== 237 ========================
238 238
239 ============================ ============== 239 ============================ ================
240 Name Removal Schedu 240 Name Removal Schedule
241 ============================ ============== 241 ============================ ================
242 Mounting with V4 filesystem September 2030 242 Mounting with V4 filesystem September 2030
243 Mounting ascii-ci filesystem September 2030 243 Mounting ascii-ci filesystem September 2030
244 ikeep/noikeep September 2025 244 ikeep/noikeep September 2025
245 attr2/noattr2 September 2025 245 attr2/noattr2 September 2025
246 ============================ ============== 246 ============================ ================
247 247
248 248
249 Removed Mount Options 249 Removed Mount Options
250 ===================== 250 =====================
251 251
252 =========================== ======= 252 =========================== =======
253 Name Removed 253 Name Removed
254 =========================== ======= 254 =========================== =======
255 delaylog/nodelaylog v4.0 255 delaylog/nodelaylog v4.0
256 ihashsize v4.0 256 ihashsize v4.0
257 irixsgid v4.0 257 irixsgid v4.0
258 osyncisdsync/osyncisosync v4.0 258 osyncisdsync/osyncisosync v4.0
259 barrier v4.19 259 barrier v4.19
260 nobarrier v4.19 260 nobarrier v4.19
261 =========================== ======= 261 =========================== =======
262 262
263 sysctls 263 sysctls
264 ======= 264 =======
265 265
266 The following sysctls are available for the XF 266 The following sysctls are available for the XFS filesystem:
267 267
268 fs.xfs.stats_clear (Min: 0 Defau 268 fs.xfs.stats_clear (Min: 0 Default: 0 Max: 1)
269 Setting this to "1" clears accumulated 269 Setting this to "1" clears accumulated XFS statistics
270 in /proc/fs/xfs/stat. It then immedia 270 in /proc/fs/xfs/stat. It then immediately resets to "0".
271 271
272 fs.xfs.xfssyncd_centisecs (Min: 100 Def 272 fs.xfs.xfssyncd_centisecs (Min: 100 Default: 3000 Max: 720000)
273 The interval at which the filesystem f 273 The interval at which the filesystem flushes metadata
274 out to disk and runs internal cache cl 274 out to disk and runs internal cache cleanup routines.
275 275
276 fs.xfs.filestream_centisecs (Min: 1 Defau 276 fs.xfs.filestream_centisecs (Min: 1 Default: 3000 Max: 360000)
277 The interval at which the filesystem a 277 The interval at which the filesystem ages filestreams cache
278 references and returns timed-out AGs b 278 references and returns timed-out AGs back to the free stream
279 pool. 279 pool.
280 280
281 fs.xfs.speculative_prealloc_lifetime 281 fs.xfs.speculative_prealloc_lifetime
282 (Units: seconds Min: 1 Default: 300 282 (Units: seconds Min: 1 Default: 300 Max: 86400)
283 The interval at which the background s 283 The interval at which the background scanning for inodes
284 with unused speculative preallocation 284 with unused speculative preallocation runs. The scan
285 removes unused preallocation from clea 285 removes unused preallocation from clean inodes and releases
286 the unused space back to the free pool 286 the unused space back to the free pool.
287 287
288 fs.xfs.speculative_cow_prealloc_lifetime 288 fs.xfs.speculative_cow_prealloc_lifetime
289 This is an alias for speculative_preal 289 This is an alias for speculative_prealloc_lifetime.
290 290
291 fs.xfs.error_level (Min: 0 Defau 291 fs.xfs.error_level (Min: 0 Default: 3 Max: 11)
292 A volume knob for error reporting when 292 A volume knob for error reporting when internal errors occur.
293 This will generate detailed messages & 293 This will generate detailed messages & backtraces for filesystem
294 shutdowns, for example. Current thres 294 shutdowns, for example. Current threshold values are:
295 295
296 XFS_ERRLEVEL_OFF: 0 296 XFS_ERRLEVEL_OFF: 0
297 XFS_ERRLEVEL_LOW: 1 297 XFS_ERRLEVEL_LOW: 1
298 XFS_ERRLEVEL_HIGH: 5 298 XFS_ERRLEVEL_HIGH: 5
299 299
300 fs.xfs.panic_mask (Min: 0 Defau 300 fs.xfs.panic_mask (Min: 0 Default: 0 Max: 511)
301 Causes certain error conditions to cal 301 Causes certain error conditions to call BUG(). Value is a bitmask;
302 OR together the tags which represent e 302 OR together the tags which represent errors which should cause panics:
303 303
304 XFS_NO_PTAG 304 XFS_NO_PTAG 0
305 XFS_PTAG_IFLUSH 305 XFS_PTAG_IFLUSH 0x00000001
306 XFS_PTAG_LOGRES 306 XFS_PTAG_LOGRES 0x00000002
307 XFS_PTAG_AILDELETE 307 XFS_PTAG_AILDELETE 0x00000004
308 XFS_PTAG_ERROR_REPORT 308 XFS_PTAG_ERROR_REPORT 0x00000008
309 XFS_PTAG_SHUTDOWN_CORRUPT 309 XFS_PTAG_SHUTDOWN_CORRUPT 0x00000010
310 XFS_PTAG_SHUTDOWN_IOERROR 310 XFS_PTAG_SHUTDOWN_IOERROR 0x00000020
311 XFS_PTAG_SHUTDOWN_LOGERROR 311 XFS_PTAG_SHUTDOWN_LOGERROR 0x00000040
312 XFS_PTAG_FSBLOCK_ZERO 312 XFS_PTAG_FSBLOCK_ZERO 0x00000080
313 XFS_PTAG_VERIFIER_ERROR 313 XFS_PTAG_VERIFIER_ERROR 0x00000100
314 314
315 This option is intended for debugging 315 This option is intended for debugging only.
316 316
317 fs.xfs.irix_symlink_mode (Min: 0 Defau 317 fs.xfs.irix_symlink_mode (Min: 0 Default: 0 Max: 1)
318 Controls whether symlinks are created 318 Controls whether symlinks are created with mode 0777 (default)
319 or whether their mode is affected by t 319 or whether their mode is affected by the umask (irix mode).
320 320
321 fs.xfs.irix_sgid_inherit (Min: 0 Defau 321 fs.xfs.irix_sgid_inherit (Min: 0 Default: 0 Max: 1)
322 Controls files created in SGID directo 322 Controls files created in SGID directories.
323 If the group ID of the new file does n 323 If the group ID of the new file does not match the effective group
324 ID or one of the supplementary group I 324 ID or one of the supplementary group IDs of the parent dir, the
325 ISGID bit is cleared if the irix_sgid_ 325 ISGID bit is cleared if the irix_sgid_inherit compatibility sysctl
326 is set. 326 is set.
327 327
328 fs.xfs.inherit_sync (Min: 0 Defau 328 fs.xfs.inherit_sync (Min: 0 Default: 1 Max: 1)
329 Setting this to "1" will cause the "sy 329 Setting this to "1" will cause the "sync" flag set
330 by the **xfs_io(8)** chattr command on 330 by the **xfs_io(8)** chattr command on a directory to be
331 inherited by files in that directory. 331 inherited by files in that directory.
332 332
333 fs.xfs.inherit_nodump (Min: 0 Defau 333 fs.xfs.inherit_nodump (Min: 0 Default: 1 Max: 1)
334 Setting this to "1" will cause the "no 334 Setting this to "1" will cause the "nodump" flag set
335 by the **xfs_io(8)** chattr command on 335 by the **xfs_io(8)** chattr command on a directory to be
336 inherited by files in that directory. 336 inherited by files in that directory.
337 337
338 fs.xfs.inherit_noatime (Min: 0 Defau 338 fs.xfs.inherit_noatime (Min: 0 Default: 1 Max: 1)
339 Setting this to "1" will cause the "no 339 Setting this to "1" will cause the "noatime" flag set
340 by the **xfs_io(8)** chattr command on 340 by the **xfs_io(8)** chattr command on a directory to be
341 inherited by files in that directory. 341 inherited by files in that directory.
342 342
343 fs.xfs.inherit_nosymlinks (Min: 0 Defau 343 fs.xfs.inherit_nosymlinks (Min: 0 Default: 1 Max: 1)
344 Setting this to "1" will cause the "no 344 Setting this to "1" will cause the "nosymlinks" flag set
345 by the **xfs_io(8)** chattr command on 345 by the **xfs_io(8)** chattr command on a directory to be
346 inherited by files in that directory. 346 inherited by files in that directory.
347 347
348 fs.xfs.inherit_nodefrag (Min: 0 Defau 348 fs.xfs.inherit_nodefrag (Min: 0 Default: 1 Max: 1)
349 Setting this to "1" will cause the "no 349 Setting this to "1" will cause the "nodefrag" flag set
350 by the **xfs_io(8)** chattr command on 350 by the **xfs_io(8)** chattr command on a directory to be
351 inherited by files in that directory. 351 inherited by files in that directory.
352 352
353 fs.xfs.rotorstep (Min: 1 Defau 353 fs.xfs.rotorstep (Min: 1 Default: 1 Max: 256)
354 In "inode32" allocation mode, this opt 354 In "inode32" allocation mode, this option determines how many
355 files the allocator attempts to alloca 355 files the allocator attempts to allocate in the same allocation
356 group before moving to the next alloca 356 group before moving to the next allocation group. The intent
357 is to control the rate at which the al 357 is to control the rate at which the allocator moves between
358 allocation groups when allocating exte 358 allocation groups when allocating extents for new files.
359 359
360 Deprecated Sysctls 360 Deprecated Sysctls
361 ================== 361 ==================
362 362
363 =========================================== 363 =========================================== ================
364 Name 364 Name Removal Schedule
365 =========================================== 365 =========================================== ================
366 fs.xfs.irix_sgid_inherit 366 fs.xfs.irix_sgid_inherit September 2025
367 fs.xfs.irix_symlink_mode 367 fs.xfs.irix_symlink_mode September 2025
368 fs.xfs.speculative_cow_prealloc_lifetime 368 fs.xfs.speculative_cow_prealloc_lifetime September 2025
369 =========================================== 369 =========================================== ================
370 370
371 371
372 Removed Sysctls 372 Removed Sysctls
373 =============== 373 ===============
374 374
375 ============================= ======= 375 ============================= =======
376 Name Removed 376 Name Removed
377 ============================= ======= 377 ============================= =======
378 fs.xfs.xfsbufd_centisec v4.0 378 fs.xfs.xfsbufd_centisec v4.0
379 fs.xfs.age_buffer_centisecs v4.0 379 fs.xfs.age_buffer_centisecs v4.0
380 ============================= ======= 380 ============================= =======
381 381
382 Error handling 382 Error handling
383 ============== 383 ==============
384 384
385 XFS can act differently according to the type 385 XFS can act differently according to the type of error found during its
386 operation. The implementation introduces the f 386 operation. The implementation introduces the following concepts to the error
387 handler: 387 handler:
388 388
389 -failure speed: 389 -failure speed:
390 Defines how fast XFS should propagate 390 Defines how fast XFS should propagate an error upwards when a specific
391 error is found during the filesystem o 391 error is found during the filesystem operation. It can propagate
392 immediately, after a defined number of 392 immediately, after a defined number of retries, after a set time period,
393 or simply retry forever. 393 or simply retry forever.
394 394
395 -error classes: 395 -error classes:
396 Specifies the subsystem the error conf 396 Specifies the subsystem the error configuration will apply to, such as
397 metadata IO or memory allocation. Diff 397 metadata IO or memory allocation. Different subsystems will have
398 different error handlers for which beh 398 different error handlers for which behaviour can be configured.
399 399
400 -error handlers: 400 -error handlers:
401 Defines the behavior for a specific er 401 Defines the behavior for a specific error.
402 402
403 The filesystem behavior during an error can be 403 The filesystem behavior during an error can be set via ``sysfs`` files. Each
404 error handler works independently - the first 404 error handler works independently - the first condition met by an error handler
405 for a specific class will cause the error to b 405 for a specific class will cause the error to be propagated rather than reset and
406 retried. 406 retried.
407 407
408 The action taken by the filesystem when the er 408 The action taken by the filesystem when the error is propagated is context
409 dependent - it may cause a shut down in the ca 409 dependent - it may cause a shut down in the case of an unrecoverable error,
410 it may be reported back to userspace, or it ma 410 it may be reported back to userspace, or it may even be ignored because
411 there's nothing useful we can with the error o 411 there's nothing useful we can with the error or anyone we can report it to (e.g.
412 during unmount). 412 during unmount).
413 413
414 The configuration files are organized into the 414 The configuration files are organized into the following hierarchy for each
415 mounted filesystem: 415 mounted filesystem:
416 416
417 /sys/fs/xfs/<dev>/error/<class>/<error>/ 417 /sys/fs/xfs/<dev>/error/<class>/<error>/
418 418
419 Where: 419 Where:
420 <dev> 420 <dev>
421 The short device name of the mounted f 421 The short device name of the mounted filesystem. This is the same device
422 name that shows up in XFS kernel error 422 name that shows up in XFS kernel error messages as "XFS(<dev>): ..."
423 423
424 <class> 424 <class>
425 The subsystem the error configuration 425 The subsystem the error configuration belongs to. As of 4.9, the defined
426 classes are: 426 classes are:
427 427
428 - "metadata": applies metadata 428 - "metadata": applies metadata buffer write IO
429 429
430 <error> 430 <error>
431 The individual error handler configura 431 The individual error handler configurations.
432 432
433 433
434 Each filesystem has "global" error configurati 434 Each filesystem has "global" error configuration options defined in their top
435 level directory: 435 level directory:
436 436
437 /sys/fs/xfs/<dev>/error/ 437 /sys/fs/xfs/<dev>/error/
438 438
439 fail_at_unmount (Min: 0 Defa 439 fail_at_unmount (Min: 0 Default: 1 Max: 1)
440 Defines the filesystem error behavior 440 Defines the filesystem error behavior at unmount time.
441 441
442 If set to a value of 1, XFS will overr 442 If set to a value of 1, XFS will override all other error configurations
443 during unmount and replace them with " 443 during unmount and replace them with "immediate fail" characteristics.
444 i.e. no retries, no retry timeout. Thi 444 i.e. no retries, no retry timeout. This will always allow unmount to
445 succeed when there are persistent erro 445 succeed when there are persistent errors present.
446 446
447 If set to 0, the configured retry beha 447 If set to 0, the configured retry behaviour will continue until all
448 retries and/or timeouts have been exha 448 retries and/or timeouts have been exhausted. This will delay unmount
449 completion when there are persistent e 449 completion when there are persistent errors, and it may prevent the
450 filesystem from ever unmounting fully 450 filesystem from ever unmounting fully in the case of "retry forever"
451 handler configurations. 451 handler configurations.
452 452
453 Note: there is no guarantee that fail_ 453 Note: there is no guarantee that fail_at_unmount can be set while an
454 unmount is in progress. It is possible 454 unmount is in progress. It is possible that the ``sysfs`` entries are
455 removed by the unmounting filesystem b 455 removed by the unmounting filesystem before a "retry forever" error
456 handler configuration causes unmount t 456 handler configuration causes unmount to hang, and hence the filesystem
457 must be configured appropriately befor 457 must be configured appropriately before unmount begins to prevent
458 unmount hangs. 458 unmount hangs.
459 459
460 Each filesystem has specific error class handl 460 Each filesystem has specific error class handlers that define the error
461 propagation behaviour for specific errors. The 461 propagation behaviour for specific errors. There is also a "default" error
462 handler defined, which defines the behaviour f 462 handler defined, which defines the behaviour for all errors that don't have
463 specific handlers defined. Where multiple retr 463 specific handlers defined. Where multiple retry constraints are configured for
464 a single error, the first retry configuration 464 a single error, the first retry configuration that expires will cause the error
465 to be propagated. The handler configurations a 465 to be propagated. The handler configurations are found in the directory:
466 466
467 /sys/fs/xfs/<dev>/error/<class>/<error>/ 467 /sys/fs/xfs/<dev>/error/<class>/<error>/
468 468
469 max_retries (Min: -1 Defa 469 max_retries (Min: -1 Default: Varies Max: INTMAX)
470 Defines the allowed number of retries 470 Defines the allowed number of retries of a specific error before
471 the filesystem will propagate the erro 471 the filesystem will propagate the error. The retry count for a given
472 error context (e.g. a specific metadat 472 error context (e.g. a specific metadata buffer) is reset every time
473 there is a successful completion of th 473 there is a successful completion of the operation.
474 474
475 Setting the value to "-1" will cause X 475 Setting the value to "-1" will cause XFS to retry forever for this
476 specific error. 476 specific error.
477 477
478 Setting the value to "0" will cause XF 478 Setting the value to "0" will cause XFS to fail immediately when the
479 specific error is reported. 479 specific error is reported.
480 480
481 Setting the value to "N" (where 0 < N 481 Setting the value to "N" (where 0 < N < Max) will make XFS retry the
482 operation "N" times before propagating 482 operation "N" times before propagating the error.
483 483
484 retry_timeout_seconds (Min: -1 Def 484 retry_timeout_seconds (Min: -1 Default: Varies Max: 1 day)
485 Define the amount of time (in seconds) 485 Define the amount of time (in seconds) that the filesystem is
486 allowed to retry its operations when t 486 allowed to retry its operations when the specific error is
487 found. 487 found.
488 488
489 Setting the value to "-1" will allow X 489 Setting the value to "-1" will allow XFS to retry forever for this
490 specific error. 490 specific error.
491 491
492 Setting the value to "0" will cause XF 492 Setting the value to "0" will cause XFS to fail immediately when the
493 specific error is reported. 493 specific error is reported.
494 494
495 Setting the value to "N" (where 0 < N 495 Setting the value to "N" (where 0 < N < Max) will allow XFS to retry the
496 operation for up to "N" seconds before 496 operation for up to "N" seconds before propagating the error.
497 497
498 **Note:** The default behaviour for a specific 498 **Note:** The default behaviour for a specific error handler is dependent on both
499 the class and error context. For example, the 499 the class and error context. For example, the default values for
500 "metadata/ENODEV" are "0" rather than "-1" so 500 "metadata/ENODEV" are "0" rather than "-1" so that this error handler defaults
501 to "fail immediately" behaviour. This is done 501 to "fail immediately" behaviour. This is done because ENODEV is a fatal,
502 unrecoverable error no matter how many times t 502 unrecoverable error no matter how many times the metadata IO is retried.
503 503
504 Workqueue Concurrency 504 Workqueue Concurrency
505 ===================== 505 =====================
506 506
507 XFS uses kernel workqueues to parallelize meta 507 XFS uses kernel workqueues to parallelize metadata update processes. This
508 enables it to take advantage of storage hardwa 508 enables it to take advantage of storage hardware that can service many IO
509 operations simultaneously. This interface exp 509 operations simultaneously. This interface exposes internal implementation
510 details of XFS, and as such is explicitly not 510 details of XFS, and as such is explicitly not part of any userspace API/ABI
511 guarantee the kernel may give userspace. Thes 511 guarantee the kernel may give userspace. These are undocumented features of
512 the generic workqueue implementation XFS uses 512 the generic workqueue implementation XFS uses for concurrency, and they are
513 provided here purely for diagnostic and tuning 513 provided here purely for diagnostic and tuning purposes and may change at any
514 time in the future. 514 time in the future.
515 515
516 The control knobs for a filesystem's workqueue 516 The control knobs for a filesystem's workqueues are organized by task at hand
517 and the short name of the data device. They a 517 and the short name of the data device. They all can be found in:
518 518
519 /sys/bus/workqueue/devices/${task}!${device} 519 /sys/bus/workqueue/devices/${task}!${device}
520 520
521 ================ =========== 521 ================ ===========
522 Task Description 522 Task Description
523 ================ =========== 523 ================ ===========
524 xfs_iwalk-$pid Inode scans of the entire fi 524 xfs_iwalk-$pid Inode scans of the entire filesystem. Currently limited to
525 mount time quotacheck. 525 mount time quotacheck.
526 xfs-gc Background garbage collectio 526 xfs-gc Background garbage collection of disk space that have been
527 speculatively allocated beyo 527 speculatively allocated beyond EOF or for staging copy on
528 write operations. 528 write operations.
529 ================ =========== 529 ================ ===========
530 530
531 For example, the knobs for the quotacheck work 531 For example, the knobs for the quotacheck workqueue for /dev/nvme0n1 would be
532 found in /sys/bus/workqueue/devices/xfs_iwalk- 532 found in /sys/bus/workqueue/devices/xfs_iwalk-1111!nvme0n1/.
533 533
534 The interesting knobs for XFS workqueues are a 534 The interesting knobs for XFS workqueues are as follows:
535 535
536 ============ =========== 536 ============ ===========
537 Knob Description 537 Knob Description
538 ============ =========== 538 ============ ===========
539 max_active Maximum number of background 539 max_active Maximum number of background threads that can be started to
540 run the work. 540 run the work.
541 cpumask CPUs upon which the threads a 541 cpumask CPUs upon which the threads are allowed to run.
542 nice Relative priority of scheduli 542 nice Relative priority of scheduling the threads. These are the
543 same nice levels that can be 543 same nice levels that can be applied to userspace processes.
544 ============ =========== 544 ============ ===========
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.