1xfs_db(8) System Manager's Manual xfs_db(8)
2
6 xfs_db - debug an XFS filesystem
7
9 xfs_db [ -c cmd ] ... [ -i|r|x|F ] [ -f ] [ -l logdev ] [ -p progname ]
10 device
11 xfs_db -V
12
14 xfs_db is used to examine an XFS filesystem. Under rare circumstances
15 it can also be used to modify an XFS filesystem, but that task is nor‐
16 mally left to xfs_repair(8) or to scripts such as xfs_admin(8) that run
17 xfs_db.
18
20 -c cmd xfs_db commands may be run interactively (the default) or as
21 arguments on the command line. Multiple -c arguments may be
22 given. The commands are run in the sequence given, then the pro‐
23 gram exits. This is the mechanism used to implement
24 xfs_check(8).
25 -f Specifies that the filesystem image to be processed is stored in
27 a regular file at device (see the mkfs.xfs(8) -d file option).
28 This might happen if an image copy of a filesystem has been made
29 into an ordinary file with xfs_copy(8).
30 -F Specifies that we want to continue even if the superblock magic
32 is not correct. For use in xfs_check and xfs_metadump.
33 -i Allows execution on a mounted filesystem, provided it is mounted
35 read-only. Useful for shell scripts such as xfs_check(8), which
36 must only operate on filesystems in a guarenteed consistent
37 state (either unmounted or mounted read-only). These semantics
38 are slightly different to that of the -r option.
39 -l logdev
41 Specifies the device where the filesystems external log resides.
42 Only for those filesystems which use an external log. See the
43 mkfs.xfs(8) -l option, and refer to xfs(5) for a detailed
44 description of the XFS log.
45 -p progname
47 Set the program name to progname for prompts and some error mes‐
48 sages, the default value is xfs_db.
49 -r Open device or filename read-only. This option is required if
51 the filesystem is mounted. It is only necessary to omit this
52 flag if a command that changes data (write, blocktrash) is to be
53 used.
54 -x Specifies expert mode. This enables the write and blocktrash
56 commands.
57 -V Prints out the current version number and exits.
59
61 xfs_db commands can be broken up into two classes. Most commands are
62 for the navigation and display of data structures in the filesystem.
63 Other commands are for scanning the filesystem in some way.
64 Commands which are used to navigate the filesystem structure take argu‐
66 ments which reflect the names of filesystem structure fields. There
67 can be multiple field names separated by dots when the underlying
68 structures are nested, as in C. The field names can be indexed (as an
69 array index) if the underlying field is an array. The array indices
70 can be specified as a range, two numbers separated by a dash.
71 xfs_db maintains a current address in the filesystem. The granularity
73 of the address is a filesystem structure. This can be a filesystem
74 block, an inode or quota (smaller than a filesystem block), or a direc‐
75 tory block (could be larger than a filesystem block). There are a
76 variety of commands to set the current address. Associated with the
77 current address is the current data type, which is the structural type
78 of this data. Commands which follow the structure of the filesystem
79 always set the type as well as the address. Commands which examine
80 pieces of an individual file (inode) need the current inode to be set,
81 this is done with the inode command.
82 The current address/type information is actually maintained in a stack
84 that can be explicitly manipulated with the push, pop, and stack com‐
85 mands. This allows for easy examination of a nested filesystem struc‐
86 ture. Also, the last several locations visited are stored in a ring
87 buffer which can be manipulated with the forward, back, and ring com‐
88 mands.
89 XFS filesystems are divided into a small number of allocation groups.
91 xfs_db maintains a notion of the current allocation group which is
92 manipulated by some commands. The initial allocation group is 0.
93
95 Many commands have extensive online help. Use the help command for more
96 details on any command.
97 a See the addr command.
99 ablock filoff
101 Set current address to the offset filoff (a filesystem block
102 number) in the attribute area of the current inode.
103 addr [field-expression]
105 Set current address to the value of the field-expression. This
106 is used to "follow" a reference in one structure to the object
107 being referred to. If no argument is given, the current address
108 is printed.
109 agf [agno]
111 Set current address to the AGF block for allocation group agno.
112 If no argument is given, use the current allocation group.
113 agfl [agno]
115 Set current address to the AGFL block for allocation group agno.
116 If no argument is given, use the current allocation group.
117 agi [agno]
119 Set current address to the AGI block for allocation group agno.
120 If no argument is given, use the current allocation group.
121 b See the back command.
123 back Move to the previous location in the position ring.
125 blockfree
127 Free block usage information collected by the last execution of
128 the blockget command. This must be done before another blockget
129 command can be given, presumably with different arguments than
130 the previous one.
131 blockget [-npvs] [-b bno] ... [-i ino] ...
133 Get block usage and check filesystem consistency. The informa‐
134 tion is saved for use by a subsequent blockuse, ncheck, or
135 blocktrash command. See xfs_check(8) for more information.
136 -b is used to specify filesystem block numbers about which
138 verbose information should be printed.
139 -i is used to specify inode numbers about which verbose
141 information should be printed.
142 -n is used to save pathnames for inodes visited, this is
144 used to support the xfs_ncheck(8) command. It also means
145 that pathnames will be printed for inodes that have prob‐
146 lems. This option uses a lot of memory so is not enabled
147 by default.
148 -p causes error messages to be prefixed with the filesystem
150 name being processed. This is useful if several copies of
151 xfs_db are run in parallel.
152 -s restricts output to severe errors only. This is useful if
154 the output is too long otherwise.
155 -v enables verbose output. Messages will be printed for
157 every block and inode processed.
158 blocktrash [-n count] [-x min] [-y max] [-s seed] [-0|1|2|3] [-t type]
160 ...
161 Trash randomly selected filesystem metadata blocks. Trashing
162 occurs to randomly selected bits in the chosen blocks. This
163 command is available only in debugging versions of xfs_db. It
164 is useful for testing xfs_repair(8) and xfs_check(8).
165 -0 | -1 | -2 | -3
167 These are used to set the operating mode for blocktrash.
168 Only one can be used: -0 changed bits are cleared; -1
169 changed bits are set; -2 changed bits are inverted; -3
170 changed bits are randomized.
171 -n supplies the count of block-trashings to perform (default
173 1).
174 -s supplies a seed to the random processing.
176 -t gives a type of blocks to be selected for trashing. Mul‐
178 tiple -t options may be given. If no -t options are given
179 then all metadata types can be trashed.
180 -x sets the minimum size of bit range to be trashed. The
182 default value is 1.
183 -y sets the maximum size of bit range to be trashed. The
185 default value is 1024.
186 blockuse [-n] [-c count]
188 Print usage for current filesystem block(s). For each block,
189 the type and (if any) inode are printed.
190 -c specifies a count of blocks to process. The default value
192 is 1 (the current block only).
193 -n specifies that file names should be printed. The prior
195 blockget command must have also specified the -n option.
196 bmap [-a] [-d] [block [len]]
198 Show the block map for the current inode. The map display can
199 be restricted to an area of the file with the block and len
200 arguments. If block is given and len is omitted then 1 is
201 assumed for len.
202 The -a and -d options are used to select the attribute or data
204 area of the inode, if neither option is given then both areas
205 are shown.
206 check See the blockget command.
208 convert type number [type number] ... type
210 Convert from one address form to another. The known types, with
211 alternate names, are:
212 agblock or agbno (filesystem block within an allocation
213 group)
214 agino or aginode (inode number within an allocation group)
215 agnumber or agno (allocation group number)
216 bboff or daddroff (byte offset in a daddr)
217 blkoff or fsboff or agboff (byte offset in a agblock or
218 fsblock)
219 byte or fsbyte (byte address in filesystem)
220 daddr or bb (disk address, 512-byte blocks)
221 fsblock or fsb or fsbno (filesystem block, see the fsblock
222 command)
223 ino or inode (inode number)
224 inoidx or offset (index of inode in filesystem block)
225 inooff or inodeoff (byte offset in inode)
226 Only conversions that "make sense" are allowed. The compound
228 form (with more than three arguments) is useful for conversions
229 such as convert agno ag agbno agb fsblock.
230 daddr [d]
232 Set current address to the daddr (512 byte block) given by d.
233 If no value for d is given, the current address is printed,
234 expressed as a daddr. The type is set to data (uninterpreted).
235 dblock filoff
237 Set current address to the offset filoff (a filesystem block
238 number) in the data area of the current inode.
239 debug [flagbits]
241 Set debug option bits. These are used for debugging xfs_db. If
242 no value is given for flagbits, print the current debug option
243 bits. These are for the use of the implementor.
244 dquot [projectid_or_userid]
246 Set current address to a project or user quota block.
247 echo [arg] ...
249 Echo the arguments to the output.
250 f See the forward command.
252 forward
254 Move forward to the next entry in the position ring.
255 frag [-adflqRrv]
257 Get file fragmentation data. This prints information about frag‐
258 mentation of file data in the filesystem (as opposed to fragmen‐
259 tation of freespace, for which see the freesp command). Every
260 file in the filesystem is examined to see how far from ideal its
261 extent mappings are. A summary is printed giving the totals.
262 -v sets verbosity, every inode has information printed for
264 it. The remaining options select which inodes and
265 extents are examined. If no options are given then all
266 are assumed set, otherwise just those given are enabled.
267 -a enables processing of attribute data.
269 -d enables processing of directory data.
271 -f enables processing of regular file data.
273 -l enables processing of symbolic link data.
275 -q enables processing of quota file data.
277 -R enables processing of realtime control file data.
279 -r enables processing of realtime file data.
281 freesp [-bcds] [-a ag] ... [-e i] [-h h1] ... [-m m]
283 Summarize free space for the filesystem. The free blocks are
284 examined and totalled, and displayed in the form of a histogram,
285 with a count of extents in each range of free extent sizes.
286 -a adds ag to the list of allocation groups to be processed.
288 If no -a options are given then all allocation groups are
289 processed.
290 -b specifies that the histogram buckets are binary-sized,
292 with the starting sizes being the powers of 2.
293 -c specifies that freesp will search the by-size (cnt) space
295 Btree instead of the default by-block (bno) space Btree.
296 -d specifies that every free extent will be displayed.
298 -e specifies that the histogram buckets are equal-sized,
300 with the size specified as i.
301 -h specifies a starting block number for a histogram bucket
303 as h1. Multiple -h's are given to specify the complete
304 set of buckets.
305 -m specifies that the histogram starting block numbers are
307 powers of m. This is the general case of -b.
308 -s specifies that a final summary of total free extents,
310 free blocks, and the average free extent size is printed.
311 fsb See the fsblock command.
313 fsblock [fsb]
315 Set current address to the fsblock value given by fsb. If no
316 value for fsb is given the current address is printed, expressed
317 as an fsb. The type is set to data (uninterpreted). XFS
318 filesystem block numbers are computed ((agno << agshift) |
319 agblock) where agshift depends on the size of an allocation
320 group. Use the convert command to convert to and from this form.
321 Block numbers given for file blocks (for instance from the bmap
322 command) are in this form.
323 hash string
325 Prints the hash value of string using the hash function of the
326 XFS directory and attribute implementation.
327 help [command]
329 Print help for one or all commands.
330 inode [inode#]
332 Set the current inode number. If no inode# is given, print the
333 current inode number.
334 label [label]
336 Set the filesystem label. The filesystem label can be used by
337 mount(8) instead of using a device special file. The maximum
338 length of an XFS label is 12 characters - use of a longer label
339 will result in truncation and a warning will be issued. If no
340 label is given, the current filesystem label is printed.
341 log [stop | start filename]
343 Start logging output to filename, stop logging, or print the
344 current logging status.
345 metadump [-egow] filename
347 Dumps metadata to a file. See xfs_metadump(8) for more informa‐
348 tion.
349 ncheck [-s] [-i ino] ...
351 Print name-inode pairs. A blockget -n command must be run first
352 to gather the information.
353 -i specifies an inode number to be printed. If no -i options
355 are given then all inodes are printed.
356 -s specifies that only setuid and setgid files are printed.
358 p See the print command.
360 pop Pop location from the stack.
362 print [field-expression] ...
364 Print field values. If no argument is given, print all fields
365 in the current structure.
366 push [command]
368 Push location to the stack. If command is supplied, set the cur‐
369 rent location to the results of command after pushing the old
370 location.
371 q See the quit command.
373 quit Exit xfs_db.
375 ring [index]
377 Show position ring (if no index argument is given), or move to a
378 specific entry in the position ring given by index.
379 sb [agno]
381 Set current address to SB header in allocation group agno. If
382 no agno is given, use the current allocation group number.
383 source source-file
385 Process commands from source-file. source commands can be
386 nested.
387 stack View the location stack.
389 type [type]
391 Set the current data type to type. If no argument is given,
392 show the current data type. The possible data types are: agf,
393 agfl, agi, attr, bmapbta, bmapbtd, bnobt, cntbt, data, dir,
394 dir2, dqblk, inobt, inode, log, rtbitmap, rtsummary, sb, symlink
395 and text. See the TYPES section below for more information on
396 these data types.
397 uuid [uuid | generate | rewrite]
399 Set the filesystem universally unique identifier (UUID). The
400 filesystem UUID can be used by mount(8) instead of using a
401 device special file. The uuid can be set directly to the
402 desired UUID, or it can be automatically generated using the
403 generate option. These options will both write the UUID into
404 every copy of the superblock in the filesystem. rewrite copies
405 the current UUID from the primary superblock to all secondary
406 copies of the superblock. If no argument is given, the current
407 filesystem UUID is printed.
408 version [feature | versionnum features2]
410 Enable selected features for a filesystem (certain features can
411 be enabled on an unmounted filesystem, after mkfs.xfs(8) has
412 created the filesystem). Support for unwritten extents can be
413 enabled using the extflg option. Support for version 2 log for‐
414 mat can be enabled using the log2 option. Support for extended
415 attributes can be enabled using the attr1 or attr2 option. Once
416 enabled, extended attributes cannot be disabled, but the user
417 may toggle between attr1 and attr2 at will (older kernels may
418 not support the newer version).
419 If no argument is given, the current version and feature bits
421 are printed. With one argument, this command will write the
422 updated version number into every copy of the superblock in the
423 filesystem. If two arguments are given, they will be used as
424 numeric values for the versionnum and features2 bits respec‐
425 tively, and their string equivalent reported (but no modifica‐
426 tions are made).
427 write [field value] ...
429 Write a value to disk. Specific fields can be set in structures
430 (struct mode), or a block can be set to data values (data mode),
431 or a block can be set to string values (string mode, for symlink
432 blocks). The operation happens immediately: there is no buffer‐
433 ing.
434 Struct mode is in effect when the current type is structural,
436 i.e. not data. For struct mode, the syntax is "write field
437 value".
438 Data mode is in effect when the current type is data. In this
440 case the contents of the block can be shifted or rotated left or
441 right, or filled with a sequence, a constant value, or a random
442 value. In this mode write with no arguments gives more informa‐
443 tion on the allowed commands.
444
446 This section gives the fields in each structure type and their mean‐
447 ings. Note that some types of block cover multiple actual structures,
448 for instance directory blocks.
449 agf The AGF block is the header for block allocation information;
451 it is in the second 512-byte block of each allocation group.
452 The following fields are defined:
453 magicnum AGF block magic number, 0x58414746 ('XAGF').
454 versionnum version number, currently 1.
455 seqno sequence number starting from 0.
456 length size in filesystem blocks of the allocation
457 group. All allocation groups except the last
458 one of the filesystem have the superblock's
459 agblocks value here.
460 bnoroot block number of the root of the Btree holding
461 free space information sorted by block num‐
462 ber.
463 cntroot block number of the root of the Btree holding
464 free space information sorted by block count.
465 bnolevel number of levels in the by-block-number
466 Btree.
467 cntlevel number of levels in the by-block-count Btree.
468 flfirst index into the AGFL block of the first active
469 entry.
470 fllast index into the AGFL block of the last active
471 entry.
472 flcount count of active entries in the AGFL block.
473 freeblks count of blocks represented in the freespace
474 Btrees.
475 longest longest free space represented in the
476 freespace Btrees.
477 btreeblks number of blocks held in the AGF Btrees.
478 agfl The AGFL block contains block numbers for use of the block
480 allocator; it is in the fourth 512-byte block of each alloca‐
481 tion group. Each entry in the active list is a block number
482 within the allocation group that can be used for any purpose
483 if space runs low. The AGF block fields flfirst, fllast, and
484 flcount designate which entries are currently active. Entry
485 space is allocated in a circular manner within the AGFL
486 block. Fields defined:
487 bno array of all block numbers. Even those which
488 are not active are printed.
489 agi The AGI block is the header for inode allocation information;
491 it is in the third 512-byte block of each allocation group.
492 Fields defined:
493 magicnum AGI block magic number, 0x58414749 ('XAGI').
494 versionnum version number, currently 1.
495 seqno sequence number starting from 0.
496 length size in filesystem blocks of the allocation
497 group.
498 count count of inodes allocated.
499 root block number of the root of the Btree holding
500 inode allocation information.
501 level number of levels in the inode allocation
502 Btree.
503 freecount count of allocated inodes that are not in
504 use.
505 newino last inode number allocated.
506 dirino unused.
507 unlinked an array of inode numbers within the alloca‐
508 tion group. The entries in the AGI block are
509 the heads of lists which run through the
510 inode next_unlinked field. These inodes are
511 to be unlinked the next time the filesystem
512 is mounted.
513 attr An attribute fork is organized as a Btree with the actual
515 data embedded in the leaf blocks. The root of the Btree is
516 found in block 0 of the fork. The index (sort order) of the
517 Btree is the hash value of the attribute name. All the
518 blocks contain a blkinfo structure at the beginning, see type
519 dir for a description. Nonleaf blocks are identical in format
520 to those for version 1 and version 2 directories, see type
521 dir for a description. Leaf blocks can refer to "local" or
522 "remote" attribute values. Local values are stored directly
523 in the leaf block. Remote values are stored in an indepen‐
524 dent block in the attribute fork (with no structure). Leaf
525 blocks contain the following fields:
526 hdr header containing a blkinfo structure info
527 (magic number 0xfbee), a count of active
528 entries, usedbytes total bytes of names and
529 values, the firstused byte in the name area,
530 holes set if the block needs compaction, and
531 array freemap as for dir leaf blocks.
532 entries array of structures containing a hashval,
533 nameidx (index into the block of the name),
534 and flags incomplete, root, and local.
535 nvlist array of structures describing the attribute
536 names and values. Fields always present: val‐
537 uelen (length of value in bytes), namelen,
538 and name. Fields present for local values:
539 value (value string). Fields present for
540 remote values: valueblk (fork block number of
541 containing the value).
542 bmapbt Files with many extents in their data or attribute fork will
544 have the extents described by the contents of a Btree for
545 that fork, instead of being stored directly in the inode.
546 Each bmap Btree starts with a root block contained within the
547 inode. The other levels of the Btree are stored in filesys‐
548 tem blocks. The blocks are linked to sibling left and right
549 blocks at each level, as well as by pointers from parent to
550 child blocks. Each block contains the following fields:
551 magic bmap Btree block magic number, 0x424d4150
552 ('BMAP').
553 level level of this block above the leaf level.
554 numrecs number of records or keys in the block.
555 leftsib left (logically lower) sibling block, 0 if
556 none.
557 rightsib right (logically higher) sibling block, 0 if
558 none.
559 recs [leaf blocks only] array of extent records.
560 Each record contains startoff, startblock,
561 blockcount, and extentflag (1 if the extent
562 is unwritten).
563 keys [nonleaf blocks only] array of key records.
564 These are the first key value of each block
565 in the level below this one. Each record con‐
566 tains startoff.
567 ptrs [nonleaf blocks only] array of child block
568 pointers. Each pointer is a filesystem block
569 number to the next level in the Btree.
570 bnobt There is one set of filesystem blocks forming the by-block-
572 number allocation Btree for each allocation group. The root
573 block of this Btree is designated by the bnoroot field in the
574 coresponding AGF block. The blocks are linked to sibling
575 left and right blocks at each level, as well as by pointers
576 from parent to child blocks. Each block has the following
577 fields:
578 magic BNOBT block magic number, 0x41425442
579 ('ABTB').
580 level level number of this block, 0 is a leaf.
581 numrecs number of data entries in the block.
582 leftsib left (logically lower) sibling block, 0 if
583 none.
584 rightsib right (logically higher) sibling block, 0 if
585 none.
586 recs [leaf blocks only] array of freespace
587 records. Each record contains startblock and
588 blockcount.
589 keys [nonleaf blocks only] array of key records.
590 These are the first value of each block in
591 the level below this one. Each record con‐
592 tains startblock and blockcount.
593 ptrs [nonleaf blocks only] array of child block
594 pointers. Each pointer is a block number
595 within the allocation group to the next level
596 in the Btree.
597 cntbt There is one set of filesystem blocks forming the by-block-
599 count allocation Btree for each allocation group. The root
600 block of this Btree is designated by the cntroot field in the
601 coresponding AGF block. The blocks are linked to sibling left
602 and right blocks at each level, as well as by pointers from
603 parent to child blocks. Each block has the following fields:
604 magic CNTBT block magic number, 0x41425443
605 ('ABTC').
606 level level number of this block, 0 is a leaf.
607 numrecs number of data entries in the block.
608 leftsib left (logically lower) sibling block, 0 if
609 none.
610 rightsib right (logically higher) sibling block, 0 if
611 none.
612 recs [leaf blocks only] array of freespace
613 records. Each record contains startblock and
614 blockcount.
615 keys [nonleaf blocks only] array of key records.
616 These are the first value of each block in
617 the level below this one. Each record con‐
618 tains blockcount and startblock.
619 ptrs [nonleaf blocks only] array of child block
620 pointers. Each pointer is a block number
621 within the allocation group to the next level
622 in the Btree.
623 data User file blocks, and other blocks whose type is unknown,
625 have this type for display purposes in xfs_db. The block
626 data is displayed in hexadecimal format.
627 dir A version 1 directory is organized as a Btree with the direc‐
629 tory data embedded in the leaf blocks. The root of the Btree
630 is found in block 0 of the file. The index (sort order) of
631 the Btree is the hash value of the entry name. All the blocks
632 contain a blkinfo structure at the beginning with the follow‐
633 ing fields:
634 forw next sibling block.
635 back previous sibling block.
636 magic magic number for this block type.
637 The non-leaf (node) blocks have the following fields:
638 hdr header containing a blkinfo structure info
639 (magic number 0xfebe), the count of active
640 entries, and the level of this block above
641 the leaves.
642 btree array of entries containing hashval and
643 before fields. The before value is a block
644 number within the directory file to the child
645 block, the hashval is the last hash value in
646 that block.
647 The leaf blocks have the following fields:
648 hdr header containing a blkinfo structure info
649 (magic number 0xfeeb), the count of active
650 entries, namebytes (total name string bytes),
651 holes flag (block needs compaction), and
652 freemap (array of base, size entries for free
653 regions).
654 entries array of structures containing hashval,
655 nameidx (byte index into the block of the
656 name string), and namelen.
657 namelist array of structures containing inumber and
658 name.
659 dir2 A version 2 directory has four kinds of blocks. Data blocks
661 start at offset 0 in the file. There are two kinds of data
662 blocks: single-block directories have the leaf information
663 embedded at the end of the block, data blocks in multi-block
664 directories do not. Node and leaf blocks start at offset
665 32GiB (with either a single leaf block or the root node
666 block). Freespace blocks start at offset 64GiB. The node
667 and leaf blocks form a Btree, with references to the data in
668 the data blocks. The freespace blocks form an index of long‐
669 est free spaces within the data blocks.
670 A single-block directory block contains the following fields:
672 bhdr header containing magic number 0x58443242
673 ('XD2B') and an array bestfree of the longest
674 3 free spaces in the block (offset, length).
675 bu array of union structures. Each element is
676 either an entry or a freespace. For entries,
677 there are the following fields: inumber,
678 namelen, name, and tag. For freespace, there
679 are the following fields: freetag [22m(0xffff),
680 length, and tag. The tag value is the byte
681 offset in the block of the start of the entry
682 it is contained in.
683 bleaf array of leaf entries containing hashval and
684 address. The address is a 64-bit word offset
685 into the file.
686 btail tail structure containing the total count of
687 leaf entries and stale count of unused leaf
688 entries.
689 A data block contains the following fields:
690 dhdr header containing magic number 0x58443244
691 ('XD2D') and an array bestfree of the longest
692 3 free spaces in the block (offset, length).
693 du array of union structures as for bu.
694 Leaf blocks have two possible forms. If the Btree consists of
695 a single leaf then the freespace information is in the leaf
696 block, otherwise it is in separate blocks and the root of the
697 Btree is a node block. A leaf block contains the following
698 fields:
699 lhdr header containing a blkinfo structure info
700 (magic number 0xd2f1 for the single leaf
701 case, 0xd2ff for the true Btree case), the
702 total count of leaf entries, and stale count
703 of unused leaf entries.
704 lents leaf entries, as for bleaf.
705 lbests [single leaf only] array of values which rep‐
706 resent the longest freespace in each data
707 block in the directory.
708 ltail [single leaf only] tail structure containing
709 bestcount count of lbests.
710 A node block is identical to that for types attr and dir.
711 A freespace block contains the following fields:
713 fhdr header containing magic number 0x58443246
714 ('XD2F'), firstdb first data block number
715 covered by this freespace block, nvalid num‐
716 ber of valid entries, and nused number of
717 entries representing real data blocks.
718 fbests array of values as for lbests.
719 dqblk The quota information is stored in files referred to by the
721 superblock uquotino and pquotino fields. Each filesystem
722 block in a quota file contains a constant number of quota
723 entries. The quota entry size is currently 136 bytes, so with
724 a 4KiB filesystem block size there are 30 quota entries per
725 block. The dquot command is used to locate these entries in
726 the filesystem. The file entries are indexed by the user or
727 project identifier to determine the block and offset. Each
728 quota entry has the following fields:
729 magic magic number, 0x4451 ('DQ').
730 version version number, currently 1.
731 flags flags, values include 0x01 for user quota,
732 0x02 for project quota.
733 id user or project identifier.
734 blk_hardlimit absolute limit on blocks in use.
735 blk_softlimit preferred limit on blocks in use.
736 ino_hardlimit absolute limit on inodes in use.
737 ino_softlimit preferred limit on inodes in use.
738 bcount blocks actually in use.
739 icount inodes actually in use.
740 itimer time when service will be refused if soft
741 limit is violated for inodes.
742 btimer time when service will be refused if soft
743 limit is violated for blocks.
744 iwarns number of warnings issued about inode
745 limit violations.
746 bwarns number of warnings issued about block
747 limit violations.
748 rtb_hardlimit absolute limit on realtime blocks in use.
749 rtb_softlimit preferred limit on realtime blocks in use.
750 rtbcount realtime blocks actually in use.
751 rtbtimer time when service will be refused if soft
752 limit is violated for realtime blocks.
753 rtbwarns number of warnings issued about realtime
754 block limit violations.
755 inobt There is one set of filesystem blocks forming the inode allo‐
757 cation Btree for each allocation group. The root block of
758 this Btree is designated by the root field in the corespond‐
759 ing AGI block. The blocks are linked to sibling left and
760 right blocks at each level, as well as by pointers from par‐
761 ent to child blocks. Each block has the following fields:
762 magic INOBT block magic number, 0x49414254
763 ('IABT').
764 level level number of this block, 0 is a leaf.
765 numrecs number of data entries in the block.
766 leftsib left (logically lower) sibling block, 0 if
767 none.
768 rightsib right (logically higher) sibling block, 0 if
769 none.
770 recs [leaf blocks only] array of inode records.
771 Each record contains startino allocation-
772 group relative inode number, freecount count
773 of free inodes in this chunk, and free bit‐
774 map, LSB corresponds to inode 0.
775 keys [nonleaf blocks only] array of key records.
776 These are the first value of each block in
777 the level below this one. Each record con‐
778 tains startino.
779 ptrs [nonleaf blocks only] array of child block
780 pointers. Each pointer is a block number
781 within the allocation group to the next level
782 in the Btree.
783 inode Inodes are allocated in "chunks" of 64 inodes each. Usually a
785 chunk is multiple filesystem blocks, although there are cases
786 with large filesystem blocks where a chunk is less than one
787 block. The inode Btree (see inobt above) refers to the inode
788 numbers per allocation group. The inode numbers directly
789 reflect the location of the inode block on disk. Use the
790 inode command to point xfs_db to a specific inode. Each inode
791 contains four regions: core, next_unlinked, u, and a. core
792 contains the fixed information. next_unlinked is separated
793 from the core due to journaling considerations, see type agi
794 field unlinked. u is a union structure that is different in
795 size and format depending on the type and representation of
796 the file data ("data fork"). a is an optional union struc‐
797 ture to describe attribute data, that is different in size,
798 format, and location depending on the presence and represen‐
799 tation of attribute data, and the size of the u data
800 ("attribute fork"). xfs_db automatically selects the proper
801 union members based on information in the inode.
802 The following are fields in the inode core:
804 magic inode magic number, 0x494e ('IN').
805 mode mode and type of file, as described in
806 chmod(2), mknod(2), and stat(2).
807 version inode version, 1 or 2.
808 format format of u union data (0: xfs_dev_t, 1:
809 local file - in-inode directory or symlink,
810 2: extent list, 3: Btree root, 4: unique id
811 [unused]).
812 nlinkv1 number of links to the file in a version 1
813 inode.
814 nlinkv2 number of links to the file in a version 2
815 inode.
816 projid_lo owner's project id (low word; version 2 inode
817 only). projid_hi owner's project id (high
818 word; version 2 inode only).
819 uid owner's user id.
820 gid owner's group id.
821 atime time last accessed (seconds and nanoseconds).
822 mtime time last modified.
823 ctime time created or inode last modified.
824 size number of bytes in the file.
825 nblocks total number of blocks in the file including
826 indirect and attribute.
827 extsize basic/minimum extent size for the file.
828 nextents number of extents in the data fork.
829 naextents number of extents in the attribute fork.
830 forkoff attribute fork offset in the inode, in 64-bit
831 words from the start of u.
832 aformat format of a data (1: local attribute data, 2:
833 extent list, 3: Btree root).
834 dmevmask DMAPI event mask.
835 dmstate DMAPI state information.
836 newrtbm file is the realtime bitmap and is "new" for‐
837 mat.
838 prealloc file has preallocated data space after EOF.
839 realtime file data is in the realtime subvolume.
840 gen inode generation number.
841 The following fields are in the u data fork union:
842 bmbt bmap Btree root. This looks like a bmapbtd
843 block with redundant information removed.
844 bmx array of extent descriptors.
845 dev dev_t for the block or character device.
846 sfdir shortform (in-inode) version 1 directory.
847 This consists of a hdr containing the parent
848 inode number and a count of active entries in
849 the directory, followed by an array list of
850 hdr.count entries. Each such entry contains
851 inumber, namelen, and name string.
852 sfdir2 shortform (in-inode) version 2 directory.
853 This consists of a hdr containing a count of
854 active entries in the directory, an i8count
855 of entries with inumbers that don't fit in a
856 32-bit value, and the parent inode number,
857 followed by an array list of hdr.count
858 entries. Each such entry contains namelen, a
859 saved offset used when the directory is con‐
860 verted to a larger form, a name string, and
861 the inumber.
862 symlink symbolic link string value.
863 The following fields are in the a attribute fork union if it
864 exists:
865 bmbt bmap Btree root, as above.
866 bmx array of extent descriptors.
867 sfattr shortform (in-inode) attribute values. This
868 consists of a hdr containing a totsize (total
869 size in bytes) and a count of active entries,
870 followed by an array list of hdr.count
871 entries. Each such entry contains namelen,
872 valuelen, root flag, name, and value.
873 log Log blocks contain the journal entries for XFS. It's not
875 useful to examine these with xfs_db, use xfs_logprint(8)
876 instead.
877 rtbitmap If the filesystem has a realtime subvolume, then the rbmino
879 field in the superblock refers to a file that contains the
880 realtime bitmap. Each bit in the bitmap file controls the
881 allocation of a single realtime extent (set == free). The
882 bitmap is processed in 32-bit words, the LSB of a word is
883 used for the first extent controlled by that bitmap word. The
884 atime field of the realtime bitmap inode contains a counter
885 that is used to control where the next new realtime file will
886 start.
887 rtsummary If the filesystem has a realtime subvolume, then the rsumino
889 field in the superblock refers to a file that contains the
890 realtime summary data. The summary file contains a two-dimen‐
891 sional array of 16-bit values. Each value counts the number
892 of free extent runs (consecutive free realtime extents) of a
893 given range of sizes that starts in a given bitmap block.
894 The size ranges are binary buckets (low size in the bucket is
895 a power of 2). There are as many size ranges as are neces‐
896 sary given the size of the realtime subvolume. The first
897 dimension is the size range, the second dimension is the
898 starting bitmap block number (adjacent entries are for the
899 same size, adjacent bitmap blocks).
900 sb There is one sb (superblock) structure per allocation group.
902 It is the first disk block in the allocation group. Only the
903 first one (block 0 of the filesystem) is actually used; the
904 other blocks are redundant information for xfs_repair(8) to
905 use if the first superblock is damaged. Fields defined:
906 magicnum superblock magic number, 0x58465342 ('XFSB').
907 blocksize filesystem block size in bytes.
908 dblocks number of filesystem blocks present in the
909 data subvolume.
910 rblocks number of filesystem blocks present in the
911 realtime subvolume.
912 rextents number of realtime extents that rblocks con‐
913 tain.
914 uuid unique identifier of the filesystem.
915 logstart starting filesystem block number of the log
916 (journal). If this value is 0 the log is
917 "external".
918 rootino root inode number.
919 rbmino realtime bitmap inode number.
920 rsumino realtime summary data inode number.
921 rextsize realtime extent size in filesystem blocks.
922 agblocks size of an allocation group in filesystem
923 blocks.
924 agcount number of allocation groups.
925 rbmblocks number of realtime bitmap blocks.
926 logblocks number of log blocks (filesystem blocks).
927 versionnum filesystem version information. This value
928 is currently 1, 2, 3, or 4 in the low 4 bits.
929 If the low bits are 4 then the other bits
930 have additional meanings. 1 is the original
931 value. 2 means that attributes were used. 3
932 means that version 2 inodes (large link
933 counts) were used. 4 is the bitmask version
934 of the version number. In this case, the
935 other bits are used as flags (0x0010:
936 attributes were used, 0x0020: version 2
937 inodes were used, 0x0040: quotas were used,
938 0x0080: inode cluster alignment is in force,
939 0x0100: data stripe alignment is in force,
940 0x0200: the shared_vn field is used, 0x1000:
941 unwritten extent tracking is on, 0x2000: ver‐
942 sion 2 directories are in use).
943 sectsize sector size in bytes, currently always 512.
944 This is the size of the superblock and the
945 other header blocks.
946 inodesize inode size in bytes.
947 inopblock number of inodes per filesystem block.
948 fname obsolete, filesystem name.
949 fpack obsolete, filesystem pack name.
950 blocklog log2 of blocksize.
951 sectlog log2 of sectsize.
952 inodelog log2 of inodesize.
953 inopblog log2 of inopblock.
954 agblklog log2 of agblocks (rounded up).
955 rextslog log2 of rextents.
956 inprogress mkfs.xfs(8) or xfs_copy(8) aborted before
957 completing this filesystem.
958 imax_pct maximum percentage of filesystem space used
959 for inode blocks.
960 icount number of allocated inodes.
961 ifree number of allocated inodes that are not in
962 use.
963 fdblocks number of free data blocks.
964 frextents number of free realtime extents.
965 uquotino user quota inode number.
966 pquotino project quota inode number; this is currently
967 unused.
968 qflags quota status flags (0x01: user quota account‐
969 ing is on, 0x02: user quota limits are
970 enforced, 0x04: quotacheck has been run on
971 user quotas, 0x08: project quota accounting
972 is on, 0x10: project quota limits are
973 enforced, 0x20: quotacheck has been run on
974 project quotas).
975 flags random flags. 0x01: only read-only mounts are
976 allowed.
977 shared_vn shared version number (shared readonly
978 filesystems).
979 inoalignmt inode chunk alignment in filesystem blocks.
980 unit stripe or RAID unit.
981 width stripe or RAID width.
982 dirblklog log2 of directory block size (filesystem
983 blocks).
984 symlink Symbolic link blocks are used only when the symbolic link
986 value does not fit inside the inode. The block content is
987 just the string value. Bytes past the logical end of the
988 symbolic link value have arbitrary values.
989 text User file blocks, and other blocks whose type is unknown,
991 have this type for display purposes in xfs_db. The block
992 data is displayed in two columns: Hexadecimal format and
993 printable ASCII chars.
994
996 Many messages can come from the check (blockget) command; these are
997 documented in xfs_check(8).
998
1000 mkfs.xfs(8), xfs_admin(8), xfs_check(8), xfs_copy(8), xfs_logprint(8),
1001 xfs_metadump(8), xfs_ncheck(8), xfs_repair(8), mount(8), chmod(2),
1002 mknod(2), stat(2), xfs(5).
1003 xfs_db(8)