1 .\" Copyright (c) 2007 The DragonFly Project. All rights reserved.
3 .\" This code is derived from software contributed to The DragonFly Project
4 .\" by Matthew Dillon <dillon@backplane.com>
6 .\" Redistribution and use in source and binary forms, with or without
7 .\" modification, are permitted provided that the following conditions
10 .\" 1. Redistributions of source code must retain the above copyright
11 .\" notice, this list of conditions and the following disclaimer.
12 .\" 2. Redistributions in binary form must reproduce the above copyright
13 .\" notice, this list of conditions and the following disclaimer in
14 .\" the documentation and/or other materials provided with the
16 .\" 3. Neither the name of The DragonFly Project nor the names of its
17 .\" contributors may be used to endorse or promote products derived
18 .\" from this software without specific, prior written permission.
20 .\" THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
21 .\" ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
22 .\" LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
23 .\" FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
24 .\" COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
25 .\" INCIDENTAL, SPECIAL, EXEMPLARY OR CONSEQUENTIAL DAMAGES (INCLUDING,
26 .\" BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
27 .\" LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
28 .\" AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
29 .\" OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
30 .\" OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
33 .\" $DragonFly: src/sbin/hammer/hammer.8,v 1.58 2008/11/13 02:04:27 dillon Exp $
40 .Nd HAMMER file system utility
46 .Op Fl f Ar blkdev[:blkdev]*
47 .\" .Op Fl s Ar linkpath
53 This manual page documents the
55 utility which provides miscellaneous functions related to managing a
58 For a general introduction to the
60 file system, its features, and
61 examples on how to set up and maintain one, see
64 The options are as follows:
65 .Bl -tag -width indent
69 Tell the mirror commands to use a 2-way protocol, which allows
70 automatic negotiation of transaction id ranges.
71 This option is automatically enabled by the
75 Specify a bandwidth limit in bytes per second for mirroring streams.
76 This option is typically used to prevent batch mirroring operations from
77 loading down the machine.
78 The bandwidth may be suffixed with
83 to specify values in kilobytes, megabytes, and gigabytes per second.
84 If no suffix is specified, bytes per second is assumed.
86 When pruning and reblocking you can instruction
88 to start at the object id stored in the specified file.
89 If the file does not exist
91 will start at the beginning.
95 specific period of time and is unable to complete the operation it will
96 write out the current object id so the next run can pick up where it left off.
99 runs to completion it will delete
101 .It Fl f Ar blkdev[:blkdev]*
102 Specify the volumes making up a
106 When maintaining a streaming mirroring this option specifies the
107 minimum delay after a batch ends before the next batch is allowed
109 The default is five seconds.
111 Decrease verboseness.
112 May be specified multiple times.
114 Specify recursion for those commands which support it.
116 When pruning and reblocking you can tell the utility to stop after a
117 certain period of time.
118 This option is used along with the
120 option to prune or reblock a portion of the file system incrementally.
122 Increase verboseness.
123 May be specified multiple times.
125 Force "yes" for any interactive question.
128 The commands are as follows:
129 .Bl -tag -width indent
130 .\" ==== synctid ====
131 .It Ar synctid Ar filesystem Op quick
132 Generates a guaranteed, formal 64 bit transaction id representing the
133 current state of the specified
136 The file system will be synced to the media.
140 keyword is specified the file system will be soft-synced, meaning that a
141 crash might still undo the state of the file system as of the transaction
142 id returned but any new modifications will occur after the returned
143 transaction id as expected.
145 .It Ar bstats Op interval
148 B-tree statistics until interrupted.
151 seconds between each display.
152 The default interval is one second.
153 .\" ==== iostats ====
154 .It Ar iostats Op interval
157 I/O statistics until interrupted.
160 seconds between each display.
161 The default interval is one second.
162 .\" ==== history ====
163 .It Ar history Ar path ...
164 Show the modification history for
166 file's inode and data.
170 This command needs the
174 .\" Dump the B-tree, record, large-data, and small-data blockmaps, showing
175 .\" physical block assignments and free space percentages.
176 .\" ==== namekey1 ====
177 .It Ar namekey1 Ar filename
180 64 bit directory hash for the specified file name, using
181 the original directory hash algorithm in version 1 of the filesystem.
182 The low 32 bits are used as an iterator for hash collisions and will be
184 .\" ==== namekey2 ====
185 .It Ar namekey2 Ar filename
188 64 bit directory hash for the specified file name, using
189 the new directory hash algorithm in version 2 of the filesystem.
190 The low 32 bits are still used as an iterator but will start out containing
191 part of the hash key.
192 .\" ==== namekey32 ====
193 .It Ar namekey32 Ar filename
194 Generate the top 32 bits of a
196 64 bit directory hash for the specified file name.
197 .\" ==== cleanup ====
198 .It Ar cleanup Op Ar filesystem ...
199 This is a meta-command which executes snapshot, pruning, and reblocking
200 commands on the specified
205 is specified this command will clean-up all
207 file systems in use, including PFS's.
208 To do this it will scan all
212 mounts, extract PFS id's, and clean-up each PFS found.
214 This command will by default access a
220 creating them if necessary.
221 The format of the configuration file is:
222 .Bd -literal -offset indent
223 snapshots <period> <retention-time> [any]
224 prune <period> <max-runtime>
225 rebalance <period> <max-runtime>
226 reblock <period> <1/3 max-runtime>
227 recopy <period> <1/3 max-runtime>
230 snapshots 1d 60d # 0d 60d for PFS /tmp, /var/tmp, /usr/obj
237 Time is given with a suffix of
243 meaning day, hour, minute and second.
245 If the snapshots directive has a period of 0 and a retention time of 0
246 then snapshot generation is disabled, removal of old snapshots are
247 disabled, and prunes will use
248 .Ar prune-everything .
249 If the snapshots directive has a period of 0 but a non-zero retention time
250 then this command will not create any new snapshots but will remove old
251 snapshots it finds based on the retention time.
253 By default only snapshots in the form: snap-yyyymmdd[-hhmm] are processed.
256 directive is specified as a third argument on the snapshots config line
257 then any softlink of the form *[- or .]yyyymmdd[-hhmm] will be processed.
259 A prune max-runtime of 0 means unlimited.
261 If period hasn't passed since the previous
264 For example a day has passed when midnight is passed (localtime).
272 The default configuration file will create a daily snapshot, do a daily
273 pruning and reblocking run and a monthly recopy run.
274 Reblocking is defragmentation with a level of 95%,
275 and recopy is full defragmentation.
277 By default prune and reblock operations are limited to 5 minutes per function,
278 and recopy operations are limited to 10 minutes per function.
279 Reblocking and recopy runs are each broken down into three separate functions
280 (btree, inodes and data)
281 and are thus by default limited to 15 and 30 minutes respectively.
282 Also note that this directive will by default disable snapshots on
289 The defaults may be adjusted by modifying the
292 The pruning and reblocking commands automatically maintain a cyclefile
293 for incremental operation.
294 If you interrupt (^C) the program the cyclefile will be updated, but a sub-command
295 may continue to run in the background for a few seconds until the
297 ioctl detects the interrupt.
300 PFS option can be set to use another location for the snapshots directory.
302 Work on this command is still in progress.
303 Expected additions: An ability to remove snapshots dynamically as the
304 file system becomes full.
305 .\" ==== snapshot ====
306 .It Ar snapshot Oo Ar filesystem Oc Ar snapshot-dir
307 Takes a snapshot of the file system either explicitly given by
309 or implicitly derived from the
311 argument and creates a symlink in the directory provided by
313 pointing to the snapshot.
316 is not a directory, it is assumed to be a format string passed to
318 with the current time as parameter.
321 refers to an existing directory, a default format string of "snap-%Y%d%m-%H%M"
322 is assumed and used as name for the newly created symlink.
324 Snapshot is a per PFS operation, so a
326 file system and each PFS in it have to be snapshot separately.
328 Example, assuming that
334 is a file system on its own, the following invocations:
335 .Bd -literal -offset indent
336 hammer snapshot /mysnapshots
338 hammer snapshot /mysnapshots/%Y-%m-%d
340 hammer snapshot /obj /mysnapshots/obj-%Y-%m-%d
343 would create symlinks similar to:
344 .Bd -literal -offset indent
345 /mysnapshots/snap-20080627-1210 -> /@@0x10d2cd05b7270d16
347 /mysnapshots/2008-06-27 -> /@@0x10d2cd05b7270d16
349 /mysnapshots/obj-2008-06-27 -> /obj@@0x10d2cd05b7270d16
352 .It Ar prune Ar softlink-dir
353 Prune the file system based on previously created snapshot softlinks.
354 Pruning is the act of deleting file system history.
358 will delete file system history such that
359 the file system state is retained for the given snapshots,
360 and all history after the latest snapshot,
361 but all other history is deleted.
363 The target directory is expected to contain softlinks pointing to
364 snapshots of the file systems you wish to retain.
365 The directory is scanned non-recursively and the mount points and
366 transaction ids stored in the softlinks are extracted and sorted.
367 The file system is then explicitly pruned according to what is found.
368 Cleaning out portions of the file system is as simple as removing a softlink
373 As a safety measure pruning only occurs if one or more softlinks are found
374 containing the @@ snapshot id extension.
375 Currently the scanned softlink directory must contain softlinks pointing
379 The softlinks may specify absolute or relative paths.
380 Softlinks must use 20-character (@@0x%016llx) transaction ids,
381 as might be returned from
382 .Dq Nm Ar synctid filesystem .
384 Pruning is a per PFS operation, so a
386 file system and each PFS in it have to be pruned separately.
388 Note that pruning a file system may not immediately free-up space,
389 though typically some space will be freed if a large number of records are
391 The file system must be reblocked to completely recover all available space.
393 Example, lets say your snapshot directory contains the following links:
394 .Bd -literal -offset indent
395 lrwxr-xr-x 1 root wheel 29 May 31 17:57 snap1 ->
396 /usr/obj/@@0x10d2cd05b7270d16
398 lrwxr-xr-x 1 root wheel 29 May 31 17:58 snap2 ->
399 /usr/obj/@@0x10d2cd13f3fde98f
401 lrwxr-xr-x 1 root wheel 29 May 31 17:59 snap3 ->
402 /usr/obj/@@0x10d2cd222adee364
405 If you were to run the
407 command on this directory, then the
410 mount will be pruned to retain the above three snapshots.
411 In addition, history for modifications made to the file system older than
412 the oldest snapshot will be destroyed and history for potentially fine-grained
413 modifications made to the file system more recently than the most recent
414 snapshot will be retained.
416 If you then delete the
418 softlink and rerun the
421 history for modifications pertaining to that snapshot would be destroyed.
422 .\" ==== prune-everything ====
423 .It Ar prune-everything Ar filesystem
424 This command will remove all historical records from the file system.
425 This directive is not normally used on a production system.
426 .\" ==== rebalance ====
427 .It Ar rebalance Ar filesystem Op Ar saturation_level
428 This command will rebalance the B-Tree, nodes with small numbers of
429 elements will be combined and element counts will be smoothed out
432 The saturation level is a percentage between 50 and 100. The default
434 .\" ==== reblock ====
435 .It Ar reblock Ar filesystem Op Ar fill_percentage
436 .It Ar reblock-btree Ar filesystem Op Ar fill_percentage
437 .It Ar reblock-inodes Ar filesystem Op Ar fill_percentage
438 .It Ar reblock-dirs Ar filesystem Op Ar fill_percentage
439 .It Ar reblock-data Ar filesystem Op Ar fill_percentage
440 Attempt to defragment and free space for reuse by reblocking a live
443 Big blocks cannot be reused by
445 until they are completely free.
446 This command also has the effect of reordering all elements, effectively
447 defragmenting the file system.
449 The default fill percentage is 100% and will cause the file system to be
450 completely defragmented.
451 All specified element types will be reallocated and rewritten.
452 If you wish to quickly free up space instead try specifying
453 a smaller fill percentage, such as 90% or 80% (the
455 suffix is not needed).
457 Since this command may rewrite the entire contents of the disk it is
458 best to do it incrementally from a
464 options to limit the run time.
465 The file system would thus be defragmented over long period of time.
467 It is recommended that separate invocations be used for each data type.
468 B-tree nodes, inodes, and directories are typically the most important
469 elements needing defragmentation.
470 Data can be defragmented over a longer period of time.
472 Reblocking is a per PFS operation, so a
474 file system and each PFS in it have to be reblocked separately.
475 .\" ==== pfs-status ====
476 .It Ar pfs-status Ar dirpath ...
477 Retrieve the mirroring configuration parameters for the specified
479 file systems or pseudo-filesystems (PFS's).
480 .\" ==== pfs-master ====
481 .It Ar pfs-master Ar dirpath Op options
482 Create a pseudo-filesystem (PFS) inside a
485 Up to 65535 such file systems can be created.
486 Each PFS uses an independent inode numbering space making it suitable
487 for use as a replication source or target.
491 directive creates a PFS that you can read, write, and use as a mirroring
494 It is recommended to use a
496 mount to access a PFS, for more information see
498 .\" ==== pfs-slave ====
499 .It Ar pfs-slave Ar dirpath Op options
500 Create a pseudo-filesystem (PFS) inside a
503 Up to 65535 such file systems can be created.
504 Each PFS uses an independent inode numbering space making it suitable
505 for use as a replication source or target.
509 directive creates a PFS that you can use as a mirroring target.
510 You will not be able to access a slave PFS until you have completed the
511 first mirroring operation with it as the target (its root directory will
512 not exist until then).
514 Access to the pfs-slave via the special softlink,
520 dynamically modify the snapshot transaction id by returning a dynamic result
525 A PFS can only be truly destroyed with the
528 Removing the softlink will not destroy the underlying PFS.
530 It is recommended to use a
532 mount to access a PFS, for more information see
534 .\" ==== pfs-update ====
535 .It Ar pfs-update Ar dirpath Op options
536 Update the configuration parameters for an existing
539 or pseudo-filesystem.
540 Options that may be specified:
541 .Bl -tag -width indent
542 .It sync-beg-tid=0x16llx
543 This is the automatic snapshot access starting transaction id for
545 This parameter is normally updated automatically by the
549 It is important to note that accessing a mirroring slave
550 with a transaction id greater than the last fully synchronized transaction
551 id can result in an unreliable snapshot since you will be accessing
552 data that is still undergoing synchronization.
554 Manually modifying this field is dangerous and can result in a broken
556 .It sync-end-tid=0x16llx
557 This is the current synchronization point for mirroring slaves.
558 This parameter is normally updated automatically by the
562 Manually modifying this field is dangerous and can result in a broken mirror.
563 .It shared-uuid=<uuid>
564 Set the shared UUID for this file system.
565 All mirrors must have the same shared UUID.
566 For safety purposes the
568 directives will refuse to operate on a target with a different shared UUID.
570 Changing the shared UUID on an existing, non-empty mirroring target,
571 including an empty but not completely pruned target,
572 can lead to corruption of the mirroring target.
573 .It unique-uuid=<uuid>
574 Set the unique UUID for this file system.
575 This UUID should not be used anywhere else,
576 even on exact copies of the file system.
578 Set a descriptive label for this file system.
579 .It snapshots=<string>
580 Specify the snapshots directory which
583 will use to manage this PFS.
584 The snapshots directory does not need to be configured for
585 PFS masters and will default to
586 .Pa <pfs>/snapshots .
588 PFS slaves are mirroring slaves so you cannot configure a snapshots
589 directory on the slave itself to be managed by the slave's machine.
590 In fact, the slave will likely have a
592 sub-directory mirrored
593 from the master, but that directory contains the configuration the master
594 is using for its copy of the file system, not the configuration that we
595 want to use for our slave.
597 It is recommended that
598 .Pa <fs>/var/slaves/<name>
599 be configured for a PFS slave, where
605 is an appropriate label.
606 You can control snapshot retention on your slave independent of the master.
608 Zero out the snapshots directory path for this PFS.
610 .\" ==== pfs-upgrade ====
611 .It Ar pfs-upgrade Ar dirpath
612 Upgrade a PFS from slave to master operation.
613 The PFS will be rolled back to the current end synchronization tid
614 (removing any partial synchronizations), and will then become writable.
618 currently supports only single masters and using
619 this command can easily result in file system corruption
620 if you don't know what you are doing.
622 This directive will refuse to run if any programs have open descriptors
623 in the PFS, including programs chdir'd into the PFS.
624 .\" ==== pfs-downgrade ====
625 .It Ar pfs-downgrade Ar dirpath
626 Downgrade a master PFS from master to slave operation
627 The PFS becomes read-only and access will be locked to its
630 This directive will refuse to run if any programs have open descriptors
631 in the PFS, including programs chdir'd into the PFS.
632 .\" ==== pfs-destroy ====
633 .It Ar pfs-destroy Ar dirpath
634 This permanently destroys a PFS.
636 This directive will refuse to run if any programs have open descriptors
637 in the PFS, including programs chdir'd into the PFS.
638 .\" ==== mirror-read ====
639 .It Ar mirror-read Ar filesystem Op Ar <begin-tid>
640 Generate a mirroring stream to stdout.
641 The stream ends when the transaction id space has been exhausted.
642 .\" ==== mirror-read-stream ====
643 .It Ar mirror-read-stream Ar filesystem Op Ar <begin-tid>
644 Generate a mirroring stream to stdout.
645 Upon completion the stream is paused until new data is synced to the
646 master, then resumed.
647 Operation continues until the pipe is broken.
648 .\" ==== mirror-write ====
649 .It Ar mirror-write Ar filesystem
650 Take a mirroring stream on stdin.
652 This command will fail if the
654 configuration field for the two file systems do not match.
656 If the target PFS does not exist this command will ask you whether
657 you want to create a compatible PFS slave for the target or not.
658 .\" ==== mirror-dump ====
664 to dump an ASCII representation of the mirroring stream.
665 .\" ==== mirror-copy ====
666 .It Ar mirror-copy Ar [[user@]host:]filesystem Ar [[user@]host:]filesystem
667 This is a shortcut which pipes a
672 If a remote host specification is made the program forks a
678 on the appropriate host.
679 The source may be a master or slave PFS, and the target must be a slave PFS.
681 This command also established full duplex communication and turns on
682 the two-way protocol feature which automatically negotiates transaction id
683 ranges without having to use a cyclefile.
684 If the operation completes successfully the target PFS's
687 Note that you must re-chdir into the target PFS to see the updated information.
688 If you do not you will still be in the previous snapshot.
690 If the target PFS does not exist this command will ask you whether
691 you want to create a compatible PFS slave for the target or not.
692 .\" ==== mirror-stream ====
693 .It Ar mirror-stream Ar [[user@]host:]filesystem Ar [[user@]host:]filesystem
694 This command works similarly to
696 but does not exit unless the pipe is broken.
697 This command will resume the mirroring operation whenever the master is synced.
698 The command is commonly used with
702 options to keep the mirroring target in sync with the source on a continuing
704 .\" ==== version ====
705 .It Ar version Ar filesystem
706 This command returns the
708 filesystem version for the specified
709 filesystem as well as the range of versions supported in the kernel.
712 option may be used to remove the summary at the end.
713 .\" ==== version-upgrade ====
714 .It Ar version-upgrade Ar filesystem Ar version Op Ar force
715 This command upgrades the
717 filesystem to the specified version.
718 Once upgraded a filesystem may not be downgraded.
719 If you wish to upgrade a filesystem to a version greater or equal to the
720 work-in-progress version number you must specify the
723 Use of WIP versions should be relegated to testing and may require wiping
724 the filesystem as development progresses, even though the WIP version might
727 NOTE! This command operates on the entire
729 filesystem and is not a per-PFS operation.
730 All PFS's will be affected.
731 .Bl -tag -width indent
734 default version, first
738 Work-in-progress version.
739 This version is developing a new directory hash key.
743 .Sh PSEUDO FILESYSTEM (PFS) NOTES
744 The root of a PFS is not hooked into the primary
746 file system as a directory.
749 creates a special softlink called "@@PFS%05d" (exactly 10 characters long)
754 then modifies the contents of the softlink as read by
756 and thus what you see with an
758 command or if you were to
761 If the PFS is a master the link reflects the current state of the PFS.
762 If the PFS is a slave the link reflects the last completed snapshot, and the
763 contents of the link will change when the next snapshot is completed, and
766 PFS support is currently very new and experimental.
769 utility employs numerous safeties to reduce user foot-shooting.
772 directive requires that the target be configured as a slave and that the
774 field of the mirroring source and target match.
776 .Bl -tag -width ".It Pa <fs>/var/slaves/<name>" -compact
778 default per PFS snapshots directory
779 .It Pa <snapshots>/config
783 .It Pa <fs>/var/slaves/<name>
784 recommended slave PFS snapshots directory
791 .Xr periodic.conf 5 ,
798 utility first appeared in
801 .An Matthew Dillon Aq dillon@backplane.com