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
47 .Op Fl C Ar cachesize Ns Op Ns Cm \&: Ns Ar readahead
50 .\" .Op Fl s Ar linkpath
57 This manual page documents the
59 utility which provides miscellaneous functions related to managing a
62 For a general introduction to the
64 file system, its features, and
65 examples on how to set up and maintain one, see
68 The options are as follows:
69 .Bl -tag -width indent
73 Tell the mirror commands to use a 2-way protocol, which allows
74 automatic negotiation of transaction id ranges.
75 This option is automatically enabled by the
81 will not attempt to break-up large initial bulk transfers into smaller pieces.
82 This can save time but if the link is lost in the middle of the
83 initial bulk transfer you will have to start over from scratch.
85 Specify a bandwidth limit in bytes per second for mirroring streams.
86 This option is typically used to prevent batch mirroring operations from
87 loading down the machine.
88 The bandwidth may be suffixed with
92 to specify values in kilobytes, megabytes, and gigabytes per second.
93 If no suffix is specified, bytes per second is assumed.
94 .It Fl C Ar cachesize Ns Op Ns Cm \&: Ns Ar readahead
95 Set the memory cache size for any raw
102 for megabytes is allowed,
103 else the cache size is specified in bytes.
105 The read-behind/read-ahead defaults to 4
109 This option is typically only used with diagnostic commands
110 as kernel-supported commands will use the kernel's buffer cache.
111 .It Fl c Ar cyclefile
112 When pruning and reblocking you can instruction
114 to start at the object id stored in the specified file.
115 If the file does not exist
117 will start at the beginning.
121 specific period of time and is unable to complete the operation it will
122 write out the current object id so the next run can pick up where it left off.
125 runs to completion it will delete
128 Specify the volumes making up a
132 is a colon-separated list of devices, each specifying a
136 When maintaining a streaming mirroring this option specifies the
137 minimum delay after a batch ends before the next batch is allowed
139 The default is five seconds.
141 This passes the -p <port> option to ssh when using a remote
142 specification for the source and/or destination.
144 Decrease verboseness.
145 May be specified multiple times.
147 Specify recursion for those commands which support it.
149 When pruning and reblocking you can tell the utility to stop after a
150 certain period of time.
151 This option is used along with the
153 option to prune or reblock a portion of the file system incrementally.
155 Increase verboseness.
156 May be specified multiple times.
158 Enable compression for any remote ssh specifications. Unfortunately
161 option has already been reserved for other purposes so we had to use
162 a different letter. This option is typically used with the
163 mirroring directives.
165 Force "yes" for any interactive question.
168 The commands are as follows:
169 .Bl -tag -width indent
170 .\" ==== synctid ====
171 .It Cm synctid Ar filesystem Op Cm quick
172 Generates a guaranteed, formal 64 bit transaction id representing the
173 current state of the specified
176 The file system will be synced to the media.
180 keyword is specified the file system will be soft-synced, meaning that a
181 crash might still undo the state of the file system as of the transaction
182 id returned but any new modifications will occur after the returned
183 transaction id as expected.
185 This operation does not create a snapshot.
186 It is meant to be used
187 to track temporary fine-grained changes to a subset of files and
188 will only remain valid for
190 snapshot access purposes for the
192 period configured for the PFS.
193 If you desire a real snapshot then the
195 directive may be what you are looking for.
197 .It Cm bstats Op Ar interval
200 B-tree statistics until interrupted.
203 seconds between each display.
204 The default interval is one second.
205 .\" ==== iostats ====
206 .It Cm iostats Op Ar interval
210 statistics until interrupted.
213 seconds between each display.
214 The default interval is one second.
215 .\" ==== history ====
216 .It Cm history Ar path ...
217 Show the modification history for
219 file's inode and data.
220 .\" ==== blockmap ====
222 Dump the blockmap for the file system.
225 blockmap is two-layer
226 blockmap representing the maximum possible file system size of 1 Exabyte.
227 Needless to say the second layer is only present for blocks which exist.
229 blockmap represents 8-Megabyte blocks, called big-blocks.
230 Each big-block has an append
231 point, a free byte count, and a typed zone id which allows content to be
232 reverse engineered to some degree.
236 allocations essentially appended to a selected big-block using
237 the append offset and deducted from the free byte count.
238 When space is freed the free byte count is adjusted but
240 does not track holes in big-blocks for reallocation.
241 A big-block must be completely freed, either
242 through normal file system operations or through reblocking, before
245 Data blocks can be shared by deducting the space used from the free byte
246 count for each shared references, though
248 does not yet make use of this feature.
249 This means the free byte count can legally go negative.
251 This command needs the
255 .It Cm show Op Ar lo Ns Cm \&: Ns Ar objid
257 By default this command will validate all B-Tree
258 linkages and CRCs, including data CRCs, and will report the most verbose
259 information it can dig up.
260 Any errors will show up with a
262 in column 1 along with various
265 If you specify a localization and object id field,
266 .Ar lo Ns Cm \&: Ns Ar objid ,
268 search for the key printing nodes as it recurses down, and then
269 will iterate forwards.
273 the command will report less information about the inode contents.
277 the command will not report the content of the inode or other typed
282 the command will not report volume header information, big-block fill
283 ratios, mirror transaction ids, or report or check data CRCs.
284 B-tree CRCs and linkages are still checked.
286 This command needs the
289 .\" ==== show-undo ====
295 This command needs the
299 .\" Dump the B-tree, record, large-data, and small-data blockmaps, showing
300 .\" physical block assignments and free space percentages.
301 .\" ==== namekey1 ====
302 .It Cm namekey1 Ar filename
305 64 bit directory hash for the specified file name, using
306 the original directory hash algorithm in version 1 of the file system.
307 The low 32 bits are used as an iterator for hash collisions and will be
309 .\" ==== namekey2 ====
310 .It Cm namekey2 Ar filename
313 64 bit directory hash for the specified file name, using
314 the new directory hash algorithm in version 2 of the file system.
315 The low 32 bits are still used as an iterator but will start out containing
316 part of the hash key.
317 .\" ==== namekey32 ====
318 .It Cm namekey32 Ar filename
319 Generate the top 32 bits of a
321 64 bit directory hash for the specified file name.
324 Shows extended information about all the mounted
327 The information is divided into sections:
329 .It Volume identification
330 General information, like the label of the
332 filesystem, the number of volumes it contains, the FSID, and the
335 .It Big block information
336 Big block statistics, such as total, used, reserved and free big blocks.
337 .It Space information
338 Information about space used on the filesystem.
339 Currently total size, used, reserved and free space are displayed.
341 Basic information about the PFSs currently present on a
346 is the ID of the PFS, with 0 being the root PFS.
348 is the current snapshot count on the PFS.
350 displays the mount point of the PFS is currently mounted on (if any).
352 .\" ==== cleanup ====
353 .It Cm cleanup Op Ar filesystem ...
354 This is a meta-command which executes snapshot, prune, rebalance and reblock
355 commands on the specified
360 is specified this command will clean-up all
362 file systems in use, including PFS's.
363 To do this it will scan all
367 mounts, extract PFS id's, and clean-up each PFS found.
369 This command will access a snapshots
370 directory and a configuration file for each
372 creating them if necessary.
373 .Bl -tag -width indent
374 .It Nm HAMMER No version 2-
375 The configuration file is
377 in the snapshots directory which defaults to
378 .Pa <pfs>/snapshots .
379 .It Nm HAMMER No version 3+
380 The configuration file is saved in file system meta-data, see
383 The snapshots directory defaults to
384 .Pa /var/hammer/<pfs>
385 .Pa ( /var/hammer/root
389 The format of the configuration file is:
390 .Bd -literal -offset indent
391 snapshots <period> <retention-time> [any]
392 prune <period> <max-runtime>
393 rebalance <period> <max-runtime>
394 reblock <period> <max-runtime>
395 recopy <period> <max-runtime>
399 .Bd -literal -offset indent
400 snapshots 1d 60d # 0d 0d for PFS /tmp, /var/tmp, /usr/obj
407 Time is given with a suffix of
413 meaning day, hour, minute and second.
417 directive has a period of 0 and a retention time of 0
418 then snapshot generation is disabled, removal of old snapshots are
419 disabled, and prunes will use
420 .Cm prune-everything .
423 directive has a period of 0 but a non-zero retention time
424 then this command will not create any new snapshots but will remove old
425 snapshots it finds based on the retention time.
427 By default only snapshots in the form
428 .Ql snap- Ns Ar yyyymmdd Ns Op - Ns Ar HHMM
432 directive is specified as a third argument on the
434 config line then any softlink of the form
435 .Ql *- Ns Ar yyyymmdd Ns Op - Ns Ar HHMM
437 .Ql *. Ns Ar yyyymmdd Ns Op - Ns Ar HHMM
440 A prune max-runtime of 0 means unlimited.
442 If period hasn't passed since the previous
445 For example a day has passed when midnight is passed (localtime).
453 The default configuration file will create a daily snapshot, do a daily
454 pruning, rebalancing and reblocking run and a monthly recopy run.
455 Reblocking is defragmentation with a level of 95%,
456 and recopy is full defragmentation.
458 By default prune and rebalance operations are time limited to 5 minutes,
459 reblock operations to a bit over 5 minutes,
460 and recopy operations to a bit over 10 minutes.
461 Reblocking and recopy runs are each broken down into four separate functions:
462 btree, inodes, dirs and data.
463 Each function is time limited to the time given in the configuration file,
464 but the btree, inodes and dirs functions usually does not take very long time,
465 full defragmentation is always used for these three functions.
466 Also note that this directive will by default disable snapshots on
473 The defaults may be adjusted by modifying the configuration file.
474 The pruning and reblocking commands automatically maintain a cyclefile
475 for incremental operation.
476 If you interrupt (^C) the program the cyclefile will be updated,
478 may continue to run in the background for a few seconds until the
480 ioctl detects the interrupt.
483 PFS option can be set to use another location for the snapshots directory.
485 Work on this command is still in progress.
487 An ability to remove snapshots dynamically as the
488 file system becomes full.
490 .It Cm config Op Ar filesystem Op Ar configfile
493 If zero or one arguments are specified this function dumps the current
494 configuration file to stdout.
495 Zero arguments specifies the PFS containing the current directory.
496 This configuration file is stored in file system meta-data.
497 If two arguments are specified this function installs a new config file.
501 versions less than 3 the configuration file is by default stored in
502 .Pa <pfs>/snapshots/config ,
503 but in all later versions the configuration file is stored in file system
505 .\" ==== viconfig ====
506 .It Cm viconfig Op Ar filesystem
509 Edit the configuration file and reinstall into file system meta-data when done.
510 Zero arguments specifies the PFS containing the current directory.
511 .\" ==== volume-add ====
512 .It Cm volume-add Ar device Ar filesystem
513 This command will format
515 and add all of its space to
519 All existing data contained on
521 will be destroyed by this operation!
526 file system, formatting will be denied.
527 You can overcome this sanity check
530 to erase the beginning sectors of the device.
531 Also remember that you have to specify
533 together with any other device that make up the file system,
538 .\" ==== volume-del ====
539 .It Cm volume-del Ar device Ar filesystem
540 This command will remove volume
545 Remember that you have to remove
547 from the colon-separated list in
551 .\" ==== snapshot ====
552 .It Cm snapshot Oo Ar filesystem Oc Ar snapshot-dir
553 .It Cm snapshot Ar filesystem Ar snapshot-dir Op Ar note
554 Takes a snapshot of the file system either explicitly given by
556 or implicitly derived from the
558 argument and creates a symlink in the directory provided by
560 pointing to the snapshot.
563 is not a directory, it is assumed to be a format string passed to
565 with the current time as parameter.
568 refers to an existing directory, a default format string of
570 is assumed and used as name for the newly created symlink.
572 Snapshot is a per PFS operation, so a
574 file system and each PFS in it have to be snapshot separately.
576 Example, assuming that
584 are file systems on their own, the following invocations:
585 .Bd -literal -offset indent
586 hammer snapshot /mysnapshots
588 hammer snapshot /mysnapshots/%Y-%m-%d
590 hammer snapshot /obj /mysnapshots/obj-%Y-%m-%d
592 hammer snapshot /usr /my/snaps/usr "note"
595 Would create symlinks similar to:
596 .Bd -literal -offset indent
597 /mysnapshots/snap-20080627-1210 -> /@@0x10d2cd05b7270d16
599 /mysnapshots/2008-06-27 -> /@@0x10d2cd05b7270d16
601 /mysnapshots/obj-2008-06-27 -> /obj@@0x10d2cd05b7270d16
603 /my/snaps/usr/snap-20080627-1210 -> /usr@@0x10d2cd05b7270d16
608 version 3+ file system the snapshot is also recorded in file system meta-data
609 along with the optional
615 .It Cm snap Ar path Op Ar note
618 Create a snapshot for the PFS containing
620 and create a snapshot softlink.
621 If the path specified is a
622 directory a standard snapshot softlink will be created in the directory.
623 The snapshot softlink points to the base of the mounted PFS.
624 .It Cm snaplo Ar path Op Ar note
627 Create a snapshot for the PFS containing
629 and create a snapshot softlink.
630 If the path specified is a
631 directory a standard snapshot softlink will be created in the directory.
632 The snapshot softlink points into the directory it is contained in.
633 .It Cm snapq Ar dir Op Ar note
636 Create a snapshot for the PFS containing the specified directory but do
637 not create a softlink.
638 Instead output a path which can be used to access
639 the directory via the snapshot.
641 An absolute or relative path may be specified.
642 The path will be used as-is as a prefix in the path output to stdout.
644 snap and snapshot directives the snapshot transaction id will be registered
645 in the file system meta-data.
646 .It Cm snaprm Ar path Ar ...
647 .It Cm snaprm Ar transid Ar ...
648 .It Cm snaprm Ar filesystem Ar transid Ar ...
651 Remove a snapshot given its softlink or transaction id.
652 If specifying a transaction id
653 the snapshot is removed from file system meta-data but you are responsible
654 for removing any related softlinks.
656 If a softlink path is specified the filesystem and transaction id
657 is derived from the contents of the softlink.
658 If just a transaction id is specified it is assumed to be a snapshot
659 in the HAMMER filesystem you are currently chdir'd into.
660 You can also specify the filesystem and transaction id explicitly.
661 .It Cm snapls Op Ar path ...
664 Dump the snapshot meta-data for PFSs containing each
666 listing all available snapshots and their notes.
667 If no arguments are specified snapshots for the PFS containing the
668 current directory are listed.
669 This is the definitive list of snapshots for the file system.
671 .It Cm prune Ar softlink-dir
672 Prune the file system based on previously created snapshot softlinks.
673 Pruning is the act of deleting file system history.
676 command will delete file system history such that
677 the file system state is retained for the given snapshots,
678 and all history after the latest snapshot.
679 By setting the per PFS parameter
681 history is guaranteed to be saved at least this time interval.
682 All other history is deleted.
684 The target directory is expected to contain softlinks pointing to
685 snapshots of the file systems you wish to retain.
686 The directory is scanned non-recursively and the mount points and
687 transaction ids stored in the softlinks are extracted and sorted.
688 The file system is then explicitly pruned according to what is found.
689 Cleaning out portions of the file system is as simple as removing a
690 snapshot softlink and then running the
694 As a safety measure pruning only occurs if one or more softlinks are found
697 snapshot id extension.
698 Currently the scanned softlink directory must contain softlinks pointing
702 The softlinks may specify absolute or relative paths.
703 Softlinks must use 20-character
705 transaction ids, as might be returned from
706 .Nm Cm synctid Ar filesystem .
708 Pruning is a per PFS operation, so a
710 file system and each PFS in it have to be pruned separately.
712 Note that pruning a file system may not immediately free-up space,
713 though typically some space will be freed if a large number of records are
715 The file system must be reblocked to completely recover all available space.
717 Example, lets say your that you didn't set
719 and snapshot directory contains the following links:
720 .Bd -literal -offset indent
721 lrwxr-xr-x 1 root wheel 29 May 31 17:57 snap1 ->
722 /usr/obj/@@0x10d2cd05b7270d16
724 lrwxr-xr-x 1 root wheel 29 May 31 17:58 snap2 ->
725 /usr/obj/@@0x10d2cd13f3fde98f
727 lrwxr-xr-x 1 root wheel 29 May 31 17:59 snap3 ->
728 /usr/obj/@@0x10d2cd222adee364
731 If you were to run the
733 command on this directory, then the
736 mount will be pruned to retain the above three snapshots.
737 In addition, history for modifications made to the file system older than
738 the oldest snapshot will be destroyed and history for potentially fine-grained
739 modifications made to the file system more recently than the most recent
740 snapshot will be retained.
742 If you then delete the
744 softlink and rerun the
747 history for modifications pertaining to that snapshot would be destroyed.
751 file system versions 3+ this command also scans the snapshots stored
752 in the file system meta-data and includes them in the prune.
753 .\" ==== prune-everything ====
754 .It Cm prune-everything Ar filesystem
755 This command will remove all historical records from the file system.
756 This directive is not normally used on a production system.
758 This command does not remove snapshot softlinks but will delete all
759 snapshots recorded in file system meta-data (for file system version 3+).
760 The user is responsible for deleting any softlinks.
761 .\" ==== rebalance ====
762 .It Cm rebalance Ar filesystem Op Ar saturation_percentage
763 This command will rebalance the B-tree, nodes with small number of
764 elements will be combined and element counts will be smoothed out
767 The saturation percentage is between 50% and 100%.
768 The default is 75% (the
770 suffix is not needed).
771 .\" ==== reblock* ====
772 .It Cm reblock Ar filesystem Op Ar fill_percentage
773 .It Cm reblock-btree Ar filesystem Op Ar fill_percentage
774 .It Cm reblock-inodes Ar filesystem Op Ar fill_percentage
775 .It Cm reblock-dirs Ar filesystem Op Ar fill_percentage
776 .It Cm reblock-data Ar filesystem Op Ar fill_percentage
777 Attempt to defragment and free space for reuse by reblocking a live
780 Big-blocks cannot be reused by
782 until they are completely free.
783 This command also has the effect of reordering all elements, effectively
784 defragmenting the file system.
786 The default fill percentage is 100% and will cause the file system to be
787 completely defragmented.
788 All specified element types will be reallocated and rewritten.
789 If you wish to quickly free up space instead try specifying
790 a smaller fill percentage, such as 90% or 80% (the
792 suffix is not needed).
794 Since this command may rewrite the entire contents of the disk it is
795 best to do it incrementally from a
801 options to limit the run time.
802 The file system would thus be defragmented over long period of time.
804 It is recommended that separate invocations be used for each data type.
805 B-tree nodes, inodes, and directories are typically the most important
806 elements needing defragmentation.
807 Data can be defragmented over a longer period of time.
809 Reblocking is a per PFS operation, so a
811 file system and each PFS in it have to be reblocked separately.
812 .\" ==== pfs-status ====
813 .It Cm pfs-status Ar dirpath ...
814 Retrieve the mirroring configuration parameters for the specified
816 file systems or pseudo-filesystems (PFS's).
817 .\" ==== pfs-master ====
818 .It Cm pfs-master Ar dirpath Op Ar options
819 Create a pseudo-filesystem (PFS) inside a
822 Up to 65535 such file systems can be created.
823 Each PFS uses an independent inode numbering space making it suitable
824 for use as a replication source or target.
828 directive creates a PFS that you can read, write, and use as a mirroring
831 It is recommended to use a
833 mount to access a PFS, for more information see
835 .\" ==== pfs-slave ====
836 .It Cm pfs-slave Ar dirpath Op Ar options
837 Create a pseudo-filesystem (PFS) inside a
840 Up to 65535 such file systems can be created.
841 Each PFS uses an independent inode numbering space making it suitable
842 for use as a replication source or target.
846 directive creates a PFS that you can use as a mirroring target.
847 You will not be able to access a slave PFS until you have completed the
848 first mirroring operation with it as the target (its root directory will
849 not exist until then).
851 Access to the pfs-slave via the special softlink, as described in the
856 dynamically modify the snapshot transaction id by returning a dynamic result
861 A PFS can only be truly destroyed with the
864 Removing the softlink will not destroy the underlying PFS.
866 It is recommended to use a
868 mount to access a PFS, for more information see
870 .\" ==== pfs-update ====
871 .It Cm pfs-update Ar dirpath Op Ar options
872 Update the configuration parameters for an existing
874 file system or pseudo-filesystem.
875 Options that may be specified:
876 .Bl -tag -width indent
877 .It Cm sync-beg-tid= Ns Ar 0x16llx
878 This is the automatic snapshot access starting transaction id for
880 This parameter is normally updated automatically by the
884 It is important to note that accessing a mirroring slave
885 with a transaction id greater than the last fully synchronized transaction
886 id can result in an unreliable snapshot since you will be accessing
887 data that is still undergoing synchronization.
889 Manually modifying this field is dangerous and can result in a broken mirror.
890 .It Cm sync-end-tid= Ns Ar 0x16llx
891 This is the current synchronization point for mirroring slaves.
892 This parameter is normally updated automatically by the
896 Manually modifying this field is dangerous and can result in a broken mirror.
897 .It Cm shared-uuid= Ns Ar uuid
898 Set the shared UUID for this file system.
899 All mirrors must have the same shared UUID.
900 For safety purposes the
902 directives will refuse to operate on a target with a different shared UUID.
904 Changing the shared UUID on an existing, non-empty mirroring target,
905 including an empty but not completely pruned target,
906 can lead to corruption of the mirroring target.
907 .It Cm unique-uuid= Ns Ar uuid
908 Set the unique UUID for this file system.
909 This UUID should not be used anywhere else,
910 even on exact copies of the file system.
911 .It Cm label= Ns Ar string
912 Set a descriptive label for this file system.
913 .It Cm snapshots= Ns Ar string
914 Specify the snapshots directory which
917 will use to manage this PFS.
918 .Bl -tag -width indent
919 .It Nm HAMMER No version 2-
920 The snapshots directory does not need to be configured for
921 PFS masters and will default to
922 .Pa <pfs>/snapshots .
924 PFS slaves are mirroring slaves so you cannot configure a snapshots
925 directory on the slave itself to be managed by the slave's machine.
926 In fact, the slave will likely have a
928 sub-directory mirrored
929 from the master, but that directory contains the configuration the master
930 is using for its copy of the file system, not the configuration that we
931 want to use for our slave.
933 It is recommended that
934 .Pa <fs>/var/slaves/<name>
935 be configured for a PFS slave, where
941 is an appropriate label.
942 .It Nm HAMMER No version 3+
943 The snapshots directory does not need to be configured for PFS masters or
945 The snapshots directory defaults to
946 .Pa /var/hammer/<pfs>
947 .Pa ( /var/hammer/root
951 You can control snapshot retention on your slave independent of the master.
952 .It Cm snapshots-clear
955 directory path for this PFS.
956 .It Cm prune-min= Ns Ar N Ns Cm d
957 .It Cm prune-min= Ns Oo Ar N Ns Cm d/ Oc Ns \
958 Ar hh Ns Op Cm \&: Ns Ar mm Ns Op Cm \&: Ns Ar ss
959 Set the minimum fine-grained data retention period.
961 always retains fine-grained history up to the most recent snapshot.
962 You can extend the retention period further by specifying a non-zero
964 Any snapshot softlinks within the retention period are ignored
965 for the purposes of pruning (the fine grained history is retained).
966 Number of days, hours, minutes and seconds are given as
971 Because the transaction id in the snapshot softlink cannot be used
972 to calculate a timestamp,
974 uses the earlier of the
978 field of the softlink to
979 determine which snapshots fall within the retention period.
980 Users must be sure to retain one of these two fields when manipulating
983 .\" ==== pfs-upgrade ====
984 .It Cm pfs-upgrade Ar dirpath
985 Upgrade a PFS from slave to master operation.
986 The PFS will be rolled back to the current end synchronization transaction id
987 (removing any partial synchronizations), and will then become writable.
991 currently supports only single masters and using
992 this command can easily result in file system corruption
993 if you don't know what you are doing.
995 This directive will refuse to run if any programs have open descriptors
996 in the PFS, including programs chdir'd into the PFS.
997 .\" ==== pfs-downgrade ====
998 .It Cm pfs-downgrade Ar dirpath
999 Downgrade a master PFS from master to slave operation
1000 The PFS becomes read-only and access will be locked to its
1003 This directive will refuse to run if any programs have open descriptors
1004 in the PFS, including programs chdir'd into the PFS.
1005 .\" ==== pfs-destroy ====
1006 .It Cm pfs-destroy Ar dirpath
1007 This permanently destroys a PFS.
1009 This directive will refuse to run if any programs have open descriptors
1010 in the PFS, including programs chdir'd into the PFS.
1011 .\" ==== mirror-read ====
1012 .It Cm mirror-read Ar filesystem Op Ar begin-tid
1013 Generate a mirroring stream to stdout.
1014 The stream ends when the transaction id space has been exhausted.
1015 .\" ==== mirror-read-stream ====
1016 .It Cm mirror-read-stream Ar filesystem Op Ar begin-tid
1017 Generate a mirroring stream to stdout.
1018 Upon completion the stream is paused until new data is synced to the
1019 master, then resumed.
1020 Operation continues until the pipe is broken.
1021 .\" ==== mirror-write ====
1022 .It Cm mirror-write Ar filesystem
1023 Take a mirroring stream on stdin.
1025 This command will fail if the
1027 configuration field for the two file systems do not match.
1029 If the target PFS does not exist this command will ask you whether
1030 you want to create a compatible PFS slave for the target or not.
1031 .\" ==== mirror-dump ====
1037 to dump an ASCII representation of the mirroring stream.
1038 .\" ==== mirror-copy ====
1039 .\".It Cm mirror-copy Ar [[user@]host:]filesystem [[user@]host:]filesystem
1040 .It Cm mirror-copy \
1041 Oo Oo Ar user Ns Cm @ Oc Ns Ar host Ns Cm \&: Oc Ns Ar filesystem \
1042 Oo Oo Ar user Ns Cm @ Oc Ns Ar host Ns Cm \&: Oc Ns Ar filesystem
1043 This is a shortcut which pipes a
1048 If a remote host specification is made the program forks a
1054 on the appropriate host.
1055 The source may be a master or slave PFS, and the target must be a slave PFS.
1057 This command also established full duplex communication and turns on
1058 the two-way protocol feature which automatically negotiates transaction id
1059 ranges without having to use a cyclefile.
1060 If the operation completes successfully the target PFS's
1063 Note that you must re-chdir into the target PFS to see the updated information.
1064 If you do not you will still be in the previous snapshot.
1066 If the target PFS does not exist this command will ask you whether
1067 you want to create a compatible PFS slave for the target or not.
1068 .\" ==== mirror-stream ====
1069 .\".It Cm mirror-stream Ar [[user@]host:]filesystem [[user@]host:]filesystem
1070 .It Cm mirror-stream \
1071 Oo Oo Ar user Ns Cm @ Oc Ns Ar host Ns Cm \&: Oc Ns Ar filesystem \
1072 Oo Oo Ar user Ns Cm @ Oc Ns Ar host Ns Cm \&: Oc Ns Ar filesystem
1073 This command works similarly to
1075 but does not exit after the initial mirroring completes.
1076 The mirroring operation will resume as changes continue to be made to the
1078 The command is commonly used with
1082 options to keep the mirroring target in sync with the source on a continuing
1085 If the pipe is broken the command will automatically retry after sleeping
1087 The time slept will be 15 seconds plus the time given in the
1091 This command also detects the initial-mirroring case and spends some
1092 time scanning the B-Tree to find good break points, allowing the initial
1093 bulk mirroring operation to be broken down into about 20 separate pieces.
1094 This means that the user can kill and restart the operation and it will
1095 not have to start from scratch once it has gotten past the first chunk.
1098 option may be used to disable this feature and perform an initial bulk
1100 .\" ==== version ====
1101 .It Cm version Ar filesystem
1102 This command returns the
1104 file system version for the specified
1106 as well as the range of versions supported in the kernel.
1109 option may be used to remove the summary at the end.
1110 .\" ==== version-upgrade ====
1111 .It Cm version-upgrade Ar filesystem Ar version Op Cm force
1112 This command upgrades the
1117 Once upgraded a file system may not be downgraded.
1118 If you wish to upgrade a file system to a version greater or equal to the
1119 work-in-progress version number you must specify the
1122 Use of WIP versions should be relegated to testing and may require wiping
1123 the file system as development progresses, even though the WIP version might
1127 This command operates on the entire
1129 file system and is not a per PFS operation.
1130 All PFS's will be affected.
1131 .Bl -tag -width indent
1134 default version, first
1139 New directory entry layout.
1140 This version is using a new directory hash key.
1143 New snapshot management, using file system meta-data for saving
1144 configuration file and snapshots (transaction ids etc.).
1145 Also default snapshots directory has changed.
1149 New undo/flush, giving faster sync.
1153 .Sh PSEUDO-FILESYSTEM (PFS) NOTES
1154 The root of a PFS is not hooked into the primary
1156 file system as a directory.
1159 creates a special softlink called
1161 (exactly 10 characters long) in the primary
1165 then modifies the contents of the softlink as read by
1167 and thus what you see with an
1169 command or if you were to
1172 If the PFS is a master the link reflects the current state of the PFS.
1173 If the PFS is a slave the link reflects the last completed snapshot, and the
1174 contents of the link will change when the next snapshot is completed, and
1179 utility employs numerous safeties to reduce user foot-shooting.
1182 directive requires that the target be configured as a slave and that the
1184 field of the mirroring source and target match.
1185 .Sh UPGRADE INSTRUCTIONS HAMMER V1 TO V2
1186 This upgrade changes the way directory entries are stored.
1187 It is possible to upgrade a V1 file system to V2 in place, but
1188 directories created prior to the upgrade will continue to use
1191 Note that the slave mirroring code in the target kernel had bugs in
1192 V1 which can create an incompatible root directory on the slave.
1195 master created after the upgrade with a
1197 slave created prior to the upgrade.
1199 Any directories created after upgrading will use a new layout.
1200 .Sh UPGRADE INSTRUCTIONS HAMMER V2 TO V3
1201 This upgrade adds meta-data elements to the B-Tree.
1202 It is possible to upgrade a V2 file system to V3 in place.
1203 After issuing the upgrade be sure to run a
1206 to perform post-upgrade tasks.
1208 After making this upgrade running a
1213 directory for each PFS mount into
1214 .Pa /var/hammer/<pfs> .
1217 root mount will migrate
1220 .Pa /var/hammer/root .
1221 Migration occurs only once and only if you have not specified
1222 a snapshots directory in the PFS configuration.
1223 If you have specified a snapshots directory in the PFS configuration no
1224 automatic migration will occur.
1226 For slaves, if you desire, you can migrate your snapshots
1227 config to the new location manually and then clear the
1228 snapshot directory configuration in the slave PFS.
1229 The new snapshots hierarchy is designed to work with
1230 both master and slave PFSs equally well.
1232 In addition, the old config file will be moved to file system meta-data,
1233 editable via the new
1237 The old config file will be deleted.
1238 Migration occurs only once.
1240 The V3 file system has new
1242 directives for creating snapshots.
1243 All snapshot directives, including the original, will create
1244 meta-data entries for the snapshots and the pruning code will
1245 automatically incorporate these entries into its list and
1246 expire them the same way it expires softlinks.
1247 If you by accident blow away your snapshot softlinks you can use the
1249 directive to get a definitive list from the file system meta-data and
1250 regenerate them from that list.
1255 to backup file systems your scripts may be using the
1257 directive to generate transaction ids.
1258 This directive does not create a snapshot.
1259 You will have to modify your scripts to use the
1261 directive to generate the linkbuf for the softlink you create, or
1262 use one of the other
1267 directive will continue to work as expected and in V3 it will also
1268 record the snapshot transaction id in file system meta-data.
1269 You may also want to make use of the new
1271 tag for the meta-data.
1274 If you used to remove snapshot softlinks with
1276 you should probably start using the
1278 directive instead to also remove the related meta-data.
1279 The pruning code scans the meta-data so just removing the
1280 softlink is not sufficient.
1281 .Sh UPGRADE INSTRUCTIONS HAMMER V3 TO V4
1282 This upgrade changes undo/flush, giving faster sync.
1283 It is possible to upgrade a V3 file system to V4 in place.
1284 This upgrade reformats the UNDO FIFO (typically 1GB), so upgrade might take
1285 a minute or two depending.
1287 Version 4 allows the UNDO FIFO to be flushed without also having
1288 to flush the volume header, removing 2 of the 4 disk syncs typically
1289 required for an fsync() and removing 1 of the 2 disk syncs typically
1290 required for a flush sequence.
1294 .Bl -tag -width ".It Pa <fs>/var/slaves/<name>" -compact
1295 .It Pa <pfs>/snapshots
1296 default per PFS snapshots directory
1299 .It Pa /var/hammer/<pfs>
1300 default per PFS snapshots directory (not root)
1303 .It Pa /var/hammer/root
1304 default snapshots directory for root directory
1307 .It Pa <snapshots>/config
1314 .It Pa <fs>/var/slaves/<name>
1315 recommended slave PFS snapshots directory
1323 .Xr periodic.conf 5 ,
1324 .Xr mount_hammer 8 ,
1330 utility first appeared in
1333 .An Matthew Dillon Aq dillon@backplane.com