1.\" Copyright (c) 2007 The DragonFly Project. All rights reserved. 2.\" 3.\" This code is derived from software contributed to The DragonFly Project 4.\" by Matthew Dillon <dillon@backplane.com> 5.\" 6.\" Redistribution and use in source and binary forms, with or without 7.\" modification, are permitted provided that the following conditions 8.\" are met: 9.\" 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 15.\" distribution. 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. 19.\" 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 31.\" SUCH DAMAGE. 32.\" 33.Dd February 5, 2015 34.Dt HAMMER 8 35.Os 36.Sh NAME 37.Nm hammer 38.Nd HAMMER file system utility 39.Sh SYNOPSIS 40.Nm 41.Fl h 42.Nm 43.Op Fl 2ABFqrvXy 44.Op Fl b Ar bandwidth 45.Op Fl C Ar cachesize Ns Op Ns Cm \&: Ns Ar readahead 46.Op Fl R Ar restrictcmd 47.Op Fl T Ar restrictpath 48.Op Fl c Ar cyclefile 49.Op Fl e Ar scoreboardfile 50.Op Fl f Ar blkdevs 51.\" .Op Fl s Ar linkpath 52.Op Fl i Ar delay 53.Op Fl p Ar ssh-port 54.Op Fl S Ar splitsize 55.Op Fl t Ar seconds 56.Op Fl m Ar memlimit 57.Ar command 58.Op Ar argument ... 59.Sh DESCRIPTION 60This manual page documents the 61.Nm 62utility which provides miscellaneous functions related to managing a 63.Nm HAMMER 64file system. 65For a general introduction to the 66.Nm HAMMER 67file system, its features, and 68examples on how to set up and maintain one, see 69.Xr HAMMER 5 . 70.Pp 71The options are as follows: 72.Bl -tag -width indent 73.It Fl 2 74Tell the mirror commands to use a 2-way protocol, which allows 75automatic negotiation of transaction id ranges. 76This option is automatically enabled by the 77.Cm mirror-copy 78command. 79.It Fl A 80Make per PFS commands perform on all PFSs if possible. 81If the command supports this option, it targets all PFSs of the 82.Nm HAMMER 83filesystem that the 84.Ar filesystem 85argument (of that command) belongs to. 86Currently 87.Cm rebalance , 88.Cm reblock , 89.Cm reblock-btree , 90.Cm reblock-inodes , 91.Cm reblock-dirs 92and 93.Cm reblock-data 94commands support this option. 95If the command does not support this option, it does nothing. 96.It Fl B 97Bulk transfer. 98.Cm Mirror-stream 99will not attempt to break-up large initial bulk transfers into smaller 100pieces. 101This can save time but if the link is lost in the middle of the 102initial bulk transfer you will have to start over from scratch. 103For more information see the 104.Fl S 105option. 106.It Fl b Ar bandwidth 107Specify a bandwidth limit in bytes per second for mirroring streams. 108This option is typically used to prevent batch mirroring operations from 109loading down the machine. 110The bandwidth may be suffixed with 111.Cm k , m , 112or 113.Cm g 114to specify values in kilobytes, megabytes, and gigabytes per second. 115If no suffix is specified, bytes per second is assumed. 116.Pp 117Unfortunately this is only applicable to the pre-compression bandwidth 118when compression is used, so a better solution would probably be to 119use a 120.Xr ipfw 8 121pipe or a 122.Xr pf 4 123queue. 124.It Fl C Ar cachesize Ns Op Ns Cm \&: Ns Ar readahead 125Set the memory cache size for any raw 126.Tn I/O . 127The default is 16MB. 128A suffix of 129.Cm k 130for kilobytes and 131.Cm m 132for megabytes is allowed, 133else the cache size is specified in bytes. 134.Pp 135The read-behind/read-ahead defaults to 4 136.Nm HAMMER 137blocks. 138.Pp 139This option is typically only used with diagnostic commands 140as kernel-supported commands will use the kernel's buffer cache. 141.It Fl R Ar restrictcmd 142This option is used by hammer ssh-remote to restrict the command later 143on in the argument list. Multiple commands may be specified, separated 144by a comma (all one argument). 145.It Fl T Ar restrictpath 146This option is used by hammer ssh-remote to restrict the filesystem path 147specified later on in the argument list. 148.It Fl c Ar cyclefile 149When pruning, rebalancing or reblocking you can tell the utility 150to start at the object id stored in the specified file. 151If the file does not exist 152.Nm 153will start at the beginning. 154If 155.Nm 156is told to run for a specific period of time 157.Pq Fl t 158and is unable to complete the operation it will write out 159the current object id so the next run can pick up where it left off. 160If 161.Nm 162runs to completion it will delete 163.Ar cyclefile . 164.It Fl e Ar scoreboardfile 165Update scoreboard file with progress, primarily used by mirror-stream. 166.It Fl F 167Force operation. 168E.g.\& 169.Cm cleanup 170will not check that time period has elapsed if this option is given. 171.It Fl f Ar blkdevs 172Specify the volumes making up a 173.Nm HAMMER 174file system. 175.Ar Blkdevs 176is a colon-separated list of devices, each specifying a 177.Nm HAMMER 178volume. 179.It Fl h 180Show usage. 181.It Fl i Ar delay 182Specify delay in seconds for 183.Cm mirror-read-stream . 184When maintaining a streaming mirroring this option specifies the 185minimum delay after a batch ends before the next batch is allowed 186to start. 187The default is five seconds. 188.It Fl m Ar memlimit 189Specify the maximum amount of memory 190.Nm 191will allocate during a dedup pass. 192Specify a suffix 'm', 'g', or 't' for megabytes, gigabytes, or terabytes. 193By default 194.Nm 195will allocate up to 1G of ram to hold CRC/SHA tables while running dedup. 196When the limit is reached the dedup code restricts the range of CRCs to 197keep memory use within bounds and runs multiple passes as necessary until 198the entire filesystem has been deduped. 199.It Fl p Ar ssh-port 200Pass the 201.Fl p Ar ssh-port 202option to 203.Xr ssh 1 204when using a remote 205specification for the source and/or destination. 206.It Fl q 207Decrease verboseness. 208May be specified multiple times. 209.It Fl r 210Specify recursion for those commands which support it. 211.It Fl S Ar splitsize 212Specify the bulk splitup size in bytes for mirroring streams. 213When a 214.Cm mirror-stream 215is first started 216.Nm 217will do an initial run-through of the data to calculate good 218transaction ids to cut up the bulk transfers, creating 219restart points in case the stream is interrupted. 220If we don't do this and the stream is interrupted it might 221have to start all over again. 222The default is a 223.Ar splitsize 224of 4GB. 225.Pp 226At the moment the run-through is disk-bandwidth-heavy but some 227future version will limit the run-through to just the B-Tree 228records and not the record data. 229.Pp 230The splitsize may be suffixed with 231.Cm k , m , 232or 233.Cm g 234to specify values in kilobytes, megabytes, or gigabytes. 235If no suffix is specified, bytes is assumed. 236.Pp 237When mirroring very large filesystems the minimum recommended 238split size is 4GB. 239A small split size may wind up generating a great deal of overhead 240but very little actual incremental data and is not recommended. 241.It Fl t Ar seconds 242Specify timeout in seconds. 243When pruning, rebalancing, reblocking or mirror-reading 244you can tell the utility to stop after a certain period of time. 245A value of 0 means unlimited. 246This option is used along with the 247.Fl c Ar cyclefile 248option to prune, rebalance or reblock incrementally. 249.It Fl v 250Increase verboseness. 251May be specified multiple times. 252.It Fl X 253Enable compression for any remote ssh specifications. 254This option is typically used with the mirroring directives. 255.It Fl y 256Force 257.Dq yes 258for interactive questions. 259.El 260.Pp 261The commands are as follows: 262.Bl -tag -width indent 263.\" ==== synctid ==== 264.It Cm synctid Ar filesystem Op Cm quick 265Generate a guaranteed, formal 64-bit transaction id representing the 266current state of the specified 267.Nm HAMMER 268file system. 269The file system will be synced to the media. 270.Pp 271If the 272.Cm quick 273keyword is specified the file system will be soft-synced, meaning that a 274crash might still undo the state of the file system as of the transaction 275id returned but any new modifications will occur after the returned 276transaction id as expected. 277.Pp 278This operation does not create a snapshot. 279It is meant to be used 280to track temporary fine-grained changes to a subset of files and 281will only remain valid for 282.Ql @@ 283access purposes for the 284.Cm prune-min 285period configured for the PFS. 286If you desire a real snapshot then the 287.Cm snapq 288directive may be what you are looking for. 289.\" ==== bstats ==== 290.It Cm bstats Op Ar interval 291Output 292.Nm HAMMER 293B-Tree statistics until interrupted. 294Pause 295.Ar interval 296seconds between each display. 297The default interval is one second. 298.\" ==== iostats ==== 299.It Cm iostats Op Ar interval 300Output 301.Nm HAMMER 302.Tn I/O 303statistics until interrupted. 304Pause 305.Ar interval 306seconds between each display. 307The default interval is one second. 308.\" ==== history ==== 309.It Cm history Ns Oo Cm @ Ns Ar offset Ns Oo Cm \&, Ns Ar length Oc Oc Ar path ... 310Show the modification history for inode and data of 311.Nm HAMMER 312files. 313If 314.Ar offset 315is given history is shown for data block at given offset, 316otherwise history is shown for inode. 317If 318.Fl v 319is specified 320.Ar length 321data bytes at given offset are dumped for each version, 322default is 32. 323.Pp 324For each 325.Ar path 326this directive shows object id and sync status, 327and for each object version it shows transaction id and time stamp. 328Files has to exist for this directive to be applicable, 329to track inodes which has been deleted or renamed see 330.Xr undo 1 . 331.\" ==== blockmap ==== 332.It Cm blockmap 333Dump the blockmap for the file system. 334The 335.Nm HAMMER 336blockmap is two-layer 337blockmap representing the maximum possible file system size of 1 Exabyte. 338Needless to say the second layer is only present for blocks which exist. 339.Nm HAMMER Ns 's 340blockmap represents 8-Megabyte blocks, called big-blocks. 341Each big-block has an append 342point, a free byte count, and a typed zone id which allows content to be 343reverse engineered to some degree. 344.Pp 345In 346.Nm HAMMER 347allocations are essentially appended to a selected big-block using 348the append offset and deducted from the free byte count. 349When space is freed the free byte count is adjusted but 350.Nm HAMMER 351does not track holes in big-blocks for reallocation. 352A big-block must be completely freed, either 353through normal file system operations or through reblocking, before 354it can be reused. 355.Pp 356Data blocks can be shared by deducting the space used from the free byte 357count for each shared references. 358This means the free byte count can legally go negative. 359.Pp 360This command needs the 361.Fl f Ar blkdevs 362option. 363.\" ==== checkmap ==== 364.It Cm checkmap 365Check the blockmap allocation count. 366.Nm 367will scan the B-Tree, collect allocation information, and 368construct a blockmap in-memory. 369It will then check that blockmap against the on-disk blockmap. 370.Pp 371This command needs the 372.Fl f Ar blkdevs 373option. 374.\" ==== show ==== 375.It Cm show Op Ar localization Ns Op Cm \&: Ns Ar object_id 376Dump the B-Tree. 377By default this command will validate all B-Tree 378linkages and CRCs, including data CRCs, and will report the most verbose 379information it can dig up. 380Any errors will show up with a 381.Ql B 382in column 1 along with various 383other error flags. 384.Pp 385If you specify 386.Ar localization 387or 388.Ar localization Ns Cm \&: Ns Ar object_id 389the dump will 390search for the key printing nodes as it recurses down, and then 391will iterate forwards. 392These fields are specified in HEX. 393Note that the pfsid is the top 16 bits of the 32-bit localization 394field so PFS #1 would be 00010000. 395.Pp 396If you use 397.Fl q 398the command will report less information about the inode contents. 399.Pp 400If you use 401.Fl qq 402the command will not report the content of the inode or other typed 403data at all. 404.Pp 405If you use 406.Fl qqq 407the command will not report volume header information, big-block fill 408ratios, mirror transaction ids, or report or check data CRCs. 409B-Tree CRCs and linkages are still checked. 410.Pp 411This command needs the 412.Fl f Ar blkdevs 413option. 414.\" ==== show-undo ==== 415.It Cm show-undo 416.Nm ( HAMMER 417VERSION 4+) 418Dump the UNDO/REDO map. 419.Pp 420This command needs the 421.Fl f Ar blkdevs 422option. 423.\" .It Ar blockmap 424.\" Dump the B-Tree, record, large-data, and small-data blockmaps, showing 425.\" physical block assignments and free space percentages. 426.\" ==== ssh-remote ==== 427.It Cm ssh-remote Ar command Ar targetdir 428Used in a ssh authorized_keys line such as 429command="/sbin/hammer ssh-remote mirror-read /fubarmount" ... to allow 430mirror-read or mirror-write access to a particular subdirectory tree. 431This way you do not have to give shell access to the remote box. 432.Nm 433will obtain the original command line from the SSH_ORIGINAL_COMMAND 434environment variable, validate it against the restriction, and then 435re-exec hammer with the validated arguments. 436.Pp 437The remote hammer command does not allow the 438.Fl c 439or 440.Fl f 441options to be passed in. 442.\" ==== recover ==== 443.It Cm recover Ar targetdir 444Recover data from a corrupted 445.Nm HAMMER 446filesystem. 447This is a low level command which operates on the filesystem image and 448attempts to locate and recover files from a corrupted filesystem. 449The entire image is scanned linearly looking for B-Tree nodes. 450Any node 451found which passes its CRC test is scanned for file, inode, and directory 452fragments and the target directory is populated with the resulting data. 453files and directories in the target directory are initially named after 454the object id and are renamed as fragmentary information is processed. 455.Pp 456This command keeps track of filename/object_id translations and may eat a 457considerably amount of memory while operating. 458.Pp 459This command is literally the last line of defense when it comes to 460recovering data from a dead filesystem. 461.Pp 462This command needs the 463.Fl f Ar blkdevs 464option. 465.\" ==== namekey1 ==== 466.It Cm namekey1 Ar filename 467Generate a 468.Nm HAMMER 46964-bit directory hash for the specified file name, using 470the original directory hash algorithm in version 1 of the file system. 471The low 32 bits are used as an iterator for hash collisions and will be 472output as 0. 473.\" ==== namekey2 ==== 474.It Cm namekey2 Ar filename 475Generate a 476.Nm HAMMER 47764-bit directory hash for the specified file name, using 478the new directory hash algorithm in version 2 of the file system. 479The low 32 bits are still used as an iterator but will start out containing 480part of the hash key. 481.\" ==== namekey32 ==== 482.It Cm namekey32 Ar filename 483Generate the top 32 bits of a 484.Nm HAMMER 48564 bit directory hash for the specified file name. 486.\" ==== info ==== 487.It Cm info Ar dirpath ... 488Show extended information about all 489.Nm HAMMER 490file systems mounted in the system or the one mounted in 491.Ar dirpath 492when this argument is specified. 493.Pp 494The information is divided into sections: 495.Bl -tag -width indent 496.It Volume identification 497General information, like the label of the 498.Nm HAMMER 499filesystem, the number of volumes it contains, the FSID, and the 500.Nm HAMMER 501version being used. 502.It Big block information 503Big block statistics, such as total, used, reserved and free big blocks. 504.It Space information 505Information about space used on the filesystem. 506Currently total size, used, reserved and free space are displayed. 507.It PFS information 508Basic information about the PFSs currently present on a 509.Nm HAMMER 510filesystem. 511.Pp 512.Dq PFS ID 513is the ID of the PFS, with 0 being the root PFS. 514.Dq Snaps 515is the current snapshot count on the PFS. 516.Dq Mounted on 517displays the mount point of the PFS is currently mounted on (if any). 518.El 519.\" ==== cleanup ==== 520.It Cm cleanup Op Ar filesystem ... 521This is a meta-command which executes snapshot, prune, rebalance, dedup 522and reblock commands on the specified 523.Nm HAMMER 524file systems. 525If no 526.Ar filesystem 527is specified this command will clean-up all 528.Nm HAMMER 529file systems in use, including PFS's. 530To do this it will scan all 531.Nm HAMMER 532and 533.Nm null 534mounts, extract PFS id's, and clean-up each PFS found. 535.Pp 536This command will access a snapshots 537directory and a configuration file for each 538.Ar filesystem , 539creating them if necessary. 540.Bl -tag -width indent 541.It Nm HAMMER No version 2- 542The configuration file is 543.Pa config 544in the snapshots directory which defaults to 545.Pa <pfs>/snapshots . 546.It Nm HAMMER No version 3+ 547The configuration file is saved in file system meta-data, see 548.Nm 549.Cm config . 550The snapshots directory defaults to 551.Pa /var/hammer/<pfs> 552.Pa ( /var/hammer/root 553for root mount). 554.El 555.Pp 556The format of the configuration file is: 557.Bd -literal -offset indent 558snapshots <period> <retention-time> [any] 559prune <period> <max-runtime> 560rebalance <period> <max-runtime> 561dedup <period> <max-runtime> 562reblock <period> <max-runtime> 563recopy <period> <max-runtime> 564.Ed 565.Pp 566Defaults are: 567.Bd -literal -offset indent 568snapshots 1d 60d # 0d 0d for PFS /tmp, /var/tmp, /usr/obj 569prune 1d 5m 570rebalance 1d 5m 571dedup 1d 5m 572reblock 1d 5m 573recopy 30d 10m 574.Ed 575.Pp 576Time is given with a suffix of 577.Cm d , 578.Cm h , 579.Cm m 580or 581.Cm s 582meaning day, hour, minute and second. 583.Pp 584If the 585.Cm snapshots 586directive has a period of 0 and a retention time of 0 587then snapshot generation is disabled, removal of old snapshots are 588disabled, and prunes will use 589.Cm prune-everything . 590.Pp 591If the 592.Cm snapshots 593directive has a period of 0 but a non-zero retention time 594then this command will not create any new snapshots but will remove old 595snapshots it finds based on the retention time. 596This form should be 597used on PFS masters where you are generating your own snapshot softlinks 598manually and on PFS slaves when all you wish to do is prune away existing 599snapshots inherited via the mirroring stream. 600.Pp 601By default only snapshots in the form 602.Ql snap- Ns Ar yyyymmdd Ns Op - Ns Ar HHMM 603are processed. 604If the 605.Cm any 606directive is specified as a third argument on the 607.Cm snapshots 608config line then any softlink of the form 609.Ql *- Ns Ar yyyymmdd Ns Op - Ns Ar HHMM 610or 611.Ql *. Ns Ar yyyymmdd Ns Op - Ns Ar HHMM 612will be processed. 613.Pp 614A period of 0 for prune, rebalance, dedup, reblock or recopy disables the directive. 615A max-runtime of 0 means unlimited. 616.Pp 617If period hasn't passed since the previous 618.Cm cleanup 619run nothing is done. 620For example a day has passed when midnight is passed (localtime). 621If the 622.Fl F 623flag is given the period is ignored. 624By default, 625.Dx 626is set up to run 627.Nm Cm cleanup 628nightly via 629.Xr periodic 8 . 630.Pp 631The default configuration file will create a daily snapshot, do a daily 632pruning, rebalancing, deduping and reblocking run and a monthly recopy run. 633Reblocking is defragmentation with a level of 95%, 634and recopy is full defragmentation. 635.Pp 636By default prune, dedup and rebalance operations are time limited to 5 minutes, 637and reblock operations to a bit over 5 minutes, 638and recopy operations to a bit over 10 minutes. 639Reblocking and recopy runs are each broken down into four separate functions: 640btree, inodes, dirs and data. 641Each function is time limited to the time given in the configuration file, 642but the btree, inodes and dirs functions usually does not take very long time, 643full defragmentation is always used for these three functions. 644Also note that this directive will by default disable snapshots on 645the following PFS's: 646.Pa /tmp , 647.Pa /var/tmp 648and 649.Pa /usr/obj . 650.Pp 651The defaults may be adjusted by modifying the configuration file. 652The pruning and reblocking commands automatically maintain a cyclefile 653for incremental operation. 654If you interrupt (^C) the program the cyclefile will be updated, 655but a sub-command 656may continue to run in the background for a few seconds until the 657.Nm HAMMER 658ioctl detects the interrupt. 659The 660.Cm snapshots 661PFS option can be set to use another location for the snapshots directory. 662.Pp 663Work on this command is still in progress. 664Expected additions: 665An ability to remove snapshots dynamically as the 666file system becomes full. 667.\" ==== config ==== 668.It Cm config Op Ar filesystem Op Ar configfile 669.Nm ( HAMMER 670VERSION 3+) 671Show or change configuration for 672.Ar filesystem . 673If zero or one arguments are specified this function dumps the current 674configuration file to stdout. 675Zero arguments specifies the PFS containing the current directory. 676This configuration file is stored in file system meta-data. 677If two arguments are specified this function installs a new config file. 678.Pp 679In 680.Nm HAMMER 681versions less than 3 the configuration file is by default stored in 682.Pa <pfs>/snapshots/config , 683but in all later versions the configuration file is stored in file system 684meta-data. 685.\" ==== viconfig ==== 686.It Cm viconfig Op Ar filesystem 687.Nm ( HAMMER 688VERSION 3+) 689Edit the configuration file and reinstall into file system meta-data when done. 690Zero arguments specifies the PFS containing the current directory. 691.\" ==== volume-add ==== 692.It Cm volume-add Ar device Ar filesystem 693Add volume 694.Ar device 695to 696.Ar filesystem . 697This will format 698.Ar device 699and add all of its space to 700.Ar filesystem . 701A 702.Nm HAMMER 703file system can use up to 256 volumes. 704.Pp 705.Em NOTE! 706All existing data contained on 707.Ar device 708will be destroyed by this operation! 709If 710.Ar device 711contains a valid 712.Nm HAMMER 713file system, formatting will be denied. 714You can overcome this sanity check by using 715.Xr dd 1 716to erase the beginning sectors of the device. 717.Pp 718Remember that you have to specify 719.Ar device , 720together with any other device that make up the file system, 721colon-separated to 722.Pa /etc/fstab 723and 724.Xr mount_hammer 8 . 725If 726.Ar filesystem 727is root file system, also remember to add 728.Ar device 729to 730.Va vfs.root.mountfrom 731in 732.Pa /boot/loader.conf , 733see 734.Xr loader 8 . 735.\" ==== volume-del ==== 736.It Cm volume-del Ar device Ar filesystem 737Remove volume 738.Ar device 739from 740.Ar filesystem . 741.Pp 742Remember that you have to remove 743.Ar device 744from the colon-separated list in 745.Pa /etc/fstab 746and 747.Xr mount_hammer 8 . 748If 749.Ar filesystem 750is root file system, also remember to remove 751.Ar device 752from 753.Va vfs.root.mountfrom 754in 755.Pa /boot/loader.conf , 756see 757.Xr loader 8 . 758.\" ==== volume-list ==== 759.It Cm volume-list Ar filesystem 760List the volumes that make up 761.Ar filesystem . 762.\" ==== snapshot ==== 763.It Cm snapshot Oo Ar filesystem Oc Ar snapshot-dir 764.It Cm snapshot Ar filesystem Ar snapshot-dir Op Ar note 765Take a snapshot of the file system either explicitly given by 766.Ar filesystem 767or implicitly derived from the 768.Ar snapshot-dir 769argument and creates a symlink in the directory provided by 770.Ar snapshot-dir 771pointing to the snapshot. 772If 773.Ar snapshot-dir 774is not a directory, it is assumed to be a format string passed to 775.Xr strftime 3 776with the current time as parameter. 777If 778.Ar snapshot-dir 779refers to an existing directory, a default format string of 780.Ql snap-%Y%m%d-%H%M 781is assumed and used as name for the newly created symlink. 782.Pp 783Snapshot is a per PFS operation, so each PFS in a 784.Nm HAMMER 785file system have to be snapshot separately. 786.Pp 787Example, assuming that 788.Pa /mysnapshots 789is on file system 790.Pa / 791and that 792.Pa /obj 793and 794.Pa /usr 795are file systems on their own, the following invocations: 796.Bd -literal -offset indent 797hammer snapshot /mysnapshots 798 799hammer snapshot /mysnapshots/%Y-%m-%d 800 801hammer snapshot /obj /mysnapshots/obj-%Y-%m-%d 802 803hammer snapshot /usr /my/snaps/usr "note" 804.Ed 805.Pp 806Would create symlinks similar to: 807.Bd -literal -offset indent 808/mysnapshots/snap-20080627-1210 -> /@@0x10d2cd05b7270d16 809 810/mysnapshots/2008-06-27 -> /@@0x10d2cd05b7270d16 811 812/mysnapshots/obj-2008-06-27 -> /obj@@0x10d2cd05b7270d16 813 814/my/snaps/usr/snap-20080627-1210 -> /usr@@0x10d2cd05b7270d16 815.Ed 816.Pp 817When run on a 818.Nm HAMMER 819version 3+ file system the snapshot is also recorded in file system meta-data 820along with the optional 821.Ar note . 822See the 823.Cm snapls 824directive. 825.\" ==== snap* ==== 826.It Cm snap Ar path Op Ar note 827.Nm ( HAMMER 828VERSION 3+) 829Create a snapshot for the PFS containing 830.Ar path 831and create a snapshot softlink. 832If the path specified is a 833directory a standard snapshot softlink will be created in the directory. 834The snapshot softlink points to the base of the mounted PFS. 835.It Cm snaplo Ar path Op Ar note 836.Nm ( HAMMER 837VERSION 3+) 838Create a snapshot for the PFS containing 839.Ar path 840and create a snapshot softlink. 841If the path specified is a 842directory a standard snapshot softlink will be created in the directory. 843The snapshot softlink points into the directory it is contained in. 844.It Cm snapq Ar dir Op Ar note 845.Nm ( HAMMER 846VERSION 3+) 847Create a snapshot for the PFS containing the specified directory but do 848not create a softlink. 849Instead output a path which can be used to access 850the directory via the snapshot. 851.Pp 852An absolute or relative path may be specified. 853The path will be used as-is as a prefix in the path output to stdout. 854As with the other 855snap and snapshot directives the snapshot transaction id will be registered 856in the file system meta-data. 857.It Cm snaprm Ar path Ar ... 858.It Cm snaprm Ar transaction_id Ar ... 859.It Cm snaprm Ar filesystem Ar transaction_id Ar ... 860.Nm ( HAMMER 861VERSION 3+) 862Remove a snapshot given its softlink or transaction id. 863If specifying a transaction id 864the snapshot is removed from file system meta-data but you are responsible 865for removing any related softlinks. 866.Pp 867If a softlink path is specified the filesystem and transaction id 868is derived from the contents of the softlink. 869If just a transaction id is specified it is assumed to be a snapshot in the 870.Nm HAMMER 871filesystem you are currently chdir'd into. 872You can also specify the filesystem and transaction id explicitly. 873.It Cm snapls Op Ar path ... 874.Nm ( HAMMER 875VERSION 3+) 876Dump the snapshot meta-data for PFSs containing each 877.Ar path 878listing all available snapshots and their notes. 879If no arguments are specified snapshots for the PFS containing the 880current directory are listed. 881This is the definitive list of snapshots for the file system. 882.\" ==== prune ==== 883.It Cm prune Ar softlink-dir 884Prune the file system based on previously created snapshot softlinks. 885Pruning is the act of deleting file system history. 886The 887.Cm prune 888command will delete file system history such that 889the file system state is retained for the given snapshots, 890and all history after the latest snapshot. 891By setting the per PFS parameter 892.Cm prune-min , 893history is guaranteed to be saved at least this time interval. 894All other history is deleted. 895.Pp 896The target directory is expected to contain softlinks pointing to 897snapshots of the file systems you wish to retain. 898The directory is scanned non-recursively and the mount points and 899transaction ids stored in the softlinks are extracted and sorted. 900The file system is then explicitly pruned according to what is found. 901Cleaning out portions of the file system is as simple as removing a 902snapshot softlink and then running the 903.Cm prune 904command. 905.Pp 906As a safety measure pruning only occurs if one or more softlinks are found 907containing the 908.Ql @@ 909snapshot id extension. 910Currently the scanned softlink directory must contain softlinks pointing 911to a single 912.Nm HAMMER 913mount. 914The softlinks may specify absolute or relative paths. 915Softlinks must use 20-character 916.Ql @@0x%016llx 917transaction ids, as might be returned from 918.Nm Cm synctid Ar filesystem . 919.Pp 920Pruning is a per PFS operation, so each PFS in a 921.Nm HAMMER 922file system have to be pruned separately. 923.Pp 924Note that pruning a file system may not immediately free-up space, 925though typically some space will be freed if a large number of records are 926pruned out. 927The file system must be reblocked to completely recover all available space. 928.Pp 929Example, lets say your that you didn't set 930.Cm prune-min , 931and snapshot directory contains the following links: 932.Bd -literal -offset indent 933lrwxr-xr-x 1 root wheel 29 May 31 17:57 snap1 -> 934/usr/obj/@@0x10d2cd05b7270d16 935 936lrwxr-xr-x 1 root wheel 29 May 31 17:58 snap2 -> 937/usr/obj/@@0x10d2cd13f3fde98f 938 939lrwxr-xr-x 1 root wheel 29 May 31 17:59 snap3 -> 940/usr/obj/@@0x10d2cd222adee364 941.Ed 942.Pp 943If you were to run the 944.Cm prune 945command on this directory, then the 946.Nm HAMMER 947.Pa /usr/obj 948mount will be pruned to retain the above three snapshots. 949In addition, history for modifications made to the file system older than 950the oldest snapshot will be destroyed and history for potentially fine-grained 951modifications made to the file system more recently than the most recent 952snapshot will be retained. 953.Pp 954If you then delete the 955.Pa snap2 956softlink and rerun the 957.Cm prune 958command, 959history for modifications pertaining to that snapshot would be destroyed. 960.Pp 961In 962.Nm HAMMER 963file system versions 3+ this command also scans the snapshots stored 964in the file system meta-data and includes them in the prune. 965.\" ==== prune-everything ==== 966.It Cm prune-everything Ar filesystem 967Remove all historical records from 968.Ar filesystem . 969Use this directive with caution on PFSs where you intend to use history. 970.Pp 971This command does not remove snapshot softlinks but will delete all 972snapshots recorded in file system meta-data (for file system version 3+). 973The user is responsible for deleting any softlinks. 974.Pp 975Pruning is a per PFS operation, so each PFS in a 976.Nm HAMMER 977file system have to be pruned separately. 978.\" ==== rebalance ==== 979.It Cm rebalance Ar filesystem Op Ar saturation_percentage 980Rebalance the B-Tree, nodes with small number of 981elements will be combined and element counts will be smoothed out 982between nodes. 983.Pp 984The saturation percentage is between 50% and 100%. 985The default is 85% (the 986.Sq % 987suffix is not needed). 988.Pp 989Rebalancing is a per PFS operation, so each PFS in a 990.Nm HAMMER 991file system have to be rebalanced separately. 992.\" ==== dedup ==== 993.It Cm dedup Ar filesystem 994.Nm ( HAMMER 995VERSION 5+) 996Perform offline (post-process) deduplication. 997Deduplication occurs at 998the block level, currently only data blocks of the same size can be 999deduped, metadata blocks can not. 1000The hash function used for comparing 1001data blocks is CRC-32 (CRCs are computed anyways as part of 1002.Nm HAMMER 1003data integrity features, so there's no additional overhead). 1004Since CRC is a weak hash function a byte-by-byte comparison is done 1005before actual deduping. 1006In case of a CRC collision (two data blocks have the same CRC 1007but different contents) the checksum is upgraded to SHA-256. 1008.Pp 1009Currently 1010.Nm HAMMER 1011reblocker may partially blow up (re-expand) dedup (reblocker's normal 1012operation is to reallocate every record, so it's possible for deduped 1013blocks to be re-expanded back). 1014.Pp 1015Deduplication is a per PFS operation, so each PFS in a 1016.Nm HAMMER 1017file system have to be deduped separately. 1018This also 1019means that if you have duplicated data in two different PFSs that data 1020won't be deduped, however the addition of such feature is planned. 1021.Pp 1022The 1023.Fl m Ar memlimit 1024option should be used to limit memory use during the dedup run if the 1025default 1G limit is too much for the machine. 1026.\" ==== dedup-simulate ==== 1027.It Cm dedup-simulate Ar filesystem 1028Shows potential space savings (simulated dedup ratio) one can get after 1029running 1030.Cm dedup 1031command. 1032If the estimated dedup ratio is greater than 1.00 you will see 1033dedup space savings. 1034Remember that this is an estimated number, in 1035practice real dedup ratio will be slightly smaller because of 1036.Nm HAMMER 1037bigblock underflows, B-Tree locking issues and other factors. 1038.Pp 1039Note that deduplication currently works only on bulk data so if you 1040try to run 1041.Cm dedup-simulate 1042or 1043.Cm dedup 1044commands on a PFS that contains metadata only (directory entries, 1045softlinks) you will get a 0.00 dedup ratio. 1046.Pp 1047The 1048.Fl m Ar memlimit 1049option should be used to limit memory use during the dedup run if the 1050default 1G limit is too much for the machine. 1051.\" ==== reblock* ==== 1052.It Cm reblock Ar filesystem Op Ar fill_percentage 1053.It Cm reblock-btree Ar filesystem Op Ar fill_percentage 1054.It Cm reblock-inodes Ar filesystem Op Ar fill_percentage 1055.It Cm reblock-dirs Ar filesystem Op Ar fill_percentage 1056.It Cm reblock-data Ar filesystem Op Ar fill_percentage 1057Attempt to defragment and free space for reuse by reblocking a live 1058.Nm HAMMER 1059file system. 1060Big-blocks cannot be reused by 1061.Nm HAMMER 1062until they are completely free. 1063This command also has the effect of reordering all elements, effectively 1064defragmenting the file system. 1065.Pp 1066The default fill percentage is 100% and will cause the file system to be 1067completely defragmented. 1068All specified element types will be reallocated and rewritten. 1069If you wish to quickly free up space instead try specifying 1070a smaller fill percentage, such as 90% or 80% (the 1071.Sq % 1072suffix is not needed). 1073.Pp 1074Since this command may rewrite the entire contents of the disk it is 1075best to do it incrementally from a 1076.Xr cron 8 1077job along with the 1078.Fl c Ar cyclefile 1079and 1080.Fl t Ar seconds 1081options to limit the run time. 1082The file system would thus be defragmented over long period of time. 1083.Pp 1084It is recommended that separate invocations be used for each data type. 1085B-Tree nodes, inodes, and directories are typically the most important 1086elements needing defragmentation. 1087Data can be defragmented over a longer period of time. 1088.Pp 1089Reblocking is a per PFS operation, so each PFS in a 1090.Nm HAMMER 1091file system have to be reblocked separately. 1092.\" ==== pfs-status ==== 1093.It Cm pfs-status Ar dirpath ... 1094Retrieve the mirroring configuration parameters for the specified 1095.Nm HAMMER 1096file systems or pseudo-filesystems (PFS's). 1097.\" ==== pfs-master ==== 1098.It Cm pfs-master Ar dirpath Op Ar options 1099Create a pseudo-filesystem (PFS) inside a 1100.Nm HAMMER 1101file system. 1102Up to 65536 PFSs can be created. 1103Each PFS uses an independent inode numbering space making it suitable 1104for replication. 1105.Pp 1106The 1107.Cm pfs-master 1108directive creates a PFS that you can read, write, and use as a mirroring 1109source. 1110.Pp 1111A PFS can only be truly destroyed with the 1112.Cm pfs-destroy 1113directive. 1114Removing the softlink will not destroy the underlying PFS. 1115.Pp 1116A PFS can only be created in the root PFS (PFS# 0), 1117not in a PFS created by 1118.Cm pfs-master 1119or 1120.Cm pfs-slave 1121(PFS# >0). 1122.Pp 1123It is recommended that 1124.Ar dirpath 1125is of the form 1126.Pa <fs>/pfs/<name> 1127(i.e.\& located in 1128.Pa pfs 1129directory at root of 1130.Nm HAMMER 1131file system). 1132.Pp 1133It is recommended to use a 1134.Nm null 1135mount to access a PFS, except for root PFS, for more information see 1136.Xr HAMMER 5 . 1137.\" ==== pfs-slave ==== 1138.It Cm pfs-slave Ar dirpath Op Ar options 1139Create a pseudo-filesystem (PFS) inside a 1140.Nm HAMMER 1141file system. 1142Up to 65536 PFSs can be created. 1143Each PFS uses an independent inode numbering space making it suitable 1144for replication. 1145.Pp 1146The 1147.Cm pfs-slave 1148directive creates a PFS that you can use as a mirroring source or target. 1149You will not be able to access a slave PFS until you have completed the 1150first mirroring operation with it as the target (its root directory will 1151not exist until then). 1152.Pp 1153Access to the pfs-slave via the special softlink, as described in the 1154.Sx PSEUDO-FILESYSTEM (PFS) NOTES 1155below, allows 1156.Nm HAMMER 1157to 1158dynamically modify the snapshot transaction id by returning a dynamic result 1159from 1160.Xr readlink 2 1161calls. 1162.Pp 1163A PFS can only be truly destroyed with the 1164.Cm pfs-destroy 1165directive. 1166Removing the softlink will not destroy the underlying PFS. 1167.Pp 1168A PFS can only be created in the root PFS (PFS# 0), 1169not in a PFS created by 1170.Cm pfs-master 1171or 1172.Cm pfs-slave 1173(PFS# >0). 1174.Pp 1175It is recommended that 1176.Ar dirpath 1177is of the form 1178.Pa <fs>/pfs/<name> 1179(i.e.\& located in 1180.Pa pfs 1181directory at root of 1182.Nm HAMMER 1183file system). 1184.Pp 1185It is recommended to use a 1186.Nm null 1187mount to access a PFS, except for root PFS, for more information see 1188.Xr HAMMER 5 . 1189.\" ==== pfs-update ==== 1190.It Cm pfs-update Ar dirpath Op Ar options 1191Update the configuration parameters for an existing 1192.Nm HAMMER 1193file system or pseudo-filesystem. 1194Options that may be specified: 1195.Bl -tag -width indent 1196.It Cm sync-beg-tid= Ns Ar 0x16llx 1197This is the automatic snapshot access starting transaction id for 1198mirroring slaves. 1199This parameter is normally updated automatically by the 1200.Cm mirror-write 1201directive. 1202.Pp 1203It is important to note that accessing a mirroring slave 1204with a transaction id greater than the last fully synchronized transaction 1205id can result in an unreliable snapshot since you will be accessing 1206data that is still undergoing synchronization. 1207.Pp 1208Manually modifying this field is dangerous and can result in a broken mirror. 1209.It Cm sync-end-tid= Ns Ar 0x16llx 1210This is the current synchronization point for mirroring slaves. 1211This parameter is normally updated automatically by the 1212.Cm mirror-write 1213directive. 1214.Pp 1215Manually modifying this field is dangerous and can result in a broken mirror. 1216.It Cm shared-uuid= Ns Ar uuid 1217Set the shared UUID for this file system. 1218All mirrors must have the same shared UUID. 1219For safety purposes the 1220.Cm mirror-write 1221directives will refuse to operate on a target with a different shared UUID. 1222.Pp 1223Changing the shared UUID on an existing, non-empty mirroring target, 1224including an empty but not completely pruned target, 1225can lead to corruption of the mirroring target. 1226.It Cm unique-uuid= Ns Ar uuid 1227Set the unique UUID for this file system. 1228This UUID should not be used anywhere else, 1229even on exact copies of the file system. 1230.It Cm label= Ns Ar string 1231Set a descriptive label for this file system. 1232.It Cm snapshots= Ns Ar string 1233Specify the snapshots directory which 1234.Nm 1235.Cm cleanup 1236will use to manage this PFS. 1237.Bl -tag -width indent 1238.It Nm HAMMER No version 2- 1239The snapshots directory does not need to be configured for 1240PFS masters and will default to 1241.Pa <pfs>/snapshots . 1242.Pp 1243PFS slaves are mirroring slaves so you cannot configure a snapshots 1244directory on the slave itself to be managed by the slave's machine. 1245In fact, the slave will likely have a 1246.Pa snapshots 1247sub-directory mirrored 1248from the master, but that directory contains the configuration the master 1249is using for its copy of the file system, not the configuration that we 1250want to use for our slave. 1251.Pp 1252It is recommended that 1253.Pa <fs>/var/slaves/<name> 1254be configured for a PFS slave, where 1255.Pa <fs> 1256is the base 1257.Nm HAMMER 1258file system, and 1259.Pa <name> 1260is an appropriate label. 1261.It Nm HAMMER No version 3+ 1262The snapshots directory does not need to be configured for PFS masters or 1263slaves. 1264The snapshots directory defaults to 1265.Pa /var/hammer/<pfs> 1266.Pa ( /var/hammer/root 1267for root mount). 1268.El 1269.Pp 1270You can control snapshot retention on your slave independent of the master. 1271.It Cm snapshots-clear 1272Zero out the 1273.Cm snapshots 1274directory path for this PFS. 1275.It Cm prune-min= Ns Ar N Ns Cm d 1276.It Cm prune-min= Ns Oo Ar N Ns Cm d/ Oc Ns \ 1277Ar hh Ns Op Cm \&: Ns Ar mm Ns Op Cm \&: Ns Ar ss 1278Set the minimum fine-grained data retention period. 1279.Nm HAMMER 1280always retains fine-grained history up to the most recent snapshot. 1281You can extend the retention period further by specifying a non-zero 1282pruning minimum. 1283Any snapshot softlinks within the retention period are ignored 1284for the purposes of pruning (i.e.\& the fine grained history is retained). 1285Number of days, hours, minutes and seconds are given as 1286.Ar N , hh , mm 1287and 1288.Ar ss . 1289.Pp 1290Because the transaction id in the snapshot softlink cannot be used 1291to calculate a timestamp, 1292.Nm HAMMER 1293uses the earlier of the 1294.Fa st_ctime 1295or 1296.Fa st_mtime 1297field of the softlink to 1298determine which snapshots fall within the retention period. 1299Users must be sure to retain one of these two fields when manipulating 1300the softlink. 1301.El 1302.\" ==== pfs-upgrade ==== 1303.It Cm pfs-upgrade Ar dirpath 1304Upgrade a PFS from slave to master operation. 1305The PFS will be rolled back to the current end synchronization transaction id 1306(removing any partial synchronizations), and will then become writable. 1307.Pp 1308.Em WARNING! 1309.Nm HAMMER 1310currently supports only single masters and using 1311this command can easily result in file system corruption 1312if you don't know what you are doing. 1313.Pp 1314This directive will refuse to run if any programs have open descriptors 1315in the PFS, including programs chdir'd into the PFS. 1316.\" ==== pfs-downgrade ==== 1317.It Cm pfs-downgrade Ar dirpath 1318Downgrade a master PFS from master to slave operation. 1319The PFS becomes read-only and access will be locked to its 1320.Cm sync-end-tid . 1321.Pp 1322This directive will refuse to run if any programs have open descriptors 1323in the PFS, including programs chdir'd into the PFS. 1324.\" ==== pfs-destroy ==== 1325.It Cm pfs-destroy Ar dirpath 1326This permanently destroys a PFS. 1327.Pp 1328This directive will refuse to run if any programs have open descriptors 1329in the PFS, including programs chdir'd into the PFS. 1330As safety measure the 1331.Fl y 1332flag have no effect on this directive. 1333.\" ==== mirror-read ==== 1334.It Cm mirror-read Ar filesystem Op Ar begin-tid 1335Generate a mirroring stream to stdout. 1336The stream ends when the transaction id space has been exhausted. 1337.Ar filesystem 1338may be a master or slave PFS. 1339.\" ==== mirror-read-stream ==== 1340.It Cm mirror-read-stream Ar filesystem Op Ar begin-tid 1341Generate a mirroring stream to stdout. 1342Upon completion the stream is paused until new data is synced to the 1343.Ar filesystem , 1344then resumed. 1345Operation continues until the pipe is broken. 1346See the 1347.Cm mirror-stream 1348command for more details. 1349.\" ==== mirror-write ==== 1350.It Cm mirror-write Ar filesystem 1351Take a mirroring stream on stdin. 1352.Ar filesystem 1353must be a slave PFS. 1354.Pp 1355This command will fail if the 1356.Cm shared-uuid 1357configuration field for the two file systems do not match. 1358See the 1359.Cm mirror-copy 1360command for more details. 1361.Pp 1362If the target PFS does not exist this command will ask you whether 1363you want to create a compatible PFS slave for the target or not. 1364.\" ==== mirror-dump ==== 1365.It Ar mirror-dump Ar [header] 1366A 1367.Cm mirror-read 1368can be piped into a 1369.Cm mirror-dump 1370to dump an ASCII representation of the mirroring stream. 1371If the keyword 1372.Ar header 1373is specified, only the header information is shown. 1374.\" ==== mirror-copy ==== 1375.\".It Cm mirror-copy Ar [[user@]host:]filesystem [[user@]host:]filesystem 1376.It Cm mirror-copy \ 1377Oo Oo Ar user Ns Cm @ Oc Ns Ar host Ns Cm \&: Oc Ns Ar filesystem \ 1378Oo Oo Ar user Ns Cm @ Oc Ns Ar host Ns Cm \&: Oc Ns Ar filesystem 1379This is a shortcut which pipes a 1380.Cm mirror-read 1381command to a 1382.Cm mirror-write 1383command. 1384If a remote host specification is made the program forks a 1385.Xr ssh 1 1386(or other program as specified by the 1387.Ev HAMMER_RSH 1388environment variable) and execs the 1389.Cm mirror-read 1390and/or 1391.Cm mirror-write 1392on the appropriate host. 1393The source may be a master or slave PFS, and the target must be a slave PFS. 1394.Pp 1395This command also establishes full duplex communication and turns on 1396the 2-way protocol feature 1397.Fl ( 2 ) 1398which automatically negotiates transaction id 1399ranges without having to use a cyclefile. 1400If the operation completes successfully the target PFS's 1401.Cm sync-end-tid 1402will be updated. 1403Note that you must re-chdir into the target PFS to see the updated information. 1404If you do not you will still be in the previous snapshot. 1405.Pp 1406If the target PFS does not exist this command will ask you whether 1407you want to create a compatible PFS slave for the target or not. 1408.\" ==== mirror-stream ==== 1409.\".It Cm mirror-stream Ar [[user@]host:]filesystem [[user@]host:]filesystem 1410.It Cm mirror-stream \ 1411Oo Oo Ar user Ns Cm @ Oc Ns Ar host Ns Cm \&: Oc Ns Ar filesystem \ 1412Oo Oo Ar user Ns Cm @ Oc Ns Ar host Ns Cm \&: Oc Ns Ar filesystem 1413This is a shortcut which pipes a 1414.Cm mirror-read-stream 1415command to a 1416.Cm mirror-write 1417command. 1418This command works similarly to 1419.Cm mirror-copy 1420but does not exit after the initial mirroring completes. 1421The mirroring operation will resume as changes continue to be made to the 1422source. 1423The command is commonly used with 1424.Fl i Ar delay 1425and 1426.Fl b Ar bandwidth 1427options to keep the mirroring target in sync with the source on a continuing 1428basis. 1429.Pp 1430If the pipe is broken the command will automatically retry after sleeping 1431for a short while. 1432The time slept will be 15 seconds plus the time given in the 1433.Fl i 1434option. 1435.Pp 1436This command also detects the initial-mirroring case and spends some 1437time scanning the B-Tree to find good break points, allowing the initial 1438bulk mirroring operation to be broken down into 4GB pieces. 1439This means that the user can kill and restart the operation and it will 1440not have to start from scratch once it has gotten past the first chunk. 1441The 1442.Fl S 1443option may be used to change the size of pieces and the 1444.Fl B 1445option may be used to disable this feature and perform an initial bulk 1446transfer instead. 1447.\" ==== version ==== 1448.It Cm version Ar filesystem 1449This command returns the 1450.Nm HAMMER 1451file system version for the specified 1452.Ar filesystem 1453as well as the range of versions supported in the kernel. 1454The 1455.Fl q 1456option may be used to remove the summary at the end. 1457.\" ==== version-upgrade ==== 1458.It Cm version-upgrade Ar filesystem Ar version Op Cm force 1459Upgrade the 1460.Nm HAMMER 1461.Ar filesystem 1462to the specified 1463.Ar version . 1464Once upgraded a file system may not be downgraded. 1465If you wish to upgrade a file system to a version greater or equal to the 1466work-in-progress (WIP) version number you must specify the 1467.Cm force 1468directive. 1469Use of WIP versions should be relegated to testing and may require wiping 1470the file system as development progresses, even though the WIP version might 1471not change. 1472.Pp 1473.Em NOTE! 1474This command operates on the entire 1475.Nm HAMMER 1476file system and is not a per PFS operation. 1477All PFS's will be affected. 1478.Bl -tag -width indent 1479.It 1 1480.Dx 2.0 1481default version, first 1482.Nm HAMMER 1483release. 1484.It 2 1485.Dx 2.3 . 1486New directory entry layout. 1487This version is using a new directory hash key. 1488.It 3 1489.Dx 2.5 . 1490New snapshot management, using file system meta-data for saving 1491configuration file and snapshots (transaction ids etc.). 1492Also default snapshots directory has changed. 1493.It 4 1494.Dx 2.6 1495default version. 1496New undo/redo/flush, giving 1497.Nm HAMMER 1498a much faster sync and fsync. 1499.It 5 1500.Dx 2.9 . 1501Deduplication support. 1502.It 6 1503.Dx 2.9 . 1504Directory hash ALG1. 1505Tends to maintain inode number / directory name entry ordering better 1506for files after minor renaming. 1507.El 1508.El 1509.Sh PSEUDO-FILESYSTEM (PFS) NOTES 1510The root of a PFS is not hooked into the primary 1511.Nm HAMMER 1512file system as a directory. 1513Instead, 1514.Nm HAMMER 1515creates a special softlink called 1516.Ql @@PFS%05d 1517(exactly 10 characters long) in the primary 1518.Nm HAMMER 1519file system. 1520.Nm HAMMER 1521then modifies the contents of the softlink as read by 1522.Xr readlink 2 , 1523and thus what you see with an 1524.Nm ls 1525command or if you were to 1526.Nm cd 1527into the link. 1528If the PFS is a master the link reflects the current state of the PFS. 1529If the PFS is a slave the link reflects the last completed snapshot, and the 1530contents of the link will change when the next snapshot is completed, and 1531so forth. 1532.Pp 1533The 1534.Nm 1535utility employs numerous safeties to reduce user foot-shooting. 1536The 1537.Cm mirror-copy 1538directive requires that the target be configured as a slave and that the 1539.Cm shared-uuid 1540field of the mirroring source and target match. 1541.Sh DOUBLE_BUFFER MODE 1542There is a limit to the number of vnodes the kernel can cache, and because 1543file buffers are associated with a vnode the related data cache can get 1544blown away when operating on large numbers of files even if the system has 1545sufficient memory to hold the file data. 1546.Pp 1547If you turn on 1548.Nm HAMMER Ns 's 1549double buffer mode by setting the 1550.Xr sysctl 8 1551node 1552.Va vfs.hammer.double_buffer 1553to 1 1554.Nm HAMMER 1555will cache file data via the block device and copy it into the per-file 1556buffers as needed. The data will be double-cached at least until the 1557buffer cache throws away the file buffer. 1558This mode is typically used in conjunction with 1559.Xr swapcache 8 1560when 1561.Va vm.swapcache.data_enable 1562is turned on in order to prevent unnecessary re-caching of file data 1563due to vnode recycling. 1564The swapcache will save the cached VM pages related to 1565.Nm HAMMER Ns 's 1566block 1567device (which doesn't recycle unless you umount the filesystem) instead 1568of the cached VM pages backing the file vnodes. 1569.\".Pp 1570.\"Double buffering should also be turned on if live dedup is enabled via 1571.\"Va vfs.hammer.live_dedup . 1572.\"This is because the live dedup must validate the contents of a potential 1573.\"duplicate file block and it must run through the block device to do that 1574.\"and not the file vnode. 1575.\"If double buffering is not enabled then live dedup will create extra disk 1576.\"reads to validate potential data duplicates. 1577.Sh UPGRADE INSTRUCTIONS HAMMER V1 TO V2 1578This upgrade changes the way directory entries are stored. 1579It is possible to upgrade a V1 file system to V2 in place, but 1580directories created prior to the upgrade will continue to use 1581the old layout. 1582.Pp 1583Note that the slave mirroring code in the target kernel had bugs in 1584V1 which can create an incompatible root directory on the slave. 1585Do not mix a 1586.Nm HAMMER 1587master created after the upgrade with a 1588.Nm HAMMER 1589slave created prior to the upgrade. 1590.Pp 1591Any directories created after upgrading will use a new layout. 1592.Sh UPGRADE INSTRUCTIONS HAMMER V2 TO V3 1593This upgrade adds meta-data elements to the B-Tree. 1594It is possible to upgrade a V2 file system to V3 in place. 1595After issuing the upgrade be sure to run a 1596.Nm 1597.Cm cleanup 1598to perform post-upgrade tasks. 1599.Pp 1600After making this upgrade running a 1601.Nm 1602.Cm cleanup 1603will move the 1604.Pa <pfs>/snapshots 1605directory for each PFS mount into 1606.Pa /var/hammer/<pfs> . 1607A 1608.Nm HAMMER 1609root mount will migrate 1610.Pa /snapshots 1611into 1612.Pa /var/hammer/root . 1613Migration occurs only once and only if you have not specified 1614a snapshots directory in the PFS configuration. 1615If you have specified a snapshots directory in the PFS configuration no 1616automatic migration will occur. 1617.Pp 1618For slaves, if you desire, you can migrate your snapshots 1619config to the new location manually and then clear the 1620snapshot directory configuration in the slave PFS. 1621The new snapshots hierarchy is designed to work with 1622both master and slave PFSs equally well. 1623.Pp 1624In addition, the old config file will be moved to file system meta-data, 1625editable via the new 1626.Nm 1627.Cm viconfig 1628directive. 1629The old config file will be deleted. 1630Migration occurs only once. 1631.Pp 1632The V3 file system has new 1633.Cm snap* 1634directives for creating snapshots. 1635All snapshot directives, including the original, will create 1636meta-data entries for the snapshots and the pruning code will 1637automatically incorporate these entries into its list and 1638expire them the same way it expires softlinks. 1639If you by accident blow away your snapshot softlinks you can use the 1640.Cm snapls 1641directive to get a definitive list from the file system meta-data and 1642regenerate them from that list. 1643.Pp 1644.Em WARNING! 1645If you are using 1646.Nm 1647to backup file systems your scripts may be using the 1648.Cm synctid 1649directive to generate transaction ids. 1650This directive does not create a snapshot. 1651You will have to modify your scripts to use the 1652.Cm snapq 1653directive to generate the linkbuf for the softlink you create, or 1654use one of the other 1655.Cm snap* 1656directives. 1657The older 1658.Cm snapshot 1659directive will continue to work as expected and in V3 it will also 1660record the snapshot transaction id in file system meta-data. 1661You may also want to make use of the new 1662.Ar note 1663tag for the meta-data. 1664.Pp 1665.Em WARNING! 1666If you used to remove snapshot softlinks with 1667.Nm rm 1668you should probably start using the 1669.Cm snaprm 1670directive instead to also remove the related meta-data. 1671The pruning code scans the meta-data so just removing the 1672softlink is not sufficient. 1673.Sh UPGRADE INSTRUCTIONS HAMMER V3 TO V4 1674This upgrade changes undo/flush, giving faster sync. 1675It is possible to upgrade a V3 file system to V4 in place. 1676This upgrade reformats the UNDO/REDO FIFO (typically 1GB), 1677so upgrade might take a minute or two depending. 1678.Pp 1679Version 4 allows the UNDO/REDO FIFO to be flushed without also having 1680to flush the volume header, removing 2 of the 4 disk syncs typically 1681required for an 1682.Fn fsync 1683and removing 1 of the 2 disk syncs typically 1684required for a flush sequence. 1685Version 4 also implements the REDO log (see 1686.Sx FSYNC FLUSH MODES 1687below) which is capable 1688of fsync()ing with either one disk flush or zero disk flushes. 1689.Sh UPGRADE INSTRUCTIONS HAMMER V4 TO V5 1690This upgrade brings in deduplication support. 1691It is possible to upgrade a V4 file system to V5 in place. 1692Technically it makes the layer2 1693.Va bytes_free 1694field a signed value instead of unsigned, allowing it to go negative. 1695A version 5 filesystem is required for dedup operation. 1696.Sh UPGRADE INSTRUCTIONS HAMMER V5 TO V6 1697It is possible to upgrade a V5 file system to V6 in place. 1698.Sh FSYNC FLUSH MODES 1699.Nm HAMMER 1700implements five different fsync flush modes via the 1701.Va vfs.hammer.fsync_mode 1702sysctl, for 1703.Nm HAMMER 1704version 4+ file systems. 1705.Pp 1706As of 1707.Dx 2.6 1708fsync mode 3 is set by default. 1709REDO operation and recovery is enabled by default. 1710.Bl -tag -width indent 1711.It mode 0 1712Full synchronous fsync semantics without REDO. 1713.Pp 1714.Nm HAMMER 1715will not generate REDOs. 1716A 1717.Fn fsync 1718will completely sync 1719the data and meta-data and double-flush the FIFO, including 1720issuing two disk synchronization commands. 1721The data is guaranteed 1722to be on the media as of when 1723.Fn fsync 1724returns. 1725Needless to say, this is slow. 1726.It mode 1 1727Relaxed asynchronous fsync semantics without REDO. 1728.Pp 1729This mode works the same as mode 0 except the last disk synchronization 1730command is not issued. 1731It is faster than mode 0 but not even remotely 1732close to the speed you get with mode 2 or mode 3. 1733.Pp 1734Note that there is no chance of meta-data corruption when using this 1735mode, it simply means that the data you wrote and then 1736.Fn fsync Ns 'd 1737might not have made it to the media if the storage system crashes at a bad 1738time. 1739.It mode 2 1740Full synchronous fsync semantics using REDO. 1741NOTE: If not running a 1742.Nm HAMMER 1743version 4 filesystem or later mode 0 is silently used. 1744.Pp 1745.Nm HAMMER 1746will generate REDOs in the UNDO/REDO FIFO based on a heuristic. 1747If this is sufficient to satisfy the 1748.Fn fsync 1749operation the blocks will be written out and 1750.Nm HAMMER 1751will wait for the I/Os to complete, 1752and then followup with a disk sync command to guarantee the data 1753is on the media before returning. 1754This is slower than mode 3 and can result in significant disk or 1755SSDs overheads, though not as bad as mode 0 or mode 1. 1756.It mode 3 1757Relaxed asynchronous fsync semantics using REDO. 1758NOTE: If not running a 1759.Nm HAMMER 1760version 4 filesystem or later mode 1 is silently used. 1761.Pp 1762.Nm HAMMER 1763will generate REDOs in the UNDO/REDO FIFO based on a heuristic. 1764If this is sufficient to satisfy the 1765.Fn fsync 1766operation the blocks 1767will be written out and 1768.Nm HAMMER 1769will wait for the I/Os to complete, 1770but will 1771.Em NOT 1772issue a disk synchronization command. 1773.Pp 1774Note that there is no chance of meta-data corruption when using this 1775mode, it simply means that the data you wrote and then 1776.Fn fsync Ns 'd 1777might 1778not have made it to the media if the storage system crashes at a bad 1779time. 1780.Pp 1781This mode is the fastest production fsyncing mode available. 1782This mode is equivalent to how the UFS fsync in the 1783.Bx Ns s 1784operates. 1785.It mode 4 1786fsync is ignored. 1787.Pp 1788Calls to 1789.Fn fsync 1790will be ignored. 1791This mode is primarily designed 1792for testing and should not be used on a production system. 1793.El 1794.Sh RESTORING FROM A SNAPSHOT BACKUP 1795You restore a snapshot by copying it over to live, but there is a caveat. 1796The mtime and atime fields for files accessed via a snapshot is locked 1797to the ctime in order to keep the snapshot consistent, because neither 1798mtime nor atime changes roll any history. 1799.Pp 1800In order to avoid unnecessary copying it is recommended that you use 1801.Nm cpdup 1802.Fl VV 1803.Fl v 1804when doing the copyback. 1805Also make sure you traverse the snapshot softlink by appending a ".", 1806as in "<snapshotpath>/.", and you match up the directory properly. 1807.Sh RESTORING A PFS FROM A MIRROR 1808A PFS can be restored from a mirror with 1809.Cm mirror-copy . 1810.Cm config 1811data must be copied separately. 1812At last the PFS can be upgraded to master using 1813.Cm pfs-upgrade . 1814.Pp 1815It is not possible to restore the root PFS (PFS# 0) by using mirroring, 1816as the root PFS is always a master PFS. 1817A normal copy (e.g.\& using 1818.Xr cpdup 1 ) 1819must be done, ignoring history. 1820If history is important, old root PFS can me restored to a new PFS, and 1821important directories/files can be 1822.Nm null 1823mounted to the new PFS. 1824.Sh ENVIRONMENT 1825The following environment variables affect the execution of 1826.Nm : 1827.Bl -tag -width ".Ev EDITOR" 1828.It Ev EDITOR 1829The editor program specified in the variable 1830.Ev EDITOR 1831will be invoked instead of the default editor, which is 1832.Xr vi 1 . 1833.It Ev HAMMER_RSH 1834The command specified in the variable 1835.Ev HAMMER_RSH 1836will be used to initiate remote operations for the mirror-copy and 1837mirror-stream commands instead of the default command, which is 1838.Xr ssh 1 . 1839The program will be invoked via 1840.Xr execvp 3 1841using a typical 1842.Xr rsh 1 1843style 1844.Cm -l user host <remote-command> 1845command line. 1846.It Ev VISUAL 1847Same effect as 1848.Ev EDITOR 1849variable. 1850.El 1851.Sh FILES 1852.Bl -tag -width ".It Pa <fs>/var/slaves/<name>" -compact 1853.It Pa <pfs>/snapshots 1854default per PFS snapshots directory 1855.Nm ( HAMMER 1856VERSION 2-) 1857.It Pa /var/hammer/<pfs> 1858default per PFS snapshots directory (not root) 1859.Nm ( HAMMER 1860VERSION 3+) 1861.It Pa /var/hammer/root 1862default snapshots directory for root directory 1863.Nm ( HAMMER 1864VERSION 3+) 1865.It Pa <snapshots>/config 1866per PFS 1867.Nm 1868.Cm cleanup 1869configuration file 1870.Nm ( HAMMER 1871VERSION 2-) 1872.It Pa <fs>/var/slaves/<name> 1873recommended slave PFS snapshots directory 1874.Nm ( HAMMER 1875VERSION 2-) 1876.It Pa <fs>/pfs 1877recommended PFS directory 1878.El 1879.Sh EXIT STATUS 1880.Ex -std 1881.Sh SEE ALSO 1882.Xr ssh 1 , 1883.Xr undo 1 , 1884.Xr HAMMER 5 , 1885.Xr periodic.conf 5 , 1886.Xr loader 8 , 1887.Xr mount_hammer 8 , 1888.Xr mount_null 8 , 1889.Xr newfs_hammer 8 , 1890.Xr swapcache 8 , 1891.Xr sysctl 8 1892.Sh HISTORY 1893The 1894.Nm 1895utility first appeared in 1896.Dx 1.11 . 1897.Sh AUTHORS 1898.An Matthew Dillon Aq Mt dillon@backplane.com 1899