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 May 18, 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.\" ==== stats ==== 309.It Cm stats Op Ar interval 310Output 311.Nm HAMMER 312B-Tree and 313.Tn I/O 314statistics until interrupted. 315Pause 316.Ar interval 317seconds between each display. 318The default interval is one second. 319.\" ==== history ==== 320.It Cm history Ns Oo Cm @ Ns Ar offset Ns Oo Cm \&, Ns Ar length Oc Oc Ar path ... 321Show the modification history for inode and data of 322.Nm HAMMER 323files. 324If 325.Ar offset 326is given history is shown for data block at given offset, 327otherwise history is shown for inode. 328If 329.Fl v 330is specified 331.Ar length 332data bytes at given offset are dumped for each version, 333default is 32. 334.Pp 335For each 336.Ar path 337this directive shows object id and sync status, 338and for each object version it shows transaction id and time stamp. 339Files has to exist for this directive to be applicable, 340to track inodes which has been deleted or renamed see 341.Xr undo 1 . 342.\" ==== blockmap ==== 343.It Cm blockmap 344Dump the blockmap for the file system. 345The 346.Nm HAMMER 347blockmap is two-layer 348blockmap representing the maximum possible file system size of 1 Exabyte. 349Needless to say the second layer is only present for blocks which exist. 350.Nm HAMMER Ns 's 351blockmap represents 8-Megabyte blocks, called big-blocks. 352Each big-block has an append 353point, a free byte count, and a typed zone id which allows content to be 354reverse engineered to some degree. 355.Pp 356In 357.Nm HAMMER 358allocations are essentially appended to a selected big-block using 359the append offset and deducted from the free byte count. 360When space is freed the free byte count is adjusted but 361.Nm HAMMER 362does not track holes in big-blocks for reallocation. 363A big-block must be completely freed, either 364through normal file system operations or through reblocking, before 365it can be reused. 366.Pp 367Data blocks can be shared by deducting the space used from the free byte 368count for each shared references. 369This means the free byte count can legally go negative. 370.Pp 371This command needs the 372.Fl f Ar blkdevs 373option. 374.\" ==== checkmap ==== 375.It Cm checkmap 376Check the blockmap allocation count. 377.Nm 378will scan the freemap, B-Tree, UNDO FIFO, then collect allocation information, 379and construct a blockmap in-memory. 380It will then check that blockmap against the on-disk blockmap. 381.Pp 382This command needs the 383.Fl f Ar blkdevs 384option. 385.\" ==== show ==== 386.It Cm show Op Ar localization Ns Op Cm \&: Ns Ar object_id 387Dump the B-Tree. 388By default this command will validate all B-Tree 389linkages and CRCs, including data CRCs, and will report the most verbose 390information it can dig up. 391Any errors will show up with a 392.Ql B 393in column 1 along with various 394other error flags. 395.Pp 396If you specify 397.Ar localization 398or 399.Ar localization Ns Cm \&: Ns Ar object_id 400the dump will 401search for the key printing nodes as it recurses down, and then 402will iterate forwards. 403These fields are specified in HEX. 404Note that the pfsid is the top 16 bits of the 32-bit localization 405field so PFS #1 would be 00010000. 406.Pp 407If you use 408.Fl q 409the command will report less information about the inode contents. 410.Pp 411If you use 412.Fl qq 413the command will not report the content of the inode or other typed 414data at all. 415.Pp 416If you use 417.Fl qqq 418the command will not report volume header information, big-block fill 419ratios, mirror transaction ids, or report or check data CRCs. 420B-Tree CRCs and linkages are still checked. 421.Pp 422This command needs the 423.Fl f Ar blkdevs 424option. 425.\" ==== show-undo ==== 426.It Cm show-undo 427.Nm ( HAMMER 428VERSION 4+) 429Dump the UNDO/REDO map. 430.Pp 431This command needs the 432.Fl f Ar blkdevs 433option. 434.\" .It Ar blockmap 435.\" Dump the B-Tree, record, large-data, and small-data blockmaps, showing 436.\" physical block assignments and free space percentages. 437.\" ==== ssh-remote ==== 438.It Cm ssh-remote Ar command Ar targetdir 439Used in a ssh authorized_keys line such as 440command="/sbin/hammer ssh-remote mirror-read /fubarmount" ... to allow 441mirror-read or mirror-write access to a particular subdirectory tree. 442This way you do not have to give shell access to the remote box. 443.Nm 444will obtain the original command line from the SSH_ORIGINAL_COMMAND 445environment variable, validate it against the restriction, and then 446re-exec hammer with the validated arguments. 447.Pp 448The remote hammer command does not allow the 449.Fl c 450or 451.Fl f 452options to be passed in. 453.\" ==== recover ==== 454.It Cm recover Ar targetdir 455Recover data from a corrupted 456.Nm HAMMER 457filesystem. 458This is a low level command which operates on the filesystem image and 459attempts to locate and recover files from a corrupted filesystem. 460The entire image is scanned linearly looking for B-Tree nodes. 461Any node 462found which passes its CRC test is scanned for file, inode, and directory 463fragments and the target directory is populated with the resulting data. 464files and directories in the target directory are initially named after 465the object id and are renamed as fragmentary information is processed. 466.Pp 467This command keeps track of filename/object_id translations and may eat a 468considerably amount of memory while operating. 469.Pp 470This command is literally the last line of defense when it comes to 471recovering data from a dead filesystem. 472.Pp 473This command needs the 474.Fl f Ar blkdevs 475option. 476.\" ==== namekey1 ==== 477.It Cm namekey1 Ar filename 478Generate a 479.Nm HAMMER 48064-bit directory hash for the specified file name, using 481the original directory hash algorithm in version 1 of the file system. 482The low 32 bits are used as an iterator for hash collisions and will be 483output as 0. 484.\" ==== namekey2 ==== 485.It Cm namekey2 Ar filename 486Generate a 487.Nm HAMMER 48864-bit directory hash for the specified file name, using 489the new directory hash algorithm in version 2 of the file system. 490The low 32 bits are still used as an iterator but will start out containing 491part of the hash key. 492.\" ==== namekey32 ==== 493.It Cm namekey32 Ar filename 494Generate the top 32 bits of a 495.Nm HAMMER 49664 bit directory hash for the specified file name. 497.\" ==== info ==== 498.It Cm info Ar dirpath ... 499Show extended information about all 500.Nm HAMMER 501file systems mounted in the system or the one mounted in 502.Ar dirpath 503when this argument is specified. 504.Pp 505The information is divided into sections: 506.Bl -tag -width indent 507.It Volume identification 508General information, like the label of the 509.Nm HAMMER 510filesystem, the number of volumes it contains, the FSID, and the 511.Nm HAMMER 512version being used. 513.It Big-block information 514Big-block statistics, such as total, used, reserved and free big-blocks. 515.It Space information 516Information about space used on the filesystem. 517Currently total size, used, reserved and free space are displayed. 518.It PFS information 519Basic information about the PFSs currently present on a 520.Nm HAMMER 521filesystem. 522.Pp 523.Dq PFS ID 524is the ID of the PFS, with 0 being the root PFS. 525.Dq Snaps 526is the current snapshot count on the PFS. 527.Dq Mounted on 528displays the mount point of the PFS is currently mounted on (if any). 529.El 530.\" ==== cleanup ==== 531.It Cm cleanup Op Ar filesystem ... 532This is a meta-command which executes snapshot, prune, rebalance, dedup 533and reblock commands on the specified 534.Nm HAMMER 535file systems. 536If no 537.Ar filesystem 538is specified this command will clean-up all 539.Nm HAMMER 540file systems in use, including PFS's. 541To do this it will scan all 542.Nm HAMMER 543and 544.Nm null 545mounts, extract PFS id's, and clean-up each PFS found. 546.Pp 547This command will access a snapshots 548directory and a configuration file for each 549.Ar filesystem , 550creating them if necessary. 551.Bl -tag -width indent 552.It Nm HAMMER No version 2- 553The configuration file is 554.Pa config 555in the snapshots directory which defaults to 556.Pa <pfs>/snapshots . 557.It Nm HAMMER No version 3+ 558The configuration file is saved in file system meta-data, see 559.Nm 560.Cm config . 561The snapshots directory defaults to 562.Pa /var/hammer/<pfs> 563.Pa ( /var/hammer/root 564for root mount). 565.El 566.Pp 567The format of the configuration file is: 568.Bd -literal -offset indent 569snapshots <period> <retention-time> [any] 570prune <period> <max-runtime> 571rebalance <period> <max-runtime> 572dedup <period> <max-runtime> 573reblock <period> <max-runtime> 574recopy <period> <max-runtime> 575.Ed 576.Pp 577Defaults are: 578.Bd -literal -offset indent 579snapshots 1d 60d # 0d 0d for PFS /tmp, /var/tmp, /usr/obj 580prune 1d 5m 581rebalance 1d 5m 582dedup 1d 5m 583reblock 1d 5m 584recopy 30d 10m 585.Ed 586.Pp 587Time is given with a suffix of 588.Cm d , 589.Cm h , 590.Cm m 591or 592.Cm s 593meaning day, hour, minute and second. 594.Pp 595If the 596.Cm snapshots 597directive has a period of 0 and a retention time of 0 598then snapshot generation is disabled, removal of old snapshots are 599disabled, and prunes will use 600.Cm prune-everything . 601.Pp 602If the 603.Cm snapshots 604directive has a period of 0 but a non-zero retention time 605then this command will not create any new snapshots but will remove old 606snapshots it finds based on the retention time. 607This form should be 608used on PFS masters where you are generating your own snapshot softlinks 609manually and on PFS slaves when all you wish to do is prune away existing 610snapshots inherited via the mirroring stream. 611.Pp 612By default only snapshots in the form 613.Ql snap- Ns Ar yyyymmdd Ns Op - Ns Ar HHMM 614are processed. 615If the 616.Cm any 617directive is specified as a third argument on the 618.Cm snapshots 619config line then any softlink of the form 620.Ql *- Ns Ar yyyymmdd Ns Op - Ns Ar HHMM 621or 622.Ql *. Ns Ar yyyymmdd Ns Op - Ns Ar HHMM 623will be processed. 624.Pp 625A period of 0 for prune, rebalance, dedup, reblock or recopy disables the directive. 626A max-runtime of 0 means unlimited. 627.Pp 628If period hasn't passed since the previous 629.Cm cleanup 630run nothing is done. 631For example a day has passed when midnight is passed (localtime). 632If the 633.Fl F 634flag is given the period is ignored. 635By default, 636.Dx 637is set up to run 638.Nm Cm cleanup 639nightly via 640.Xr periodic 8 . 641.Pp 642The default configuration file will create a daily snapshot, do a daily 643pruning, rebalancing, deduping and reblocking run and a monthly recopy run. 644Reblocking is defragmentation with a level of 95%, 645and recopy is full defragmentation. 646.Pp 647By default prune, dedup and rebalance operations are time limited to 5 minutes, 648and reblock operations to a bit over 5 minutes, 649and recopy operations to a bit over 10 minutes. 650Reblocking and recopy runs are each broken down into four separate functions: 651btree, inodes, dirs and data. 652Each function is time limited to the time given in the configuration file, 653but the btree, inodes and dirs functions usually does not take very long time, 654full defragmentation is always used for these three functions. 655Also note that this directive will by default disable snapshots on 656the following PFS's: 657.Pa /tmp , 658.Pa /var/tmp 659and 660.Pa /usr/obj . 661.Pp 662The defaults may be adjusted by modifying the configuration file. 663The pruning and reblocking commands automatically maintain a cyclefile 664for incremental operation. 665If you interrupt (^C) the program the cyclefile will be updated, 666but a sub-command 667may continue to run in the background for a few seconds until the 668.Nm HAMMER 669ioctl detects the interrupt. 670The 671.Cm snapshots 672PFS option can be set to use another location for the snapshots directory. 673.Pp 674Work on this command is still in progress. 675Expected additions: 676An ability to remove snapshots dynamically as the 677file system becomes full. 678.\" ==== abort-cleanup ==== 679.It Cm abort-cleanup 680This command will terminate all active 681.Cm cleanup 682processes. 683.\" ==== config ==== 684.It Cm config Op Ar filesystem Op Ar configfile 685.Nm ( HAMMER 686VERSION 3+) 687Show or change configuration for 688.Ar filesystem . 689If zero or one arguments are specified this function dumps the current 690configuration file to stdout. 691Zero arguments specifies the PFS containing the current directory. 692This configuration file is stored in file system meta-data. 693If two arguments are specified this function installs a new config file. 694.Pp 695In 696.Nm HAMMER 697versions less than 3 the configuration file is by default stored in 698.Pa <pfs>/snapshots/config , 699but in all later versions the configuration file is stored in file system 700meta-data. 701.\" ==== viconfig ==== 702.It Cm viconfig Op Ar filesystem 703.Nm ( HAMMER 704VERSION 3+) 705Edit the configuration file and reinstall into file system meta-data when done. 706Zero arguments specifies the PFS containing the current directory. 707.\" ==== volume-add ==== 708.It Cm volume-add Ar device Ar filesystem 709Add volume 710.Ar device 711to 712.Ar filesystem . 713This will format 714.Ar device 715and add all of its space to 716.Ar filesystem . 717A 718.Nm HAMMER 719file system can use up to 256 volumes. 720.Pp 721.Em NOTE! 722All existing data contained on 723.Ar device 724will be destroyed by this operation! 725If 726.Ar device 727contains a valid 728.Nm HAMMER 729file system, formatting will be denied. 730You can overcome this sanity check by using 731.Xr dd 1 732to erase the beginning sectors of the device. 733.Pp 734Remember that you have to specify 735.Ar device , 736together with any other device that make up the file system, 737colon-separated to 738.Pa /etc/fstab 739and 740.Xr mount_hammer 8 . 741If 742.Ar filesystem 743is root file system, also remember to add 744.Ar device 745to 746.Va vfs.root.mountfrom 747in 748.Pa /boot/loader.conf , 749see 750.Xr loader 8 . 751.\" ==== volume-del ==== 752.It Cm volume-del Ar device Ar filesystem 753Remove volume 754.Ar device 755from 756.Ar filesystem . 757.Pp 758Remember that you have to remove 759.Ar device 760from the colon-separated list in 761.Pa /etc/fstab 762and 763.Xr mount_hammer 8 . 764If 765.Ar filesystem 766is root file system, also remember to remove 767.Ar device 768from 769.Va vfs.root.mountfrom 770in 771.Pa /boot/loader.conf , 772see 773.Xr loader 8 . 774.Pp 775It is not possible to remove the 776.Ar root-volume 777as it contains 778.Ar filesystem 779meta data such as 780.Nm HAMMER Ns 's 781layer1 blockmap and UNDO FIFO. 782.Pp 783This command may 784.Cm reblock 785filesystem before it attemps to remove the volume if the volume is not empty. 786.\" ==== volume-list ==== 787.It Cm volume-list Ar filesystem 788List the volumes that make up 789.Ar filesystem . 790.\" ==== volume-blkdevs ==== 791.It Cm volume-blkdevs Ar filesystem 792List the volumes that make up 793.Ar filesystem 794in 795.Ar blkdevs 796format. 797.\" ==== snapshot ==== 798.It Cm snapshot Oo Ar filesystem Oc Ar snapshot-dir 799.It Cm snapshot Ar filesystem Ar snapshot-dir Op Ar note 800Take a snapshot of the file system either explicitly given by 801.Ar filesystem 802or implicitly derived from the 803.Ar snapshot-dir 804argument and creates a symlink in the directory provided by 805.Ar snapshot-dir 806pointing to the snapshot. 807If 808.Ar snapshot-dir 809is not a directory, it is assumed to be a format string passed to 810.Xr strftime 3 811with the current time as parameter. 812If 813.Ar snapshot-dir 814refers to an existing directory, a default format string of 815.Ql snap-%Y%m%d-%H%M 816is assumed and used as name for the newly created symlink. 817.Pp 818Snapshot is a per PFS operation, so each PFS in a 819.Nm HAMMER 820file system have to be snapshot separately. 821.Pp 822Example, assuming that 823.Pa /mysnapshots 824is on file system 825.Pa / 826and that 827.Pa /obj 828and 829.Pa /usr 830are file systems on their own, the following invocations: 831.Bd -literal -offset indent 832hammer snapshot /mysnapshots 833 834hammer snapshot /mysnapshots/%Y-%m-%d 835 836hammer snapshot /obj /mysnapshots/obj-%Y-%m-%d 837 838hammer snapshot /usr /my/snaps/usr "note" 839.Ed 840.Pp 841Would create symlinks similar to: 842.Bd -literal -offset indent 843/mysnapshots/snap-20080627-1210 -> /@@0x10d2cd05b7270d16 844 845/mysnapshots/2008-06-27 -> /@@0x10d2cd05b7270d16 846 847/mysnapshots/obj-2008-06-27 -> /obj@@0x10d2cd05b7270d16 848 849/my/snaps/usr/snap-20080627-1210 -> /usr@@0x10d2cd05b7270d16 850.Ed 851.Pp 852When run on a 853.Nm HAMMER 854version 3+ file system the snapshot is also recorded in file system meta-data 855along with the optional 856.Ar note . 857See the 858.Cm snapls 859directive. 860.\" ==== snap* ==== 861.It Cm snap Ar path Op Ar note 862.Nm ( HAMMER 863VERSION 3+) 864Create a snapshot for the PFS containing 865.Ar path 866and create a snapshot softlink. 867If the path specified is a 868directory a standard snapshot softlink will be created in the directory. 869The snapshot softlink points to the base of the mounted PFS. 870.It Cm snaplo Ar path Op Ar note 871.Nm ( HAMMER 872VERSION 3+) 873Create a snapshot for the PFS containing 874.Ar path 875and create a snapshot softlink. 876If the path specified is a 877directory a standard snapshot softlink will be created in the directory. 878The snapshot softlink points into the directory it is contained in. 879.It Cm snapq Ar dir Op Ar note 880.Nm ( HAMMER 881VERSION 3+) 882Create a snapshot for the PFS containing the specified directory but do 883not create a softlink. 884Instead output a path which can be used to access 885the directory via the snapshot. 886.Pp 887An absolute or relative path may be specified. 888The path will be used as-is as a prefix in the path output to stdout. 889As with the other 890snap and snapshot directives the snapshot transaction id will be registered 891in the file system meta-data. 892.It Cm snaprm Ar path Ar ... 893.It Cm snaprm Ar transaction_id Ar ... 894.It Cm snaprm Ar filesystem Ar transaction_id Ar ... 895.Nm ( HAMMER 896VERSION 3+) 897Remove a snapshot given its softlink or transaction id. 898If specifying a transaction id 899the snapshot is removed from file system meta-data but you are responsible 900for removing any related softlinks. 901.Pp 902If a softlink path is specified the filesystem and transaction id 903is derived from the contents of the softlink. 904If just a transaction id is specified it is assumed to be a snapshot in the 905.Nm HAMMER 906filesystem you are currently chdir'd into. 907You can also specify the filesystem and transaction id explicitly. 908.It Cm snapls Op Ar path ... 909.Nm ( HAMMER 910VERSION 3+) 911Dump the snapshot meta-data for PFSs containing each 912.Ar path 913listing all available snapshots and their notes. 914If no arguments are specified snapshots for the PFS containing the 915current directory are listed. 916This is the definitive list of snapshots for the file system. 917.\" ==== prune ==== 918.It Cm prune Ar softlink-dir 919Prune the file system based on previously created snapshot softlinks. 920Pruning is the act of deleting file system history. 921The 922.Cm prune 923command will delete file system history such that 924the file system state is retained for the given snapshots, 925and all history after the latest snapshot. 926By setting the per PFS parameter 927.Cm prune-min , 928history is guaranteed to be saved at least this time interval. 929All other history is deleted. 930.Pp 931The target directory is expected to contain softlinks pointing to 932snapshots of the file systems you wish to retain. 933The directory is scanned non-recursively and the mount points and 934transaction ids stored in the softlinks are extracted and sorted. 935The file system is then explicitly pruned according to what is found. 936Cleaning out portions of the file system is as simple as removing a 937snapshot softlink and then running the 938.Cm prune 939command. 940.Pp 941As a safety measure pruning only occurs if one or more softlinks are found 942containing the 943.Ql @@ 944snapshot id extension. 945Currently the scanned softlink directory must contain softlinks pointing 946to a single 947.Nm HAMMER 948mount. 949The softlinks may specify absolute or relative paths. 950Softlinks must use 20-character 951.Ql @@0x%016llx 952transaction ids, as might be returned from 953.Nm Cm synctid Ar filesystem . 954.Pp 955Pruning is a per PFS operation, so each PFS in a 956.Nm HAMMER 957file system have to be pruned separately. 958.Pp 959Note that pruning a file system may not immediately free-up space, 960though typically some space will be freed if a large number of records are 961pruned out. 962The file system must be reblocked to completely recover all available space. 963.Pp 964Example, lets say your that you didn't set 965.Cm prune-min , 966and snapshot directory contains the following links: 967.Bd -literal -offset indent 968lrwxr-xr-x 1 root wheel 29 May 31 17:57 snap1 -> 969/usr/obj/@@0x10d2cd05b7270d16 970 971lrwxr-xr-x 1 root wheel 29 May 31 17:58 snap2 -> 972/usr/obj/@@0x10d2cd13f3fde98f 973 974lrwxr-xr-x 1 root wheel 29 May 31 17:59 snap3 -> 975/usr/obj/@@0x10d2cd222adee364 976.Ed 977.Pp 978If you were to run the 979.Cm prune 980command on this directory, then the 981.Nm HAMMER 982.Pa /usr/obj 983mount will be pruned to retain the above three snapshots. 984In addition, history for modifications made to the file system older than 985the oldest snapshot will be destroyed and history for potentially fine-grained 986modifications made to the file system more recently than the most recent 987snapshot will be retained. 988.Pp 989If you then delete the 990.Pa snap2 991softlink and rerun the 992.Cm prune 993command, 994history for modifications pertaining to that snapshot would be destroyed. 995.Pp 996In 997.Nm HAMMER 998file system versions 3+ this command also scans the snapshots stored 999in the file system meta-data and includes them in the prune. 1000.\" ==== prune-everything ==== 1001.It Cm prune-everything Ar filesystem 1002Remove all historical records from 1003.Ar filesystem . 1004Use this directive with caution on PFSs where you intend to use history. 1005.Pp 1006This command does not remove snapshot softlinks but will delete all 1007snapshots recorded in file system meta-data (for file system version 3+). 1008The user is responsible for deleting any softlinks. 1009.Pp 1010Pruning is a per PFS operation, so each PFS in a 1011.Nm HAMMER 1012file system have to be pruned separately. 1013.\" ==== rebalance ==== 1014.It Cm rebalance Ar filesystem Op Ar saturation_percentage 1015Rebalance the B-Tree, nodes with small number of 1016elements will be combined and element counts will be smoothed out 1017between nodes. 1018.Pp 1019The saturation percentage is between 50% and 100%. 1020The default is 85% (the 1021.Sq % 1022suffix is not needed). 1023.Pp 1024Rebalancing is a per PFS operation, so each PFS in a 1025.Nm HAMMER 1026file system have to be rebalanced separately. 1027.\" ==== dedup ==== 1028.It Cm dedup Ar filesystem 1029.Nm ( HAMMER 1030VERSION 5+) 1031Perform offline (post-process) deduplication. 1032Deduplication occurs at 1033the block level, currently only data blocks of the same size can be 1034deduped, metadata blocks can not. 1035The hash function used for comparing 1036data blocks is CRC-32 (CRCs are computed anyways as part of 1037.Nm HAMMER 1038data integrity features, so there's no additional overhead). 1039Since CRC is a weak hash function a byte-by-byte comparison is done 1040before actual deduping. 1041In case of a CRC collision (two data blocks have the same CRC 1042but different contents) the checksum is upgraded to SHA-256. 1043.Pp 1044Currently 1045.Nm HAMMER 1046reblocker may partially blow up (re-expand) dedup (reblocker's normal 1047operation is to reallocate every record, so it's possible for deduped 1048blocks to be re-expanded back). 1049.Pp 1050Deduplication is a per PFS operation, so each PFS in a 1051.Nm HAMMER 1052file system have to be deduped separately. 1053This also 1054means that if you have duplicated data in two different PFSs that data 1055won't be deduped, however the addition of such feature is planned. 1056.Pp 1057The 1058.Fl m Ar memlimit 1059option should be used to limit memory use during the dedup run if the 1060default 1G limit is too much for the machine. 1061.\" ==== dedup-simulate ==== 1062.It Cm dedup-simulate Ar filesystem 1063Shows potential space savings (simulated dedup ratio) one can get after 1064running 1065.Cm dedup 1066command. 1067If the estimated dedup ratio is greater than 1.00 you will see 1068dedup space savings. 1069Remember that this is an estimated number, in 1070practice real dedup ratio will be slightly smaller because of 1071.Nm HAMMER 1072big-block underflows, B-Tree locking issues and other factors. 1073.Pp 1074Note that deduplication currently works only on bulk data so if you 1075try to run 1076.Cm dedup-simulate 1077or 1078.Cm dedup 1079commands on a PFS that contains metadata only (directory entries, 1080softlinks) you will get a 0.00 dedup ratio. 1081.Pp 1082The 1083.Fl m Ar memlimit 1084option should be used to limit memory use during the dedup run if the 1085default 1G limit is too much for the machine. 1086.\" ==== reblock* ==== 1087.It Cm reblock Ar filesystem Op Ar fill_percentage 1088.It Cm reblock-btree Ar filesystem Op Ar fill_percentage 1089.It Cm reblock-inodes Ar filesystem Op Ar fill_percentage 1090.It Cm reblock-dirs Ar filesystem Op Ar fill_percentage 1091.It Cm reblock-data Ar filesystem Op Ar fill_percentage 1092Attempt to defragment and free space for reuse by reblocking a live 1093.Nm HAMMER 1094file system. 1095Big-blocks cannot be reused by 1096.Nm HAMMER 1097until they are completely free. 1098This command also has the effect of reordering all elements, effectively 1099defragmenting the file system. 1100.Pp 1101The default fill percentage is 100% and will cause the file system to be 1102completely defragmented. 1103All specified element types will be reallocated and rewritten. 1104If you wish to quickly free up space instead try specifying 1105a smaller fill percentage, such as 90% or 80% (the 1106.Sq % 1107suffix is not needed). 1108.Pp 1109Since this command may rewrite the entire contents of the disk it is 1110best to do it incrementally from a 1111.Xr cron 8 1112job along with the 1113.Fl c Ar cyclefile 1114and 1115.Fl t Ar seconds 1116options to limit the run time. 1117The file system would thus be defragmented over long period of time. 1118.Pp 1119It is recommended that separate invocations be used for each data type. 1120B-Tree nodes, inodes, and directories are typically the most important 1121elements needing defragmentation. 1122Data can be defragmented over a longer period of time. 1123.Pp 1124Reblocking is a per PFS operation, so each PFS in a 1125.Nm HAMMER 1126file system have to be reblocked separately. 1127.\" ==== pfs-status ==== 1128.It Cm pfs-status Ar dirpath ... 1129Retrieve the mirroring configuration parameters for the specified 1130.Nm HAMMER 1131file systems or pseudo-filesystems (PFS's). 1132.\" ==== pfs-master ==== 1133.It Cm pfs-master Ar dirpath Op Ar options 1134Create a pseudo-filesystem (PFS) inside a 1135.Nm HAMMER 1136file system. 1137Up to 65536 PFSs can be created. 1138Each PFS uses an independent inode numbering space making it suitable 1139for replication. 1140.Pp 1141The 1142.Cm pfs-master 1143directive creates a PFS that you can read, write, and use as a mirroring 1144source. 1145.Pp 1146A PFS can only be truly destroyed with the 1147.Cm pfs-destroy 1148directive. 1149Removing the softlink will not destroy the underlying PFS. 1150.Pp 1151A PFS can only be created in the root PFS (PFS# 0), 1152not in a PFS created by 1153.Cm pfs-master 1154or 1155.Cm pfs-slave 1156(PFS# >0). 1157.Pp 1158It is recommended that 1159.Ar dirpath 1160is of the form 1161.Pa <fs>/pfs/<name> 1162(i.e.\& located in 1163.Pa pfs 1164directory at root of 1165.Nm HAMMER 1166file system). 1167.Pp 1168It is recommended to use a 1169.Nm null 1170mount to access a PFS, except for root PFS, for more information see 1171.Xr HAMMER 5 . 1172.\" ==== pfs-slave ==== 1173.It Cm pfs-slave Ar dirpath Op Ar options 1174Create a pseudo-filesystem (PFS) inside a 1175.Nm HAMMER 1176file system. 1177Up to 65536 PFSs can be created. 1178Each PFS uses an independent inode numbering space making it suitable 1179for replication. 1180.Pp 1181The 1182.Cm pfs-slave 1183directive creates a PFS that you can use as a mirroring source or target. 1184You will not be able to access a slave PFS until you have completed the 1185first mirroring operation with it as the target (its root directory will 1186not exist until then). 1187.Pp 1188Access to the pfs-slave via the special softlink, as described in the 1189.Sx PSEUDO-FILESYSTEM (PFS) NOTES 1190below, allows 1191.Nm HAMMER 1192to 1193dynamically modify the snapshot transaction id by returning a dynamic result 1194from 1195.Xr readlink 2 1196calls. 1197.Pp 1198A PFS can only be truly destroyed with the 1199.Cm pfs-destroy 1200directive. 1201Removing the softlink will not destroy the underlying PFS. 1202.Pp 1203A PFS can only be created in the root PFS (PFS# 0), 1204not in a PFS created by 1205.Cm pfs-master 1206or 1207.Cm pfs-slave 1208(PFS# >0). 1209.Pp 1210It is recommended that 1211.Ar dirpath 1212is of the form 1213.Pa <fs>/pfs/<name> 1214(i.e.\& located in 1215.Pa pfs 1216directory at root of 1217.Nm HAMMER 1218file system). 1219.Pp 1220It is recommended to use a 1221.Nm null 1222mount to access a PFS, except for root PFS, for more information see 1223.Xr HAMMER 5 . 1224.\" ==== pfs-update ==== 1225.It Cm pfs-update Ar dirpath Op Ar options 1226Update the configuration parameters for an existing 1227.Nm HAMMER 1228file system or pseudo-filesystem. 1229Options that may be specified: 1230.Bl -tag -width indent 1231.It Cm sync-beg-tid= Ns Ar 0x16llx 1232This is the automatic snapshot access starting transaction id for 1233mirroring slaves. 1234This parameter is normally updated automatically by the 1235.Cm mirror-write 1236directive. 1237.Pp 1238It is important to note that accessing a mirroring slave 1239with a transaction id greater than the last fully synchronized transaction 1240id can result in an unreliable snapshot since you will be accessing 1241data that is still undergoing synchronization. 1242.Pp 1243Manually modifying this field is dangerous and can result in a broken mirror. 1244.It Cm sync-end-tid= Ns Ar 0x16llx 1245This is the current synchronization point for mirroring slaves. 1246This parameter is normally updated automatically by the 1247.Cm mirror-write 1248directive. 1249.Pp 1250Manually modifying this field is dangerous and can result in a broken mirror. 1251.It Cm shared-uuid= Ns Ar uuid 1252Set the shared UUID for this file system. 1253All mirrors must have the same shared UUID. 1254For safety purposes the 1255.Cm mirror-write 1256directives will refuse to operate on a target with a different shared UUID. 1257.Pp 1258Changing the shared UUID on an existing, non-empty mirroring target, 1259including an empty but not completely pruned target, 1260can lead to corruption of the mirroring target. 1261.It Cm unique-uuid= Ns Ar uuid 1262Set the unique UUID for this file system. 1263This UUID should not be used anywhere else, 1264even on exact copies of the file system. 1265.It Cm label= Ns Ar string 1266Set a descriptive label for this file system. 1267.It Cm snapshots= Ns Ar string 1268Specify the snapshots directory which 1269.Nm 1270.Cm cleanup 1271will use to manage this PFS. 1272.Bl -tag -width indent 1273.It Nm HAMMER No version 2- 1274The snapshots directory does not need to be configured for 1275PFS masters and will default to 1276.Pa <pfs>/snapshots . 1277.Pp 1278PFS slaves are mirroring slaves so you cannot configure a snapshots 1279directory on the slave itself to be managed by the slave's machine. 1280In fact, the slave will likely have a 1281.Pa snapshots 1282sub-directory mirrored 1283from the master, but that directory contains the configuration the master 1284is using for its copy of the file system, not the configuration that we 1285want to use for our slave. 1286.Pp 1287It is recommended that 1288.Pa <fs>/var/slaves/<name> 1289be configured for a PFS slave, where 1290.Pa <fs> 1291is the base 1292.Nm HAMMER 1293file system, and 1294.Pa <name> 1295is an appropriate label. 1296.It Nm HAMMER No version 3+ 1297The snapshots directory does not need to be configured for PFS masters or 1298slaves. 1299The snapshots directory defaults to 1300.Pa /var/hammer/<pfs> 1301.Pa ( /var/hammer/root 1302for root mount). 1303.El 1304.Pp 1305You can control snapshot retention on your slave independent of the master. 1306.It Cm snapshots-clear 1307Zero out the 1308.Cm snapshots 1309directory path for this PFS. 1310.It Cm prune-min= Ns Ar N Ns Cm d 1311.It Cm prune-min= Ns Oo Ar N Ns Cm d/ Oc Ns \ 1312Ar hh Ns Op Cm \&: Ns Ar mm Ns Op Cm \&: Ns Ar ss 1313Set the minimum fine-grained data retention period. 1314.Nm HAMMER 1315always retains fine-grained history up to the most recent snapshot. 1316You can extend the retention period further by specifying a non-zero 1317pruning minimum. 1318Any snapshot softlinks within the retention period are ignored 1319for the purposes of pruning (i.e.\& the fine grained history is retained). 1320Number of days, hours, minutes and seconds are given as 1321.Ar N , hh , mm 1322and 1323.Ar ss . 1324.Pp 1325Because the transaction id in the snapshot softlink cannot be used 1326to calculate a timestamp, 1327.Nm HAMMER 1328uses the earlier of the 1329.Fa st_ctime 1330or 1331.Fa st_mtime 1332field of the softlink to 1333determine which snapshots fall within the retention period. 1334Users must be sure to retain one of these two fields when manipulating 1335the softlink. 1336.El 1337.\" ==== pfs-upgrade ==== 1338.It Cm pfs-upgrade Ar dirpath 1339Upgrade a PFS from slave to master operation. 1340The PFS will be rolled back to the current end synchronization transaction id 1341(removing any partial synchronizations), and will then become writable. 1342.Pp 1343.Em WARNING! 1344.Nm HAMMER 1345currently supports only single masters and using 1346this command can easily result in file system corruption 1347if you don't know what you are doing. 1348.Pp 1349This directive will refuse to run if any programs have open descriptors 1350in the PFS, including programs chdir'd into the PFS. 1351.\" ==== pfs-downgrade ==== 1352.It Cm pfs-downgrade Ar dirpath 1353Downgrade a master PFS from master to slave operation. 1354The PFS becomes read-only and access will be locked to its 1355.Cm sync-end-tid . 1356.Pp 1357This directive will refuse to run if any programs have open descriptors 1358in the PFS, including programs chdir'd into the PFS. 1359.\" ==== pfs-destroy ==== 1360.It Cm pfs-destroy Ar dirpath 1361This permanently destroys a PFS. 1362.Pp 1363This directive will refuse to run if any programs have open descriptors 1364in the PFS, including programs chdir'd into the PFS. 1365As safety measure the 1366.Fl y 1367flag have no effect on this directive. 1368.\" ==== mirror-read ==== 1369.It Cm mirror-read Ar filesystem Op Ar begin-tid 1370Generate a mirroring stream to stdout. 1371The stream ends when the transaction id space has been exhausted. 1372.Ar filesystem 1373may be a master or slave PFS. 1374.\" ==== mirror-read-stream ==== 1375.It Cm mirror-read-stream Ar filesystem Op Ar begin-tid 1376Generate a mirroring stream to stdout. 1377Upon completion the stream is paused until new data is synced to the 1378.Ar filesystem , 1379then resumed. 1380Operation continues until the pipe is broken. 1381See the 1382.Cm mirror-stream 1383command for more details. 1384.\" ==== mirror-write ==== 1385.It Cm mirror-write Ar filesystem 1386Take a mirroring stream on stdin. 1387.Ar filesystem 1388must be a slave PFS. 1389.Pp 1390This command will fail if the 1391.Cm shared-uuid 1392configuration field for the two file systems do not match. 1393See the 1394.Cm mirror-copy 1395command for more details. 1396.Pp 1397If the target PFS does not exist this command will ask you whether 1398you want to create a compatible PFS slave for the target or not. 1399.\" ==== mirror-dump ==== 1400.It Ar mirror-dump Ar [header] 1401A 1402.Cm mirror-read 1403can be piped into a 1404.Cm mirror-dump 1405to dump an ASCII representation of the mirroring stream. 1406If the keyword 1407.Ar header 1408is specified, only the header information is shown. 1409.\" ==== mirror-copy ==== 1410.\".It Cm mirror-copy Ar [[user@]host:]filesystem [[user@]host:]filesystem 1411.It Cm mirror-copy \ 1412Oo Oo Ar user Ns Cm @ Oc Ns Ar host Ns Cm \&: Oc Ns Ar filesystem \ 1413Oo Oo Ar user Ns Cm @ Oc Ns Ar host Ns Cm \&: Oc Ns Ar filesystem 1414This is a shortcut which pipes a 1415.Cm mirror-read 1416command to a 1417.Cm mirror-write 1418command. 1419If a remote host specification is made the program forks a 1420.Xr ssh 1 1421(or other program as specified by the 1422.Ev HAMMER_RSH 1423environment variable) and execs the 1424.Cm mirror-read 1425and/or 1426.Cm mirror-write 1427on the appropriate host. 1428The source may be a master or slave PFS, and the target must be a slave PFS. 1429.Pp 1430This command also establishes full duplex communication and turns on 1431the 2-way protocol feature 1432.Fl ( 2 ) 1433which automatically negotiates transaction id 1434ranges without having to use a cyclefile. 1435If the operation completes successfully the target PFS's 1436.Cm sync-end-tid 1437will be updated. 1438Note that you must re-chdir into the target PFS to see the updated information. 1439If you do not you will still be in the previous snapshot. 1440.Pp 1441If the target PFS does not exist this command will ask you whether 1442you want to create a compatible PFS slave for the target or not. 1443.\" ==== mirror-stream ==== 1444.\".It Cm mirror-stream Ar [[user@]host:]filesystem [[user@]host:]filesystem 1445.It Cm mirror-stream \ 1446Oo Oo Ar user Ns Cm @ Oc Ns Ar host Ns Cm \&: Oc Ns Ar filesystem \ 1447Oo Oo Ar user Ns Cm @ Oc Ns Ar host Ns Cm \&: Oc Ns Ar filesystem 1448This is a shortcut which pipes a 1449.Cm mirror-read-stream 1450command to a 1451.Cm mirror-write 1452command. 1453This command works similarly to 1454.Cm mirror-copy 1455but does not exit after the initial mirroring completes. 1456The mirroring operation will resume as changes continue to be made to the 1457source. 1458The command is commonly used with 1459.Fl i Ar delay 1460and 1461.Fl b Ar bandwidth 1462options to keep the mirroring target in sync with the source on a continuing 1463basis. 1464.Pp 1465If the pipe is broken the command will automatically retry after sleeping 1466for a short while. 1467The time slept will be 15 seconds plus the time given in the 1468.Fl i 1469option. 1470.Pp 1471This command also detects the initial-mirroring case and spends some 1472time scanning the B-Tree to find good break points, allowing the initial 1473bulk mirroring operation to be broken down into 4GB pieces. 1474This means that the user can kill and restart the operation and it will 1475not have to start from scratch once it has gotten past the first chunk. 1476The 1477.Fl S 1478option may be used to change the size of pieces and the 1479.Fl B 1480option may be used to disable this feature and perform an initial bulk 1481transfer instead. 1482.\" ==== version ==== 1483.It Cm version Ar filesystem 1484This command returns the 1485.Nm HAMMER 1486file system version for the specified 1487.Ar filesystem 1488as well as the range of versions supported in the kernel. 1489The 1490.Fl q 1491option may be used to remove the summary at the end. 1492.\" ==== version-upgrade ==== 1493.It Cm version-upgrade Ar filesystem Ar version Op Cm force 1494Upgrade the 1495.Nm HAMMER 1496.Ar filesystem 1497to the specified 1498.Ar version . 1499Once upgraded a file system may not be downgraded. 1500If you wish to upgrade a file system to a version greater or equal to the 1501work-in-progress (WIP) version number you must specify the 1502.Cm force 1503directive. 1504Use of WIP versions should be relegated to testing and may require wiping 1505the file system as development progresses, even though the WIP version might 1506not change. 1507.Pp 1508.Em NOTE! 1509This command operates on the entire 1510.Nm HAMMER 1511file system and is not a per PFS operation. 1512All PFS's will be affected. 1513.Bl -tag -width indent 1514.It 1 1515.Dx 2.0 1516default version, first 1517.Nm HAMMER 1518release. 1519.It 2 1520.Dx 2.3 . 1521New directory entry layout. 1522This version is using a new directory hash key. 1523.It 3 1524.Dx 2.5 . 1525New snapshot management, using file system meta-data for saving 1526configuration file and snapshots (transaction ids etc.). 1527Also default snapshots directory has changed. 1528.It 4 1529.Dx 2.6 1530default version. 1531New undo/redo/flush, giving 1532.Nm HAMMER 1533a much faster sync and fsync. 1534.It 5 1535.Dx 2.9 . 1536Deduplication support. 1537.It 6 1538.Dx 2.9 . 1539Directory hash ALG1. 1540Tends to maintain inode number / directory name entry ordering better 1541for files after minor renaming. 1542.El 1543.El 1544.Sh PSEUDO-FILESYSTEM (PFS) NOTES 1545The root of a PFS is not hooked into the primary 1546.Nm HAMMER 1547file system as a directory. 1548Instead, 1549.Nm HAMMER 1550creates a special softlink called 1551.Ql @@PFS%05d 1552(exactly 10 characters long) in the primary 1553.Nm HAMMER 1554file system. 1555.Nm HAMMER 1556then modifies the contents of the softlink as read by 1557.Xr readlink 2 , 1558and thus what you see with an 1559.Nm ls 1560command or if you were to 1561.Nm cd 1562into the link. 1563If the PFS is a master the link reflects the current state of the PFS. 1564If the PFS is a slave the link reflects the last completed snapshot, and the 1565contents of the link will change when the next snapshot is completed, and 1566so forth. 1567.Pp 1568The 1569.Nm 1570utility employs numerous safeties to reduce user foot-shooting. 1571The 1572.Cm mirror-copy 1573directive requires that the target be configured as a slave and that the 1574.Cm shared-uuid 1575field of the mirroring source and target match. 1576.Sh DOUBLE_BUFFER MODE 1577There is a limit to the number of vnodes the kernel can cache, and because 1578file buffers are associated with a vnode the related data cache can get 1579blown away when operating on large numbers of files even if the system has 1580sufficient memory to hold the file data. 1581.Pp 1582If you turn on 1583.Nm HAMMER Ns 's 1584double buffer mode by setting the 1585.Xr sysctl 8 1586node 1587.Va vfs.hammer.double_buffer 1588to 1 1589.Nm HAMMER 1590will cache file data via the block device and copy it into the per-file 1591buffers as needed. The data will be double-cached at least until the 1592buffer cache throws away the file buffer. 1593This mode is typically used in conjunction with 1594.Xr swapcache 8 1595when 1596.Va vm.swapcache.data_enable 1597is turned on in order to prevent unnecessary re-caching of file data 1598due to vnode recycling. 1599The swapcache will save the cached VM pages related to 1600.Nm HAMMER Ns 's 1601block 1602device (which doesn't recycle unless you umount the filesystem) instead 1603of the cached VM pages backing the file vnodes. 1604.Pp 1605Double buffering is normally desireable when working with large filesystems, 1606particularly when swapcache is used. 1607The swapcache can only back active VM objects, including the block device, 1608and large filesystems often have far more inodes than the kernel can support. 1609In addition, when using this mode, you may wish to reduce the 1610.Va kern.maxvnodes 1611setting for the system to force the system to do less caching of logical 1612file buffers and more caching of device buffers, since the device buffers 1613are backing the logical file buffers. 1614.Sh UPGRADE INSTRUCTIONS HAMMER V1 TO V2 1615This upgrade changes the way directory entries are stored. 1616It is possible to upgrade a V1 file system to V2 in place, but 1617directories created prior to the upgrade will continue to use 1618the old layout. 1619.Pp 1620Note that the slave mirroring code in the target kernel had bugs in 1621V1 which can create an incompatible root directory on the slave. 1622Do not mix a 1623.Nm HAMMER 1624master created after the upgrade with a 1625.Nm HAMMER 1626slave created prior to the upgrade. 1627.Pp 1628Any directories created after upgrading will use a new layout. 1629.Sh UPGRADE INSTRUCTIONS HAMMER V2 TO V3 1630This upgrade adds meta-data elements to the B-Tree. 1631It is possible to upgrade a V2 file system to V3 in place. 1632After issuing the upgrade be sure to run a 1633.Nm 1634.Cm cleanup 1635to perform post-upgrade tasks. 1636.Pp 1637After making this upgrade running a 1638.Nm 1639.Cm cleanup 1640will move the 1641.Pa <pfs>/snapshots 1642directory for each PFS mount into 1643.Pa /var/hammer/<pfs> . 1644A 1645.Nm HAMMER 1646root mount will migrate 1647.Pa /snapshots 1648into 1649.Pa /var/hammer/root . 1650Migration occurs only once and only if you have not specified 1651a snapshots directory in the PFS configuration. 1652If you have specified a snapshots directory in the PFS configuration no 1653automatic migration will occur. 1654.Pp 1655For slaves, if you desire, you can migrate your snapshots 1656config to the new location manually and then clear the 1657snapshot directory configuration in the slave PFS. 1658The new snapshots hierarchy is designed to work with 1659both master and slave PFSs equally well. 1660.Pp 1661In addition, the old config file will be moved to file system meta-data, 1662editable via the new 1663.Nm 1664.Cm viconfig 1665directive. 1666The old config file will be deleted. 1667Migration occurs only once. 1668.Pp 1669The V3 file system has new 1670.Cm snap* 1671directives for creating snapshots. 1672All snapshot directives, including the original, will create 1673meta-data entries for the snapshots and the pruning code will 1674automatically incorporate these entries into its list and 1675expire them the same way it expires softlinks. 1676If you by accident blow away your snapshot softlinks you can use the 1677.Cm snapls 1678directive to get a definitive list from the file system meta-data and 1679regenerate them from that list. 1680.Pp 1681.Em WARNING! 1682If you are using 1683.Nm 1684to backup file systems your scripts may be using the 1685.Cm synctid 1686directive to generate transaction ids. 1687This directive does not create a snapshot. 1688You will have to modify your scripts to use the 1689.Cm snapq 1690directive to generate the linkbuf for the softlink you create, or 1691use one of the other 1692.Cm snap* 1693directives. 1694The older 1695.Cm snapshot 1696directive will continue to work as expected and in V3 it will also 1697record the snapshot transaction id in file system meta-data. 1698You may also want to make use of the new 1699.Ar note 1700tag for the meta-data. 1701.Pp 1702.Em WARNING! 1703If you used to remove snapshot softlinks with 1704.Nm rm 1705you should probably start using the 1706.Cm snaprm 1707directive instead to also remove the related meta-data. 1708The pruning code scans the meta-data so just removing the 1709softlink is not sufficient. 1710.Sh UPGRADE INSTRUCTIONS HAMMER V3 TO V4 1711This upgrade changes undo/flush, giving faster sync. 1712It is possible to upgrade a V3 file system to V4 in place. 1713This upgrade reformats the UNDO/REDO FIFO (typically 1GB), 1714so upgrade might take a minute or two depending. 1715.Pp 1716Version 4 allows the UNDO/REDO FIFO to be flushed without also having 1717to flush the volume header, removing 2 of the 4 disk syncs typically 1718required for an 1719.Fn fsync 1720and removing 1 of the 2 disk syncs typically 1721required for a flush sequence. 1722Version 4 also implements the REDO log (see 1723.Sx FSYNC FLUSH MODES 1724below) which is capable 1725of fsync()ing with either one disk flush or zero disk flushes. 1726.Sh UPGRADE INSTRUCTIONS HAMMER V4 TO V5 1727This upgrade brings in deduplication support. 1728It is possible to upgrade a V4 file system to V5 in place. 1729Technically it makes the layer2 1730.Va bytes_free 1731field a signed value instead of unsigned, allowing it to go negative. 1732A version 5 filesystem is required for dedup operation. 1733.Sh UPGRADE INSTRUCTIONS HAMMER V5 TO V6 1734It is possible to upgrade a V5 file system to V6 in place. 1735.Sh FSYNC FLUSH MODES 1736.Nm HAMMER 1737implements five different fsync flush modes via the 1738.Va vfs.hammer.fsync_mode 1739sysctl, for 1740.Nm HAMMER 1741version 4+ file systems. 1742.Pp 1743As of 1744.Dx 2.6 1745fsync mode 3 is set by default. 1746REDO operation and recovery is enabled by default. 1747.Bl -tag -width indent 1748.It mode 0 1749Full synchronous fsync semantics without REDO. 1750.Pp 1751.Nm HAMMER 1752will not generate REDOs. 1753A 1754.Fn fsync 1755will completely sync 1756the data and meta-data and double-flush the FIFO, including 1757issuing two disk synchronization commands. 1758The data is guaranteed 1759to be on the media as of when 1760.Fn fsync 1761returns. 1762Needless to say, this is slow. 1763.It mode 1 1764Relaxed asynchronous fsync semantics without REDO. 1765.Pp 1766This mode works the same as mode 0 except the last disk synchronization 1767command is not issued. 1768It is faster than mode 0 but not even remotely 1769close to the speed you get with mode 2 or mode 3. 1770.Pp 1771Note that there is no chance of meta-data corruption when using this 1772mode, it simply means that the data you wrote and then 1773.Fn fsync Ns 'd 1774might not have made it to the media if the storage system crashes at a bad 1775time. 1776.It mode 2 1777Full synchronous fsync semantics using REDO. 1778NOTE: If not running a 1779.Nm HAMMER 1780version 4 filesystem or later mode 0 is silently used. 1781.Pp 1782.Nm HAMMER 1783will generate REDOs in the UNDO/REDO FIFO based on a heuristic. 1784If this is sufficient to satisfy the 1785.Fn fsync 1786operation the blocks will be written out and 1787.Nm HAMMER 1788will wait for the I/Os to complete, 1789and then followup with a disk sync command to guarantee the data 1790is on the media before returning. 1791This is slower than mode 3 and can result in significant disk or 1792SSDs overheads, though not as bad as mode 0 or mode 1. 1793.It mode 3 1794Relaxed asynchronous fsync semantics using REDO. 1795NOTE: If not running a 1796.Nm HAMMER 1797version 4 filesystem or later mode 1 is silently used. 1798.Pp 1799.Nm HAMMER 1800will generate REDOs in the UNDO/REDO FIFO based on a heuristic. 1801If this is sufficient to satisfy the 1802.Fn fsync 1803operation the blocks 1804will be written out and 1805.Nm HAMMER 1806will wait for the I/Os to complete, 1807but will 1808.Em NOT 1809issue a disk synchronization command. 1810.Pp 1811Note that there is no chance of meta-data corruption when using this 1812mode, it simply means that the data you wrote and then 1813.Fn fsync Ns 'd 1814might 1815not have made it to the media if the storage system crashes at a bad 1816time. 1817.Pp 1818This mode is the fastest production fsyncing mode available. 1819This mode is equivalent to how the UFS fsync in the 1820.Bx Ns s 1821operates. 1822.It mode 4 1823fsync is ignored. 1824.Pp 1825Calls to 1826.Fn fsync 1827will be ignored. 1828This mode is primarily designed 1829for testing and should not be used on a production system. 1830.El 1831.Sh RESTORING FROM A SNAPSHOT BACKUP 1832You restore a snapshot by copying it over to live, but there is a caveat. 1833The mtime and atime fields for files accessed via a snapshot is locked 1834to the ctime in order to keep the snapshot consistent, because neither 1835mtime nor atime changes roll any history. 1836.Pp 1837In order to avoid unnecessary copying it is recommended that you use 1838.Nm cpdup 1839.Fl VV 1840.Fl v 1841when doing the copyback. 1842Also make sure you traverse the snapshot softlink by appending a ".", 1843as in "<snapshotpath>/.", and you match up the directory properly. 1844.Sh RESTORING A PFS FROM A MIRROR 1845A PFS can be restored from a mirror with 1846.Cm mirror-copy . 1847.Cm config 1848data must be copied separately. 1849At last the PFS can be upgraded to master using 1850.Cm pfs-upgrade . 1851.Pp 1852It is not possible to restore the root PFS (PFS# 0) by using mirroring, 1853as the root PFS is always a master PFS. 1854A normal copy (e.g.\& using 1855.Xr cpdup 1 ) 1856must be done, ignoring history. 1857If history is important, old root PFS can me restored to a new PFS, and 1858important directories/files can be 1859.Nm null 1860mounted to the new PFS. 1861.Sh ENVIRONMENT 1862The following environment variables affect the execution of 1863.Nm : 1864.Bl -tag -width ".Ev EDITOR" 1865.It Ev EDITOR 1866The editor program specified in the variable 1867.Ev EDITOR 1868will be invoked instead of the default editor, which is 1869.Xr vi 1 . 1870.It Ev HAMMER_RSH 1871The command specified in the variable 1872.Ev HAMMER_RSH 1873will be used to initiate remote operations for the mirror-copy and 1874mirror-stream commands instead of the default command, which is 1875.Xr ssh 1 . 1876The program will be invoked via 1877.Xr execvp 3 1878using a typical 1879.Xr rsh 1 1880style 1881.Cm -l user host <remote-command> 1882command line. 1883.It Ev VISUAL 1884Same effect as 1885.Ev EDITOR 1886variable. 1887.El 1888.Sh FILES 1889.Bl -tag -width ".It Pa <fs>/var/slaves/<name>" -compact 1890.It Pa <pfs>/snapshots 1891default per PFS snapshots directory 1892.Nm ( HAMMER 1893VERSION 2-) 1894.It Pa /var/hammer/<pfs> 1895default per PFS snapshots directory (not root) 1896.Nm ( HAMMER 1897VERSION 3+) 1898.It Pa /var/hammer/root 1899default snapshots directory for root directory 1900.Nm ( HAMMER 1901VERSION 3+) 1902.It Pa <snapshots>/config 1903per PFS 1904.Nm 1905.Cm cleanup 1906configuration file 1907.Nm ( HAMMER 1908VERSION 2-) 1909.It Pa <fs>/var/slaves/<name> 1910recommended slave PFS snapshots directory 1911.Nm ( HAMMER 1912VERSION 2-) 1913.It Pa <fs>/pfs 1914recommended PFS directory 1915.El 1916.Sh EXIT STATUS 1917.Ex -std 1918.Sh SEE ALSO 1919.Xr ssh 1 , 1920.Xr undo 1 , 1921.Xr HAMMER 5 , 1922.Xr periodic.conf 5 , 1923.Xr loader 8 , 1924.Xr mount_hammer 8 , 1925.Xr mount_null 8 , 1926.Xr newfs_hammer 8 , 1927.Xr swapcache 8 , 1928.Xr sysctl 8 1929.Sh HISTORY 1930The 1931.Nm 1932utility first appeared in 1933.Dx 1.11 . 1934.Sh AUTHORS 1935.An Matthew Dillon Aq Mt dillon@backplane.com 1936