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 March 4, 2016 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 Ns Oo Cm @ Ns Ar offset Ns Oo Cm \&, Ns Ar length Oc Oc Ar ... 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.Pp 343Different 344.Ar offset 345and 346.Ar length 347can be used for each 348.Ar path 349by specifying 350.Ar offset 351and 352.Ar length 353for each 354.Ar path . 355.\" ==== blockmap ==== 356.It Cm blockmap 357Dump the blockmap for the file system. 358The 359.Nm HAMMER 360blockmap is two-layer 361blockmap representing the maximum possible file system size of 1 Exabyte. 362Needless to say the second layer is only present for blocks which exist. 363.Nm HAMMER Ns 's 364blockmap represents 8-Megabyte blocks, called big-blocks. 365Each big-block has an append 366point, a free byte count, and a typed zone id which allows content to be 367reverse engineered to some degree. 368.Pp 369In 370.Nm HAMMER 371allocations are essentially appended to a selected big-block using 372the append offset and deducted from the free byte count. 373When space is freed the free byte count is adjusted but 374.Nm HAMMER 375does not track holes in big-blocks for reallocation. 376A big-block must be completely freed, either 377through normal file system operations or through reblocking, before 378it can be reused. 379.Pp 380Data blocks can be shared by deducting the space used from the free byte 381count for each shared references. 382This means the free byte count can legally go negative. 383.Pp 384This command needs the 385.Fl f Ar blkdevs 386option. 387.\" ==== checkmap ==== 388.It Cm checkmap 389Check the blockmap allocation count. 390.Nm 391will scan the freemap, B-Tree, UNDO FIFO, then collect allocation information, 392and construct a blockmap in-memory. 393It will then check that blockmap against the on-disk blockmap. 394.Pp 395This command needs the 396.Fl f Ar blkdevs 397option. 398.\" ==== show ==== 399.It Cm show Op Ar localization Ns Op Cm \&: Ns Ar object_id Ns Op Cm \&: Ns Ar rec_type Ns Op Cm \&: Ns Ar key Ns Op Cm \&: Ns Ar create_tid 400Dump the B-Tree. 401By default this command will validate all B-Tree 402linkages and CRCs, including data CRCs, and will report the most verbose 403information it can dig up. 404Any errors will show up with a 405.Ql B 406in column 1 along with various 407other error flags. 408.Pp 409If you specify 410.Ar localization 411or 412.Ar localization Ns Cm \&: Ns Ar object_id 413or 414.Ar localization Ns Cm \&: Ns Ar object_id Ns Cm \&: Ns Ar rec_type 415or 416.Ar localization Ns Cm \&: Ns Ar object_id Ns Cm \&: Ns Ar rec_type Ns Cm \&: Ns Ar key 417or 418.Ar localization Ns Cm \&: Ns Ar object_id Ns Cm \&: Ns Ar rec_type Ns Cm \&: Ns Ar key Ns Cm \&: Ns Ar create_tid 419the dump will 420search for the key printing nodes as it recurses down, and then 421will iterate forwards. 422These fields are specified in HEX. 423Note that the PFS id is the top 16 bits of the 32-bit localization 424field so PFS #1 would be 00010000. 425.Pp 426If you use 427.Fl q 428the command will report less information about the inode contents. 429.Pp 430If you use 431.Fl qq 432the command will not report the content of the inode or other typed 433data at all. 434.Pp 435If you use 436.Fl qqq 437the command will not report volume header information, big-block fill 438ratios, mirror transaction ids, or report or check data CRCs. 439B-Tree CRCs and linkages are still checked. 440.Pp 441This command needs the 442.Fl f Ar blkdevs 443option. 444.\" ==== show-undo ==== 445.It Cm show-undo 446.Nm ( HAMMER 447VERSION 4+) 448Dump the UNDO/REDO map. 449.Pp 450This command needs the 451.Fl f Ar blkdevs 452option. 453.\" .It Ar blockmap 454.\" Dump the B-Tree, record, large-data, and small-data blockmaps, showing 455.\" physical block assignments and free space percentages. 456.\" ==== ssh-remote ==== 457.It Cm ssh-remote Ar command Ar targetdir 458Used in a ssh authorized_keys line such as 459command="/sbin/hammer ssh-remote mirror-read /fubarmount" ... to allow 460mirror-read or mirror-write access to a particular subdirectory tree. 461This way you do not have to give shell access to the remote box. 462.Nm 463will obtain the original command line from the SSH_ORIGINAL_COMMAND 464environment variable, validate it against the restriction, and then 465re-exec hammer with the validated arguments. 466.Pp 467The remote hammer command does not allow the 468.Fl c 469or 470.Fl f 471options to be passed in. 472.\" ==== recover ==== 473.It Cm recover Ar targetdir 474Recover data from a corrupted 475.Nm HAMMER 476filesystem. 477This is a low level command which operates on the filesystem image and 478attempts to locate and recover files from a corrupted filesystem. 479The entire image is scanned linearly looking for B-Tree nodes. 480Any node 481found which passes its CRC test is scanned for file, inode, and directory 482fragments and the target directory is populated with the resulting data. 483files and directories in the target directory are initially named after 484the object id and are renamed as fragmentary information is processed. 485.Pp 486This command keeps track of filename/object_id translations and may eat a 487considerably amount of memory while operating. 488.Pp 489This command is literally the last line of defense when it comes to 490recovering data from a dead filesystem. 491.Pp 492This command needs the 493.Fl f Ar blkdevs 494option. 495.\" ==== namekey1 ==== 496.It Cm namekey1 Ar filename 497Generate a 498.Nm HAMMER 49964-bit directory hash for the specified file name, using 500the original directory hash algorithm in version 1 of the file system. 501The low 32 bits are used as an iterator for hash collisions and will be 502output as 0. 503.\" ==== namekey2 ==== 504.It Cm namekey2 Ar filename 505Generate a 506.Nm HAMMER 50764-bit directory hash for the specified file name, using 508the new directory hash algorithm in version 2 of the file system. 509The low 32 bits are still used as an iterator but will start out containing 510part of the hash key. 511.\" ==== namekey32 ==== 512.It Cm namekey32 Ar filename 513Generate the top 32 bits of a 514.Nm HAMMER 51564 bit directory hash for the specified file name. 516.\" ==== info ==== 517.It Cm info Ar dirpath ... 518Show extended information about all 519.Nm HAMMER 520file systems mounted in the system or the one mounted in 521.Ar dirpath 522when this argument is specified. 523.Pp 524The information is divided into sections: 525.Bl -tag -width indent 526.It Volume identification 527General information, like the label of the 528.Nm HAMMER 529filesystem, the number of volumes it contains, the FSID, and the 530.Nm HAMMER 531version being used. 532.It Big-block information 533Big-block statistics, such as total, used, reserved and free big-blocks. 534.It Space information 535Information about space used on the filesystem. 536Currently total size, used, reserved and free space are displayed. 537.It PFS information 538Basic information about the PFSs currently present on a 539.Nm HAMMER 540filesystem. 541.Pp 542.Dq PFS ID 543is the ID of the PFS, with 0 being the root PFS. 544.Dq Snaps 545is the current snapshot count on the PFS. 546.Dq Mounted on 547displays the mount point of the PFS is currently mounted on (if any). 548.El 549.\" ==== cleanup ==== 550.It Cm cleanup Op Ar filesystem ... 551This is a meta-command which executes snapshot, prune, rebalance, dedup 552and reblock commands on the specified 553.Nm HAMMER 554file systems. 555If no 556.Ar filesystem 557is specified this command will clean-up all 558.Nm HAMMER 559file systems in use, including PFS's. 560To do this it will scan all 561.Nm HAMMER 562and 563.Nm null 564mounts, extract PFS id's, and clean-up each PFS found. 565.Pp 566This command will access a snapshots 567directory and a configuration file for each 568.Ar filesystem , 569creating them if necessary. 570.Bl -tag -width indent 571.It Nm HAMMER No version 2- 572The configuration file is 573.Pa config 574in the snapshots directory which defaults to 575.Pa <pfs>/snapshots . 576.It Nm HAMMER No version 3+ 577The configuration file is saved in file system meta-data, see 578.Nm 579.Cm config . 580The snapshots directory defaults to 581.Pa /var/hammer/<pfs> 582.Pa ( /var/hammer/root 583for root mount). 584.El 585.Pp 586The format of the configuration file is: 587.Bd -literal -offset indent 588snapshots <period> <retention-time> [any] 589prune <period> <max-runtime> 590rebalance <period> <max-runtime> 591dedup <period> <max-runtime> 592reblock <period> <max-runtime> 593recopy <period> <max-runtime> 594.Ed 595.Pp 596Defaults are: 597.Bd -literal -offset indent 598snapshots 1d 60d # 0d 0d for PFS /tmp, /var/tmp, /usr/obj 599prune 1d 5m 600rebalance 1d 5m 601dedup 1d 5m 602reblock 1d 5m 603recopy 30d 10m 604.Ed 605.Pp 606Time is given with a suffix of 607.Cm d , 608.Cm h , 609.Cm m 610or 611.Cm s 612meaning day, hour, minute and second. 613.Pp 614If the 615.Cm snapshots 616directive has a period of 0 and a retention time of 0 617then snapshot generation is disabled, removal of old snapshots are 618disabled, and prunes will use 619.Cm prune-everything . 620.Pp 621If the 622.Cm snapshots 623directive has a period of 0 but a non-zero retention time 624then this command will not create any new snapshots but will remove old 625snapshots it finds based on the retention time. 626This form should be 627used on PFS masters where you are generating your own snapshot softlinks 628manually and on PFS slaves when all you wish to do is prune away existing 629snapshots inherited via the mirroring stream. 630.Pp 631By default only snapshots in the form 632.Ql snap- Ns Ar yyyymmdd Ns Op - Ns Ar HHMM 633are processed. 634If the 635.Cm any 636directive is specified as a third argument on the 637.Cm snapshots 638config line then any softlink of the form 639.Ql *- Ns Ar yyyymmdd Ns Op - Ns Ar HHMM 640or 641.Ql *. Ns Ar yyyymmdd Ns Op - Ns Ar HHMM 642will be processed. 643.Pp 644A period of 0 for prune, rebalance, dedup, reblock or recopy disables the directive. 645A max-runtime of 0 means unlimited. 646.Pp 647If period hasn't passed since the previous 648.Cm cleanup 649run nothing is done. 650For example a day has passed when midnight is passed (localtime). 651If the 652.Fl F 653flag is given the period is ignored. 654By default, 655.Dx 656is set up to run 657.Nm Cm cleanup 658nightly via 659.Xr periodic 8 . 660.Pp 661The default configuration file will create a daily snapshot, do a daily 662pruning, rebalancing, deduping and reblocking run and a monthly recopy run. 663Reblocking is defragmentation with a level of 95%, 664and recopy is full defragmentation. 665.Pp 666By default prune, dedup and rebalance operations are time limited to 5 minutes, 667and reblock operations to a bit over 5 minutes, 668and recopy operations to a bit over 10 minutes. 669Reblocking and recopy runs are each broken down into four separate functions: 670B-Tree, inodes, dirs and data. 671Each function is time limited to the time given in the configuration file, 672but the B-Tree, inodes and dirs functions usually does not take very long time, 673full defragmentation is always used for these three functions. 674Also note that this directive will by default disable snapshots on 675the following PFS's: 676.Pa /tmp , 677.Pa /var/tmp 678and 679.Pa /usr/obj . 680.Pp 681The defaults may be adjusted by modifying the configuration file. 682The pruning and reblocking commands automatically maintain a cyclefile 683for incremental operation. 684If you interrupt (^C) the program the cyclefile will be updated, 685but a sub-command 686may continue to run in the background for a few seconds until the 687.Nm HAMMER 688ioctl detects the interrupt. 689The 690.Cm snapshots 691PFS option can be set to use another location for the snapshots directory. 692.Pp 693Work on this command is still in progress. 694Expected additions: 695An ability to remove snapshots dynamically as the 696file system becomes full. 697.\" ==== abort-cleanup ==== 698.It Cm abort-cleanup 699This command will terminate all active 700.Cm cleanup 701processes. 702.\" ==== config ==== 703.It Cm config Op Ar filesystem Op Ar configfile 704.Nm ( HAMMER 705VERSION 3+) 706Show or change configuration for 707.Ar filesystem . 708If zero or one arguments are specified this function dumps the current 709configuration file to stdout. 710Zero arguments specifies the PFS containing the current directory. 711This configuration file is stored in file system meta-data. 712If two arguments are specified this function installs a new configuration file 713.Ar configfile . 714.Pp 715In 716.Nm HAMMER 717versions less than 3 the configuration file is by default stored in 718.Pa <pfs>/snapshots/config , 719but in all later versions the configuration file is stored in file system 720meta-data. 721.\" ==== viconfig ==== 722.It Cm viconfig Op Ar filesystem 723.Nm ( HAMMER 724VERSION 3+) 725Edit the configuration file and reinstall into file system meta-data when done. 726Zero arguments specifies the PFS containing the current directory. 727.\" ==== volume-add ==== 728.It Cm volume-add Ar device Ar filesystem 729Add volume 730.Ar device 731to 732.Ar filesystem . 733This will format 734.Ar device 735and add all of its space to 736.Ar filesystem . 737A 738.Nm HAMMER 739file system can use up to 256 volumes. 740.Pp 741.Em NOTE! 742All existing data contained on 743.Ar device 744will be destroyed by this operation! 745If 746.Ar device 747contains a valid 748.Nm HAMMER 749file system, formatting will be denied. 750You can overcome this sanity check by using 751.Xr dd 1 752to erase the beginning sectors of the device. 753.Pp 754Remember that you have to specify 755.Ar device , 756together with any other device that make up the file system, 757colon-separated to 758.Pa /etc/fstab 759and 760.Xr mount_hammer 8 . 761If 762.Ar filesystem 763is root file system, also remember to add 764.Ar device 765to 766.Va vfs.root.mountfrom 767in 768.Pa /boot/loader.conf , 769see 770.Xr loader 8 . 771.\" ==== volume-del ==== 772.It Cm volume-del Ar device Ar filesystem 773Remove volume 774.Ar device 775from 776.Ar filesystem . 777.Pp 778Remember that you have to remove 779.Ar device 780from the colon-separated list in 781.Pa /etc/fstab 782and 783.Xr mount_hammer 8 . 784If 785.Ar filesystem 786is root file system, also remember to remove 787.Ar device 788from 789.Va vfs.root.mountfrom 790in 791.Pa /boot/loader.conf , 792see 793.Xr loader 8 . 794.Pp 795It is not possible to remove the 796.Ar root-volume 797as it contains 798.Ar filesystem 799meta data such as 800.Nm HAMMER Ns 's 801layer1 blockmap and UNDO FIFO. 802.Pp 803This command may 804.Cm reblock 805filesystem before it attempts to remove the volume if the volume is not empty. 806.\" ==== volume-list ==== 807.It Cm volume-list Ar filesystem 808List the volumes that make up 809.Ar filesystem . 810.\" ==== volume-blkdevs ==== 811.It Cm volume-blkdevs Ar filesystem 812List the volumes that make up 813.Ar filesystem 814in 815.Ar blkdevs 816format. 817.\" ==== snapshot ==== 818.It Cm snapshot Oo Ar filesystem Oc Ar snapshot-dir 819.It Cm snapshot Ar filesystem Ar snapshot-dir Op Ar note 820Take a snapshot of the file system either explicitly given by 821.Ar filesystem 822or implicitly derived from the 823.Ar snapshot-dir 824argument and creates a symlink in the directory provided by 825.Ar snapshot-dir 826pointing to the snapshot. 827If 828.Ar snapshot-dir 829is not a directory, it is assumed to be a format string passed to 830.Xr strftime 3 831with the current time as parameter. 832If 833.Ar snapshot-dir 834refers to an existing directory, a default format string of 835.Ql snap-%Y%m%d-%H%M 836is assumed and used as name for the newly created symlink. 837.Pp 838Snapshot is a per PFS operation, so each PFS in a 839.Nm HAMMER 840file system have to be snapshot separately. 841.Pp 842Example, assuming that 843.Pa /mysnapshots 844is on file system 845.Pa / 846and that 847.Pa /obj 848and 849.Pa /usr 850are file systems on their own, the following invocations: 851.Bd -literal -offset indent 852hammer snapshot /mysnapshots 853 854hammer snapshot /mysnapshots/%Y-%m-%d 855 856hammer snapshot /obj /mysnapshots/obj-%Y-%m-%d 857 858hammer snapshot /usr /my/snaps/usr "note" 859.Ed 860.Pp 861Would create symlinks similar to: 862.Bd -literal -offset indent 863/mysnapshots/snap-20080627-1210 -> /@@0x10d2cd05b7270d16 864 865/mysnapshots/2008-06-27 -> /@@0x10d2cd05b7270d16 866 867/mysnapshots/obj-2008-06-27 -> /obj@@0x10d2cd05b7270d16 868 869/my/snaps/usr/snap-20080627-1210 -> /usr@@0x10d2cd05b7270d16 870.Ed 871.Pp 872When run on a 873.Nm HAMMER 874version 3+ file system the snapshot is also recorded in file system meta-data 875along with the optional 876.Ar note . 877See the 878.Cm snapls 879directive. 880.\" ==== snap* ==== 881.It Cm snap Ar path Op Ar note 882.Nm ( HAMMER 883VERSION 3+) 884Create a snapshot for the PFS containing 885.Ar path 886and create a snapshot softlink. 887If the path specified is a 888directory a standard snapshot softlink will be created in the directory. 889The snapshot softlink points to the base of the mounted PFS. 890.It Cm snaplo Ar path Op Ar note 891.Nm ( HAMMER 892VERSION 3+) 893Create a snapshot for the PFS containing 894.Ar path 895and create a snapshot softlink. 896If the path specified is a 897directory a standard snapshot softlink will be created in the directory. 898The snapshot softlink points into the directory it is contained in. 899.It Cm snapq Ar dir Op Ar note 900.Nm ( HAMMER 901VERSION 3+) 902Create a snapshot for the PFS containing the specified directory but do 903not create a softlink. 904Instead output a path which can be used to access 905the directory via the snapshot. 906.Pp 907An absolute or relative path may be specified. 908The path will be used as-is as a prefix in the path output to stdout. 909As with the other 910snap and snapshot directives the snapshot transaction id will be registered 911in the file system meta-data. 912.It Cm snaprm Ar path Ar ... 913.It Cm snaprm Ar transaction_id Ar ... 914.It Cm snaprm Ar filesystem Ar transaction_id Ar ... 915.Nm ( HAMMER 916VERSION 3+) 917Remove a snapshot given its softlink or transaction id. 918If specifying a transaction id 919the snapshot is removed from file system meta-data but you are responsible 920for removing any related softlinks. 921.Pp 922If a softlink path is specified the filesystem and transaction id 923is derived from the contents of the softlink. 924If just a transaction id is specified it is assumed to be a snapshot in the 925.Nm HAMMER 926filesystem you are currently chdir'd into. 927You can also specify the filesystem and transaction id explicitly. 928.It Cm snapls Op Ar path ... 929.Nm ( HAMMER 930VERSION 3+) 931Dump the snapshot meta-data for PFSs containing each 932.Ar path 933listing all available snapshots and their notes. 934If no arguments are specified snapshots for the PFS containing the 935current directory are listed. 936This is the definitive list of snapshots for the file system. 937.\" ==== prune ==== 938.It Cm prune Ar softlink-dir 939Prune the file system based on previously created snapshot softlinks. 940Pruning is the act of deleting file system history. 941The 942.Cm prune 943command will delete file system history such that 944the file system state is retained for the given snapshots, 945and all history after the latest snapshot. 946By setting the per PFS parameter 947.Cm prune-min , 948history is guaranteed to be saved at least this time interval. 949All other history is deleted. 950.Pp 951The target directory is expected to contain softlinks pointing to 952snapshots of the file systems you wish to retain. 953The directory is scanned non-recursively and the mount points and 954transaction ids stored in the softlinks are extracted and sorted. 955The file system is then explicitly pruned according to what is found. 956Cleaning out portions of the file system is as simple as removing a 957snapshot softlink and then running the 958.Cm prune 959command. 960.Pp 961As a safety measure pruning only occurs if one or more softlinks are found 962containing the 963.Ql @@ 964snapshot id extension. 965Currently the scanned softlink directory must contain softlinks pointing 966to a single 967.Nm HAMMER 968mount. 969The softlinks may specify absolute or relative paths. 970Softlinks must use 20-character 971.Ql @@0x%016llx 972transaction ids, as might be returned from 973.Nm Cm synctid Ar filesystem . 974.Pp 975Pruning is a per PFS operation, so each PFS in a 976.Nm HAMMER 977file system have to be pruned separately. 978.Pp 979Note that pruning a file system may not immediately free-up space, 980though typically some space will be freed if a large number of records are 981pruned out. 982The file system must be reblocked to completely recover all available space. 983.Pp 984Example, lets say your that you didn't set 985.Cm prune-min , 986and snapshot directory contains the following links: 987.Bd -literal -offset indent 988lrwxr-xr-x 1 root wheel 29 May 31 17:57 snap1 -> 989/usr/obj/@@0x10d2cd05b7270d16 990 991lrwxr-xr-x 1 root wheel 29 May 31 17:58 snap2 -> 992/usr/obj/@@0x10d2cd13f3fde98f 993 994lrwxr-xr-x 1 root wheel 29 May 31 17:59 snap3 -> 995/usr/obj/@@0x10d2cd222adee364 996.Ed 997.Pp 998If you were to run the 999.Cm prune 1000command on this directory, then the 1001.Nm HAMMER 1002.Pa /usr/obj 1003mount will be pruned to retain the above three snapshots. 1004In addition, history for modifications made to the file system older than 1005the oldest snapshot will be destroyed and history for potentially fine-grained 1006modifications made to the file system more recently than the most recent 1007snapshot will be retained. 1008.Pp 1009If you then delete the 1010.Pa snap2 1011softlink and rerun the 1012.Cm prune 1013command, 1014history for modifications pertaining to that snapshot would be destroyed. 1015.Pp 1016In 1017.Nm HAMMER 1018file system versions 3+ this command also scans the snapshots stored 1019in the file system meta-data and includes them in the prune. 1020.\" ==== prune-everything ==== 1021.It Cm prune-everything Ar filesystem 1022Remove all historical records from 1023.Ar filesystem . 1024Use this directive with caution on PFSs where you intend to use history. 1025.Pp 1026This command does not remove snapshot softlinks but will delete all 1027snapshots recorded in file system meta-data (for file system version 3+). 1028The user is responsible for deleting any softlinks. 1029.Pp 1030Pruning is a per PFS operation, so each PFS in a 1031.Nm HAMMER 1032file system have to be pruned separately. 1033.\" ==== rebalance ==== 1034.It Cm rebalance Ar filesystem Op Ar saturation_percentage 1035Rebalance the B-Tree, nodes with small number of 1036elements will be combined and element counts will be smoothed out 1037between nodes. 1038.Pp 1039The saturation percentage is between 50% and 100%. 1040The default is 85% (the 1041.Sq % 1042suffix is not needed). 1043.Pp 1044Rebalancing is a per PFS operation, so each PFS in a 1045.Nm HAMMER 1046file system have to be rebalanced separately. 1047.\" ==== dedup ==== 1048.It Cm dedup Ar filesystem 1049.Nm ( HAMMER 1050VERSION 5+) 1051Perform offline (post-process) deduplication. 1052Deduplication occurs at 1053the block level, currently only data blocks of the same size can be 1054deduped, metadata blocks can not. 1055The hash function used for comparing 1056data blocks is CRC-32 (CRCs are computed anyways as part of 1057.Nm HAMMER 1058data integrity features, so there's no additional overhead). 1059Since CRC is a weak hash function a byte-by-byte comparison is done 1060before actual deduping. 1061In case of a CRC collision (two data blocks have the same CRC 1062but different contents) the checksum is upgraded to SHA-256. 1063.Pp 1064Currently 1065.Nm HAMMER 1066reblocker may partially blow up (re-expand) dedup (reblocker's normal 1067operation is to reallocate every record, so it's possible for deduped 1068blocks to be re-expanded back). 1069.Pp 1070Deduplication is a per PFS operation, so each PFS in a 1071.Nm HAMMER 1072file system have to be deduped separately. 1073This also 1074means that if you have duplicated data in two different PFSs that data 1075won't be deduped, however the addition of such feature is planned. 1076.Pp 1077The 1078.Fl m Ar memlimit 1079option should be used to limit memory use during the dedup run if the 1080default 1G limit is too much for the machine. 1081.\" ==== dedup-simulate ==== 1082.It Cm dedup-simulate Ar filesystem 1083Shows potential space savings (simulated dedup ratio) one can get after 1084running 1085.Cm dedup 1086command. 1087If the estimated dedup ratio is greater than 1.00 you will see 1088dedup space savings. 1089Remember that this is an estimated number, in 1090practice real dedup ratio will be slightly smaller because of 1091.Nm HAMMER 1092big-block underflows, B-Tree locking issues and other factors. 1093.Pp 1094Note that deduplication currently works only on bulk data so if you 1095try to run 1096.Cm dedup-simulate 1097or 1098.Cm dedup 1099commands on a PFS that contains metadata only (directory entries, 1100softlinks) you will get a 0.00 dedup ratio. 1101.Pp 1102The 1103.Fl m Ar memlimit 1104option should be used to limit memory use during the dedup run if the 1105default 1G limit is too much for the machine. 1106.\" ==== reblock* ==== 1107.It Cm reblock Ar filesystem Op Ar fill_percentage 1108.It Cm reblock-btree Ar filesystem Op Ar fill_percentage 1109.It Cm reblock-inodes Ar filesystem Op Ar fill_percentage 1110.It Cm reblock-dirs Ar filesystem Op Ar fill_percentage 1111.It Cm reblock-data Ar filesystem Op Ar fill_percentage 1112Attempt to defragment and free space for reuse by reblocking a live 1113.Nm HAMMER 1114file system. 1115Big-blocks cannot be reused by 1116.Nm HAMMER 1117until they are completely free. 1118This command also has the effect of reordering all elements, effectively 1119defragmenting the file system. 1120.Pp 1121The default fill percentage is 100% and will cause the file system to be 1122completely defragmented. 1123All specified element types will be reallocated and rewritten. 1124If you wish to quickly free up space instead try specifying 1125a smaller fill percentage, such as 90% or 80% (the 1126.Sq % 1127suffix is not needed). 1128.Pp 1129Since this command may rewrite the entire contents of the disk it is 1130best to do it incrementally from a 1131.Xr cron 8 1132job along with the 1133.Fl c Ar cyclefile 1134and 1135.Fl t Ar seconds 1136options to limit the run time. 1137The file system would thus be defragmented over long period of time. 1138.Pp 1139It is recommended that separate invocations be used for each data type. 1140B-Tree nodes, inodes, and directories are typically the most important 1141elements needing defragmentation. 1142Data can be defragmented over a longer period of time. 1143.Pp 1144Reblocking is a per PFS operation, so each PFS in a 1145.Nm HAMMER 1146file system have to be reblocked separately. 1147.\" ==== pfs-status ==== 1148.It Cm pfs-status Ar dirpath ... 1149Retrieve the mirroring configuration parameters for the specified 1150.Nm HAMMER 1151file systems or pseudo-filesystems (PFS's). 1152.\" ==== pfs-master ==== 1153.It Cm pfs-master Ar dirpath Op Ar options 1154Create a pseudo-filesystem (PFS) inside a 1155.Nm HAMMER 1156file system. 1157Up to 65536 PFSs can be created. 1158Each PFS uses an independent inode numbering space making it suitable 1159for replication. 1160.Pp 1161The 1162.Cm pfs-master 1163directive creates a PFS that you can read, write, and use as a mirroring 1164source. 1165.Pp 1166A PFS can only be truly destroyed with the 1167.Cm pfs-destroy 1168directive. 1169Removing the softlink will not destroy the underlying PFS. 1170.Pp 1171A PFS can only be created in the root PFS (PFS# 0), 1172not in a PFS created by 1173.Cm pfs-master 1174or 1175.Cm pfs-slave 1176(PFS# >0). 1177.Pp 1178It is recommended that 1179.Ar dirpath 1180is of the form 1181.Pa <fs>/pfs/<name> 1182(i.e.\& located in 1183.Pa PFS 1184directory at root of 1185.Nm HAMMER 1186file system). 1187.Pp 1188It is recommended to use a 1189.Nm null 1190mount to access a PFS, except for root PFS, for more information see 1191.Xr HAMMER 5 . 1192.\" ==== pfs-slave ==== 1193.It Cm pfs-slave Ar dirpath Op Ar options 1194Create a pseudo-filesystem (PFS) inside a 1195.Nm HAMMER 1196file system. 1197Up to 65536 PFSs can be created. 1198Each PFS uses an independent inode numbering space making it suitable 1199for replication. 1200.Pp 1201The 1202.Cm pfs-slave 1203directive creates a PFS that you can use as a mirroring source or target. 1204You will not be able to access a slave PFS until you have completed the 1205first mirroring operation with it as the target (its root directory will 1206not exist until then). 1207.Pp 1208Access to the pfs-slave via the special softlink, as described in the 1209.Sx PSEUDO-FILESYSTEM (PFS) NOTES 1210below, allows 1211.Nm HAMMER 1212to 1213dynamically modify the snapshot transaction id by returning a dynamic result 1214from 1215.Xr readlink 2 1216calls. 1217.Pp 1218A PFS can only be truly destroyed with the 1219.Cm pfs-destroy 1220directive. 1221Removing the softlink will not destroy the underlying PFS. 1222.Pp 1223A PFS can only be created in the root PFS (PFS# 0), 1224not in a PFS created by 1225.Cm pfs-master 1226or 1227.Cm pfs-slave 1228(PFS# >0). 1229.Pp 1230It is recommended that 1231.Ar dirpath 1232is of the form 1233.Pa <fs>/pfs/<name> 1234(i.e.\& located in 1235.Pa PFS 1236directory at root of 1237.Nm HAMMER 1238file system). 1239.Pp 1240It is recommended to use a 1241.Nm null 1242mount to access a PFS, except for root PFS, for more information see 1243.Xr HAMMER 5 . 1244.\" ==== pfs-update ==== 1245.It Cm pfs-update Ar dirpath Op Ar options 1246Update the configuration parameters for an existing 1247.Nm HAMMER 1248file system or pseudo-filesystem. 1249Options that may be specified: 1250.Bl -tag -width indent 1251.It Cm sync-beg-tid= Ns Ar 0x16llx 1252This is the automatic snapshot access starting transaction id for 1253mirroring slaves. 1254This parameter is normally updated automatically by the 1255.Cm mirror-write 1256directive. 1257.Pp 1258It is important to note that accessing a mirroring slave 1259with a transaction id greater than the last fully synchronized transaction 1260id can result in an unreliable snapshot since you will be accessing 1261data that is still undergoing synchronization. 1262.Pp 1263Manually modifying this field is dangerous and can result in a broken mirror. 1264.It Cm sync-end-tid= Ns Ar 0x16llx 1265This is the current synchronization point for mirroring slaves. 1266This parameter is normally updated automatically by the 1267.Cm mirror-write 1268directive. 1269.Pp 1270Manually modifying this field is dangerous and can result in a broken mirror. 1271.It Cm shared-uuid= Ns Ar uuid 1272Set the shared UUID for this file system. 1273All mirrors must have the same shared UUID. 1274For safety purposes the 1275.Cm mirror-write 1276directives will refuse to operate on a target with a different shared UUID. 1277.Pp 1278Changing the shared UUID on an existing, non-empty mirroring target, 1279including an empty but not completely pruned target, 1280can lead to corruption of the mirroring target. 1281.It Cm unique-uuid= Ns Ar uuid 1282Set the unique UUID for this file system. 1283This UUID should not be used anywhere else, 1284even on exact copies of the file system. 1285.It Cm label= Ns Ar string 1286Set a descriptive label for this file system. 1287.It Cm snapshots= Ns Ar string 1288Specify the snapshots directory which 1289.Nm 1290.Cm cleanup 1291will use to manage this PFS. 1292.Bl -tag -width indent 1293.It Nm HAMMER No version 2- 1294The snapshots directory does not need to be configured for 1295PFS masters and will default to 1296.Pa <pfs>/snapshots . 1297.Pp 1298PFS slaves are mirroring slaves so you cannot configure a snapshots 1299directory on the slave itself to be managed by the slave's machine. 1300In fact, the slave will likely have a 1301.Pa snapshots 1302sub-directory mirrored 1303from the master, but that directory contains the configuration the master 1304is using for its copy of the file system, not the configuration that we 1305want to use for our slave. 1306.Pp 1307It is recommended that 1308.Pa <fs>/var/slaves/<name> 1309be configured for a PFS slave, where 1310.Pa <fs> 1311is the base 1312.Nm HAMMER 1313file system, and 1314.Pa <name> 1315is an appropriate label. 1316.It Nm HAMMER No version 3+ 1317The snapshots directory does not need to be configured for PFS masters or 1318slaves. 1319The snapshots directory defaults to 1320.Pa /var/hammer/<pfs> 1321.Pa ( /var/hammer/root 1322for root mount). 1323.El 1324.Pp 1325You can control snapshot retention on your slave independent of the master. 1326.It Cm snapshots-clear 1327Zero out the 1328.Cm snapshots 1329directory path for this PFS. 1330.It Cm prune-min= Ns Ar N Ns Cm d 1331.It Cm prune-min= Ns Oo Ar N Ns Cm d/ Oc Ns \ 1332Ar hh Ns Op Cm \&: Ns Ar mm Ns Op Cm \&: Ns Ar ss 1333Set the minimum fine-grained data retention period. 1334.Nm HAMMER 1335always retains fine-grained history up to the most recent snapshot. 1336You can extend the retention period further by specifying a non-zero 1337pruning minimum. 1338Any snapshot softlinks within the retention period are ignored 1339for the purposes of pruning (i.e.\& the fine grained history is retained). 1340Number of days, hours, minutes and seconds are given as 1341.Ar N , hh , mm 1342and 1343.Ar ss . 1344.Pp 1345Because the transaction id in the snapshot softlink cannot be used 1346to calculate a timestamp, 1347.Nm HAMMER 1348uses the earlier of the 1349.Fa st_ctime 1350or 1351.Fa st_mtime 1352field of the softlink to 1353determine which snapshots fall within the retention period. 1354Users must be sure to retain one of these two fields when manipulating 1355the softlink. 1356.El 1357.\" ==== pfs-upgrade ==== 1358.It Cm pfs-upgrade Ar dirpath 1359Upgrade a PFS from slave to master operation. 1360The PFS will be rolled back to the current end synchronization transaction id 1361(removing any partial synchronizations), and will then become writable. 1362.Pp 1363.Em WARNING! 1364.Nm HAMMER 1365currently supports only single masters and using 1366this command can easily result in file system corruption 1367if you don't know what you are doing. 1368.Pp 1369This directive will refuse to run if any programs have open descriptors 1370in the PFS, including programs chdir'd into the PFS. 1371.\" ==== pfs-downgrade ==== 1372.It Cm pfs-downgrade Ar dirpath 1373Downgrade a master PFS from master to slave operation. 1374The PFS becomes read-only and access will be locked to its 1375.Cm sync-end-tid . 1376.Pp 1377This directive will refuse to run if any programs have open descriptors 1378in the PFS, including programs chdir'd into the PFS. 1379.\" ==== pfs-destroy ==== 1380.It Cm pfs-destroy Ar dirpath 1381This permanently destroys a PFS. 1382.Pp 1383This directive will refuse to run if any programs have open descriptors 1384in the PFS, including programs chdir'd into the PFS. 1385As safety measure the 1386.Fl y 1387flag have no effect on this directive. 1388.\" ==== mirror-read ==== 1389.It Cm mirror-read Ar filesystem Op Ar begin-tid 1390Generate a mirroring stream to stdout. 1391The stream ends when the transaction id space has been exhausted. 1392.Ar filesystem 1393may be a master or slave PFS. 1394.\" ==== mirror-read-stream ==== 1395.It Cm mirror-read-stream Ar filesystem Op Ar begin-tid 1396Generate a mirroring stream to stdout. 1397Upon completion the stream is paused until new data is synced to the 1398.Ar filesystem , 1399then resumed. 1400Operation continues until the pipe is broken. 1401See the 1402.Cm mirror-stream 1403command for more details. 1404.\" ==== mirror-write ==== 1405.It Cm mirror-write Ar filesystem 1406Take a mirroring stream on stdin. 1407.Ar filesystem 1408must be a slave PFS. 1409.Pp 1410This command will fail if the 1411.Cm shared-uuid 1412configuration field for the two file systems do not match. 1413See the 1414.Cm mirror-copy 1415command for more details. 1416.Pp 1417If the target PFS does not exist this command will ask you whether 1418you want to create a compatible PFS slave for the target or not. 1419.\" ==== mirror-dump ==== 1420.It Ar mirror-dump Ar [header] 1421A 1422.Cm mirror-read 1423can be piped into a 1424.Cm mirror-dump 1425to dump an ASCII representation of the mirroring stream. 1426If the keyword 1427.Ar header 1428is specified, only the header information is shown. 1429.\" ==== mirror-copy ==== 1430.\".It Cm mirror-copy Ar [[user@]host:]filesystem [[user@]host:]filesystem 1431.It Cm mirror-copy \ 1432Oo Oo Ar user Ns Cm @ Oc Ns Ar host Ns Cm \&: Oc Ns Ar filesystem \ 1433Oo Oo Ar user Ns Cm @ Oc Ns Ar host Ns Cm \&: Oc Ns Ar filesystem 1434This is a shortcut which pipes a 1435.Cm mirror-read 1436command to a 1437.Cm mirror-write 1438command. 1439If a remote host specification is made the program forks a 1440.Xr ssh 1 1441(or other program as specified by the 1442.Ev HAMMER_RSH 1443environment variable) and execs the 1444.Cm mirror-read 1445and/or 1446.Cm mirror-write 1447on the appropriate host. 1448The source may be a master or slave PFS, and the target must be a slave PFS. 1449.Pp 1450This command also establishes full duplex communication and turns on 1451the 2-way protocol feature 1452.Fl ( 2 ) 1453which automatically negotiates transaction id 1454ranges without having to use a cyclefile. 1455If the operation completes successfully the target PFS's 1456.Cm sync-end-tid 1457will be updated. 1458Note that you must re-chdir into the target PFS to see the updated information. 1459If you do not you will still be in the previous snapshot. 1460.Pp 1461If the target PFS does not exist this command will ask you whether 1462you want to create a compatible PFS slave for the target or not. 1463.\" ==== mirror-stream ==== 1464.\".It Cm mirror-stream Ar [[user@]host:]filesystem [[user@]host:]filesystem 1465.It Cm mirror-stream \ 1466Oo Oo Ar user Ns Cm @ Oc Ns Ar host Ns Cm \&: Oc Ns Ar filesystem \ 1467Oo Oo Ar user Ns Cm @ Oc Ns Ar host Ns Cm \&: Oc Ns Ar filesystem 1468This is a shortcut which pipes a 1469.Cm mirror-read-stream 1470command to a 1471.Cm mirror-write 1472command. 1473This command works similarly to 1474.Cm mirror-copy 1475but does not exit after the initial mirroring completes. 1476The mirroring operation will resume as changes continue to be made to the 1477source. 1478The command is commonly used with 1479.Fl i Ar delay 1480and 1481.Fl b Ar bandwidth 1482options to keep the mirroring target in sync with the source on a continuing 1483basis. 1484.Pp 1485If the pipe is broken the command will automatically retry after sleeping 1486for a short while. 1487The time slept will be 15 seconds plus the time given in the 1488.Fl i 1489option. 1490.Pp 1491This command also detects the initial-mirroring case and spends some 1492time scanning the B-Tree to find good break points, allowing the initial 1493bulk mirroring operation to be broken down into 4GB pieces. 1494This means that the user can kill and restart the operation and it will 1495not have to start from scratch once it has gotten past the first chunk. 1496The 1497.Fl S 1498option may be used to change the size of pieces and the 1499.Fl B 1500option may be used to disable this feature and perform an initial bulk 1501transfer instead. 1502.\" ==== version ==== 1503.It Cm version Ar filesystem 1504This command returns the 1505.Nm HAMMER 1506file system version for the specified 1507.Ar filesystem 1508as well as the range of versions supported in the kernel. 1509The 1510.Fl q 1511option may be used to remove the summary at the end. 1512.\" ==== version-upgrade ==== 1513.It Cm version-upgrade Ar filesystem Ar version Op Cm force 1514Upgrade the 1515.Nm HAMMER 1516.Ar filesystem 1517to the specified 1518.Ar version . 1519Once upgraded a file system may not be downgraded. 1520If you wish to upgrade a file system to a version greater or equal to the 1521work-in-progress (WIP) version number you must specify the 1522.Cm force 1523directive. 1524Use of WIP versions should be relegated to testing and may require wiping 1525the file system as development progresses, even though the WIP version might 1526not change. 1527.Pp 1528.Em NOTE! 1529This command operates on the entire 1530.Nm HAMMER 1531file system and is not a per PFS operation. 1532All PFS's will be affected. 1533.Bl -tag -width indent 1534.It 1 1535.Dx 2.0 1536default version, first 1537.Nm HAMMER 1538release. 1539.It 2 1540.Dx 2.3 . 1541New directory entry layout. 1542This version is using a new directory hash key. 1543.It 3 1544.Dx 2.5 . 1545New snapshot management, using file system meta-data for saving 1546configuration file and snapshots (transaction ids etc.). 1547Also default snapshots directory has changed. 1548.It 4 1549.Dx 2.6 1550default version. 1551New undo/redo/flush, giving 1552.Nm HAMMER 1553a much faster sync and fsync. 1554.It 5 1555.Dx 2.9 . 1556Deduplication support. 1557.It 6 1558.Dx 2.9 . 1559Directory hash ALG1. 1560Tends to maintain inode number / directory name entry ordering better 1561for files after minor renaming. 1562.El 1563.El 1564.Sh PSEUDO-FILESYSTEM (PFS) NOTES 1565The root of a PFS is not hooked into the primary 1566.Nm HAMMER 1567file system as a directory. 1568Instead, 1569.Nm HAMMER 1570creates a special softlink called 1571.Ql @@PFS%05d 1572(exactly 10 characters long) in the primary 1573.Nm HAMMER 1574file system. 1575.Nm HAMMER 1576then modifies the contents of the softlink as read by 1577.Xr readlink 2 , 1578and thus what you see with an 1579.Nm ls 1580command or if you were to 1581.Nm cd 1582into the link. 1583If the PFS is a master the link reflects the current state of the PFS. 1584If the PFS is a slave the link reflects the last completed snapshot, and the 1585contents of the link will change when the next snapshot is completed, and 1586so forth. 1587.Pp 1588The 1589.Nm 1590utility employs numerous safeties to reduce user foot-shooting. 1591The 1592.Cm mirror-copy 1593directive requires that the target be configured as a slave and that the 1594.Cm shared-uuid 1595field of the mirroring source and target match. 1596The 1597.Cm pfs-master 1598and 1599.Cm pfs-slave 1600directives require that the PFS softlink be created under the main 1601.Nm 1602filesystem mount. You may only access PFS softlinks via the main 1603.Nm 1604filesystem mount. 1605.Sh DOUBLE_BUFFER MODE 1606There is a limit to the number of vnodes the kernel can cache, and because 1607file buffers are associated with a vnode the related data cache can get 1608blown away when operating on large numbers of files even if the system has 1609sufficient memory to hold the file data. 1610.Pp 1611If you turn on 1612.Nm HAMMER Ns 's 1613double buffer mode by setting the 1614.Xr sysctl 8 1615node 1616.Va vfs.hammer.double_buffer 1617to 1 1618.Nm HAMMER 1619will cache file data via the block device and copy it into the per-file 1620buffers as needed. The data will be double-cached at least until the 1621buffer cache throws away the file buffer. 1622This mode is typically used in conjunction with 1623.Xr swapcache 8 1624when 1625.Va vm.swapcache.data_enable 1626is turned on in order to prevent unnecessary re-caching of file data 1627due to vnode recycling. 1628The swapcache will save the cached VM pages related to 1629.Nm HAMMER Ns 's 1630block 1631device (which doesn't recycle unless you umount the filesystem) instead 1632of the cached VM pages backing the file vnodes. 1633.Pp 1634Double buffering is normally desirable when working with large filesystems, 1635particularly when swapcache is used. 1636The swapcache can only back active VM objects, including the block device, 1637and large filesystems often have far more inodes than the kernel can support. 1638In addition, when using this mode, you may wish to reduce the 1639.Va kern.maxvnodes 1640setting for the system to force the system to do less caching of logical 1641file buffers and more caching of device buffers, since the device buffers 1642are backing the logical file buffers. 1643.Sh UPGRADE INSTRUCTIONS HAMMER V1 TO V2 1644This upgrade changes the way directory entries are stored. 1645It is possible to upgrade a V1 file system to V2 in place, but 1646directories created prior to the upgrade will continue to use 1647the old layout. 1648.Pp 1649Note that the slave mirroring code in the target kernel had bugs in 1650V1 which can create an incompatible root directory on the slave. 1651Do not mix a 1652.Nm HAMMER 1653master created after the upgrade with a 1654.Nm HAMMER 1655slave created prior to the upgrade. 1656.Pp 1657Any directories created after upgrading will use a new layout. 1658.Sh UPGRADE INSTRUCTIONS HAMMER V2 TO V3 1659This upgrade adds meta-data elements to the B-Tree. 1660It is possible to upgrade a V2 file system to V3 in place. 1661After issuing the upgrade be sure to run a 1662.Nm 1663.Cm cleanup 1664to perform post-upgrade tasks. 1665.Pp 1666After making this upgrade running a 1667.Nm 1668.Cm cleanup 1669will move the 1670.Pa <pfs>/snapshots 1671directory for each PFS mount into 1672.Pa /var/hammer/<pfs> . 1673A 1674.Nm HAMMER 1675root mount will migrate 1676.Pa /snapshots 1677into 1678.Pa /var/hammer/root . 1679Migration occurs only once and only if you have not specified 1680a snapshots directory in the PFS configuration. 1681If you have specified a snapshots directory in the PFS configuration no 1682automatic migration will occur. 1683.Pp 1684For slaves, if you desire, you can migrate your snapshots 1685config to the new location manually and then clear the 1686snapshot directory configuration in the slave PFS. 1687The new snapshots hierarchy is designed to work with 1688both master and slave PFSs equally well. 1689.Pp 1690In addition, the old config file will be moved to file system meta-data, 1691editable via the new 1692.Nm 1693.Cm viconfig 1694directive. 1695The old config file will be deleted. 1696Migration occurs only once. 1697.Pp 1698The V3 file system has new 1699.Cm snap* 1700directives for creating snapshots. 1701All snapshot directives, including the original, will create 1702meta-data entries for the snapshots and the pruning code will 1703automatically incorporate these entries into its list and 1704expire them the same way it expires softlinks. 1705If you by accident blow away your snapshot softlinks you can use the 1706.Cm snapls 1707directive to get a definitive list from the file system meta-data and 1708regenerate them from that list. 1709.Pp 1710.Em WARNING! 1711If you are using 1712.Nm 1713to backup file systems your scripts may be using the 1714.Cm synctid 1715directive to generate transaction ids. 1716This directive does not create a snapshot. 1717You will have to modify your scripts to use the 1718.Cm snapq 1719directive to generate the linkbuf for the softlink you create, or 1720use one of the other 1721.Cm snap* 1722directives. 1723The older 1724.Cm snapshot 1725directive will continue to work as expected and in V3 it will also 1726record the snapshot transaction id in file system meta-data. 1727You may also want to make use of the new 1728.Ar note 1729tag for the meta-data. 1730.Pp 1731.Em WARNING! 1732If you used to remove snapshot softlinks with 1733.Nm rm 1734you should probably start using the 1735.Cm snaprm 1736directive instead to also remove the related meta-data. 1737The pruning code scans the meta-data so just removing the 1738softlink is not sufficient. 1739.Sh UPGRADE INSTRUCTIONS HAMMER V3 TO V4 1740This upgrade changes undo/flush, giving faster sync. 1741It is possible to upgrade a V3 file system to V4 in place. 1742This upgrade reformats the UNDO/REDO FIFO (typically 1GB), 1743so upgrade might take a minute or two depending. 1744.Pp 1745Version 4 allows the UNDO/REDO FIFO to be flushed without also having 1746to flush the volume header, removing 2 of the 4 disk syncs typically 1747required for an 1748.Fn fsync 1749and removing 1 of the 2 disk syncs typically 1750required for a flush sequence. 1751Version 4 also implements the REDO log (see 1752.Sx FSYNC FLUSH MODES 1753below) which is capable 1754of fsync()ing with either one disk flush or zero disk flushes. 1755.Sh UPGRADE INSTRUCTIONS HAMMER V4 TO V5 1756This upgrade brings in deduplication support. 1757It is possible to upgrade a V4 file system to V5 in place. 1758Technically it makes the layer2 1759.Va bytes_free 1760field a signed value instead of unsigned, allowing it to go negative. 1761A version 5 filesystem is required for dedup operation. 1762.Sh UPGRADE INSTRUCTIONS HAMMER V5 TO V6 1763It is possible to upgrade a V5 file system to V6 in place. 1764.Sh FSYNC FLUSH MODES 1765.Nm HAMMER 1766implements five different fsync flush modes via the 1767.Va vfs.hammer.fsync_mode 1768sysctl, for 1769.Nm HAMMER 1770version 4+ file systems. 1771.Pp 1772As of 1773.Dx 2.6 1774fsync mode 3 is set by default. 1775REDO operation and recovery is enabled by default. 1776.Bl -tag -width indent 1777.It mode 0 1778Full synchronous fsync semantics without REDO. 1779.Pp 1780.Nm HAMMER 1781will not generate REDOs. 1782A 1783.Fn fsync 1784will completely sync 1785the data and meta-data and double-flush the FIFO, including 1786issuing two disk synchronization commands. 1787The data is guaranteed 1788to be on the media as of when 1789.Fn fsync 1790returns. 1791Needless to say, this is slow. 1792.It mode 1 1793Relaxed asynchronous fsync semantics without REDO. 1794.Pp 1795This mode works the same as mode 0 except the last disk synchronization 1796command is not issued. 1797It is faster than mode 0 but not even remotely 1798close to the speed you get with mode 2 or mode 3. 1799.Pp 1800Note that there is no chance of meta-data corruption when using this 1801mode, it simply means that the data you wrote and then 1802.Fn fsync Ns 'd 1803might not have made it to the media if the storage system crashes at a bad 1804time. 1805.It mode 2 1806Full synchronous fsync semantics using REDO. 1807NOTE: If not running a 1808.Nm HAMMER 1809version 4 filesystem or later mode 0 is silently used. 1810.Pp 1811.Nm HAMMER 1812will generate REDOs in the UNDO/REDO FIFO based on a heuristic. 1813If this is sufficient to satisfy the 1814.Fn fsync 1815operation the blocks will be written out and 1816.Nm HAMMER 1817will wait for the I/Os to complete, 1818and then followup with a disk sync command to guarantee the data 1819is on the media before returning. 1820This is slower than mode 3 and can result in significant disk or 1821SSDs overheads, though not as bad as mode 0 or mode 1. 1822.It mode 3 1823Relaxed asynchronous fsync semantics using REDO. 1824NOTE: If not running a 1825.Nm HAMMER 1826version 4 filesystem or later mode 1 is silently used. 1827.Pp 1828.Nm HAMMER 1829will generate REDOs in the UNDO/REDO FIFO based on a heuristic. 1830If this is sufficient to satisfy the 1831.Fn fsync 1832operation the blocks 1833will be written out and 1834.Nm HAMMER 1835will wait for the I/Os to complete, 1836but will 1837.Em NOT 1838issue a disk synchronization command. 1839.Pp 1840Note that there is no chance of meta-data corruption when using this 1841mode, it simply means that the data you wrote and then 1842.Fn fsync Ns 'd 1843might 1844not have made it to the media if the storage system crashes at a bad 1845time. 1846.Pp 1847This mode is the fastest production fsyncing mode available. 1848This mode is equivalent to how the UFS fsync in the 1849.Bx Ns s 1850operates. 1851.It mode 4 1852fsync is ignored. 1853.Pp 1854Calls to 1855.Fn fsync 1856will be ignored. 1857This mode is primarily designed 1858for testing and should not be used on a production system. 1859.El 1860.Sh RESTORING FROM A SNAPSHOT BACKUP 1861You restore a snapshot by copying it over to live, but there is a caveat. 1862The mtime and atime fields for files accessed via a snapshot is locked 1863to the ctime in order to keep the snapshot consistent, because neither 1864mtime nor atime changes roll any history. 1865.Pp 1866In order to avoid unnecessary copying it is recommended that you use 1867.Nm cpdup 1868.Fl VV 1869.Fl v 1870when doing the copyback. 1871Also make sure you traverse the snapshot softlink by appending a ".", 1872as in "<snapshotpath>/.", and you match up the directory properly. 1873.Sh RESTORING A PFS FROM A MIRROR 1874A PFS can be restored from a mirror with 1875.Cm mirror-copy . 1876.Cm config 1877data must be copied separately. 1878At last the PFS can be upgraded to master using 1879.Cm pfs-upgrade . 1880.Pp 1881It is not possible to restore the root PFS (PFS# 0) by using mirroring, 1882as the root PFS is always a master PFS. 1883A normal copy (e.g.\& using 1884.Xr cpdup 1 ) 1885must be done, ignoring history. 1886If history is important, old root PFS can me restored to a new PFS, and 1887important directories/files can be 1888.Nm null 1889mounted to the new PFS. 1890.Sh ENVIRONMENT 1891The following environment variables affect the execution of 1892.Nm : 1893.Bl -tag -width ".Ev EDITOR" 1894.It Ev EDITOR 1895The editor program specified in the variable 1896.Ev EDITOR 1897will be invoked instead of the default editor, which is 1898.Xr vi 1 . 1899.It Ev HAMMER_RSH 1900The command specified in the variable 1901.Ev HAMMER_RSH 1902will be used to initiate remote operations for the mirror-copy and 1903mirror-stream commands instead of the default command, which is 1904.Xr ssh 1 . 1905The program will be invoked via 1906.Xr execvp 3 1907using a typical 1908.Xr rsh 1 1909style 1910.Cm -l user host <remote-command> 1911command line. 1912.It Ev VISUAL 1913Same effect as 1914.Ev EDITOR 1915variable. 1916.El 1917.Sh FILES 1918.Bl -tag -width ".It Pa <fs>/var/slaves/<name>" -compact 1919.It Pa <pfs>/snapshots 1920default per PFS snapshots directory 1921.Nm ( HAMMER 1922VERSION 2-) 1923.It Pa /var/hammer/<pfs> 1924default per PFS snapshots directory (not root) 1925.Nm ( HAMMER 1926VERSION 3+) 1927.It Pa /var/hammer/root 1928default snapshots directory for root directory 1929.Nm ( HAMMER 1930VERSION 3+) 1931.It Pa <snapshots>/config 1932per PFS 1933.Nm 1934.Cm cleanup 1935configuration file 1936.Nm ( HAMMER 1937VERSION 2-) 1938.It Pa <fs>/var/slaves/<name> 1939recommended slave PFS snapshots directory 1940.Nm ( HAMMER 1941VERSION 2-) 1942.It Pa <fs>/pfs 1943recommended PFS directory 1944.El 1945.Sh EXIT STATUS 1946.Ex -std 1947.Sh SEE ALSO 1948.Xr ssh 1 , 1949.Xr undo 1 , 1950.Xr HAMMER 5 , 1951.Xr periodic.conf 5 , 1952.Xr loader 8 , 1953.Xr mount_hammer 8 , 1954.Xr mount_null 8 , 1955.Xr newfs_hammer 8 , 1956.Xr swapcache 8 , 1957.Xr sysctl 8 1958.Sh HISTORY 1959The 1960.Nm 1961utility first appeared in 1962.Dx 1.11 . 1963.Sh AUTHORS 1964.An Matthew Dillon Aq Mt dillon@backplane.com 1965