1.\" Copyright (c) 2015 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 26, 2015 34.Dt HAMMER2 8 35.Os 36.Sh NAME 37.Nm hammer2 38.Nd hammer2 file system utility 39.Sh SYNOPSIS 40.Nm 41.Fl h 42.Nm 43.Op Fl s Ar path 44.Op Fl t Ar type 45.Op Fl u Ar uuid 46.Ar command 47.Op Ar argument ... 48.Sh DESCRIPTION 49The 50.Nm 51utility provides miscellaneous support functions for a 52HAMMER2 file system. 53.Pp 54The options are as follows: 55.Bl -tag -width indent 56.It Fl s Ar path 57Specify the path to a mounted HAMMER2 filesystem. 58At least one PFS on a HAMMER2 filesystem must be mounted for the system 59to act on all PFSs managed by it. 60Every HAMMER2 filesystem typically has a PFS called "LOCAL" for this purpose. 61.It Fl t Ar type 62Specify the type when creating, upgrading, or downgrading a PFS. 63Supported types are MASTER, SLAVE, SOFT_MASTER, SOFT_SLAVE, CACHE, and DUMMY. 64If not specified the pfs-create directive will default to MASTER if no 65uuid is specified, and SLAVE if a uuid is specified. 66.It Fl u Ar uuid 67Specify the cluster uuid when creating a PFS. If not specified, a unique, 68random uuid will be generated. 69Note that every PFS also has a unique pfs_id which is always generated 70and cannot be overridden with an option. 71The { pfs_clid, pfs_fsid } tuple uniquely identifies a component of a cluster. 72.El 73.Pp 74.Nm 75directives are as shown below. 76Note that most directives require you to either be CD'd into a hammer2 77filesystem, specify a path to a mounted hammer2 filesystem via the 78.Fl s 79option, or specify a path after the directive. 80It depends on the directive. 81All hammer2 filesystem have a PFS called "LOCAL" which is typically mounted 82locally on the host in order to be able to issue commands for other PFSs 83on the filesystem. 84The mount also enables PFS configuration scanning for that filesystem. 85.Bl -tag -width indent 86.\" ==== connect ==== 87.It Cm connect Ar target 88Add a cluster link entry to the volume header. 89The volume header can support up to 255 link entries. 90This feature is not currently used. 91.\" ==== disconnect ==== 92.It Cm disconnect Ar target 93Delete a cluster link entry from the volume header. 94This feature is not currently used. 95.\" ==== info ==== 96.It Cm info Op devpath 97Access and print the status and super-root entries for all HAMMER2 98partitions found in /dev/serno or the specified device path(s). 99The partitions do not have to be mounted. 100Note that only mounted partitions will be under active management. 101This is accomplished by mounting at least one PFS within the partition. 102Typically at least the @LOCAL PFS is mounted. 103.\" ==== mountall ==== 104.It Cm mountall Op devpath 105This directive mounts the @LOCAL PFS on all HAMMER2 partitions found 106in /dev/serno, or the specified device path(s). 107The partitions are mounted as /var/hammer2/LOCAL.<id>. 108Mounts are executed in the background and this command will wait a 109limited amount of time for the mounts to complete before returning. 110.\" ==== status ==== 111.It Cm status Ar path... 112Dump a list of all cluster link entries configured in the volume header. 113.\" ==== hash ==== 114.It Cm hash Ar filename... 115Compute and print the directory hash for any number of filenames. 116.\" ==== pfs-list ==== 117.It Cm pfs-list Op path... 118List all local PFSs available on a mounted HAMMER2 filesystem, their type, 119and their current status. 120You must mount at least one PFS in order to be able to access the whole list. 121.\" ==== pfs-clid ==== 122.It Cm pfs-clid Ar label 123Print the cluster id for a PFS specified by name. 124.\" ==== pfs-fsid ==== 125.It Cm pfs-fsid Ar label 126Print the unique filesystem id for a PFS specified by name. 127.\" ==== pfs-create ==== 128.It Cm pfs-create Ar label 129Create a local PFS on a mounted HAMMER2 filesystem. 130If no uuid is specified the pfs-type defaults to MASTER. 131If a uuid is specified via the 132.Fl u 133option the pfs-type defaults to SLAVE. 134Other types can be specified with the 135.Fl t 136option. 137.Pp 138If you wish to add a MASTER to an existing cluster, you must first add it as 139a SLAVE and then upgrade it to MASTER to properly synchronize it. 140.Pp 141The DUMMY pfs-type is used to tie network-accessible clusters into the local 142machine when no local storage is desired. 143This type should be used on minimal H2 partitions or entirely in ram for 144netboot-centric systems to provide a tie-in point for the mount command, 145or on more complex systems where you need to also access network-centric 146clusters. 147.Pp 148The CACHE or SLAVE pfs-type is typically used when the main store is on 149the network but local storage is desired to improve performance. 150SLAVE is also used when a backup is desired. 151.Pp 152Generally speaking, you can mount any PFS element of a cluster in order to 153access the cluster via the full cluster protocol. 154There are two exceptions. 155If you mount a SOFT_SLAVE or a SOFT_MASTER then soft quorum semantics are 156employed... the soft slave or soft master's current state will always be used 157and the quorum protocol will not be used. The soft PFS will still be 158synchronized to masters in the background when available. 159Also, you can use 'mount -o local' to mount ONLY a local HAMMER2 PFS and 160not run any network or quorum protocols for the mount. 161All such mounts except for a SOFT_MASTER mount will be read-only. 162Other than that, you will be mounting the whole cluster when you mount any 163PFS within the cluster. 164.Pp 165DUMMY - Create a PFS skeleton intended to be the mount point for a 166more complex cluster, probably one that is entirely network based. 167No data will be synchronized to this PFS so it is suitable for use 168in a network boot image or memory filesystem. 169This allows you to create placeholders for mount points on your local 170disk, SSD, or memory disk. 171.Pp 172CACHE - Create a PFS for caching portions of the cluster piecemeal. 173This is similar to a SLAVE but does not synchronize the entire contents of 174the cluster to the PFS. 175Elements found in the CACHE PFS which are validated against the cluster 176will be read, presumably a faster access than having to go to the cluster. 177Only local CACHEs will be updated. 178Network-accessible CACHE PFSs might be read but will not be written to. 179If you have a large hard-drive-based cluster you can set up localized 180SSD CACHE PFSs to improve performance. 181.Pp 182SLAVE - Create a PFS which maintains synchronization with and provides a 183read-only copy of the cluster. 184HAMMER2 will prioritize local SLAVEs for data retrieval after validating 185their transaction id against the cluster. 186The difference between a CACHE and a SLAVE is that the SLAVE is synchronized 187to a full copy of the cluster and thus can serve as a backup or be staged 188for use as a MASTER later on. 189.Pp 190SOFT_SLAVE - Create a PFS which maintains synchronization with and provides 191a read-only copy of the cluster. 192This is one of the special mount cases. A SOFT_SLAVE will synchronize with 193the cluster when the cluster is available, but can still be accessed when 194the cluster is not available. 195.Pp 196MASTER - Create a PFS which will hold a master copy of the cluster. 197If you create several MASTER PFSs with the same cluster id you are 198effectively creating a multi-master cluster and causing a quorum and 199cache coherency protocol to be used to validate operations. 200The total number of masters is stored in each PFSs making up the cluster. 201Filesystem operations will stall for normal mounts if a quorum cannot be 202obtained to validate the operation. 203MASTER nodes which go offline and return later will synchronize in the 204background. 205Note that when adding a MASTER to an existing cluster you must add the 206new PFS as a SLAVE and then upgrade it to a MASTER. 207.Pp 208SOFT_MASTER - Create a PFS which maintains synchronization with and provides 209a read-write copy of the cluster. 210This is one of the special mount cases. A SOFT_MASTER will synchronize with 211the cluster when the cluster is available, but can still be read AND written 212to even when the cluster is not available. 213Modifications made to a SOFT_MASTER will be automatically flushed to the 214cluster when it becomes accessible again, and vise-versa. 215Manual intervention may be required if a conflict occurs during 216synchronization. 217.\" ==== pfs-delete ==== 218.It Cm pfs-delete Ar label 219Delete a local PFS on a mounted HAMMER2 filesystem. 220Deleting a PFS of type MASTER requires first downgrading it to a SLAVE (XXX). 221.\" ==== snapshot ==== 222.It Cm snapshot Ar path Op label 223Create a snapshot of a directory. 224This can only be used on a local PFS, and is only really useful if the PFS 225contains a complete copy of what you desire to snapshot so that typically 226means a local MASTER, SOFT_MASTER, SLAVE, or SOFT_SLAVE must be present. 227Snapshots are created simply by flushing a PFS mount to disk and then copying 228the directory inode to the PFS. 229The topology is snapshotted without having to be copied or scanned. 230Snapshots are effectively separate from the cluster they came from 231and can be used as a starting point for a new cluster. 232So unless you build a new cluster from the snapshot, it will stay local 233to the machine it was made on. 234.\" ==== service ==== 235.It Cm service 236Start the 237.Nm 238service daemon. 239This daemon is also automatically started when you run 240.Xr mount_hammer2 8 . 241The hammer2 service daemon handles incoming TCP connections and maintains 242outgoing TCP connections. It will interconnect available services on the 243machine (e.g. hammer2 mounts and xdisks) to the network. 244.\" ==== stat ==== 245.It Cm stat Op path... 246Print the inode statistics, compression, and other meta-data associated 247with a list of paths. 248.\" ==== leaf ==== 249.It Cm leaf 250XXX 251.\" ==== shell ==== 252.It Cm shell 253Start a debug shell to the local hammer2 service daemon via the DMSG protocol. 254.\" ==== debugspan ==== 255.It Cm debugspan 256(do not use) 257.\" ==== rsainit ==== 258.It Cm rsainit 259Create the 260.Pa /etc/hammer2 261directory and initialize a public/private keypair in that directory for 262use by the network cluster protocols. 263.\" ==== show ==== 264.It Cm show Ar devpath 265Dump the radix tree for the HAMMER2 filesystem by scanning a 266block device directly. No mount is required. 267.\" ==== freemap ==== 268Dump the freemap tree for the HAMMER2 filesystem by scanning a 269block device directly. No mount is required. 270.It Cm freemap Ar devpath 271.\" ==== setcomp ==== 272.It Cm setcomp Ar mode[:level] Op path... 273Set the compression mode as specified for any newly created elements at or 274under the path if not overridden by deeper elements. 275Available modes are none, autozero, lz4, or zlib. 276When zlib is used the compression level can be set. 277The default will be 6 which is the best trade-off between performance and 278time. 279.Pp 280newfs_hammer2 will set the default compression to lz4 which prioritizes 281speed over performance. 282Also note that HAMMER2 contains a heuristic and will not attempt to 283compress every block if it detects a sufficient amount of uncompressable 284data. 285.Pp 286Hammer2 compression is only effective when it can reduce the size of dataset 287(typically a 64KB block) by one or more powers of 2. A 64K block which 288only compresses to 40K will not yield any storage improvement. 289.Pp 290Generally speaking you do not want to set the compression mode to 'none', 291as this will cause blocks of all-zeros to be written as all-zero blocks, 292instead of holes. The 'autozero' compression mode detects blocks of all-zeros 293and writes them as holes. However, HAMMER2 will rewrite data in-place if 294the compression mode is set to 'none' and the check code is set to 295'disabled'. Formal snapshots will still snapshot such files. However, 296de-duplication will no longer function on the data blocks. 297.\" ==== setcheck ==== 298.It Cm setcheck Ar check Op path... 299Set the check code as specified for any newly created elements at or under 300the path if not overridden by deeper elements. 301Available codes are default, disabled, crc32, xxhash64, or sha192. 302.\" ==== clrcheck ==== 303.It Cm clrcheck Op path... 304Clear the check code override for the specified paths. 305Overrides may still be present in deeper elements. 306.\" ==== setcrc32 ==== 307.It Cm setcrc32 Op path... 308Set the check code to the ISCSI 32-bit CRC for any newly created elements 309at or under the path if not overridden by deeper elements. 310.\" ==== setxxhash64 ==== 311.It Cm setxxhash64 Op path... 312Set the check code to XXHASH64, a fast 64-bit hash 313.\" ==== setsha192 ==== 314.It Cm setsha192 Op path... 315Set the check code to SHA192 for any newly created elements at or under 316the path if not overridden by deeper elements. 317.\" ==== bulkfree ==== 318.It Cm bulkfree Op path... 319Run a bulkfree pass on a HAMMER2 mount. 320You can specify any PFS for the mount, the bulkfree pass is run on the 321entire partition. 322Note that it takes two passes to actually free space. 323.El 324.Sh SYSCTLS 325.Bl -tag -width indent 326.It Va vfs.hammer2.dedup_enable (default on) 327Enables live de-duplication. Any recently read data that is on-media 328(already synchronized to media) is tested against pending writes for 329compatibility. If a match is found, the write will reference the 330existing on-media data instead of writing new data. 331.It Va vfs.hammer2.always_compress (default off) 332This disables the H2 compression heuristic and forces H2 to always 333try to compress data blocks, even if they look uncompressable. 334Enabling this option reduces performance but has higher de-duplication 335repeatability. 336.It Va vfs.hammer2.cluster_read (default 4) 337Set the amount of read-ahead clustering to perform. 338.It Va vfs.hammer2.cluster_write (default 0) 339Set the amount of write-behind clustering to perform. This is disabled by 340default in order to give temporary files a chance to be deleted before 341media writes are committed. Enabling this reduces buffer cache stress 342but causes file writes to flush to media more quickly. 343.El 344.Sh SETTING UP /etc/hammer2 345The 'rsainit' directive will create the 346.Pa /etc/hammer2 347directory with appropriate permissions and also generate a public key 348pair in this directory for the machine. These files will be 349.Pa rsa.pub 350and 351.Pa rsa.prv 352and needless to say, the private key shouldn't leave the host. 353.Pp 354The service daemon will also scan the 355.Pa /etc/hammer2/autoconn 356file which contains a list of hosts which it will automatically maintain 357connections to to form your cluster. 358The service daemon will automatically reconnect on any failure and will 359also monitor the file for changes. 360.Pp 361When the service daemon receives a connection it expects to find a 362public key for that connection in a file in 363.Pa /etc/hammer2/remote/ 364called 365.Pa <IPADDR>.pub . 366You normally copy the 367.Pa rsa.pub 368key from the host in question to this file. 369The IP address must match exactly or the connection will not be allowed. 370.Pp 371If you want to use an unencrypted connection you can create empty, 372dummy files in the remote directory in the form 373.Pa <IPADDR>.none . 374We do not recommend using unencrypted connections. 375.Sh CLUSTER SERVICES 376Currently there are two services which use the cluster network infrastructure, 377HAMMER2 mounts and XDISK. 378Any HAMMER2 mount will make all PFSs for that filesystem available to the 379cluster. 380And if the XDISK kernel module is loaded, the hammer2 service daemon will make 381your machine's block devices available to the cluster (you must load the 382xdisk.ko kernel module before starting the hammer2 service). 383They will show up as 384.Pa /dev/xa* 385and 386.Pa /dev/serno/* 387devices on the remote machines making up the cluster. 388Remote block devices are just what they appear to be... direct access to a 389block device on a remote machine. If the link goes down remote accesses 390will stall until it comes back up again, then automatically requeue any 391pending I/O and resume as if nothing happened. 392However, if the server hosting the physical disks crashes or is rebooted, 393any remote opens to its devices will see a permanent I/O failure requiring a 394close and open sequence to re-establish. 395The latter is necessary because the server's drives might not have committed 396the data before the crash, but had already acknowledged the transfer. 397.Pp 398Data commits work exactly the same as they do for real block devices. 399The originater must issue a BUF_CMD_FLUSH. 400.Sh ADDING A NEW MASTER TO A CLUSTER 401When you 402.Xr newfs_hammer2 8 403a HAMMER2 filesystem or use the 'pfs-create' directive on one already mounted 404to create a new PFS, with no special options, you wind up with a PFS 405typed as a MASTER and a unique cluster uuid, but because there is only one 406PFS for that cluster (for each PFS you create via pfs-create), it will 407act just like a normal filesystem would act and does not require any special 408protocols to operate. 409.Pp 410If you use the 'pfs-create' directive along with the 411.Fl u 412option to specify a cluster uuid that already exists in the cluster, 413you are adding a PFS to an existing cluster and this can trigger a whole 414series of events in the background. 415When you specify the 416.Fl u 417option in a 'pfs-create', 418.Nm 419will by default create a SLAVE PFS. 420In fact, this is what must be created first even if you want to add a new 421MASTER to your cluster. 422.Pp 423The most common action a system admin will want to take is to upgrade or 424downgrade a PFS. 425A new MASTER can be added to the cluster by upgrading an existing SLAVE 426to MASTER. 427A MASTER can be removed from the cluster by downgrading it to a SLAVE. 428Upgrades and downgrades will put nodes in the cluster in a transition state 429until the operation is complete. 430For downgrades the transition state is fleeting unless one or more other 431masters has not acknowledged the change. 432For upgrades a background synchronization process must complete before the 433transition can be said to be complete, and the node remains (really) a SLAVE 434until that transition is complete. 435.Sh USE CASES FOR A SOFT_MASTER 436The SOFT_MASTER PFS type is a special type which must be specifically 437mounted by a machine. 438It is a R/W mount which does not use the quorum protocol and is not 439cache coherent with the cluster, but which synchronizes from the cluster 440and allows modifying operations which will synchronize to the cluster. 441The most common case is to use a SOFT_MASTER PFS in a laptop allowing you 442to work on your laptop when you are on the road and not connected to 443your main servers, and for the laptop to synchronize when a connection is 444available. 445.Sh USE CASES FOR A SOFT_SLAVE 446A SOFT_SLAVE PFS type is a special type which must be specifically mounted 447by a machine. 448It is a RO mount which does not use the quorum protocol and is not 449cache coherent with the cluster. It will receive synchronization from 450the cluster when network connectivity is available but will not stall if 451network connectivity is lost. 452.Sh FSYNC FLUSH MODES 453TODO. 454.Sh RESTORING FROM A SNAPSHOT BACKUP 455TODO. 456.Sh PERFORMANCE TUNING 457Because HAMMER2 implements compression, decompression, and deup natively, 458it always double-buffers file data. This means that the file data is 459cached via the device vnode (in compressed / dedupped-form) and the same 460data is also cached by the file vnode (in decompressed / non-dedupped form). 461.Pp 462While HAMMER2 will try to age the logical file buffers on its, some 463additional performance tuning may be necessary for optimal operation 464whether swapcache is used or not. Our recommendation is to reduce the 465number of vnodes (and thus also the logical buffer cache behind the 466vnodes) that the system caches via the 467.Va kern.maxvnodes 468sysctl. 469.Pp 470Too-large a value will result in excessive double-caching and can cause 471unnecessary read disk I/O. 472We recommend a number between 25000 and 250000 vnodes, depending on your 473use case. 474Keep in mind that even though the vnode cache is smaller, this will make 475room for a great deal more device-level buffer caching which can encompasses 476far more data and meta-data than the vnode-level caching. 477.Sh ENVIRONMENT 478TODO. 479.Sh FILES 480.Bl -tag -width ".It Pa <fs>/abc/defghi/<name>" -compact 481.It Pa /etc/hammer2/ 482.It Pa /etc/hammer2/rsa.pub 483.It Pa /etc/hammer2/rsa.prv 484.It Pa /etc/hammer2/autoconn 485.It Pa /etc/hammer2/remote/<IP>.pub 486.It Pa /etc/hammer2/remote/<IP>.none 487.El 488.Sh EXIT STATUS 489.Ex -std 490.Sh SEE ALSO 491.Xr mount_hammer2 8 , 492.Xr mount_null 8 , 493.Xr newfs_hammer2 8 , 494.Xr swapcache 8 , 495.Xr sysctl 8 496.Sh HISTORY 497The 498.Nm 499utility first appeared in 500.Dx 4.1 . 501.Sh AUTHORS 502.An Matthew Dillon Aq Mt dillon@backplane.com 503