1 .\" Copyright (c) 2015-2019 The DragonFly Project. All rights reserved.
3 .\" This code is derived from software contributed to The DragonFly Project
4 .\" by Matthew Dillon <dillon@backplane.com>
6 .\" Redistribution and use in source and binary forms, with or without
7 .\" modification, are permitted provided that the following conditions
10 .\" 1. Redistributions of source code must retain the above copyright
11 .\" notice, this list of conditions and the following disclaimer.
12 .\" 2. Redistributions in binary form must reproduce the above copyright
13 .\" notice, this list of conditions and the following disclaimer in
14 .\" the documentation and/or other materials provided with the
16 .\" 3. Neither the name of The DragonFly Project nor the names of its
17 .\" contributors may be used to endorse or promote products derived
18 .\" from this software without specific, prior written permission.
20 .\" THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
21 .\" ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
22 .\" LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
23 .\" FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
24 .\" COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
25 .\" INCIDENTAL, SPECIAL, EXEMPLARY OR CONSEQUENTIAL DAMAGES (INCLUDING,
26 .\" BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
27 .\" LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
28 .\" AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
29 .\" OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
30 .\" OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
38 .Nd hammer2 file system utility
52 utility provides miscellaneous support functions for a
55 The options are as follows:
56 .Bl -tag -width indent
58 Specify the path to a mounted HAMMER2 filesystem.
59 At least one PFS on a HAMMER2 filesystem must be mounted for the system
60 to act on all PFSs managed by it.
61 Every HAMMER2 filesystem typically has a PFS called "LOCAL" for this purpose.
63 Specify the type when creating, upgrading, or downgrading a PFS.
64 Supported types are MASTER, SLAVE, SOFT_MASTER, SOFT_SLAVE, CACHE, and DUMMY.
65 If not specified the pfs-create directive will default to MASTER if no
66 UUID is specified, and SLAVE if a UUID is specified.
68 Specify the cluster UUID when creating a PFS.
69 If not specified, a unique, random UUID will be generated.
70 Note that every PFS also has a unique pfs_id which is always generated
71 and cannot be overridden with an option.
72 The { pfs_clid, pfs_fsid } tuple uniquely identifies a component of a cluster.
74 Specify how much tracking memory to use for certain directives.
75 At the moment, this option is only applicable to the
77 directive, allowing it to operate in fewer passes when given more memory.
78 A nominal value for a 4TB drive with a ton of stuff on it would be around
83 directives are as shown below.
84 Note that most directives require you to either be CD'd into a hammer2
85 filesystem, specify a path to a mounted hammer2 filesystem via the
87 option, or specify a path after the directive.
88 It depends on the directive.
89 All hammer2 filesystem have a PFS called "LOCAL" which is typically mounted
90 locally on the host in order to be able to issue commands for other PFSs
92 The mount also enables PFS configuration scanning for that filesystem.
93 .Bl -tag -width indent
95 .It Cm cleanup Op path
96 Perform manual cleanup passes on paths or all mounted partitions.
98 .It Cm connect Ar target
99 Add a cluster link entry to the volume header.
100 The volume header can support up to 255 link entries.
101 This feature is not currently used.
102 .\" ==== destroy ====
103 .It Cm destroy Ar path...
104 Destroy the specified directory entry in a hammer2 filesystem.
106 all normal checks and will unconditionally destroy the directory entry.
107 The underlying inode is not checked and, if it does exist, its nlinks count
109 This directive should only be used to destroy a corrupted directory entry
110 which no longer has a working inode.
112 Note that this command may desynchronize the system namecache for the
114 If this happens, you may have to unmount and remount the filesystem.
115 .\" ==== destroy-inum ====
116 .It Cm destroy-inum Ar path...
117 Destroy the specified inode in a hammer2 filesystem.
118 .\" ==== disconnect ====
119 .It Cm disconnect Ar target
120 Delete a cluster link entry from the volume header.
121 This feature is not currently used.
122 .\" ==== emergency-mode-enable ===
123 .It Cm emergency-mode-enable Ar target
124 Flag emergency operations mode in the filesystem.
125 This mode may be used
126 as a last resort to delete files and directories from a full filesystem.
127 Inode creation, file writes, and certain meta-data cleanups are disallowed
128 while emergency mode is active.
129 File and directory removal and mode/attr setting is still allowed.
130 This mode is extremely dangerous and should only be used as a last resort.
132 This mode allows the filesystem to modify blocks in-place when it is unable
134 Thus it is possible to chflags and remove files and
135 directories even when the filesystem is completely full.
136 However, there is a price.
137 This mode of operation WILL LIKELY CORRUPT ANY SNAPSHOTS related
139 The filesystem will report this condition if it encounters
140 it but if you are forced to use this mode to fix a filesystem full condition
141 your snapshots can get a bit dicey.
142 It is usually safest to delete any related snapshots when using this mode.
144 You can detect whether related snapshots have been corrupted by running
145 a bulkfree pass and checking the console output for reported CRC errors.
146 If no errors are reported, your snapshots are fine.
147 If errors are reported
148 you should delete related snapshots until bulkfree reports no further errors.
150 The emergency mode will also make meta-data updates unsafe due to the lack of
151 copy-on-write, causing potential harm if the system unexpectedly panics or
153 GREAT CARE MUST BE TAKEN WHILE THIS MODE IS ACTIVE.
156 Determine that you are unable to recover space with normal file and directory
157 removal commands due to
159 errors being returned by 'rm', or through the
160 removal of snapshots (if any). The 'bulkfree' directive must be issued to
161 scan the filesystem and free up the actual space, then check with 'df'.
162 Continue if you still have insufficient space and are unable to remove items
165 If you need any related snapshots, this is a good time to copy them elsewhere.
167 Idle or kill any processes trying to use the filesystem.
169 Issue the emergency-mode-enable directive on the filesystem.
170 Once enabled, run 'sync' to update any dirty inodes which may still
171 be dirty due to not being able to flush.
172 Please remember that this
173 directive is a LAST RESORT, is dangerous, and will likely corrupt any
174 other snapshots you have based on the filesystem you are removing files
177 Remove file trees as necessary with 'rm -rf' to free space, being cognizant
178 of any warnings issued by the kernel on the console (via 'dmesg') while
181 Issue the 'bulkfree' directive to actually free the space and check that
182 sufficient space has been freed with 'df'.
184 If bulkfree reports CHECK errors, or if you have snapshots and insufficient
185 space has been freed, you will need to delete snapshots.
186 Re-run bulkfree and delete snapshots until no errors are reported.
188 Issue the emergency-mode-disable directive when done.
190 good idea to reboot after using this mode, but theoretically you should not
193 Restore services using the filesystem.
195 .\" ==== emergency-mode-disable ===
196 .It Cm emergency-mode-disable Ar target
197 Turn off the emergency operations mode on a filesystem, restoring normal
200 .It Cm info Op devpath...
201 Access and print the status and super-root entries for all HAMMER2
202 partitions found in /dev/serno or the specified device path(s).
203 The partitions do not have to be mounted.
204 Note that only mounted partitions will be under active management.
205 This is accomplished by mounting at least one PFS within the partition.
206 Typically at least the @LOCAL PFS is mounted.
207 .\" ==== mountall ====
208 .It Cm mountall Op devpath...
209 This directive mounts the @LOCAL PFS on all HAMMER2 partitions found
210 in /dev/serno, or the specified device path(s).
211 The partitions are mounted as /var/hammer2/LOCAL.<id>.
212 Mounts are executed in the background and this command will wait a
213 limited amount of time for the mounts to complete before returning.
215 .It Cm status Op path...
216 Dump a list of all cluster link entries configured in the volume header.
218 .It Cm hash Op filename...
219 Compute and print the directory hash for any number of filenames.
221 .It Cm dhash Op filename...
222 Compute and print the data hash for long directory entry for any number of filenames.
223 .\" ==== pfs-list ====
224 .It Cm pfs-list Op path...
225 List all PFSs associated with all mounted hammer2 storage devices.
226 The list may be restricted to a particular filesystem using
229 Note that hammer2 PFSs associated with storage devices which have not been
230 mounted in any fashion will not be listed. At least one hammer2 label must
231 be mounted for the PFSs on that device to be visible.
232 .\" ==== pfs-clid ====
233 .It Cm pfs-clid Ar label
234 Print the cluster id for a PFS specified by name.
235 .\" ==== pfs-fsid ====
236 .It Cm pfs-fsid Ar label
237 Print the unique filesystem id for a PFS specified by name.
238 .\" ==== pfs-create ====
239 .It Cm pfs-create Ar label
240 Create a local PFS on the mounted HAMMER2 filesystem represented
241 by the current directory, or specified via
243 If no UUID is specified the pfs-type defaults to MASTER.
244 If a UUID is specified via the
246 option the pfs-type defaults to SLAVE.
247 Other types can be specified with the
251 If you wish to add a MASTER to an existing cluster, you must first add it as
252 a SLAVE and then upgrade it to MASTER to properly synchronize it.
254 The DUMMY pfs-type is used to tie network-accessible clusters into the local
255 machine when no local storage is desired.
256 This type should be used on minimal H2 partitions or entirely in ram for
257 netboot-centric systems to provide a tie-in point for the mount command,
258 or on more complex systems where you need to also access network-centric
261 The CACHE or SLAVE pfs-type is typically used when the main store is on
262 the network but local storage is desired to improve performance.
263 SLAVE is also used when a backup is desired.
265 Generally speaking, you can mount any PFS element of a cluster in order to
266 access the cluster via the full cluster protocol.
267 There are two exceptions.
268 If you mount a SOFT_SLAVE or a SOFT_MASTER then soft quorum semantics are
269 employed... the soft slave or soft master's current state will always be used
270 and the quorum protocol will not be used.
271 The soft PFS will still be
272 synchronized to masters in the background when available.
275 to mount ONLY a local HAMMER2 PFS and
276 not run any network or quorum protocols for the mount.
277 All such mounts except for a SOFT_MASTER mount will be read-only.
278 Other than that, you will be mounting the whole cluster when you mount any
279 PFS within the cluster.
281 DUMMY - Create a PFS skeleton intended to be the mount point for a
282 more complex cluster, probably one that is entirely network based.
283 No data will be synchronized to this PFS so it is suitable for use
284 in a network boot image or memory filesystem.
285 This allows you to create placeholders for mount points on your local
286 disk, SSD, or memory disk.
288 CACHE - Create a PFS for caching portions of the cluster piecemeal.
289 This is similar to a SLAVE but does not synchronize the entire contents of
290 the cluster to the PFS.
291 Elements found in the CACHE PFS which are validated against the cluster
292 will be read, presumably a faster access than having to go to the cluster.
293 Only local CACHEs will be updated.
294 Network-accessible CACHE PFSs might be read but will not be written to.
295 If you have a large hard-drive-based cluster you can set up localized
296 SSD CACHE PFSs to improve performance.
298 SLAVE - Create a PFS which maintains synchronization with and provides a
299 read-only copy of the cluster.
300 HAMMER2 will prioritize local SLAVEs for data retrieval after validating
301 their transaction id against the cluster.
302 The difference between a CACHE and a SLAVE is that the SLAVE is synchronized
303 to a full copy of the cluster and thus can serve as a backup or be staged
304 for use as a MASTER later on.
306 SOFT_SLAVE - Create a PFS which maintains synchronization with and provides
307 a read-only copy of the cluster.
308 This is one of the special mount cases.
309 A SOFT_SLAVE will synchronize with
310 the cluster when the cluster is available, but can still be accessed when
311 the cluster is not available.
313 MASTER - Create a PFS which will hold a master copy of the cluster.
314 If you create several MASTER PFSs with the same cluster id you are
315 effectively creating a multi-master cluster and causing a quorum and
316 cache coherency protocol to be used to validate operations.
317 The total number of masters is stored in each PFSs making up the cluster.
318 Filesystem operations will stall for normal mounts if a quorum cannot be
319 obtained to validate the operation.
320 MASTER nodes which go offline and return later will synchronize in the
322 Note that when adding a MASTER to an existing cluster you must add the
323 new PFS as a SLAVE and then upgrade it to a MASTER.
325 SOFT_MASTER - Create a PFS which maintains synchronization with and provides
326 a read-write copy of the cluster.
327 This is one of the special mount cases.
328 A SOFT_MASTER will synchronize with
329 the cluster when the cluster is available, but can still be read AND written
330 to even when the cluster is not available.
331 Modifications made to a SOFT_MASTER will be automatically flushed to the
332 cluster when it becomes accessible again, and vise-versa.
333 Manual intervention may be required if a conflict occurs during
335 .\" ==== pfs-delete ====
336 .It Cm pfs-delete Op label...
337 Destroy a PFS by name. All hammer2 mount points will be checked, however
338 this directive will refuse to delete a PFS whos name is duplicated on
339 multiple mount points. A specific mount point may be specified to restrict
343 .\" ==== snapshot ====
344 .It Cm snapshot Ar path Op label
345 Create a snapshot of a directory.
346 The snapshot will be created on the same hammer2 storage device as the
348 This can only be used on a local PFS, and is only really useful if the PFS
349 contains a complete copy of what you desire to snapshot so that typically
350 means a local MASTER, SOFT_MASTER, SLAVE, or SOFT_SLAVE must be present.
351 Snapshots are created simply by flushing a PFS mount to disk and then copying
352 the directory inode to the PFS.
353 The topology is snapshotted without having to be copied or scanned and
354 take no additional space.
355 However, bulkfree scans may take longer.
356 Snapshots are effectively separate from the cluster they came from
357 and can be used as a starting point for a new cluster.
358 So unless you build a new cluster from the snapshot, it will stay local
359 to the machine it was made on.
361 Snapshots can be maintained automatically with
365 for details of enabling and configuring the functionality.
366 .\" ==== snapshot-debug ====
367 .It Cm snapshot-debug Ar path Op label
368 Snapshot without filesystem sync.
369 .\" ==== service ====
374 This daemon is also automatically started when you run
375 .Xr mount_hammer2 8 .
376 The hammer2 service daemon handles incoming TCP connections and maintains
377 outgoing TCP connections.
378 It will interconnect available services on the
379 machine (e.g. hammer2 mounts and xdisks) to the network.
381 .It Cm stat Op path...
382 Print the inode statistics, compression, and other meta-data associated
383 with a list of paths.
389 Start a debug shell to the local hammer2 service daemon via the DMSG protocol.
390 .\" ==== debugspan ====
391 .It Cm debugspan Ar target
393 .\" ==== rsainit ====
394 .It Cm rsainit Op path
397 directory and initialize a public/private keypair in that directory for
398 use by the network cluster protocols.
400 .It Cm show Ar devpath
401 Dump the radix tree for the HAMMER2 filesystem by scanning a
402 block device directly.
403 No mount is required.
404 .\" ==== freemap ====
405 .It Cm freemap Ar devpath
406 Dump the freemap tree for the HAMMER2 filesystem by scanning a
407 block device directly.
408 No mount is required.
410 .It Cm volhdr Ar devpath
411 Dump the volume header for the HAMMER2 filesystem by scanning a
412 block device directly.
413 No mount is required.
414 .\" ==== setcomp ====
415 .It Cm setcomp Ar mode[:level] Ar path...
416 Set the compression mode as specified for any newly created elements at or
417 under the path if not overridden by deeper elements.
418 Available modes are none, autozero, lz4, or zlib.
419 When zlib is used the compression level can be set.
420 The default will be 6 which is the best trade-off between performance and
423 newfs_hammer2 will set the default compression to lz4 which prioritizes
424 speed over performance.
425 Also note that HAMMER2 contains a heuristic and will not attempt to
426 compress every block if it detects a sufficient amount of uncompressable
429 Hammer2 compression is only effective when it can reduce the size of dataset
430 (typically a 64KB block) by one or more powers of 2. A 64K block which
431 only compresses to 40K will not yield any storage improvement.
433 Generally speaking you do not want to set the compression mode to
435 as this will cause blocks of all-zeros to be written as all-zero blocks,
439 compression mode detects blocks of all-zeros
440 and writes them as holes.
441 .\" ==== setcheck ====
442 .It Cm setcheck Ar check Ar path...
443 Set the check code as specified for any newly created elements at or under
444 the path if not overridden by deeper elements.
445 Available codes are default, disabled, crc32, xxhash64, or sha192.
447 Normally HAMMER2 does not overwrite data blocks on the media in order
448 to ensure snapshot integrity. Replacement data blocks will be reallocated.
449 However, if the compression mode is set to
451 and the check code is set to
453 HAMMER2 will overwrite data on the media in-place.
454 In this mode of operation,
455 snapshots will not be able to snapshot the data against later changes
456 made to the file, and de-duplication will no longer function on any
457 data related to the file.
458 However, you can still recover the most recent data from previously
459 taken snapshots if you accidentally remove the file.
460 .\" ==== clrcheck ====
461 .It Cm clrcheck Op path...
462 Clear the check code override for the specified paths.
463 Overrides may still be present in deeper elements.
464 .\" ==== setcrc32 ====
465 .It Cm setcrc32 Op path...
466 Set the check code to the ISCSI 32-bit CRC for any newly created elements
467 at or under the path if not overridden by deeper elements.
468 .\" ==== setxxhash64 ====
469 .It Cm setxxhash64 Op path...
470 Set the check code to XXHASH64, a fast 64-bit hash
471 .\" ==== setsha192 ====
472 .It Cm setsha192 Op path...
473 Set the check code to SHA192 for any newly created elements at or under
474 the path if not overridden by deeper elements.
475 .\" ==== bulkfree ====
476 .It Cm bulkfree Ar path
477 Run a bulkfree pass on a HAMMER2 mount.
478 You can specify any PFS for the mount, the bulkfree pass is run on the
480 Note that it takes two passes to actually free space.
481 By default this directive will use up to 1/16 physical memory to track
483 The amount of memory used may be overridden with the
486 .\" ==== printinode ====
487 .It Cm printinode Ar path
489 .\" ==== dumpchain ====
490 .It Cm dumpchain Op path Op chnflags
491 Dump in-memory chain topology.
494 .Bl -tag -width indent
495 .It Va vfs.hammer2.dedup_enable (default on)
496 Enables live de-duplication.
497 Any recently read data that is on-media
498 (already synchronized to media) is tested against pending writes for
500 If a match is found, the write will reference the
501 existing on-media data instead of writing new data.
502 .It Va vfs.hammer2.always_compress (default off)
503 This disables the H2 compression heuristic and forces H2 to always
504 try to compress data blocks, even if they look uncompressable.
505 Enabling this option reduces performance but has higher de-duplication
507 .It Va vfs.hammer2.cluster_data_read (default 4)
508 .It Va vfs.hammer2.cluster_meta_read (default 1)
509 Set the amount of read-ahead clustering to perform on data and meta-data
511 .It Va vfs.hammer2.cluster_write (default 4)
512 Set the amount of write-behind clustering to perform in buffers.
513 Each buffer represents 64KB.
514 The default is 4 and higher values typically do not improve performance.
515 A value of 0 disables clustered writes.
516 This variable applies to the underlying media device, not to logical
517 file writes, so it should not interfere with temporary file optimization.
518 Generally speaking you want this enabled to generate smoothly pipelined
520 .It Va vfs.hammer2.bulkfree_tps (default 5000)
521 Set bulkfree's maximum scan rate.
522 This is primarily intended to limit
523 I/O utilization on SSDs and CPU utilization when the meta-data is mostly
526 .Sh SETTING UP /etc/hammer2
529 directive will create the
531 directory with appropriate permissions and also generate a public key
532 pair in this directory for the machine.
537 and needless to say, the private key shouldn't leave the host.
539 The service daemon will also scan the
540 .Pa /etc/hammer2/autoconn
541 file which contains a list of hosts which it will automatically maintain
542 connections to to form your cluster.
543 The service daemon will automatically reconnect on any failure and will
544 also monitor the file for changes.
546 When the service daemon receives a connection it expects to find a
547 public key for that connection in a file in
548 .Pa /etc/hammer2/remote/
551 You normally copy the
553 key from the host in question to this file.
554 The IP address must match exactly or the connection will not be allowed.
556 If you want to use an unencrypted connection you can create empty,
557 dummy files in the remote directory in the form
559 We do not recommend using unencrypted connections.
561 Currently there are two services which use the cluster network infrastructure,
562 HAMMER2 mounts and XDISK.
563 Any HAMMER2 mount will make all PFSs for that filesystem available to the
565 And if the XDISK kernel module is loaded, the hammer2 service daemon will make
566 your machine's block devices available to the cluster (you must load the
567 xdisk.ko kernel module before starting the hammer2 service).
572 devices on the remote machines making up the cluster.
573 Remote block devices are just what they appear to be... direct access to a
574 block device on a remote machine.
575 If the link goes down remote accesses
576 will stall until it comes back up again, then automatically requeue any
577 pending I/O and resume as if nothing happened.
578 However, if the server hosting the physical disks crashes or is rebooted,
579 any remote opens to its devices will see a permanent I/O failure requiring a
580 close and open sequence to re-establish.
581 The latter is necessary because the server's drives might not have committed
582 the data before the crash, but had already acknowledged the transfer.
584 Data commits work exactly the same as they do for real block devices.
585 The originater must issue a BUF_CMD_FLUSH.
586 .Sh ADDING A NEW MASTER TO A CLUSTER
589 a HAMMER2 filesystem or use the
591 directive on one already mounted
592 to create a new PFS, with no special options, you wind up with a PFS
593 typed as a MASTER and a unique cluster UUID, but because there is only one
594 PFS for that cluster (for each PFS you create via pfs-create), it will
595 act just like a normal filesystem would act and does not require any special
596 protocols to operate.
600 directive along with the
602 option to specify a cluster UUID that already exists in the cluster,
603 you are adding a PFS to an existing cluster and this can trigger a whole
604 series of events in the background.
610 will by default create a SLAVE PFS.
611 In fact, this is what must be created first even if you want to add a new
612 MASTER to your cluster.
614 The most common action a system admin will want to take is to upgrade or
616 A new MASTER can be added to the cluster by upgrading an existing SLAVE
618 A MASTER can be removed from the cluster by downgrading it to a SLAVE.
619 Upgrades and downgrades will put nodes in the cluster in a transition state
620 until the operation is complete.
621 For downgrades the transition state is fleeting unless one or more other
622 masters has not acknowledged the change.
623 For upgrades a background synchronization process must complete before the
624 transition can be said to be complete, and the node remains (really) a SLAVE
625 until that transition is complete.
626 .Sh USE CASES FOR A SOFT_MASTER
627 The SOFT_MASTER PFS type is a special type which must be specifically
628 mounted by a machine.
629 It is a R/W mount which does not use the quorum protocol and is not
630 cache coherent with the cluster, but which synchronizes from the cluster
631 and allows modifying operations which will synchronize to the cluster.
632 The most common case is to use a SOFT_MASTER PFS in a laptop allowing you
633 to work on your laptop when you are on the road and not connected to
634 your main servers, and for the laptop to synchronize when a connection is
636 .Sh USE CASES FOR A SOFT_SLAVE
637 A SOFT_SLAVE PFS type is a special type which must be specifically mounted
639 It is a RO mount which does not use the quorum protocol and is not
640 cache coherent with the cluster.
641 It will receive synchronization from
642 the cluster when network connectivity is available but will not stall if
643 network connectivity is lost.
644 .Sh FSYNC FLUSH MODES
646 .Sh RESTORING FROM A SNAPSHOT BACKUP
648 .Sh PERFORMANCE TUNING
649 Because HAMMER2 implements compression, decompression, and dedup natively,
650 it always double-buffers file data.
651 This means that the file data is
652 cached via the device vnode (in compressed / dedupped-form) and the same
653 data is also cached by the file vnode (in decompressed / non-dedupped form).
655 While HAMMER2 will try to age the logical file buffers on its, some
656 additional performance tuning may be necessary for optimal operation
657 whether swapcache is used or not.
658 Our recommendation is to reduce the
659 number of vnodes (and thus also the logical buffer cache behind the
660 vnodes) that the system caches via the
664 Too-large a value will result in excessive double-caching and can cause
665 unnecessary read disk I/O.
666 We recommend a number between 25000 and 250000 vnodes, depending on your
668 Keep in mind that even though the vnode cache is smaller, this will make
669 room for a great deal more device-level buffer caching which can encompasses
670 far more data and meta-data than the vnode-level caching.
674 .Bl -tag -width ".It Pa <fs>/abc/defghi/<name>" -compact
676 .It Pa /etc/hammer2/rsa.pub
677 .It Pa /etc/hammer2/rsa.prv
678 .It Pa /etc/hammer2/autoconn
679 .It Pa /etc/hammer2/remote/<IP>.pub
680 .It Pa /etc/hammer2/remote/<IP>.none
685 .Xr periodic.conf 5 ,
686 .Xr mount_hammer2 8 ,
688 .Xr newfs_hammer2 8 ,
694 utility first appeared in
697 .An Matthew Dillon Aq Mt dillon@backplane.com