Previously, ZFS scaled maxinflight_bytes based on total number of
disks in the pool. A 3-wide mirror was receiving a queue depth of 3
disks, which it should not, since it reads from all the disks inside.
For wide raidz the situation was slightly better, but still a 3-wide
raidz1 received a depth of 3 disks instead of 2.
The new code counts only unique data disks, i.e. 1 disk for mirrors
and non-parity disks for raidz/draid. For draid the math is still
imperfect, since vdev_get_nparity() returns number of parity disks
per group, not per vdev, but still some better than it was.
This should slightly reduce scrub influence on payload for some pool
topologies by avoiding excessive queuing.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Ryan Moeller <ryan@iXsystems.com>
Signed-off-by: Alexander Motin <mav@FreeBSD.org>
Sponsored-By: iXsystems, Inc.
Closing #12046
A plain rewrite of the shell version, and generates identical
units, save for replacing some empty lines with nothing, having fewer
meaningless spaces in After=s and different spacing in the lock scripts,
for a clean git diff -w
This is a gain of anywhere from 0m0.336s vs 0m0.022s (15.27x)
to 0m0.202s vs 0m0.006s (33.67x), depending on the hardware,
a.k.a. from "absolutely unusable" to "perfectly fine"
This also properly deals with canmount=noauto units across multiple
pools
See PR for detailed timings (of an early version) and diffs
Reviewed-by: Antonio Russo <aerusso@aerusso.net>
Reviewed-by: Richard Laager <rlaager@wiktel.com>
Reviewed-by: InsanePrawn <insane.prawny@gmail.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Issue #11915Closes#11917
No changes to the text itself
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Closes#12111
Linux man-pages' mount(8) points at fcntl(2), as does mount(2),
and support for it is little-used, deprecated, and configurable
since 4.5.
As far as I can tell, FreeBSD doesn't support nbmand at all ‒
mandatory locks are mostly dead
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Closes#12111
zstreamdump(8) was in quite a bad state,
and the wrapper didn't work if invoked without /sbin in $PATH
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Closes#12015
Reviewed-by: John Kennedy <john.kennedy@delphix.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Lauri Tirkkonen <lauri@hacktheplanet.fi>
Closes#12064
While use of dynamic taskqs allows to reduce number of idle threads,
hardcoded 8 taskqs of each kind is a big overkill for small systems,
complicating CPU scheduling, increasing I/O reorder, etc, while
providing no real locking benefits, just not needed there.
On another side, 12*8 worker threads per kind are able to overload
almost any system nowadays. For example, pool of several fast SSDs
with SHA256 checksum makes system barely responsive during scrub, or
with dedup enabled barely responsive during large file deletion.
To address both problems this patch introduces ZTI_SCALE macro, alike
to ZTI_BATCH, but with multiple taskqs, depending on number of CPUs,
to be used in places where lock scalability is needed, while request
ordering is not so much. The code is made to create new taskq for
~6 worker threads (less for small systems, but more for very large)
up to 80% of CPU cores (previous 75% was not good for rounding down).
Both number of threads and threads per taskq are now tunable in case
somebody really wants to use all of system power for ZFS.
While obviously some benchmarks show small peak performance reduction
(not so big really, especially on systems with SMT, where use of the
second threads does not give as much performance as the first ones),
they also show dramatic latency reduction and much more smooth user-
space operation in case of high CPU usage by ZFS.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Alexander Motin <mav@FreeBSD.org>
Sponsored-By: iXsystems, Inc.
Closes#11966
FreeBSD historically has not cared about the xattr property; it was
always treated as xattr=on. With xattr=on, xattrs are stored as files
in a hidden xattr directory. With xattr=sa, xattrs are stored as
system attributes and get cached in nvlists during xattr operations.
This makes SA xattrs simpler and more efficient to manipulate. FreeBSD
needs to implement the SA xattr operations for feature parity with
Linux and to ensure that SA xattrs are accessible when migrated or
replicated from Linux.
Following the example set by Linux, refactor our existing extattr vnops
to split off the parts handling dir style xattrs, and add the
corresponding SA handling parts.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Alexander Motin <mav@FreeBSD.org>
Signed-off-by: Ryan Moeller <ryan@iXsystems.com>
Closes#11997
mandoc: ./man/man8/zfs-mount-generator.8.in:188:2:
ERROR: skipping end of block that is not open: RE
mandoc: ./man/man8/zfs_ids_to_path.8:38:2:
ERROR: skipping unknown macro: .LP
mandoc: ./man/man8/zfs_ids_to_path.8:48:2:
ERROR: inserting missing end of block: Sh breaks Bl
mandoc: ./man/man8/zfs-wait.8:69:2:
ERROR: skipping end of block that is not open: El
mandoc: ./man/man8/zfs-program.8:460:2:
ERROR: inserting missing end of block: It breaks Bd
mandoc: ./man/man8/zfs-mount-generator.8:188:2:
ERROR: skipping end of block that is not open: RE
mandoc: ./man/man8/zstream.8:43:2:
ERROR: skipping unknown macro: .LP
mandoc: ./man/man8/zstream.8:107:2:
ERROR: inserting missing end of block: Sh breaks Bl
mandoc: ./man/man8/zstream.8:107:2:
ERROR: inserting missing end of block: Sh breaks Bl
make: *** [Makefile:1529: mancheck] Error 1
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Issue #12017
Add support for http and https to the keylocation properly to
allow encryption keys to be fetched from the specified URL.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Ryan Moeller <ryan@ixsystems.com>
Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Issue #9543Closes#9947Closes#11956
Reviewed-by: Adam Moss <c@yotes.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Igor Kozhukhov <igor@dilos.org>
Signed-off-by: Ryan Moeller <ryan@iXsystems.com>
Closes#11972
zfs_zevent_console committed multiple printk()s per line without
properly continuing them ‒ a single event could easily be fragmented
across over thirty lines, making it useless for direct application
zfs_zevent_cols exists purely to wrap the output from zfs_zevent_console
The niche this was supposed to fill can be better served by something
akin to the all-syslog ZEDLET
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Closes#7082Closes#11996
Afterward, git grep ZoL matches:
* README.md: * [ZoL Site](https://zfsonlinux.org)
- Correct
* etc/default/zfs.in:# ZoL userland configuration.
- Changing this would induce a needless upgrade-check,
if the user has modified the configuration;
this can be updated the next time the defaults change
* module/zfs/dmu_send.c: * ZoL < 0.7 does not handle [...]
- Before 0.7 is ZoL, so fair enough
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Issue #11956
Changed the default specified for zfs_dbgmsg_enable, added
clarification of interaction with zfs_flags.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Ryan Moeller <ryan@ixsystems.com>
Signed-off-by: Rich Ercolani <rincebrain@gmail.com>
Closes#11984Closes#11986
The special_small_blocks section directed readers to zpool(8) for
documentation on special allocation classes, while they are actually
documented in zpoolconcepts(8).
Reviewed-by: George Melikov <mail@gmelikov.ru>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Daniel Stevenson <daniel@dstev.net>
Closes#11918
Before #11710 the flags in zfs-send(8) were sorted.
Restore order and bump the date.
Reviewed-by: George Melikov <mail@gmelikov.ru>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Ryan Moeller <ryan@iXsystems.com>
Closes#11905
Traversal code, traverse_visitbp() does visit blocks recursively.
Indirect (Non L0) Block of size 128k could contain, 1024 block pointers
of 128 bytes. In case of full traverse OR incremental traverse, where
all blocks were modified, it could traverse large number of blocks
pointed by indirect. Traversal code does issue prefetch of blocks
traversed below indirect. This could result into large number of
async reads queued on vdev queue. So, account for prefetch issued for
blocks pointed by indirect and limit max prefetch in one go.
Module Param:
zfs_traverse_indirect_prefetch_limit: Limit of prefetch while traversing
an indirect block.
Local counters:
prefetched: Local counter to account for number prefetch done.
pidx: Index for which next prefetch to be issued.
ptidx: Index at which next prefetch to be triggered.
Keep "ptidx" somewhere in the middle of blocks prefetched, so that
blocks prefetch read gets the enough time window before their demand
read is issued.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Alexander Motin <mav@FreeBSD.org>
Signed-off-by: Jitendra Patidar <jitendra.patidar@nutanix.com>
Closes#11802Closes#11803
Several improvements to the operation of the 'compatibility' property:
1) Improved handling of unrecognized features:
Change the way unrecognized features in compatibility files are handled.
* invalid features in files under /usr/share/zfs/compatibility.d
only get a warning (as these may refer to future features not yet in
the library),
* invalid features in files under /etc/zfs/compatibility.d
get an error (as these are presumed to refer to the current system).
2) Improved error reporting from zpool_load_compat.
Note: slight ABI change to zpool_load_compat for better error reporting.
3) compatibility=legacy inhibits all 'zpool upgrade' operations.
4) Detect when features are enabled outside current compatibility set
* zpool set compatibility=foo <-- print a warning
* zpool set feature@xxx=enabled <-- error
* zpool status <-- indicate this state
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Colm Buckley <colm@tuatha.org>
Closes#11861
It's been observed in the CI that the required 25% of obsolete bytes
in the mapping can be to high a threshold for this test resulting in
condensing never being triggered and a test failure. To prevent these
failures make the existing zfs_condense_indirect_obsolete_pct tuning
available so the obsolete percentage can be reduced from 25% to 5%
during this test.
Reviewed-by: Ryan Moeller <ryan@iXsystems.com>
Reviewed-by: George Melikov <mail@gmelikov.ru>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#11869
A tentative implementation and discussion was done in #5285.
According to it a send --skip-missing|-s flag has been added.
In a replication stream, when there are snapshots missing in
the hierarchy, if -s is provided print a warning and ignore
dataset (and its children) instead of throwing an error
Reviewed-by: Paul Dagnelie <pcd@delphix.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Pablo Correa Gómez <ablocorrea@hotmail.com>
Closes#11710
Just as delay zevents can flood the zevent pipe when a vdev becomes
unresponsive, so do the deadman zevents.
Ratelimit deadman zevents according to the same tunable as for delay
zevents.
Enable deadman tests on FreeBSD and add a test for deadman event
ratelimiting.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Don Brady <don.brady@delphix.com>
Signed-off-by: Ryan Moeller <ryan@iXsystems.com>
Closes#11786
These events should currently never be generated.
Also untag _zed_event_add_nvpair() from merge with
zpool_do_events_nvprint() ‒ they serve different purposes (machine,
usually script vs human consumption) and format the output differently
as it stands
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Closes#11834
And add a note on /why/ ZEDLETs need to be owned by root
Quoth chown(2), Linux man-pages project:
Only a privileged process (Linux: one with the CAP_CHOWN capability)
may change the owner of a file.
Quoth chown(2), FreeBSD:
[EPERM] The operation would change the ownership,
but the effective user ID is not the super-user.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Closes#11834
There simply isn't a need for one, since the flags the daemon takes
are all short (mostly just toggles) and administrative in nature,
and are therefore better served by the age-old tradition of sourcing an
environment file and preparing the cmdline in the init-specific handler
itself, if needed at all
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Closes#11834
The exact limitations on what features are supported when booting
vary considerably depending on the environment. In order to minimize
confusion avoid categorical statements which assume GRUB2 is being
used. The supported GRUB2 features are covered earlier in this man
page for easy reference.
Reviewed-by: George Melikov <mail@gmelikov.ru>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Closes#11842
In addition, html doc will have working hyperlinks.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Ryan Moeller <ryan@iXsystems.com>
Signed-off-by: George Melikov <mail@gmelikov.ru>
Closes#11845
Correct an assortment of typos throughout the code base.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Matthew Ahrens <mahrens@delphix.com>
Reviewed-by: Ryan Moeller <ryan@iXsystems.com>
Signed-off-by: Andrea Gelmini <andrea.gelmini@gelma.net>
Closes#11774
200ms time-out is relatively long, but if we already hit the cap,
then we'll likely be able to spawn multiple new jobs when we wake up
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Closes#11807
The lower bound for this scaling to too low and the upper bound is too
high. Use a fixed default length of 512 instead, which is a reasonable
value on any system.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Ryan Moeller <ryan@iXsystems.com>
Closes#11822
Update the fsck.zfs helper to bubble up some already-known-about
errors if they are detected in the pool.
health=degraded => 4/"Filesystem errors left uncorrected"
health=faulted && dataset in /etc/fstab => 8/"Operational error"
pool not found => 8/"Operational error"
everything else => 0
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Closes#11806
When a child process is killed waitpid() must be called on the
pid the reap the zombie process.
Update BUGS section to reflect reality by replacing "zedlets
aren't time limited with "zedlets can be interrupted".
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Closes#11769Closes#11798
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: George Melikov <mail@gmelikov.ru>
Signed-off-by: Ryan Moeller <ryan@iXsystems.com>
Closes#11770
The RAIDZ and DRAID code is responsible for reporting checksum errors on
their child vdevs. Checksum errors represent events where a disk
returned data or parity that should have been correct, but was not. In
other words, these are instances of silent data corruption. The
checksum errors show up in the vdev stats (and thus `zpool status`'s
CKSUM column), and in the event log (`zpool events`).
Note, this is in contrast with the more common "noisy" errors where a
disk goes offline, in which case ZFS knows that the disk is bad and
doesn't try to read it, or the device returns an error on the requested
read or write operation.
RAIDZ/DRAID generate checksum errors via three code paths:
1. When RAIDZ/DRAID reconstructs a damaged block, checksum errors are
reported on any children whose data was not used during the
reconstruction. This is handled in `raidz_reconstruct()`. This is the
most common type of RAIDZ/DRAID checksum error.
2. When RAIDZ/DRAID is not able to reconstruct a damaged block, that
means that the data has been lost. The zio fails and an error is
returned to the consumer (e.g. the read(2) system call). This would
happen if, for example, three different disks in a RAIDZ2 group are
silently damaged. Since the damage is silent, it isn't possible to know
which three disks are damaged, so a checksum error is reported against
every child that returned data or parity for this read. (For DRAID,
typically only one "group" of children is involved in each io.) This
case is handled in `vdev_raidz_cksum_finish()`. This is the next most
common type of RAIDZ/DRAID checksum error.
3. If RAIDZ/DRAID is not able to reconstruct a damaged block (like in
case 2), but there happens to be additional copies of this block due to
"ditto blocks" (i.e. multiple DVA's in this blkptr_t), and one of those
copies is good, then RAIDZ/DRAID compares each sector of the data or
parity that it retrieved with the good data from the other DVA, and if
they differ then it reports a checksum error on this child. This
differs from case 2 in that the checksum error is reported on only the
subset of children that actually have bad data or parity. This case
happens very rarely, since normally only metadata has ditto blocks. If
the silent damage is extensive, there will be many instances of case 2,
and the pool will likely be unrecoverable.
The code for handling case 3 is considerably more complicated than the
other cases, for two reasons:
1. It needs to run after the main raidz read logic has completed. The
data RAIDZ read needs to be preserved until after the alternate DVA has
been read, which necessitates refcounts and callbacks managed by the
non-raidz-specific zio layer.
2. It's nontrivial to map the sections of data read by RAIDZ to the
correct data. For example, the correct data does not include the parity
information, so the parity must be recalculated based on the correct
data, and then compared to the parity that was read from the RAIDZ
children.
Due to the complexity of case 3, the rareness of hitting it, and the
minimal benefit it provides above case 2, this commit removes the code
for case 3. These types of errors will now be handled the same as case
2, i.e. the checksum error will be reported against all children that
returned data or parity.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Matthew Ahrens <mahrens@delphix.com>
Closes#11735
The man page was missing these two permissions.
Add the missing permissions to the man page.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Jeremy Faulkner <gldisater@gldis.ca>
Closes#11727
To make use of zfs_refcount_held tunable it should be a module
parameter in open-zfs. Also, since the macros will auto-generate OS
specific tunables, removed the existing zfs_refcount_held reference
in module/os/freebsd/zfs/sysctl_os.c.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Ryan Moeller <ryan@iXsystems.com>
Reviewed-by: Allan Jude <allan@klarasystems.com>
Signed-off-by: Don Brady <don.brady@delphix.com>
Closes#11753
The manual page change in ecc277c has introduced whitespace on
line ends.
Reviewed-by: Ryan Moeller <ryan@iXsystems.com>
Reviewed-by: George Melikov <mail@gmelikov.ru>
Signed-off-by: Martin Matuska <mm@FreeBSD.org>
Closes#11722
Docs for send and receive do not explain behavior when sending a
compressed stream then receiving on a host that overrides compression
with -o compress=value.
The data from the send stream is written as it was from the send is
the compressed form but the compression algorithm set on the receiver
is the overridden version which causes some confusion as to what
algorithm was actually used.
Updated man docs to clarify behavior
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed By: Allan Jude <allanjude@freebsd.org>
Signed-off-by: manfromafar <manfromafar@outlook.com>
Closes#11690
On Linux increase the maximum allowed size of the src nvlist which
can be passed to the /dev/zfs ioctl. Originally, this was set
to a maximum of KMALLOC_MAX_SIZE (4M) because it was kmalloc'd.
Since that time it's been converted to a vmalloc so that's no
longer a hard limit, and it's desirable for `zfs send/recv` to
allow larger nvlists so more snapshots can be sent at once.
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#6572Closes#11638
Property to allow sets of features to be specified; for compatibility
with specific versions / releases / external systems. Influences
the behavior of 'zpool upgrade' and 'zpool create'. Initial man
page changes and test cases included.
Brief synopsis:
zpool create -o compatibility=off|legacy|file[,file...] pool vdev...
compatibility = off : disable compatibility mode (enable all features)
compatibility = legacy : request that no features be enabled
compatibility = file[,file...] : read features from specified files.
Only features present in *all* files will be enabled on the
resulting pool. Filenames may be absolute, or relative to
/etc/zfs/compatibility.d or /usr/share/zfs/compatibility.d (/etc
checked first).
Only affects zpool create, zpool upgrade and zpool status.
ABI changes in libzfs:
* New function "zpool_load_compat" to load and parse compat sets.
* Add "zpool_compat_status_t" typedef for compatibility parse status.
* Add ZPOOL_PROP_COMPATIBILITY to the pool properties enum
* Add ZPOOL_STATUS_COMPATIBILITY_ERR to the pool status enum
An initial set of base compatibility sets are included in
cmd/zpool/compatibility.d, and the Makefile for cmd/zpool is
modified to install these in $pkgdatadir/compatibility.d and to
create symbolic links to a reasonable set of aliases.
Reviewed-by: ericloewe
Reviewed-by: Matthew Ahrens <mahrens@delphix.com>
Reviewed-by: Richard Laager <rlaager@wiktel.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Colm Buckley <colm@tuatha.org>
Closes#11468
Clarify how to include snapshots in the `zpool list` output by
referencing the full name of the `listsnapshots` pool property,
and the `zpool list -t snapshot` option.
Reviewed-by: Matthew Ahrens <mahrens@delphix.com>
Reviewed-by: George Melikov <mail@gmelikov.ru>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#11562Closes#11565
While you can use zdb -R poolname vdev:offset:[<lsize>/]<psize>[:flags]
to extract individual DVAs from a vdev, it would be handy for be able
copy an entire file out of the pool.
Given a file or object number, add support to copy the contents to a
file. Useful for debugging and recovery.
Reviewed-by: Jorgen Lundman <lundman@lundman.net>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Allan Jude <allan@klarasystems.com>
Closes#11027
zfs-program.8 has an orphan link, fix it.
https://svnweb.freebsd.org/base?view=revision&revision=360080
Obtained from: FreeBSD
Reviewed-by: George Melikov <mail@gmelikov.ru>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Alan Somers <asomers@gmail.com>
Closes#11529
Mixing ZIL and normal allocations has several problems:
1. The ZIL allocations are allocated, written to disk, and then a few
seconds later freed. This leaves behind holes (free segments) where the
ZIL blocks used to be, which increases fragmentation, which negatively
impacts performance.
2. When under moderate load, ZIL allocations are of 128KB. If the pool
is fairly fragmented, there may not be many free chunks of that size.
This causes ZFS to load more metaslabs to locate free segments of 128KB
or more. The loading happens synchronously (from zil_commit()), and can
take around a second even if the metaslab's spacemap is cached in the
ARC. All concurrent synchronous operations on this filesystem must wait
while the metaslab is loading. This can cause a significant performance
impact.
3. If the pool is very fragmented, there may be zero free chunks of
128KB or more. In this case, the ZIL falls back to txg_wait_synced(),
which has an enormous performance impact.
These problems can be eliminated by using a dedicated log device
("slog"), even one with the same performance characteristics as the
normal devices.
This change sets aside one metaslab from each top-level vdev that is
preferentially used for ZIL allocations (vdev_log_mg,
spa_embedded_log_class). From an allocation perspective, this is
similar to having a dedicated log device, and it eliminates the
above-mentioned performance problems.
Log (ZIL) blocks can be allocated from the following locations. Each
one is tried in order until the allocation succeeds:
1. dedicated log vdevs, aka "slog" (spa_log_class)
2. embedded slog metaslabs (spa_embedded_log_class)
3. other metaslabs in normal vdevs (spa_normal_class)
The space required for the embedded slog metaslabs is usually between
0.5% and 1.0% of the pool, and comes out of the existing 3.2% of "slop"
space that is not available for user data.
On an all-ssd system with 4TB storage, 87% fragmentation, 60% capacity,
and recordsize=8k, testing shows a ~50% performance increase on random
8k sync writes. On even more fragmented systems (which hit problem #3
above and call txg_wait_synced()), the performance improvement can be
arbitrarily large (>100x).
Reviewed-by: Serapheim Dimitropoulos <serapheim@delphix.com>
Reviewed-by: George Wilson <gwilson@delphix.com>
Reviewed-by: Don Brady <don.brady@delphix.com>
Reviewed-by: Mark Maybee <mark.maybee@delphix.com>
Signed-off-by: Matthew Ahrens <mahrens@delphix.com>
Closes#11389
On a system with very high fragmentation, we may need to do lots of gang
allocations (e.g. most indirect block allocations (~50KB) may need to
gang). Before failing a "normal" allocation and resorting to ganging, we
try every metaslab. This has the impact of loading every metaslab (not
a huge deal since we now typically keep all metaslabs loaded), and also
iterating over every metaslab for every failing allocation. If there are
many metaslabs (more than the typical ~200, e.g. due to vdev expansion
or very large vdevs), the CPU cost of this iteration can be very
impactful. This iteration is done with the mg_lock held, creating long
hold times and high lock contention for concurrent allocations,
ultimately causing long txg sync times and poor application performance.
To address this, this commit changes the behavior of "normal" (not
try_hard, not ZIL) allocations. These will now only examine the 100
best metaslabs (as determined by their ms_weight). If none of these
have a large enough free segment, then the allocation will fail and
we'll fall back on ganging.
To accomplish this, we will now (normally) gang before doing a
`try_hard` allocation. Non-try_hard allocations will only examine the
100 best metaslabs of each vdev. In summary, we will first try normal
allocation. If that fails then we will do a gang allocation. If that
fails then we will do a "try hard" gang allocation. If that fails then
we will have a multi-layer gang block.
Reviewed-by: Paul Dagnelie <pcd@delphix.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Matthew Ahrens <mahrens@delphix.com>
Closes#11327
