mirror_zfs/module/zfs
Matthew Ahrens 7261fc2e81
Improve zfs receive performance by batching writes
For each WRITE record in the stream, `zfs receive` creates a DMU
transaction (`dmu_tx_create()`) and writes this block's data into the
object.  If per-block overheads (as opposed to per-byte overheads)
dominate performance (as is often the case with small recordsize), the
per-dmu-transaction overheads can be significant.  For example, in some
workloads the `receieve_writer` thread is 100% on CPU, and more than
half of its CPU time is in these per-tx routines (e.g.
dmu_tx_hold_write, dmu_tx_assign, dmu_tx_commit).

To improve performance of `zfs receive`, this commit batches WRITE
records which are to nearby offsets of the same object, and uses one DMU
transaction to write them all.  By default the batch size is 1MB, which
for recordsize=8K reduces the number of DMU transactions by 128x for
full send streams (incrementals will depend on how "clumpy" the changed
blocks are).

This commit improves the performance of `dd if=stream | zfs recv`
from 78,800 blocks/sec to 98,100 blocks/sec (25% improvement).

Reviewed-by: Paul Dagnelie <pcd@delphix.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Matthew Ahrens <mahrens@delphix.com>
Closes #10099
2020-03-16 11:51:56 -07:00
..
aggsum.c Reduce number of atomic_add() calls in aggsum 2020-02-06 13:21:06 -08:00
arc.c Prevent deadlock in arc_read in Linux memory reclaim callback 2020-03-12 10:24:43 -07:00
blkptr.c Undo c89 workarounds to match with upstream 2017-11-04 13:25:13 -07:00
bplist.c Fast Clone Deletion 2019-07-26 10:54:14 -07:00
bpobj.c Add subcommand to wait for background zfs activity to complete 2019-09-13 18:09:06 -07:00
bptree.c Implement Redacted Send/Receive 2019-06-19 09:48:12 -07:00
bqueue.c Implement Redacted Send/Receive 2019-06-19 09:48:12 -07:00
btree.c Replace ASSERTV macro with compiler annotation 2019-12-05 12:37:00 -08:00
cityhash.c OpenZFS 8484 - Implement aggregate sum and use for arc counters 2018-06-06 09:35:59 -07:00
dataset_kstats.c Fix panic on DilOS with kstat per dataset statistics 2019-09-03 12:12:31 -07:00
dbuf_stats.c Make module tunables cross platform 2019-09-05 14:49:49 -07:00
dbuf.c Consolidate arc_buf allocation checks 2020-02-27 17:12:44 -08:00
ddt_zap.c Refactor dnode dirty context from dbuf_dirty 2020-02-26 16:09:17 -08:00
ddt.c Reduce loaded range tree memory usage 2019-10-09 10:36:03 -07:00
dmu_diff.c Mark write_record static 2019-12-03 09:51:44 -08:00
dmu_object.c Make module tunables cross platform 2019-09-05 14:49:49 -07:00
dmu_objset.c dmu_objset_from_ds must be called with dp_config_rwlock held 2020-03-12 10:55:02 -07:00
dmu_recv.c Improve zfs receive performance by batching writes 2020-03-16 11:51:56 -07:00
dmu_redact.c dmu_objset_from_ds must be called with dp_config_rwlock held 2020-03-12 10:55:02 -07:00
dmu_send.c dmu_objset_from_ds must be called with dp_config_rwlock held 2020-03-12 10:55:02 -07:00
dmu_traverse.c Make module tunables cross platform 2019-09-05 14:49:49 -07:00
dmu_tx.c Refactor dnode dirty context from dbuf_dirty 2020-02-26 16:09:17 -08:00
dmu_zfetch.c Make module tunables cross platform 2019-09-05 14:49:49 -07:00
dmu.c Remove zfs_getattr and convoff dead code 2020-02-24 15:38:22 -08:00
dnode_sync.c Convert dbuf dirty record record list to a list_t 2020-02-05 11:07:19 -08:00
dnode.c Prevent race condition in dnode_dest (#10101) 2020-03-12 10:25:56 -07:00
dsl_bookmark.c Refactor dnode dirty context from dbuf_dirty 2020-02-26 16:09:17 -08:00
dsl_crypt.c dmu_objset_from_ds must be called with dp_config_rwlock held 2020-03-12 10:55:02 -07:00
dsl_dataset.c Refactor dnode dirty context from dbuf_dirty 2020-02-26 16:09:17 -08:00
dsl_deadlist.c Reduce loaded range tree memory usage 2019-10-09 10:36:03 -07:00
dsl_deleg.c Reduce loaded range tree memory usage 2019-10-09 10:36:03 -07:00
dsl_destroy.c Replace ASSERTV macro with compiler annotation 2019-12-05 12:37:00 -08:00
dsl_dir.c Refactor dnode dirty context from dbuf_dirty 2020-02-26 16:09:17 -08:00
dsl_pool.c Eliminate Linux specific inode usage from common code 2019-12-11 11:53:57 -08:00
dsl_prop.c Refactor dnode dirty context from dbuf_dirty 2020-02-26 16:09:17 -08:00
dsl_scan.c Fix infinite scan on a pool with only special allocations 2020-03-12 10:52:03 -07:00
dsl_synctask.c Fix typos in module/zfs/ 2019-09-02 17:56:41 -07:00
dsl_userhold.c Fix strdup conflict on other platforms 2019-10-10 09:47:06 -07:00
edonr_zfs.c DLPX-44812 integrate EP-220 large memory scalability 2016-11-29 14:34:27 -08:00
fm.c Enable zpool events tunables and tests on FreeBSD 2020-02-18 11:22:56 -08:00
gzip.c OpenZFS restructuring - move platform specific sources 2019-09-06 11:26:26 -07:00
hkdf.c Encryption patch follow-up 2017-10-11 16:54:48 -04:00
lz4.c Enable clang to use intrinsics for lz4 2019-10-01 13:17:32 -07:00
lzjb.c Change KM_PUSHPAGE -> KM_SLEEP 2015-01-16 14:41:26 -08:00
Makefile.in Support setting user properties in a channel program 2020-02-14 13:41:42 -08:00
metaslab.c Replace ASSERTV macro with compiler annotation 2019-12-05 12:37:00 -08:00
mmp.c Improve zfs destroy performance with zio_t-free zio_free() 2020-02-28 14:49:44 -08:00
multilist.c Enable use of DTRACE_PROBE* macros in "spl" module 2019-11-01 13:13:43 -07:00
objlist.c Implement Redacted Send/Receive 2019-06-19 09:48:12 -07:00
pathname.c Disable unused pathname::pn_path* (unneeded in Linux) 2019-07-15 13:57:56 -07:00
range_tree.c Function name and comment updates 2019-10-11 10:13:21 -07:00
refcount.c Prevent race in blkptr_verify against device removal 2019-08-13 21:24:43 -06:00
rrwlock.c Enable use of DTRACE_PROBE* macros in "spl" module 2019-11-01 13:13:43 -07:00
sa.c Replace ASSERTV macro with compiler annotation 2019-12-05 12:37:00 -08:00
sha256.c OpenZFS restructuring - move platform specific sources 2019-09-06 11:26:26 -07:00
skein_zfs.c DLPX-44812 integrate EP-220 large memory scalability 2016-11-29 14:34:27 -08:00
spa_boot.c Add linux kernel module support 2010-08-31 13:41:58 -07:00
spa_checkpoint.c Refactor dnode dirty context from dbuf_dirty 2020-02-26 16:09:17 -08:00
spa_config.c Add zfs_file_* interface, remove vnodes 2019-11-21 09:32:57 -08:00
spa_errlog.c Fix typos in module/zfs/ 2019-09-02 17:56:41 -07:00
spa_history.c Make spa_history_zone platform-dependent in kernel 2020-03-02 09:43:30 -08:00
spa_log_spacemap.c Make module tunables cross platform 2019-09-05 14:49:49 -07:00
spa_misc.c Share some code for spa deadman tunables 2020-02-10 13:11:30 -08:00
spa.c Add trim support to zpool wait 2020-03-04 15:07:11 -08:00
space_map.c Reduce loaded range tree memory usage 2019-10-09 10:36:03 -07:00
space_reftree.c Reduce loaded range tree memory usage 2019-10-09 10:36:03 -07:00
THIRDPARTYLICENSE.cityhash OpenZFS 8484 - Implement aggregate sum and use for arc counters 2018-06-06 09:35:59 -07:00
THIRDPARTYLICENSE.cityhash.descrip OpenZFS 8484 - Implement aggregate sum and use for arc counters 2018-06-06 09:35:59 -07:00
txg.c Replace ASSERTV macro with compiler annotation 2019-12-05 12:37:00 -08:00
uberblock.c MMP interval and fail_intervals in uberblock 2019-03-21 12:47:57 -07:00
unique.c Reduce loaded range tree memory usage 2019-10-09 10:36:03 -07:00
vdev_cache.c Replace ASSERTV macro with compiler annotation 2019-12-05 12:37:00 -08:00
vdev_indirect_births.c Fixes: #8934 Large kmem_alloc 2019-07-10 15:54:49 -07:00
vdev_indirect_mapping.c Replace ASSERTV macro with compiler annotation 2019-12-05 12:37:00 -08:00
vdev_indirect.c Replace ASSERTV macro with compiler annotation 2019-12-05 12:37:00 -08:00
vdev_initialize.c Linux 5.6 compat: time_t 2020-02-27 09:31:02 -08:00
vdev_label.c Add zfs_file_* interface, remove vnodes 2019-11-21 09:32:57 -08:00
vdev_mirror.c Avoid some crashes when importing a pool with corrupt metadata 2019-12-26 10:57:05 -08:00
vdev_missing.c Update vdev_ops_t from illumos 2019-06-20 18:29:02 -07:00
vdev_queue.c Reduce loaded range tree memory usage 2019-10-09 10:36:03 -07:00
vdev_raidz_math_aarch64_neon_common.h Minor performance fix for NEON RAID-Z 2019-12-17 19:34:52 -08:00
vdev_raidz_math_aarch64_neon.c Linux 5.0 compat: SIMD compatibility 2019-07-12 09:31:20 -07:00
vdev_raidz_math_aarch64_neonx2.c Linux 5.0 compat: SIMD compatibility 2019-07-12 09:31:20 -07:00
vdev_raidz_math_avx2.c OpenZFS restructuring - move platform specific headers 2019-09-05 09:34:54 -07:00
vdev_raidz_math_avx512bw.c OpenZFS restructuring - move platform specific headers 2019-09-05 09:34:54 -07:00
vdev_raidz_math_avx512f.c Make clang happy with vdev_raidz_ code 2019-10-10 09:45:37 -07:00
vdev_raidz_math_impl.h Fix const-correctness in raidz math 2020-02-03 10:52:41 -08:00
vdev_raidz_math_powerpc_altivec_common.h Add AltiVec RAID-Z 2020-01-23 11:01:24 -08:00
vdev_raidz_math_powerpc_altivec.c Add AltiVec RAID-Z 2020-01-23 11:01:24 -08:00
vdev_raidz_math_scalar.c Linux 5.3: Fix switch() fall though compiler errors 2019-08-21 09:29:23 -07:00
vdev_raidz_math_sse2.c Make clang happy with vdev_raidz_ code 2019-10-10 09:45:37 -07:00
vdev_raidz_math_ssse3.c OpenZFS restructuring - move platform specific headers 2019-09-05 09:34:54 -07:00
vdev_raidz_math.c Add AltiVec RAID-Z 2020-01-23 11:01:24 -08:00
vdev_raidz.c Reduce loaded range tree memory usage 2019-10-09 10:36:03 -07:00
vdev_removal.c Cancel initialize and TRIM before vdev_metaslab_fini() 2019-12-26 10:50:23 -08:00
vdev_root.c Update vdev_ops_t from illumos 2019-06-20 18:29:02 -07:00
vdev_trim.c Add trim support to zpool wait 2020-03-04 15:07:11 -08:00
vdev.c Refactor dnode dirty context from dbuf_dirty 2020-02-26 16:09:17 -08:00
zap_leaf.c Refactor dnode dirty context from dbuf_dirty 2020-02-26 16:09:17 -08:00
zap_micro.c Reduce loaded range tree memory usage 2019-10-09 10:36:03 -07:00
zap.c Make module tunables cross platform 2019-09-05 14:49:49 -07:00
zcp_get.c Refactor dnode dirty context from dbuf_dirty 2020-02-26 16:09:17 -08:00
zcp_global.c OpenZFS 8600 - ZFS channel programs - snapshot 2018-02-08 15:29:24 -08:00
zcp_iter.c Fix typos in module/zfs/ 2019-09-02 17:56:41 -07:00
zcp_set.c Support setting user properties in a channel program 2020-02-14 13:41:42 -08:00
zcp_synctask.c Refactor dnode dirty context from dbuf_dirty 2020-02-26 16:09:17 -08:00
zcp.c Refactor dnode dirty context from dbuf_dirty 2020-02-26 16:09:17 -08:00
zfeature.c Replace ASSERTV macro with compiler annotation 2019-12-05 12:37:00 -08:00
zfs_byteswap.c Fix typos in module/zfs/ 2019-09-02 17:56:41 -07:00
zfs_fm.c Add zpool status -s (slow I/Os) and -p (parseable) 2018-11-08 16:47:24 -08:00
zfs_fuid.c Relocate common quota functions to shared code 2019-12-11 12:12:08 -08:00
zfs_ioctl.c Re-share zfsdev_getminor and zfs_onexit_fd_hold 2020-02-28 14:50:32 -08:00
zfs_log.c Don't call zrele on passed zp in zfs_xattr_owner_unlinked on FreeBSD 2020-02-28 14:53:18 -08:00
zfs_onexit.c Re-share zfsdev_getminor and zfs_onexit_fd_hold 2020-02-28 14:50:32 -08:00
zfs_quota.c Relocate common quota functions to shared code 2019-12-11 12:12:08 -08:00
zfs_ratelimit.c Change checksum & IO delay ratelimit values 2018-03-04 17:34:51 -08:00
zfs_replay.c Simplify FreeBSD's locking requirements in zfs_replay.c 2020-01-22 17:55:56 -08:00
zfs_rlock.c Move platform specific parts of zfs_znode.h to platform code 2019-11-06 10:54:25 -08:00
zfs_sa.c Add inode accessors to common code 2019-10-02 09:15:12 -07:00
zil.c Improve logging of 128KB writes 2019-11-11 09:27:59 -08:00
zio_checksum.c Disable EDONR on FreeBSD 2019-12-05 13:10:29 -08:00
zio_compress.c zio_decompress_data always ASSERTs successful decompression 2019-12-10 15:51:58 -08:00
zio_inject.c Replace ASSERTV macro with compiler annotation 2019-12-05 12:37:00 -08:00
zio.c zio: dprintf_bp() if errors > 0 in zfs_blkptr_verify() 2020-03-04 15:08:41 -08:00
zle.c Fix zle_decompress out of bound access 2018-02-09 10:08:05 -08:00
zrlock.c Enable use of DTRACE_PROBE* macros in "spl" module 2019-11-01 13:13:43 -07:00
zthr.c Fast Clone Deletion 2019-07-26 10:54:14 -07:00
zvol.c async zvol minor node creation interferes with receive 2020-02-03 09:33:14 -08:00