zfs/include
Ned Bass 68cbd56e18 Backfill metadnode more intelligently
Only attempt to backfill lower metadnode object numbers if at least
4096 objects have been freed since the last rescan, and at most once
per transaction group. This avoids a pathology in dmu_object_alloc()
that caused O(N^2) behavior for create-heavy workloads and
substantially improves object creation rates.  As summarized by
@mahrens in #4636:

"Normally, the object allocator simply checks to see if the next
object is available. The slow calls happened when dmu_object_alloc()
checks to see if it can backfill lower object numbers. This happens
every time we move on to a new L1 indirect block (i.e. every 32 *
128 = 4096 objects).  When re-checking lower object numbers, we use
the on-disk fill count (blkptr_t:blk_fill) to quickly skip over
indirect blocks that don’t have enough free dnodes (defined as an L2
with at least 393,216 of 524,288 dnodes free). Therefore, we may
find that a block of dnodes has a low (or zero) fill count, and yet
we can’t allocate any of its dnodes, because they've been allocated
in memory but not yet written to disk. In this case we have to hold
each of the dnodes and then notice that it has been allocated in
memory.

The end result is that allocating N objects in the same TXG can
require CPU usage proportional to N^2."

Add a tunable dmu_rescan_dnode_threshold to define the number of
objects that must be freed before a rescan is performed. Don't bother
to export this as a module option because testing doesn't show a
compelling reason to change it. The vast majority of the performance
gain comes from limit the rescan to at most once per TXG.

Signed-off-by: Ned Bass <bass6@llnl.gov>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2016-06-24 13:13:12 -07:00
..
linux Linux 4.7 compat: handler->set() takes both dentry and inode 2016-06-01 18:10:06 -07:00
sys Backfill metadnode more intelligently 2016-06-24 13:13:12 -07:00
Makefile.am Kernel header installation should respect --prefix 2014-10-28 09:37:06 -07:00
libnvpair.h Support custom build directories and move includes 2010-09-08 12:38:56 -07:00
libuutil.h Support custom build directories and move includes 2010-09-08 12:38:56 -07:00
libuutil_common.h Support custom build directories and move includes 2010-09-08 12:38:56 -07:00
libuutil_impl.h Support custom build directories and move includes 2010-09-08 12:38:56 -07:00
libzfs.h Add -lhHpw options to "zpool iostat" for avg latency, histograms, & queues 2016-05-12 12:36:32 -07:00
libzfs_core.h Illumos 5027 - zfs large block support 2015-05-11 12:23:16 -07:00
libzfs_impl.h Remove libzfs_graph.c 2016-06-16 13:53:15 -07:00
zfeature_common.h Illumos 5959 - clean up per-dataset feature count code 2015-12-04 14:20:20 -08:00
zfs_comutil.h Illumos #2882, #2883, #2900 2013-09-04 15:49:00 -07:00
zfs_deleg.h Illumos 4368, 4369. 2014-07-29 10:55:29 -07:00
zfs_fletcher.h Implementation of AVX2 optimized Fletcher-4 2016-06-02 14:30:51 -07:00
zfs_namecheck.h Illumos 4368, 4369. 2014-07-29 10:55:29 -07:00
zfs_prop.h Check the dataset type more rigorously when fetching properties. 2014-05-06 10:41:46 -07:00
zpios-ctl.h Add large block support to zpios(1) benchmark 2015-09-22 09:13:20 -07:00
zpios-internal.h Add large block support to zpios(1) benchmark 2015-09-22 09:13:20 -07:00