Summary:
This fixes OOMs that we (logdevice) are currently having in production.
SkipListRep constructor does a couple small allocations from ConcurrentArena (see InlineSkipList constructor). ConcurrentArena would sometimes allocate an entire block for that, which is a few megabytes (we use Options::arena_block_size = 4 MB). So an empty memtable can take take 4 MB of memory. We have ~40k column families (spread across 15 DB instances), so 4 MB per empty memtable easily OOMs a machine for us.
This PR makes ConcurrentArena always allocate from Arena's inline block when possible. So as long as InlineSkipList's initial allocations are below 2 KB there would be no blocks allocated for empty memtables.
Closes https://github.com/facebook/rocksdb/pull/2569
Differential Revision: D5404029
Pulled By: al13n321
fbshipit-source-id: 568ec22a3fd1a485c06123f6b2dfc5e9ef67cd23
Summary:
Improve write buffer manager in several ways:
1. Size is tracked when arena block is allocated, rather than every allocation, so that it can better track actual memory usage and the tracking overhead is slightly lower.
2. We start to trigger memtable flush when 7/8 of the memory cap hits, instead of 100%, and make 100% much harder to hit.
3. Allow a cache object to be passed into buffer manager and the size allocated by memtable can be costed there. This can help users have one single memory cap across block cache and memtable.
Closes https://github.com/facebook/rocksdb/pull/2350
Differential Revision: D5110648
Pulled By: siying
fbshipit-source-id: b4238113094bf22574001e446b5d88523ba00017
Summary:
Moved the logic for core-local array out of ConcurrentArena and into a separate class because I want to reuse it for core-local stats.
Closes https://github.com/facebook/rocksdb/pull/2256
Differential Revision: D5011518
Pulled By: ajkr
fbshipit-source-id: a75a7b8f7b7a42fd6273489ada405f14c6be196a
Summary: Adding "__attribute__((__unused__))" after padding fields will pass CLANG build but will fail gcc 4.8.1. Fix it by not generating it under GCC 4.8.1.
Test Plan: Build under four combinations of USE_CLANG=0,1 and ROCKSDB_FBCODE_BUILD_WITH_481=0.1.
Reviewers: yhchiang, rven, ngbronson, anthony, IslamAbdelRahman
Reviewed By: anthony
Subscribers: leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D52371
Summary: Fix some CLANG errors introduced in 7d87f02799
Test Plan: Build with both of CLANG and gcc
Reviewers: rven, yhchiang, kradhakrishnan, anthony, IslamAbdelRahman, ngbronson
Subscribers: leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D52329
Summary:
This diff adds support for concurrent adds to the skiplist memtable
implementations. Memory allocation is made thread-safe by the addition of
a spinlock, with small per-core buffers to avoid contention. Concurrent
memtable writes are made via an additional method and don't impose a
performance overhead on the non-concurrent case, so parallelism can be
selected on a per-batch basis.
Write thread synchronization is an increasing bottleneck for higher levels
of concurrency, so this diff adds --enable_write_thread_adaptive_yield
(default off). This feature causes threads joining a write batch
group to spin for a short time (default 100 usec) using sched_yield,
rather than going to sleep on a mutex. If the timing of the yield calls
indicates that another thread has actually run during the yield then
spinning is avoided. This option improves performance for concurrent
situations even without parallel adds, although it has the potential to
increase CPU usage (and the heuristic adaptation is not yet mature).
Parallel writes are not currently compatible with
inplace updates, update callbacks, or delete filtering.
Enable it with --allow_concurrent_memtable_write (and
--enable_write_thread_adaptive_yield). Parallel memtable writes
are performance neutral when there is no actual parallelism, and in
my experiments (SSD server-class Linux and varying contention and key
sizes for fillrandom) they are always a performance win when there is
more than one thread.
Statistics are updated earlier in the write path, dropping the number
of DB mutex acquisitions from 2 to 1 for almost all cases.
This diff was motivated and inspired by Yahoo's cLSM work. It is more
conservative than cLSM: RocksDB's write batch group leader role is
preserved (along with all of the existing flush and write throttling
logic) and concurrent writers are blocked until all memtable insertions
have completed and the sequence number has been advanced, to preserve
linearizability.
My test config is "db_bench -benchmarks=fillrandom -threads=$T
-batch_size=1 -memtablerep=skip_list -value_size=100 --num=1000000/$T
-level0_slowdown_writes_trigger=9999 -level0_stop_writes_trigger=9999
-disable_auto_compactions --max_write_buffer_number=8
-max_background_flushes=8 --disable_wal --write_buffer_size=160000000
--block_size=16384 --allow_concurrent_memtable_write" on a two-socket
Xeon E5-2660 @ 2.2Ghz with lots of memory and an SSD hard drive. With 1
thread I get ~440Kops/sec. Peak performance for 1 socket (numactl
-N1) is slightly more than 1Mops/sec, at 16 threads. Peak performance
across both sockets happens at 30 threads, and is ~900Kops/sec, although
with fewer threads there is less performance loss when the system has
background work.
Test Plan:
1. concurrent stress tests for InlineSkipList and DynamicBloom
2. make clean; make check
3. make clean; DISABLE_JEMALLOC=1 make valgrind_check; valgrind db_bench
4. make clean; COMPILE_WITH_TSAN=1 make all check; db_bench
5. make clean; COMPILE_WITH_ASAN=1 make all check; db_bench
6. make clean; OPT=-DROCKSDB_LITE make check
7. verify no perf regressions when disabled
Reviewers: igor, sdong
Reviewed By: sdong
Subscribers: MarkCallaghan, IslamAbdelRahman, anthony, yhchiang, rven, sdong, guyg8, kradhakrishnan, dhruba
Differential Revision: https://reviews.facebook.net/D50589
Mike Kolupaev