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
Summary: We should never set max_open_files to be bigger than the system's ulimit. Otherwise we will get "Too many open files" errors. See an example in this Travis run: https://travis-ci.org/facebook/rocksdb/jobs/79591566
Test Plan:
make check
I will also verify that max_max_open_files is reasonable.
Reviewers: anthony, kradhakrishnan, IslamAbdelRahman, sdong
Reviewed By: sdong
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D46551
Summary: We want to keep Env a think layer for better portability. Less platform dependent codes should be moved out of Env. In this patch, I create a wrapper of file readers and writers, and put rate limiting, write buffering, as well as most perf context instrumentation and random kill out of Env. It will make it easier to maintain multiple Env in the future.
Test Plan: Run all existing unit tests.
Reviewers: anthony, kradhakrishnan, IslamAbdelRahman, yhchiang, igor
Reviewed By: igor
Subscribers: leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D42321
Summary: So iOS size_t is 32-bit, so we need to static_cast<size_t> any uint64_t :(
Test Plan: TARGET_OS=IOS make static_lib
Reviewers: dhruba, ljin, yhchiang, rven, sdong
Reviewed By: sdong
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D28743
Summary:
A generic rate limiter that can be shared by threads and rocksdb
instances. Will use this to smooth out write traffic generated by
compaction and flush. This will help us get better p99 behavior on flash
storage.
Test Plan:
unit test output
==== Test RateLimiterTest.Rate
request size [1 - 1023], limit 10 KB/sec, actual rate: 10.374969 KB/sec, elapsed 2002265
request size [1 - 2047], limit 20 KB/sec, actual rate: 20.771242 KB/sec, elapsed 2002139
request size [1 - 4095], limit 40 KB/sec, actual rate: 41.285299 KB/sec, elapsed 2202424
request size [1 - 8191], limit 80 KB/sec, actual rate: 81.371605 KB/sec, elapsed 2402558
request size [1 - 16383], limit 160 KB/sec, actual rate: 162.541268 KB/sec, elapsed 3303500
Reviewers: yhchiang, igor, sdong
Reviewed By: sdong
Subscribers: leveldb
Differential Revision: https://reviews.facebook.net/D19359
Summary:
This diff adds timeout_hint_us to WriteOptions. If it's non-zero, then
1) writes associated with this options MAY be aborted when it has been
waiting for longer than the specified time. If an abortion happens,
associated writes will return Status::TimeOut.
2) the stall time of the associated write caused by flush or compaction
will be limited by timeout_hint_us.
The default value of timeout_hint_us is 0 (i.e., OFF.)
The statistics of timeout writes will be recorded in WRITE_TIMEDOUT.
Test Plan:
export ROCKSDB_TESTS=WriteTimeoutAndDelayTest
make db_test
./db_test
Reviewers: igor, ljin, haobo, sdong
Reviewed By: sdong
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D18837
Summary:
Add TimedWait() API to CondVar, which will be used in the future to
support TimedOut Write API and Rate limiter.
Test Plan: make db_test -j32
Reviewers: sdong, ljin
Reviewed By: ljin
Subscribers: leveldb
Differential Revision: https://reviews.facebook.net/D19431
Some platforms, particularly Windows, do not have a single method that can
release both a held reader lock and a held writer lock; instead, a
separate method (ReleaseSRWLockShared or ReleaseSRWLockExclusive) must be
called in each case.
This may also be necessary to back MutexRW with a shared_mutex in C++14;
the current language proposal includes both an unlock() and a
shared_unlock() method.
Summary:
Using ThreadLocalPtr as a flag to determine if a mutex is locked or not enables us to implement AssertNotHeld(). It also makes AssertHeld() actually correct.
I had to remove port::Mutex as a dependency for util/thread_local.h, but that's fine since we can just use std::mutex :)
Test Plan: make check
Reviewers: ljin, dhruba, haobo, sdong, yhchiang
Reviewed By: ljin
CC: leveldb
Differential Revision: https://reviews.facebook.net/D18171
Summary:
AssertHeld() was a no-op before. Now it does things.
Also, this change caught a bad bug in SuperVersion::Init(). The method is calling db->mutex.AssertHeld(), but db variable is not initialized yet! I also fixed that issue.
Test Plan: make check
Reviewers: dhruba, haobo, ljin, sdong, yhchiang
Reviewed By: haobo
CC: leveldb
Differential Revision: https://reviews.facebook.net/D17193
Summary:
Change namespace from leveldb to rocksdb. This allows a single
application to link in open-source leveldb code as well as
rocksdb code into the same process.
Test Plan: compile rocksdb
Reviewers: emayanke
Reviewed By: emayanke
CC: leveldb
Differential Revision: https://reviews.facebook.net/D13287
Summary:
This diff adds an option to specify whether PTHREAD_MUTEX_ADAPTIVE_NP will be enabled for the rocksdb single big kernel lock. db_bench also have this option now.
Quickly tested 8 thread cpu bound 100 byte random read.
No fast mutex: ~750k/s ops
With fast mutex: ~880k/s ops
Test Plan: make check; db_bench; db_stress
Reviewers: dhruba
CC: MarkCallaghan, leveldb
Differential Revision: https://reviews.facebook.net/D11031
Summary:
Implement ReadWrite locks for leveldb. These will be helpful
to implement a read-modify-write operation (e.g. atomic increments).
Test Plan: does not modify any existing code
Reviewers: heyongqiang
Reviewed By: heyongqiang
CC: MarkCallaghan
Differential Revision: https://reviews.facebook.net/D5787
- Replace raw slice comparison with a call to user comparator.
Added test for custom comparators.
- Fix end of namespace comments.
- Fixed bug in picking inputs for a level-0 compaction.
When finding overlapping files, the covered range may expand
as files are added to the input set. We now correctly expand
the range when this happens instead of continuing to use the
old range. For example, suppose L0 contains files with the
following ranges:
F1: a .. d
F2: c .. g
F3: f .. j
and the initial compaction target is F3. We used to search
for range f..j which yielded {F2,F3}. However we now expand
the range as soon as another file is added. In this case,
when F2 is added, we expand the range to c..j and restart the
search. That picks up file F1 as well.
This change fixes a bug related to deleted keys showing up
incorrectly after a compaction as described in Issue 44.
(Sync with upstream @25072954)
