Summary:
Clock-based cache implemenetation aim to have better concurreny than
default LRU cache. See inline comments for implementation details.
Test Plan:
Update cache_test to run on both LRUCache and ClockCache. Adding some
new tests to catch some of the bugs that I fixed while implementing the
cache.
Reviewers: kradhakrishnan, sdong
Reviewed By: sdong
Subscribers: andrewkr, dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D61647
Summary:
This is a proof of concept of a RocksDB blob log file. The actual value of the Put() is appended to a blob log using normal data block format, and the handle of the block is written as the value of the key in RocksDB.
The prototype only supports Put() and Get(). It doesn't support DB restart, garbage collection, Write() call, iterator, snapshots, etc.
Test Plan: Add unit tests.
Reviewers: arahut
Reviewed By: arahut
Subscribers: kradhakrishnan, leveldb, andrewkr, dhruba
Differential Revision: https://reviews.facebook.net/D61485
Summary: As title.
Test Plan: Run the benchmark with and without the parameter.
Reviewers: yiwu, andrewkr, lightmark, IslamAbdelRahman
Reviewed By: IslamAbdelRahman
Subscribers: leveldb, andrewkr, dhruba
Differential Revision: https://reviews.facebook.net/D61491
Summary: Add a benchmark to `db_bench`. In this benchmark, a write thread will populate time series data in the format of 'id | timestamp', and multiple read threads will randomly retrieve all data from one id at a time.
Test Plan: Run the benchmark: `num=134217728;bpl=536870912;mb=67108864;overlap=10;mcz=2;del=300000000;levels=6;ctrig=4;delay=8;stop=12;wbn=3;mbc=20;wbs=134217728;dds=0;sync=0;t=32;vs=800;bs=4096;cs=17179869184;of=500000;wps=0;si=10000000; kir=100000; dir=/data/users/jhli/test/; ./db_bench --benchmarks=timeseries --disable_seek_compaction=1 --mmap_read=0 --statistics=1 --histogram=1 --num=$num --threads=$t --value_size=$vs --block_size=$bs --cache_size=$cs --bloom_bits=10 --cache_numshardbits=6 --open_files=$of --verify_checksum=1 --db=$dir --sync=$sync --disable_wal=0 --compression_type=none --stats_interval=$si --compression_ratio=1 --disable_data_sync=$dds --write_buffer_size=$wbs --target_file_size_base=$mb --max_write_buffer_number=$wbn --max_background_compactions=$mbc --level0_file_num_compaction_trigger=$ctrig --level0_slowdown_writes_trigger=$delay --level0_stop_writes_trigger=$stop --num_levels=$levels --delete_obsolete_files_period_micros=$del --min_level_to_compress=$mcz --max_grandparent_overlap_factor=$overlap --stats_per_interval=1 --max_bytes_for_level_base=$bpl --use_existing_db=0 --key_id_range=$kir`
Reviewers: andrewkr, sdong
Reviewed By: sdong
Subscribers: lgalanis, andrewkr, dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D60651
Summary: Extend the option memtable_prefix_bloom_huge_page_tlb_size from just putting memtable bloom filter to huge page to memtable itself too.
Test Plan: Run all existing tests.
Reviewers: IslamAbdelRahman, yhchiang, andrewkr
Reviewed By: andrewkr
Subscribers: leveldb, andrewkr, dhruba
Differential Revision: https://reviews.facebook.net/D60513
* Fixed Windows build error in CMakeLists.txt and perf_level error in db_bench_tool.cc
* Changed hard-coded perf levels in db_bench_tool.cc to enum values from perf_level.h
* Replaced remaining FLAGS_perf_level > 0
Summary:
This diff makes the following improvement in regression_test.sh:
1. Add NUM_OPS and DELETE_TEST_PATH to regression_test.sh:
* NUM_OPS: The number of operations that will be issued
in EACH thread.
Default: $NUM_KEYS / $NUM_THREADS
* DELETE_TEST_PATH: If true, then the test directory
will be deleted after the script ends.
Default: 0
2. Add more information in SUMMARY.csv
3. Fix a bug in regression_test.sh where each thread in fillseq will all issue $NUM_KEYS writes.
4. Add --deletes in db_bench, which allows us to control the number of deletes instead of must using FLAGS_num.
Test Plan: run regression test with and without DELETE_TEST_PATH and NUM_OPS
Reviewers: yiwu, sdong, IslamAbdelRahman, gunnarku
Reviewed By: gunnarku
Subscribers: andrewkr, dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D60039
Summary: filter_deltes is not a frequently used feature. Remove it.
Test Plan: Run all test suites.
Reviewers: igor, yhchiang, IslamAbdelRahman
Reviewed By: IslamAbdelRahman
Subscribers: leveldb, andrewkr, dhruba
Differential Revision: https://reviews.facebook.net/D59427
Summary:
memtable_prefix_bloom_probes is not a critical option. Remove it to reduce number of options.
It's easier for users to make mistakes with memtable_prefix_bloom_bits, turn it to memtable_prefix_bloom_bits_ratio
Test Plan: Run all existing tests
Reviewers: yhchiang, igor, IslamAbdelRahman
Reviewed By: IslamAbdelRahman
Subscribers: gunnarku, yoshinorim, MarkCallaghan, leveldb, andrewkr, dhruba
Differential Revision: https://reviews.facebook.net/D59199
Summary:
Added an option, --env_uri. When provided, it is used as an argument to
NewEnvFromUri(), which instantiates an Env based on it.
Test Plan:
built a simple binary that registers ChrootEnv for prefix "/", then
ran:
$ ./tmp --env_uri /tmp/ --db /abcde
/tmp/ is the chroot directory and /abcde is the db_name. Then I verified
db_bench uses /tmp/abcde
Reviewers: sdong, kradhakrishnan, lightmark
Reviewed By: lightmark
Subscribers: andrewkr, dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D59325
Summary:
This patch allows db_bench to initialize it's RocksDB Options via a
options file, specified by the --options_file flag. Note that if
--options_file flag is set, then it has higher priority than the
command-line argument.
Test Plan: db_bench_tool_test
Reviewers: sdong, IslamAbdelRahman, kradhakrishnan, yiwu, andrewkr
Reviewed By: andrewkr
Subscribers: andrewkr, dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D58533
Summary:
Added a new abstraction to cache page to RocksDB designed for the read
cache use.
RocksDB current block cache is more of an object cache. For the persistent read cache
project, what we need is a page cache equivalent. This changes adds a cache
abstraction to RocksDB to cache pages called PersistentCache. PersistentCache can cache
uncompressed pages or raw pages (content as in filesystem). The user can
choose to operate PersistentCache either in COMPRESSED or UNCOMPRESSED mode.
Blame Rev:
Test Plan: Run unit tests
Reviewers: sdong
Subscribers: andrewkr, dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D55707
Summary:
This diff includes an initial script running a set of benchmarks for
regression test. The script does the following things:
checkout the specified rocksdb commit (or origin/master as default)
make clean && DEBUG_LEVEL=0 make db_bench
setup test directories
run set of benchmarks and store results
Currently, the script will run couple benchmarks, store all the benchmark
output, extract micros per op and percentile information for each benchmark
and store them in a single SUMMARY.csv file. The SUMMARY.csv will make the
follow-up regression detection easier.
In addition, the current script only takes env arguments to set important
attributes of db_bench. Will follow-up with a patch that allows db_bench
to construct options from an options file.
Test Plan:
NUM_KEYS=100 ./tools/regression_test.sh
Sample SUMMARY.csv file:
commit id, benchmark, ms-per-op, p50, p75, p99, p99.9, p99.99
7e23ddf575890510e7d2fc7a79b31a1bbf317917, fillseq, 15.28, 54.66, 77.14, 5000.00, 17900.00, 18483.00
7e23ddf575890510e7d2fc7a79b31a1bbf317917, overwrite, 13.54, 57.69, 86.39, 3000.00, 15600.00, 17013.00
7e23ddf575890510e7d2fc7a79b31a1bbf317917, readrandom, 1.04, 0.80, 1.67, 293.33, 395.00, 504.00
7e23ddf575890510e7d2fc7a79b31a1bbf317917, readwhilewriting, 2.75, 1.01, 1.87, 200.00, 460.00, 485.00
7e23ddf575890510e7d2fc7a79b31a1bbf317917, deleterandom, 3.64, 48.12, 70.09, 200.00, 336.67, 347.00
7e23ddf575890510e7d2fc7a79b31a1bbf317917, seekrandom, 24.31, 391.87, 513.69, 872.73, 990.00, 1048.00
7e23ddf575890510e7d2fc7a79b31a1bbf317917, seekrandomwhilewriting, 14.02, 185.14, 294.15, 700.00, 1440.00, 1527.00
Reviewers: sdong, IslamAbdelRahman, kradhakrishnan, yiwu, andrewkr, gunnarku
Reviewed By: gunnarku
Subscribers: gunnarku, MarkCallaghan, andrewkr, dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D57597
Summary:
Introduced option to dump malloc statistics using new option flag.
Added new command line option to db_bench tool to enable this
funtionality.
Also extended build to support environments with/without jemalloc.
Test Plan:
1) Build rocksdb using `make` command. Launch the following command
`./db_bench --benchmarks=fillrandom --dump_malloc_stats=true
--num=10000000` end verified that jemalloc dump is present in LOG file.
2) Build rocksdb using `DISABLE_JEMALLOC=1 make db_bench -j32` and ran
the same db_bench tool and found the following message in LOG file:
"Please compile with jemalloc to enable malloc dump".
3) Also built rocksdb using `make` command on MacOS to verify behavior
in non-FB environment.
Also to debug build configuration change temporary changed
AM_DEFAULT_VERBOSITY = 1 in Makefile to see compiler and build
tools output. For case 1) -DROCKSDB_JEMALLOC was present in compiler
command line. For both 2) and 3) this flag was not present.
Reviewers: sdong
Reviewed By: sdong
Subscribers: andrewkr, dhruba
Differential Revision: https://reviews.facebook.net/D57321
Comparable with Snappy on comp ratio.
Implemented using Windows API, does not require external package.
Avaiable since Windows 8 and server 2012.
Use -DXPRESS=1 with CMake to enable.
Summary: It is useful to print out IO stats in flush jobs too. Extend options.compaction_measure_io_stats to flush jobs and raname it.
Test Plan: Try db_bench and see the stats are printed out.
Reviewers: yhchiang
Reviewed By: yhchiang
Subscribers: kradhakrishnan, yiwu, IslamAbdelRahman, leveldb, andrewkr, dhruba
Differential Revision: https://reviews.facebook.net/D56769
* Update db_bench_tool.cc
I fixed the wrong letters, LevelDB -> rocksDB, because I thought of LevelDB as the wrong presentation.
the following show my fix :
fprintf(stderr, "LevelDB: version %d.%d\n",
kMajorVersion, kMinorVersion);
----------------->
fprintf(stderr, "rocksDB: version %d.%d\n",
kMajorVersion, kMinorVersion);
* Update db_bench_tool.cc
* Update db_bench_tool.cc
Summary:
When a block based table file is opened, if prefetch_index_and_filter is true, it will prefetch the index and filter blocks, putting them into the block cache.
What this feature adds: when a L0 block based table file is opened, if pin_l0_filter_and_index_blocks_in_cache is true in the options (and prefetch_index_and_filter is true), then the filter and index blocks aren't released back to the block cache at the end of BlockBasedTableReader::Open(). Instead the table reader takes ownership of them, hence pinning them, ie. the LRU cache will never push them out. Meanwhile in the table reader, further accesses will not hit the block cache, thus avoiding lock contention.
Test Plan:
'export TEST_TMPDIR=/dev/shm/ && DISABLE_JEMALLOC=1 OPT=-g make all valgrind_check -j32' is OK.
I didn't run the Java tests, I don't have Java set up on my devserver.
Reviewers: sdong
Reviewed By: sdong
Subscribers: andrewkr, dhruba
Differential Revision: https://reviews.facebook.net/D56133
Summary:
When a block based table file is opened, if prefetch_index_and_filter is true, it will prefetch the index and filter blocks, putting them into the block cache.
What this feature adds: when a L0 block based table file is opened, if pin_l0_filter_and_index_blocks_in_cache is true in the options (and prefetch_index_and_filter is true), then the filter and index blocks aren't released back to the block cache at the end of BlockBasedTableReader::Open(). Instead the table reader takes ownership of them, hence pinning them, ie. the LRU cache will never push them out. Meanwhile in the table reader, further accesses will not hit the block cache, thus avoiding lock contention.
When the table reader is destroyed, it releases the pinned blocks (if there were any). This has to happen before the cache is destroyed, so I had to introduce a TableReader::Close(), to guarantee the order of destruction.
Test Plan:
Added two unit tests for this. Existing unit tests run fine (default is pin_l0_filter_and_index_blocks_in_cache=false).
DISABLE_JEMALLOC=1 OPT=-g make all valgrind_check -j32
Mac: OK.
Linux: with D55287 patched in it's OK.
Reviewers: sdong
Reviewed By: sdong
Subscribers: andrewkr, leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D54801
Summary: Refactored db_bench transaction stress tests so that they can be called from unit tests as well.
Test Plan: run new unit test as well as db_bench
Reviewers: yhchiang, IslamAbdelRahman, sdong
Reviewed By: IslamAbdelRahman
Subscribers: andrewkr, dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D55203
Test Plan:
Built and ran with gflags:
% ./db_bench
LevelDB: version 4.5
Date: Tue Feb 16 12:04:23 2016
CPU: 40 * Intel(R) Xeon(R) CPU E5-2660 v2 @ 2.20GHz
...
And without gflags:
% ./db_bench
Please install gflags to run rocksdb tools
%
Reviewers: sdong, igor
Reviewed By: igor
Subscribers: igor, dhruba
Differential Revision: https://reviews.facebook.net/D54243
Summary: It is reported that in compress benchmark in db_bench, zlib will cause an OOM. The suggestd fix was to clear the buffer.
Test Plan: Build and run compress benchmark.
Reviewers: IslamAbdelRahman, yhchiang, rven, andrewkr, kradhakrishnan, anthony
Reviewed By: anthony
Subscribers: leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D52857
Summary:
1) changes tools/{benchmark,run_flash_bench}.sh to optionally use the write rate limit
2) removes code for --writes_per_second and switches the 'background' write rate limit
to use --benchmark_write_rate_limit
Replaces https://reviews.facebook.net/D49113
Task ID: #9555881
Blame Rev:
Test Plan:
tools/run_flash_bench.sh
Revert Plan:
Database Impact:
Memcache Impact:
Other Notes:
EImportant:
- begin *PUBLIC* platform impact section -
Bugzilla: #
- end platform impact -
Reviewers: igor
Reviewed By: igor
Subscribers: dhruba
Differential Revision: https://reviews.facebook.net/D52485
Summary: After removing two move operations, we can make CLANG 3.7 build pass under GCC 4.8.1.
Test Plan: USE_CLANG=1 ROCKSDB_FBCODE_BUILD_WITH_481=1 make all -j32
Reviewers: yhchiang, IslamAbdelRahman, rven, anthony
Reviewed By: anthony
Subscribers: leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D52365
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 now have a mechanism to further slowdown writes. Double default options.delayed_write_rate to try to keep the default behavior closer to it used to be.
Test Plan: Run all tests.
Reviewers: IslamAbdelRahman
Reviewed By: IslamAbdelRahman
Subscribers: yhchiang, kradhakrishnan, rven, leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D52281
Summary: Fix a bug that options.soft_pending_compaction_bytes_limit is not actually set with --soft_pending_compaction_bytes_limit
Test Plan: Run db_bench with this parameter and make sure the parameter is set correctly.
Reviewers: anthony, kradhakrishnan, yhchiang, IslamAbdelRahman, igor, rven
Reviewed By: rven
Subscribers: leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D52125
Summary:
List of changes:
1) Fix the snprintf() usage in cases where wrong variable was used to determine the output buffer size.
2) Remove unnecessary checks before calling delete operator.
3) Increase code correctness by using size_t type when getting vector's size.
4) Unify the coding style by removing namespace::std usage at the top of the file to confirm to the majority usage.
5) Fix various lint errors pointed out by 'arc lint'.
Test Plan:
Code review and build:
git diff
make clean
make -j 32 commit-prereq
arc lint
Reviewers: kradhakrishnan, sdong, rven, anthony, yhchiang, igor
Reviewed By: igor
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D51849
Summary: Deprecate options.soft_rate_limit, which is hard to tune, with options.soft_pending_compaction_bytes_limit, which would trigger the slowdown if estimated pending compaction bytes exceeds the threshold. The hope is to make it more striaght-forward to tune.
Test Plan: Modify DBTest.SoftLimit to cover options.soft_pending_compaction_bytes_limit instead; run all unit tests.
Reviewers: IslamAbdelRahman, yhchiang, rven, kradhakrishnan, igor, anthony
Reviewed By: anthony
Subscribers: leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D51117
Summary: Now in benchmark "uncompress" in db_bench, we get size from compressed stream for all other compression types except Snappy, where we allocate memory based on parameter. Change it to match to behavior of other compression types.
Test Plan: Run ./db_bench --benchmarks=uncompress with snappy and other compression types.
Reviewers: yhchiang, kradhakrishnan, anthony, IslamAbdelRahman, igor
Reviewed By: igor
Subscribers: leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D51681
Enable C4307 'operator' : integral constant overflow
Longs and ints on Windows are 32-bit hence the overflow
Enable C4309 'conversion' : truncation of constant value
Enable C4512 'class' : assignment operator could not be generated
Enable C4701 Potentially uninitialized local variable 'name' used
Summary: Currently db_bnech's --compaction_pri default is set to be rocksdb::Options().compaction_style. Change it to rocksdb::Options().compaction_pri. Although, for now both is 0.
Test Plan: Build db_bench
Reviewers: anthony, rven, IslamAbdelRahman, kradhakrishnan, igor
Reviewed By: igor
Subscribers: leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D49773
Yi Wu