Summary:
Here's my proposal for making our LOGs easier to read by machines.

The idea is to dump all events as JSON objects. JSON is easy to read by humans, but more importantly, it's easy to read by machines. That way, we can parse this, load into SQLite/mongo and then query or visualize.

I started with table_create and table_delete events, but if everybody agrees, I'll continue by adding more events (flush/compaction/etc etc)

Test Plan:
Ran db_bench. Observed:
2015/01/15-14:13:25.788019 1105ef000 EVENT_LOG_v1 {"time_micros": 1421360005788015, "event": "table_file_creation", "file_number": 12, "file_size": 1909699}
2015/01/15-14:13:25.956500 110740000 EVENT_LOG_v1 {"time_micros": 1421360005956498, "event": "table_file_deletion", "file_number": 12}

Reviewers: yhchiang, rven, dhruba, MarkCallaghan, lgalanis, sdong

Reviewed By: sdong

Subscribers: dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D31647

[RocksJava] Fix JTest issues with enabled assertions.

Closes https://github.com/facebook/rocksdb/issues/523

[RocksJava] Fix broken jacoco report

Summary:
we are using uint64_t for Wnew this is not correct since this value can be negative
https://github.com/facebook/rocksdb/issues/535

Test Plan: run db_bench and check what happens when Wnew is -ve

Reviewers: sdong, igor

Reviewed By: igor

Subscribers: dhruba

Differential Revision: https://reviews.facebook.net/D34935

[RocksJava] Fix NativeLibraryLoader

With the last folder layout change a bug was introduced which
prevents Jacoco from working correctly.

- Resolve problem while using a temporary data folder
- Fix test

Summary:
The functions and global symbols in xxhash.h and xxhash.cc were not in any namespace.
This caused issues when rocksdb library was being used along with other uses of libraries
with the same name

Test Plan:
unit tests

Reviewers:

CC:

Task ID: #

Blame Rev:

Summary: Remove unnecessary rocksdb::kInlineSize, since it's not used and there is rocksdb::Arena::kInlineSize.

Test Plan: make all check

Reviewers: igor

Reviewed By: igor

Subscribers: dhruba

Differential Revision: https://reviews.facebook.net/D34905

Summary:
The tests using sync_point for intent to shutdown stop
compaction and this results in stalls if too many rows are written. We
now limit the number of rows written to prevent stalls, since the focus
of the test is to cancel background work, which is being correctly
tested. This fixes a Jenkins issue.

Test Plan: DBTest.PreShutdown*

Reviewers: sdong, igor

Reviewed By: igor

Subscribers: dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D34893

Summary:
Fixing segmentation fault when running db_bench

This seg fault happens because num_created is used without being initialized

Test Plan:
running db_bench using these arguments
bpl=10485760;overlap=10;mcz=2;del=300000000;levels=6;ctrig=4; delay=8; stop=12; wbn=3; mbc=20; mb=67108864;wbs=134217728; dds=0; sync=0; r=1000000; t=1; vs=800; bs=65536; cs=1048576; of=500000; si=1000000; ./db_bench --benchmarks=overwrite --disable_seek_compaction=1 --mmap_read=0 --statistics=1 --histogram=1 --num=$r --threads=$t --value_size=$vs --block_size=$bs --cache_size=$cs --bloom_bits=10 --cache_numshardbits=4 --open_files=$of --verify_checksum=1 --db=/home/tec/koko/ --sync=$sync --disable_wal=1 --compression_type=zlib --stats_interval=$si --compression_ratio=0.5 --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=1

Reviewers: sdong, igor

Reviewed By: igor

Subscribers: dhruba

Differential Revision: https://reviews.facebook.net/D34881

Change the way options.compression_per_level is used when options.level_compaction_dynamic_level_bytes=true

Summary:
Change the way options.compression_per_level is used when options.level_compaction_dynamic_level_bytes=true so that options.compression_per_level[1] determines compression for the level L0 is merged to, options.compression_per_level[2] to the level after that, etc.

Test Plan: run all tests

Reviewers: rven, yhchiang, kradhakrishnan, igor

Reviewed By: igor

Subscribers: yoshinorim, leveldb, dhruba

Differential Revision: https://reviews.facebook.net/D34431

Summary: Fixed a bug where CompactFiles won't delete obsolete files until flush.

Test Plan:
./compact_files_test
export ROCKSDB_TESTS=CompactFiles
./db_test

Reviewers: rven, sdong, igor

Reviewed By: igor

Subscribers: dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D34671

Summary:
Fixed a compile error in db_bench in mac.

Test Plan:
make db_bench

Summary:
The build process now requires new source files to be added to
src.mk. Adding convenience.cc to src.mk

Test Plan: Build rocksdb

Reviewers: igor, sdong

Reviewed By: sdong

Subscribers: dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D34815

Summary: Use a better way to map from a key with locality to a random location. Now with the same -read_random_exp_range setting, hit rate drops, which it is expected.

Test Plan: ./db_bench --benchmarks=readrandom -statistics -use_existing_db -cache_size=5000000 --read_random_exp_range=<multiple_values>

Reviewers: MarkCallaghan, kradhakrishnan, igor

Reviewed By: igor

Subscribers: leveldb, dhruba

Differential Revision: https://reviews.facebook.net/D34761

Summary:
Provide an API which enables users to infor Rocksdb that it is
shutting down.

Test Plan: db_test

Reviewers: sdong, igor

Reviewed By: igor

Subscribers: dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D34617

Summary: Got it working by some voodoo programming

Test Plan: works!

Reviewers: sdong

Reviewed By: sdong

Subscribers: dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D34611

Summary:
One of the commands in the script make_new_version.sh was
incorrect. Fixed it.

Test Plan: Try the script out to verify that it works.

Reviewers: igor, sdong

Reviewed By: sdong

Subscribers: dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D34791

Summary: Allow GetThreadList() to report the start time of the current operation.

Test Plan:
./db_bench --benchmarks=fillrandom --num=100000 --threads=40 \
  --max_background_compactions=10 --max_background_flushes=3 \
  --thread_status_per_interval=1000 --key_size=16 --value_size=1000 \
  --num_column_families=10

Sample output:
          ThreadID ThreadType                    cfName    Operation        OP_StartTime         State
   140338840797248   High Pri column_family_name_000003        Flush 2015/03/09-17:49:59
   140338844991552   High Pri column_family_name_000004        Flush 2015/03/09-17:49:59
   140338849185856    Low Pri
   140338983403584    Low Pri
   140339008569408    Low Pri
   140338861768768    Low Pri
   140338924683328    Low Pri
   140338899517504    Low Pri
   140338853380160    Low Pri
   140338882740288    Low Pri
   140338865963072   High Pri column_family_name_000006        Flush 2015/03/09-17:49:59
   140338954043456    Low Pri
   140338857574464    Low Pri

Reviewers: igor, rven, sdong

Reviewed By: sdong

Subscribers: lgalanis, dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D34689

Summary: Introduce parameter -read_random_exp_range in db_bench to provide some key skewness in readrandom and multireadrandom benchmarks. It will helpful to cover block cache better.

Test Plan:
Run benchmarks with this new parameter. I can clearly see block cache hit rate change while I increase this value (DB size is about 66MB):

./db_bench --benchmarks=readrandom -statistics -use_existing_db -cache_size=5000000 --read_random_exp_range=0.0
rocksdb.block.cache.data.miss COUNT : 958418
rocksdb.block.cache.data.hit COUNT : 41582

./db_bench --benchmarks=readrandom -statistics -use_existing_db -cache_size=5000000 --read_random_exp_range=5.0
rocksdb.block.cache.data.miss COUNT : 819518
rocksdb.block.cache.data.hit COUNT : 180482

./db_bench --benchmarks=readrandom -statistics -use_existing_db -cache_size=5000000 --read_random_exp_range=10.0
rocksdb.block.cache.data.miss COUNT : 450479
rocksdb.block.cache.data.hit COUNT : 549521

./db_bench --benchmarks=readrandom -statistics -use_existing_db -cache_size=5000000 --read_random_exp_range=20.0
rocksdb.block.cache.data.miss COUNT : 223192
rocksdb.block.cache.data.hit COUNT : 776808

Reviewers: MarkCallaghan, kradhakrishnan, yhchiang, rven, igor

Reviewed By: igor

Subscribers: leveldb, dhruba

Differential Revision: https://reviews.facebook.net/D34629

Summary: I want to be able to set this through mongo config.

Test Plan: added unit test

Reviewers: sdong

Reviewed By: sdong

Subscribers: dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D34599

Single threaded tests -> sync=0 Multi threaded tests -> sync=1 by default unless DB_BENCH_NO_SYNC is defined

Summary:
Single threaded tests -> sync=0 Multi threaded tests -> sync=1 by default unless DB_BENCH_NO_SYNC is defined.

Also added updaterandom and mergerandom with putOperator. I am waiting for some results from udb on this.

Test Plan:
DB_BENCH_NO_SYNC=1 WAL_DIR=/tmp OUTPUT_DIR=/tmp/b DB_DIR=/tmp ./tools/benchmark.sh debug,bulkload,fillseq,overwrite,filluniquerandom,readrandom,readwhilewriting,updaterandom,mergerandom

WAL_DIR=/tmp OUTPUT_DIR=/tmp/b DB_DIR=/tmp ./tools/benchmark.sh debug,bulkload,fillseq,overwrite,filluniquerandom,readrandom,readwhilewriting,updaterandom,mergerandom

Verify sync settings

Reviewers: sdong, MarkCallaghan, igor, rven

Reviewed By: igor, rven

Subscribers: dhruba

Differential Revision: https://reviews.facebook.net/D34185

Summary:
Add --thread_status_per_interval to db_bench, which allows
db_bench to optionally enable print the current thread status
periodically.

Test Plan:
./db_bench --benchmarks=fillrandom --num=100000 --threads=40 --max_background_compactions=10 --max_background_flushes=3 --thread_status_per_interval=1000 --key_size=16 --value_size=1000 --num_column_families=10

Sample output:
          ThreadID ThreadType                         dbName                     cfName       Operation           State
   140281571770432    Low Pri
   140281575964736   High Pri  /tmp/rocksdbtest-5297/dbbench  column_family_name_000001           Flush
   140281710182464    Low Pri  /tmp/rocksdbtest-5297/dbbench  column_family_name_000008      Compaction
   140281638879296    Low Pri  /tmp/rocksdbtest-5297/dbbench  column_family_name_000007      Compaction
   140281592741952    Low Pri
   140281580159040   High Pri  /tmp/rocksdbtest-5297/dbbench  column_family_name_000002           Flush
   140281676628032    Low Pri  /tmp/rocksdbtest-5297/dbbench  column_family_name_000006      Compaction
   140281584353344    Low Pri
   140281622102080    Low Pri  /tmp/rocksdbtest-5297/dbbench  column_family_name_000009      Compaction
   140281605324864    Low Pri  /tmp/rocksdbtest-5297/dbbench  column_family_name_000004      Compaction
   140281601130560   High Pri  /tmp/rocksdbtest-5297/dbbench                    default           Flush
   140281596936256    Low Pri
   140281588547648    Low Pri

Reviewers: igor, rven, sdong

Reviewed By: rven

Subscribers: dhruba

Differential Revision: https://reviews.facebook.net/D34515

Summary:
* src.mk (LIB_SOURCES): Add this file:
utilities/document/json_document_builder.cc
to avoid link errors.

Blame Rev: D33849

Test Plan: run "make"

Reviewers: ljin, igor.sugak, rven, sdong, igor

Reviewed By: igor

Subscribers: dhruba

Differential Revision: https://reviews.facebook.net/D34593

Summary:
Any time one would modify a dependent of any *test*.cc file,
"make" would fail to rebuild the affected test binaries,
e.g., db_test.  That was due to the fact that we deliberately
excluded those test-related files from the definition of SOURCES
and only $(SOURCES) was used to create the automatically-generated
.d dependency files.  The fix is to generate a .d file for every
source file.
* src.mk: New file.  Defines LIB_SOURCES, MOCK_SOURCES
and TEST_BENCH_SOURCES.
* Makefile: Include src.mk.
Reflect s/SOURCES/LIB_SOURCES/ renaming.
* build_tools/build_detect_platform: Remove the code
that was used to generate SOURCES= and MOCK_SOURCES=
definitions in make_config.mk. Those lists of files
are now hard-coded in src.mk. Hard-coding this list of
sources is desirable, because without that, one risks
including stray .cc files in a build.  Not reproducible.

Test Plan:
Touch a file used by db_test's dependent .o files and ensure that
they are all recompiled.  Before, none would be:

  $ touch db/db_impl.h && make db_test
    CC       db/db_test.o
    CC       db/column_family.o
    CC       db/db_filesnapshot.o
    CC       db/db_impl.o
    CC       db/db_impl_debug.o
    CC       db/db_impl_readonly.o
    CC       db/forward_iterator.o
    CC       db/internal_stats.o
    CC       db/managed_iterator.o
    CC       db/repair.o
    CC       db/write_batch.o
    CC       utilities/compacted_db/compacted_db_impl.o
    CC       utilities/ttl/db_ttl_impl.o
    CC       util/ldb_cmd.o
    CC       util/ldb_tool.o
    CC       util/sst_dump_tool.o
    CC       util/xfunc.o
    CCLD     db_test

Reviewers: ljin, igor.sugak, igor, rven, sdong

Reviewed By: sdong

Subscribers: yhchiang, adamretter, fyrz, dhruba

Differential Revision: https://reviews.facebook.net/D33849

Summary:
The problem appears to be caused by a bug in Mac OS X compiler
(http://llvm.org/bugs/show_bug.cgi?id=15337). We need explicitly construct the
base object std::ostream(std::streambuf*) with nullptr. Otherwise, ostream will
try to delete the underlying streambuf* which apparently is undefined in the
Mac OS X compiler.

https://github.com/facebook/rocksdb/issues/525

Test Plan:
unit test in fbson
make all check
document_db_test (on mac)

Reviewers: IslamAbdelRahman, igor

Reviewed By: igor

Subscribers: dhruba

Differential Revision: https://reviews.facebook.net/D34587

Summary:
Cleaned up some tests regarding disposal order and tests
which were failing when C++ assertions were enabled.

Test Plan:
- Enable C++ Assertions (remove e.g. -DNDebug in rocksdbjava target)
- make rocksdbjava jtest

Reviewers: adamretter, ankgup87, yhchiang

Subscribers: dhruba

Differential Revision: https://reviews.facebook.net/D34569

In the current implementation DefaultColumnFamily will not disown
the native handle. As the database handles the lease on the native
handle this should be the case.

Cleanup column family memory cleanup

Summary: Fix a bug in stall time counter.  Improve its output format.

Test Plan:
export ROCKSDB_TESTS=Timeout
./db_test

./db_bench --benchmarks=fillrandom --stats_interval=10000 --statistics=true --stats_per_interval=1 --num=1000000 --threads=4 --level0_stop_writes_trigger=3 --level0_slowdown_writes_trigger=2

sample output:
    Uptime(secs): 35.8 total, 0.0 interval
    Cumulative writes: 359590 writes, 359589 keys, 183047 batches, 2.0 writes per batch, 0.04 GB user ingest, stall seconds: 1786.008 ms
    Cumulative WAL: 359591 writes, 183046 syncs, 1.96 writes per sync, 0.04 GB written
    Interval writes: 253 writes, 253 keys, 128 batches, 2.0 writes per batch, 0.0 MB user ingest, stall time: 0 us
    Interval WAL: 253 writes, 128 syncs, 1.96 writes per sync, 0.00 MB written

Reviewers: MarkCallaghan, igor, sdong

Reviewed By: sdong

Subscribers: dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D34275

Summary:
For some reason, libstdc++ implements steady_clock::now() using syscall instead of VDSO optimized clock_gettime() when using glibc 2.16 and earlier. This leads to significant performance degradation for users with older glibcs. See bug reported here: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=59177

We observed this behavior when testing mongo on AWS hosts. Facebook hosts are unaffected since we use glibc2.17 and 2.20.

Revert "Fix timing"
This reverts commit 965d9d50.

Revert "Use chrono for timing"
This reverts commit 001ce64d.

Test Plan: make check

Reviewers: MarkCallaghan, yhchiang, rven, meyering, sdong

Reviewed By: sdong

Subscribers: dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D34371

options.level_compaction_dynamic_level_bytes to allow RocksDB to pick size bases of levels dynamically.

Summary:
When having fixed max_bytes_for_level_base, the ratio of size of largest level and the second one can range from 0 to the multiplier. This makes LSM tree frequently irregular and unpredictable. It can also cause poor space amplification in some cases.

In this improvement (proposed by Igor Kabiljo), we introduce a parameter option.level_compaction_use_dynamic_max_bytes. When turning it on, RocksDB is free to pick a level base in the range of (options.max_bytes_for_level_base/options.max_bytes_for_level_multiplier, options.max_bytes_for_level_base] so that real level ratios are close to options.max_bytes_for_level_multiplier.

Test Plan: New unit tests and pass tests suites including valgrind.

Reviewers: MarkCallaghan, rven, yhchiang, igor, ikabiljo

Reviewed By: ikabiljo

Subscribers: yoshinorim, ikabiljo, dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D31437

Summary:
Pre-fetching is a common operation performed by data stores for
disk/flash based systems as part of database startup.

This is part of task 5197184.

Test Plan: Run the newly added unit test

Reviewers: rven, igor, sdong

Reviewed By: sdong

Subscribers: dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D33933

Summary:
fix typo

Task ID: #

Blame Rev:

Test Plan:
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/D34251

Summary: As title

Test Plan: compiles

Reviewers: sdong

Reviewed By: sdong

Subscribers: dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D34191