Commit Graph

4561 Commits

Author SHA1 Message Date
Dmitri Smirnov
cb9459f85c Fix empty vector write in ForwardIterator 2015-11-16 13:58:10 -08:00
Islam AbdelRahman
a163cc2d5a Lint everything
Summary:
```
arc2 lint --everything
```

run the linter on the whole code repo to fix exisitng lint issues

Test Plan: make check -j64

Reviewers: sdong, rven, anthony, kradhakrishnan, yhchiang

Reviewed By: yhchiang

Subscribers: dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D50769
2015-11-16 12:56:21 -08:00
Dmitri Smirnov
8f01f25414 Remove forward_iter_bench from Win build. 2015-11-16 11:10:58 -08:00
sdong
dac5b248b1 UniversalCompactionPicker::PickCompaction(): avoid to form compactions if there is no file
Summary:
Currently RocksDB may break in lines like this:

for (size_t i = sorted_runs.size() - 1; i >= first_index_after; i--) {

if options.level0_file_num_compaction_trigger=0.

Fix it by not executing the logic of picking compactions if there is no file (sorted_runs.size() = 0). Also internally set options.level0_file_num_compaction_trigger=1 if users give a 0. 0 is a value makes no sense in RocksDB.

Test Plan: Run all tests. Will add a unit test too.

Reviewers: yhchiang, IslamAbdelRahman, anthony, kradhakrishnan, rven

Reviewed By: rven

Subscribers: leveldb, dhruba

Differential Revision: https://reviews.facebook.net/D50727
2015-11-16 10:32:45 -08:00
Venkatesh Radhakrishnan
d06b63e99f Fix Rocksdb lite build failure in forward_iterator_bench
Summary:
Fixed Rocksdb lite build failure in forward_iterator_bench by
defining main for the ROCKSDB_LITE case

Test Plan: build ROCKSDB_LITE

Reviewers: anthony, yhchiang, IslamAbdelRahman, sdong

Reviewed By: sdong

Subscribers: dhruba

Differential Revision: https://reviews.facebook.net/D50733
2015-11-16 09:57:08 -08:00
Venkatesh Radhakrishnan
7824444bfc Reuse file iterators in tailing iterator when memtable is flushed
Summary:
Under a tailing workload, there were increased block cache
misses when a memtable was flushed because we were rebuilding iterators
in that case since the version set changed. This was exacerbated in the
case of iterate_upper_bound, since file iterators which were over the
iterate_upper_bound would have been deleted and are now brought back as
part of the Rebuild, only to be deleted again. We now renew the iterators
and only build iterators for files which are added and delete file
iterators for files which are deleted.
Refer to https://reviews.facebook.net/D50463 for previous version

Test Plan: DBTestTailingIterator.TailingIteratorTrimSeekToNext

Reviewers: anthony, IslamAbdelRahman, igor, tnovak, yhchiang, sdong

Reviewed By: sdong

Subscribers: yhchiang, march, dhruba, leveldb, lovro

Differential Revision: https://reviews.facebook.net/D50679
2015-11-13 15:50:59 -08:00
Venkatesh Radhakrishnan
2ae4d7d708 Make sure that CompactFiles does not run two parallel Level 0 compactions
Summary:
Since level 0 files can overlap, two level 0 compactions cannot
run in parallel. Compact files needs to check this before running a
compaction.

Test Plan: CompactFilesTest.L0ConflictsFiles

Reviewers: igor, IslamAbdelRahman, anthony, sdong, yhchiang

Reviewed By: yhchiang

Subscribers: dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D50079
2015-11-13 12:01:00 -08:00
Yueh-Hsuan Chiang
d781da8164 Add CheckOptionsCompatibility() API to options_util
Summary:
Add CheckOptionsCompatibility() API to options_util that returns
Status::OK if the input DBOptions and ColumnFamilyDescriptors
are compatible with the latest options stored in the specified DB path.

Test Plan: Added tests in options_util_test

Reviewers: igor, anthony, IslamAbdelRahman, rven, sdong

Reviewed By: sdong

Subscribers: dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D50649
2015-11-12 16:52:51 -08:00
Siying Dong
2391b459b8 Merge pull request #824 from yuslepukhin/try_ci_tests_on_daily
Try running db_test during integration build
2015-11-12 15:52:41 -08:00
yuslepukhin
2ab3e2df2c Fix a build break so tests can run 2015-11-12 14:53:19 -08:00
yuslepukhin
247c49a40f Merge branch 'master' into try_ci_tests_on_daily 2015-11-12 14:51:19 -08:00
yuslepukhin
935d1495c5 Run tests imporvements
Add sequential rerun for any failed tests. Add env_test case.
  Limit concurrency
  Allow to specify individual tests
  Take $Limit into account when displaying number of tests
2015-11-12 14:42:38 -08:00
Yueh-Hsuan Chiang
5ac16300b0 Fixed valgrind error in options_util_test
Summary:
Fixed valgrind error in options_util_test by deleting the
compaction_filter allocated from RandomInitCFOptions().

Test Plan: valgrind --error-exitcode=2 --leak-check=full ./options_util_test

Reviewers: anthony, sdong

Reviewed By: sdong

Subscribers: dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D50661
2015-11-12 14:12:27 -08:00
Nathan Bronson
6ce42dd075 Don't merge WriteBatch-es if WAL is disabled
Summary:
There's no need for WriteImpl to flatten the write batch group
into a single WriteBatch if the WAL is disabled.  This diff moves the
flattening into the WAL step, and skips flattening entirely if it isn't
needed.  It's good for about 5% speedup on a multi-threaded workload
with no WAL.

This diff also adds clarifying comments about the chance for partial
failure of WriteBatchInternal::InsertInto, and always sets bg_error_ if
the memtable state diverges from the logged state or if a WriteBatch
succeeds only partially.

Benchmark for speedup:
  db_bench -benchmarks=fillrandom -threads=16 -batch_size=1 -memtablerep=skip_list -value_size=0 --num=200000 -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

Test Plan: asserts + make check

Reviewers: sdong, igor

Reviewed By: igor

Subscribers: dhruba

Differential Revision: https://reviews.facebook.net/D50583
2015-11-12 10:50:38 -08:00
Yueh-Hsuan Chiang
56245ddcf5 Fixed DBCompactionTest.SkipStatsUpdateTest
Summary:
DBCompactionTest.SkipStatsUpdateTest relies on the number
of files opened during the DB::Open process, but the persisting
options file support altered this number and thus makes
DBCompactionTest.SkipStatsUpdateTest in certain environment.

This patch fixed this test failure.

Test Plan: db_compaction_test

Reviewers: igor, sdong, anthony, IslamAbdelRahman

Subscribers: dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D50637
2015-11-12 07:45:53 -08:00
Yueh-Hsuan Chiang
e11f676e34 Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file.  Note that most pointer-typed options such as
merge_operator will not be constructed.

With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:

  DBOptions db_options;
  std::vector<std::string> cf_names;
  std::vector<ColumnFamilyOptions> cf_opts;

  // Load primitive-typed options from an existing DB
  OptionsUtil::LoadLatestOptionsFromDB(
      dbname, &db_options, &cf_names, &cf_opts);

  // Initialize necessary pointer-typed options
  cf_opts[0].merge_operator.reset(new MyMergeOperator());
  ...

  // Construct the vector of ColumnFamilyDescriptor
  std::vector<ColumnFamilyDescriptor> cf_descs;
  for (size_t i = 0; i < cf_opts.size(); ++i) {
    cf_descs.emplace_back(cf_names[i], cf_opts[i]);
  }

  // Open the DB
  DB* db = nullptr;
  std::vector<ColumnFamilyHandle*> cf_handles;
  auto s = DB::Open(db_options, dbname, cf_descs,
                    &handles, &db);

Test Plan:
Augment existing tests in column_family_test
options_test
db_test

Reviewers: igor, IslamAbdelRahman, sdong, anthony

Reviewed By: anthony

Subscribers: dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 06:52:43 -08:00
Yueh-Hsuan Chiang
e78389b554 Fixed build failure of RocksDBLite test on options_file_test.cc
Summary: Fixed build failure of RocksDBLite test

Test Plan: options_file_test

Reviewers: igor, sdong, anthony, IslamAbdelRahman

Subscribers: dhruba

Differential Revision: https://reviews.facebook.net/D50595
2015-11-10 23:23:36 -08:00
Yueh-Hsuan Chiang
e114f0abb8 Enable RocksDB to persist Options file.
Summary:
This patch allows rocksdb to persist options into a file on
DB::Open, SetOptions, and Create / Drop ColumnFamily.
Options files are created under the same directory as the rocksdb
instance.

In addition, this patch also adds a fail_if_missing_options_file in DBOptions
that makes any function call return non-ok status when it is not able to
persist options properly.

  // If true, then DB::Open / CreateColumnFamily / DropColumnFamily
  // / SetOptions will fail if options file is not detected or properly
  // persisted.
  //
  // DEFAULT: false
  bool fail_if_missing_options_file;

Options file names are formatted as OPTIONS-<number>, and RocksDB
will always keep the latest two options files.

Test Plan:
Add options_file_test.

options_test
column_family_test

Reviewers: igor, IslamAbdelRahman, sdong, anthony

Reviewed By: anthony

Subscribers: dhruba

Differential Revision: https://reviews.facebook.net/D48285
2015-11-10 22:58:01 -08:00
Siying Dong
7ed2c3e45b Merge pull request #823 from yuslepukhin/fix_off_t_type
Make use of portable `uint64_t` type to make possible 64-bit file access
2015-11-10 18:57:17 -08:00
Dmitri Smirnov
7f59e33b19 Make CI build debug/optimized 2015-11-10 17:28:56 -08:00
Dmitri Smirnov
ae2dfe404b Try running db_test during integration build 2015-11-10 17:23:16 -08:00
Dmitri Smirnov
720af2269a Merge branch 'fix_off_t_type' of https://github.com/yuslepukhin/rocksdb into fix_off_t_type 2015-11-10 17:04:23 -08:00
Dmitri Smirnov
5270b33bd3 Make use of portable uint64_t type to make possible file access
in 64-bit.

  Currently, a signed off_t type is being used for the following
  interfaces for both offset and the length in bytes:
  * `Allocate`
  * `RangeSync`

  On Linux `off_t` is automatically either 32 or 64-bit depending on
  the platform. On Windows it is always a 32-bit signed long which
  limits file access and in particular space pre-allocation
  to effectively 2 Gb.

  Proposal is to replace off_t with uint64_t as a portable type
  always access files with 64-bit interfaces.

  May need to modify posix code but lack resources to test it.
2015-11-10 17:03:42 -08:00
Nathan Bronson
631863c63b track WriteBatch contents
Summary:
Parallel writes will only be possible for certain combinations of
flags and WriteBatch contents.  Traversing the WriteBatch at write time
to check these conditions would be expensive, but it is very cheap to
keep track of when building WriteBatch-es.  When loading WriteBatch-es
during recovery, a deferred computation state is used so that the flags
never need to be computed.

Test Plan:
1. add asserts and EXPECT_EQ-s
2. make check

Reviewers: sdong, igor

Reviewed By: igor

Subscribers: dhruba

Differential Revision: https://reviews.facebook.net/D50337
2015-11-10 16:56:06 -08:00
Nathan Bronson
505accda38 remove constexpr from util/random.h for MSVC compat
Summary:
Scoped anonymous enums seem to be better supported than static
constexpr at the moment, so this diff replaces the latter with the former.
Also, this diff removes an incorrect inclusion of pthread.h.  MSVC build
was broken starting with D50439.

Test Plan:
1. build
2. observe proper skiplist behavior by absence of pathological slowdown
3. push diff to tmp_try_windows branch to tickle AppVeyor
4. wait for contbuild before committing to master

Reviewers: sdong

Reviewed By: sdong

Subscribers: dhruba

Differential Revision: https://reviews.facebook.net/D50517
2015-11-10 16:41:23 -08:00
Dmitri Smirnov
5421c9728b Make use of portable uint64_t type to make possible file access
in 64-bit.

  Currently, a signed off_t type is being used for the following
  interfaces for both offset and the length in bytes:
  * `Allocate`
  * `RangeSync`

  On Linux `off_t` is automatically either 32 or 64-bit depending on
  the platform. On Windows it is always a 32-bit signed long which
  limits file access and in particular space pre-allocation
  to effectively 2 Gb.

  Proposal is to replace off_t with uint64_t as a portable type
  always access files with 64-bit interfaces.

  May need to modify posix code but lack resources to test it.
2015-11-10 12:58:39 -08:00
Nathan Bronson
b81b430987 Switch to thread-local random for skiplist
Summary:
Using a TLS random instance for skiplist makes it smaller
(useful for hash_skiplist_rep) and prepares skiplist for concurrent
adds.  This diff also modifies the branching factor math to avoid an
unnecessary division.

This diff has the effect of changing the sequence of skip list node
height choices made by tests, so it has the potential to cause unit
test failures for tests that implicitly rely on the exact structure
of the skip list.  Tests that try to exactly trigger a compaction are
likely suspects for this problem (these tests have always been brittle to
changes in the skiplist details).  I've minimizes this risk by reseeding
the main thread's Random at the beginning of each test, increasing the
universal compaction size_ratio limit from 101% to 105% for some tests,
and verifying that the tests pass many times.

Test Plan: for i in `seq 0 9`; do make check; done

Reviewers: sdong, igor

Reviewed By: igor

Subscribers: dhruba

Differential Revision: https://reviews.facebook.net/D50439
2015-11-09 19:25:22 -08:00
Islam AbdelRahman
75a8bad2ab Merge pull request #821 from yuslepukhin/continue_windows_warnings
Enable C4200, C4702, C4305 windows warnings
2015-11-09 13:34:49 -08:00
Nathan Bronson
986230b8c8 Revert "Fix TSAN build for fbcode"
Summary:
Reverting c745f1d2c4 because it
was based on an incorrect understanding of the correct way to enable
TSAN tests (it assumes "make COMPILE_WITH_TSAN=1 check" but in fact only
"COMPILE_WITH_TSAN=1 make check" is supported).

Test Plan: COMPILE_WITH_TSAN=1 make check

Reviewers: kradhakrishnan

Reviewed By: kradhakrishnan

Subscribers: dhruba

Differential Revision: https://reviews.facebook.net/D50445
2015-11-09 11:51:10 -08:00
Yueh-Hsuan Chiang
f3ca28ab03 Correct the comment of GetApproximateMemoryUsageByType
Summary: Correct the comment of GetApproximateMemoryUsageByType.

Test Plan: No code change.

Reviewers: igor, sdong, anthony, IslamAbdelRahman

Subscribers: dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D50409
2015-11-08 09:02:35 -08:00
Islam AbdelRahman
838676c17b Revert "Adding new table properties"
Summary:
Reverting https://reviews.facebook.net/D34269 for now
after I landed it a flaky test started continuously failing, I am almost sure this patch is not related to the test but I will revert it until I figure out why it's failing

Test Plan: make check

Reviewers: kradhakrishnan

Subscribers: dhruba

Differential Revision: https://reviews.facebook.net/D50385
2015-11-06 16:49:38 -08:00
Dmitri Smirnov
7c86d50497 Enable C4305 'identifier' : truncation from 'type1' to 'type2' 2015-11-06 16:44:39 -08:00
Dmitri Smirnov
85a2ce9c19 Enable C4702 unreachable code 2015-11-06 15:50:11 -08:00
Dmitri Smirnov
62aa1b1b78 Enable C4200 warning
nonstandard extension used : zero-sized array in struct/union
2015-11-06 15:32:45 -08:00
Islam AbdelRahman
5b9ce1a323 Merge pull request #820 from yuslepukhin/enable_compiler_warnings
Enable Windows warnings C4307 C4309 C4512 C4701
2015-11-06 12:08:25 -08:00
Dmitri Smirnov
20f57b1715 Enable Windows warnings C4307 C4309 C4512 C4701
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
2015-11-06 11:34:06 -08:00
Islam AbdelRahman
8be568a9c2 Adding new table properties
Summary:
This diff introduce new table properties that will be written for block based tables
These properties are
  - comparator name
  - merge operator name
  - property collectors names

Test Plan:
  - Added a new unit test to verify that these tests are written/read correctly
  - Running all other tests right now (wont land until all tests finish)

Reviewers: rven, kradhakrishnan, igor, sdong, anthony, yhchiang

Reviewed By: yhchiang

Subscribers: dhruba

Differential Revision: https://reviews.facebook.net/D34269
2015-11-06 11:19:01 -08:00
Nathan Bronson
2b42000f43 incorrect batch group size computation for write throttling
Summary:
When a write batch can't join a batch group due to the total
size of the contained batches, the write controller's GetDelay is passed
a size value that includes the rejected batch.

Test Plan: make check

Reviewers: sdong, igor

Reviewed By: igor

Subscribers: dhruba

Differential Revision: https://reviews.facebook.net/D50343
2015-11-06 09:23:55 -08:00
Nathan Bronson
c745f1d2c4 Fix TSAN build for fbcode
Summary:
TSAN builds for gcc 4.9 need a PIC version of the libraries
taken from the fbcode platform.  This is accomplished by assuming every
.a has a _pic.a sibling, and by fixing the third-party2 zlib build.

Test Plan: make COMPILE_WITH_TSAN=1 check

Reviewers: sdong, igor

Reviewed By: igor

Subscribers: dhruba

Differential Revision: https://reviews.facebook.net/D50331
2015-11-06 09:05:12 -08:00
agiardullo
fe789c5f2b Document SingleDelete
Summary: Docuemented what is currently supported by SingleDelete based on its current implementation.

Test Plan: n/a

Reviewers: sdong, kradhakrishnan, yhchiang, rven

Reviewed By: rven

Subscribers: dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D50205
2015-11-05 17:24:07 -08:00
Islam AbdelRahman
e89e5b253c Merge pull request #818 from yuslepukhin/improve_test_concurrency
Improve concurrency when running tests
2015-11-05 17:21:34 -08:00
Venkatesh Radhakrishnan
ae7940b628 Fix regression failure in PrefixTest.PrefixValid
Summary: Use IterKey to store prefix_start_ so that it doesn't get freed

Test Plan: PrefixTest.PrefixValid

Reviewers: anthony, IslamAbdelRahman

Reviewed By: IslamAbdelRahman

Subscribers: dhruba

Differential Revision: https://reviews.facebook.net/D50289
2015-11-05 16:43:54 -08:00
Dmitri Smirnov
3277d172ba Improve concurrency when running tests
PowerShell seems to have a hard time when a flood of async tasks is
  scheduled at the same time.
  I speculated that WaitForMultipleObjects() in Windows can only take up
  to 64 process handles and if you want to handle more than you should write
  some additional code which can be sub-optimal. I.e to implement Wait-Job -Any.
  I decided to test that suggestion and introduced a $Concurrency parameter with a default value of 62.
  So in the new version the script fires up up to $Concurrency value
  and wait for anything to complete before starting any more processes.
  This improved matters greatly. Individual tests against ramdrive now
  run in 8 minutes and all of the 200+ db_tests run in 9 minutes with concurrency
  values of 8-16. About 48 is required to load a CPU on my box running against HD
  but that does not improve running times much.

  Other changes include respect -EnableJE for the individual test exes.
  Enforce exclusions for the individual tests.
2015-11-05 14:03:41 -08:00
Satnam Singh
c8e01ef982 Delete test iterators
Summary:
Valgrind reports an issue with the test for GeoIterator.
This diff explicitly deletes the two iterators used in this test.

Test Plan: This diff is for a test. The test still passes.

Reviewers: IslamAbdelRahman

Reviewed By: IslamAbdelRahman

Subscribers: dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D50193
2015-11-05 13:30:51 -08:00
Venkatesh Radhakrishnan
9d50afc3b9 Prefix-based iterating only shows keys in prefix
Summary:
MyRocks testing found an issue that while iterating over keys
that are outside the prefix, sometimes wrong results were seen for keys
outside the prefix. We now tighten the range of keys seen with a new
read option called prefix_seen_at_start. This remembers the starting
prefix and then compares it on a Next for equality of prefix. If they
are from a different prefix, it sets valid to false.

Test Plan: PrefixTest.PrefixValid

Reviewers: IslamAbdelRahman, sdong, yhchiang, anthony

Reviewed By: anthony

Subscribers: spetrunia, hermanlee4, yoshinorim, dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D50211
2015-11-05 13:24:05 -08:00
Igor Canadi
14c6e1a04c Add write_stress to RocksDB Legocastle runs
Summary: As title. Let's run it for 1 hour.

Test Plan: How can I test legocastle changes?

Reviewers: IslamAbdelRahman, yhchiang, rven, sdong, anthony, kradhakrishnan

Reviewed By: kradhakrishnan

Subscribers: dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D49653
2015-11-05 12:07:39 -08:00
Yueh-Hsuan Chiang
db3f5e4948 Update HISTORY.md
Summary:
Move some of the items in public API changes section to to new
features section.

Test Plan: no code change.

Reviewers: igor, sdong, anthony, IslamAbdelRahman

Subscribers: dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D50229
2015-11-04 23:52:09 -08:00
Islam AbdelRahman
042fb053fd Fix clang
Summary: Fix build for clang

Test Plan:
USE_CLANG=1 make all -j64
make clean
make check -j64

Reviewers: yhchiang

Subscribers: dhruba

Differential Revision: https://reviews.facebook.net/D50217
2015-11-04 21:02:20 -08:00
Islam AbdelRahman
2419f435a1 Merge pull request #816 from SherlockNoMad/GeoDBTestFix
Fix appveyor build failure
2015-11-04 20:40:57 -08:00
SherlockNoMad
2e45409910 Fix appveyor build failure 2015-11-04 20:10:16 -08:00