- 06 2月, 2016 5 次提交
-
-
由 sdong 提交于
Summary: If users turn on concurrent insert but the memtable doesn't support it, they might see unexcepted crash. Fix it by explicitly fail. Test Plan: Run different setting of stress_test and make sure it fails correctly. Will add a unit test too. Reviewers: anthony, kradhakrishnan, IslamAbdelRahman, yhchiang, andrewkr, ngbronson Reviewed By: ngbronson Subscribers: leveldb, dhruba Differential Revision: https://reviews.facebook.net/D53895
-
由 Gunnar Kudrjavets 提交于
Summary: This set of changes is part of the work to introduce benchmark for universal style compaction in RocksDB. It's conceptually separate from the compaction work, so sending it out as a separate diff to get it out of the way. Test Plan: - Run `./tools/run_flash_bench.sh`. - Look at the contents of `report.txt` and `report2.txt` to make sure that data is reported and attributed correctly. - During `db_bench` execution time make sure that the correct flags are passed to `--disable_wal` depending on the benchmark being executed. Reviewers: MarkCallaghan Reviewed By: MarkCallaghan Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D53865
-
由 sdong 提交于
Summary: Time to cut branch for release 4.5. Change the versions. Test Plan: Not needed Reviewers: IslamAbdelRahman, yhchiang, kradhakrishnan, andrewkr, anthony Reviewed By: anthony Subscribers: leveldb, dhruba Differential Revision: https://reviews.facebook.net/D53883
-
由 reid horuff 提交于
Summary: copy from task 8196669: 1) Optimistic transactions do not support batching writes from different threads. 2) Pessimistic transactions do not support batching writes if an expiration time is set. In these 2 cases, we currently do not do any write batching in DBImpl::WriteImpl() because there is a WriteCallback that could decide at the last minute to abort the write. But we could support batching write operations with callbacks if we make sure to process the callbacks correctly. To do this, we would first need to modify write_thread.cc to stop preventing writes with callbacks from being batched together. Then we would need to change DBImpl::WriteImpl() to call all WriteCallback's in a batch, only write the batches that succeed, and correctly set the state of each batch's WriteThread::Writer. Test Plan: Added test WriteWithCallbackTest to write_callback_test.cc which creates multiple client threads and verifies that writes are batched and executed properly. Reviewers: hermanlee4, anthony, ngbronson Subscribers: leveldb, dhruba Differential Revision: https://reviews.facebook.net/D52863
-
由 Islam AbdelRahman 提交于
Summary: Add a new option to BlockBasedTableOptions that will allow us to change the restart interval for the index block Test Plan: unit tests Reviewers: yhchiang, anthony, andrewkr, sdong Reviewed By: sdong Subscribers: march, dhruba Differential Revision: https://reviews.facebook.net/D53721
-
- 05 2月, 2016 2 次提交
-
-
由 sdong 提交于
Summary: Add an option of --allow_concurrent_memtable_write in stress test and cover it in crash test Test Plan: Run crash test and make sure three combinations of the two options show up randomly. Reviewers: IslamAbdelRahman, yhchiang, andrewkr, anthony, kradhakrishnan Reviewed By: kradhakrishnan Subscribers: leveldb, dhruba Differential Revision: https://reviews.facebook.net/D53811
-
由 Siying Dong 提交于
Move out of release build some test code
-
- 04 2月, 2016 10 次提交
-
-
由 sdong 提交于
Summary: Unity build because perf_context.cc and iostats_context.cc define a different OUTPUT macro. Fix it. Test Plan: Build unity Reviewers: kradhakrishnan, anthony, yhchiang, andrewkr, IslamAbdelRahman Reviewed By: IslamAbdelRahman Subscribers: divchenko, leveldb, dhruba Differential Revision: https://reviews.facebook.net/D53769
-
https://github.com/facebook/rocksdb/issues/935由 Dmitri Smirnov 提交于
testutil.cc and testharness.cc could not be moved out at this time as they are used by 4 benchmarks in release builds.
-
由 Jonathan Wiepert 提交于
Summary: Remove obolete references to files in src.mk Fix incorrect path for reference in source.mk Test Plan: Ran build to ensure changes do not break anything. Reviewers: leveldb, sdong Reviewed By: sdong Subscribers: dhruba Differential Revision: https://reviews.facebook.net/D53733
-
由 krad 提交于
Summary: When making environment specific changes, it is better to run all CI tests. This diff provides a mechanism to do that Format is: ROCKSDB_CHECK_ALL=1 arc diff Test Plan: Submit request for diff Reviewers: sdong Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D53631
-
由 Igor Canadi 提交于
Threaded tests for WriteBatch
-
由 Adam Retter 提交于
-
由 krad 提交于
Summary: unit_481 is misspelt. Fixing it. Test Plan: Running make commit_prereq Reviewers: leveldb Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D53757
-
由 Nathan Bronson 提交于
Summary: InlineSkipList::InsertConcurrently should invalidate the sequential-insertion cache prev_[] for all inserts of multi-level nodes, not just those that increase the height of the skip list. The invariant for prev_ is that prev_[i] (i > 0) is supposed to be the predecessor of prev_[0] at level i. Before this diff InsertConcurrently could violate this constraint when inserting a multi-level node after prev_[i] but before prev_[0]. This diff also reenables kConcurrentSkipList as db_test's MultiThreaded/MultiThreadedDBTest.MultiThreaded/29. Test Plan: 1. unit tests 2. temporarily hack kConcurrentSkipList timing so that it is fast but has a 1.5% failure rate on my dev box (1ms stagger on thread launch, 1s test duration, failure rate baseline over 1000 runs) 3. observe 1000 passes post-fix Reviewers: igor, sdong Reviewed By: sdong Subscribers: MarkCallaghan, dhruba Differential Revision: https://reviews.facebook.net/D53751
-
由 Siying Dong 提交于
Adding support for Windows JNI build
-
由 Gunnar Kudrjavets 提交于
Summary: Some of the tests aren't considered to be critical when it comes to getting key benchmarking data for RocksDB. Therefore we'll introduce an environment variable `SKIP_LOW_PRI_TESTS` which enables skipping those test cases. By default all the tests will be run. If you want to optimize the test-case execution then do the following: ` $ export SKIP_LOW_PRI_TESTS=1 $ ./tools/run_flash_bench.sh ` Test Plan: Verified that when `SKIP_LOW_PRI_TESTS` is not set then `benchmark.sh` is called for all the scenarios and when `SKIP_LOW_PRI_TESTS` is set to `1` then `benchmark.sh` is called only for the test-cases which are critical. Reviewers: MarkCallaghan Reviewed By: MarkCallaghan Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D53739
-
- 03 2月, 2016 6 次提交
-
-
由 Andrew Kryczka 提交于
Summary: Before this diff, there were duplicated constants to refer to properties (user- facing API had strings and InternalStats had an enum). I noticed these were inconsistent in terms of which constants are provided, names of constants, and documentation of constants. Overall it seemed annoying/error-prone to maintain these duplicated constants. So, this diff gets rid of InternalStats's constants and replaces them with a map keyed on the user-facing constant. The value in that map contains a function pointer to get the property value, so we don't need to do string matching while holding db->mutex_. This approach has a side benefit of making many small handler functions rather than a giant switch-statement. Test Plan: db_properties_test passes, running "make commit-prereq -j32" Reviewers: sdong, yhchiang, kradhakrishnan, IslamAbdelRahman, rven, anthony Reviewed By: anthony Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D53253
-
-
Summary: Doing inline checking of transaction expiration instead of using a callback. Test Plan: To be added Reviewers: anthony Reviewed By: anthony Subscribers: leveldb, dhruba Differential Revision: https://reviews.facebook.net/D53673
-
由 Nathan Bronson 提交于
Summary: Disable test that is intermittently failing Test Plan: unit tests Reviewers: igor, andrewkr, sdong Reviewed By: sdong Subscribers: dhruba Differential Revision: https://reviews.facebook.net/D53715
-
由 Andrew Kryczka 提交于
Summary: db/c_test.c uses the functions in db/c.cc. If we have tags generated for one but not the other, it's easy to make mistakes like updating a function signature and missing a call site. Test Plan: $ make tags in vim: :cscope find s rocksdb_options_set_compression_options ... 3 325 db/c_test.c <<main>> rocksdb_options_set_compression_options(options, -14, -1, 0); Reviewers: sdong, yhchiang, IslamAbdelRahman Reviewed By: IslamAbdelRahman Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D53685
-
由 Siying Dong 提交于
Making use of GetSystemTimePreciseAsFileTime dynamic
-
- 02 2月, 2016 15 次提交
-
-
由 Tomas Kolda 提交于
-
由 Tomas Kolda 提交于
-
由 Tomas Kolda 提交于
break compatibility with Windows 7. The issue with rotated logs was fixed other way.
-
由 Tomas Kolda 提交于
-
由 Tomas Kolda 提交于
-
由 Nathan Bronson 提交于
Summary: Concurrent memtable adds were incorrectly computing the last sequence number for a write batch group when the write batches were not solitary. This is the cause of https://github.com/facebook/mysql-5.6/issues/155 Test Plan: 1. unit tests 2. new unit test 3. parallel db_bench stress tests with batch size of 10 and asserts enabled Reviewers: igor, sdong Reviewed By: sdong Subscribers: IslamAbdelRahman, MarkCallaghan, dhruba Differential Revision: https://reviews.facebook.net/D53595
-
由 sdong 提交于
Summary: Travis CI fails most of the times because of timing out. To unblock it, disable LITE tests in Travis CI. Test Plan: Will see. Reviewers: anthony, yhchiang, kradhakrishnan, IslamAbdelRahman, andrewkr Reviewed By: andrewkr Subscribers: leveldb, dhruba Differential Revision: https://reviews.facebook.net/D53679
-
由 Siying Dong 提交于
Add histogram for value size per operation
-
由 Siying Dong 提交于
Enable per-request buffer allocation in RandomAccessFile
-
由 Islam AbdelRahman 提交于
Summary: Fix current clang failure https://ci-builds.fb.com/view/rocksdb/job/rocksdb_clang_build/1398/console Test Plan: make sure that both clang and g++ compilation succeed USE_CLANG=1 make check -j64 make check -j64 Reviewers: anthony, yhchiang, sdong Reviewed By: sdong Subscribers: dhruba Differential Revision: https://reviews.facebook.net/D53667
-
由 sdong 提交于
Summary: Add unit tests: (1) insert entries of 8MB key and 3GB value to DB (2) insert entry of 3GB key and 3GB value into write batch and make sure we can read it. (3) insert 3 billions of key-value pairs into write batch and make sure we can read it. Disable them because not all platform can run it. Test Plan: Run the tests Reviewers: IslamAbdelRahman, yhchiang, kradhakrishnan, andrewkr, anthony Reviewed By: anthony Subscribers: leveldb, dhruba Differential Revision: https://reviews.facebook.net/D53619
-
由 Andrew Kryczka 提交于
Summary: Following up on D53493, we can still enable the filter-skipping optimization for last file in L0. It's correct to assume the key will be present in the last L0 file when we're hit-optimized and L0 is deepest. The FilePicker encapsulates the state for traversing each level's files, so I needed to make it expose whether the returned file is last in its level. Test Plan: verified below test fails before this patch and passes afterwards. The change to how the test memtable is populated is needed so file 1 has keys (0, 30, 60), file 2 has keys (10, 40, 70), etc. $ ./db_universal_compaction_test --gtest_filter=UniversalCompactionNumLevels/DBTestUniversalCompaction.OptimizeFiltersForHits/* Reviewers: sdong Reviewed By: sdong Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D53583
-
由 Dmytro Ivchenko 提交于
Test Plan: Added unit test to check w/ w/o zeros scenarios Reviewers: yhchiang Reviewed By: yhchiang Subscribers: sdong, dhruba Differential Revision: https://reviews.facebook.net/D52809
-
由 Dmitri Smirnov 提交于
This change impacts only non-buffered I/O on Windows. Currently, there is a buffer per RandomAccessFile instance that is protected by a lock. The reason we maintain the buffer is non-buffered I/O requires an aligned buffer to work. XPerf traces demonstrate that we accumulate a considerable wait time while waiting for that lock. This change enables to set random access buffer size to zero which would indicate a per request allocation. We are expecting that allocation expense would be much less than I/O costs plus wait time due to the fact that the memory heap would tend to re-use page aligned allocations especially with the use of Jemalloc. This change does not affect buffer use as a read_ahead_buffer for compaction purposes.
-
由 Yueh-Hsuan Chiang 提交于
Summary: Fixed the asan error on column_family_test caused by not disabling SyncPoint. Test Plan: column_family_test Reviewers: anthony, rven, kradhakrishnan, sdong, IslamAbdelRahman Reviewed By: IslamAbdelRahman Subscribers: dhruba, leveldb Differential Revision: https://reviews.facebook.net/D53505
-
- 01 2月, 2016 1 次提交
-
-
由 SherlockNoMad 提交于
-
- 30 1月, 2016 1 次提交
-
-
由 Venkatesh Radhakrishnan 提交于
Summary: If options.base_background_compactions is given, we try to schedule number of compactions not existing this number, only when L0 files increase to certain number, or pending compaction bytes more than certain threshold, we schedule compactions based on options.max_background_compactions. The watermarks are calculated based on slowdown thresholds. Test Plan: Add new test cases in column_family_test. Adding more unit tests. Reviewers: IslamAbdelRahman, yhchiang, kradhakrishnan, rven, anthony Reviewed By: anthony Subscribers: leveldb, dhruba, yoshinorim Differential Revision: https://reviews.facebook.net/D53409
-