1. 15 7月, 2017 3 次提交
    • I
      Make TARGETS file portable · 13201336
      Islam AbdelRahman 提交于
      Summary:
      Instead of hard coding the path of the internal repo.
      Make TARGETS file work anywhere in fbcode
      Closes https://github.com/facebook/rocksdb/pull/2586
      
      Differential Revision: D5428122
      
      Pulled By: IslamAbdelRahman
      
      fbshipit-source-id: 21adec82bfbff14ea93532bee789b5f5bbee5b01
      13201336
    • D
      Set CACHE_LINE_SIZE for s390, PPC, ARM64 · ccf5f08f
      Daniel Black 提交于
      Summary: Closes https://github.com/facebook/rocksdb/pull/2579
      
      Differential Revision: D5427667
      
      Pulled By: maysamyabandeh
      
      fbshipit-source-id: cd0b076aa0cd38d3554516f01723c548713ece61
      ccf5f08f
    • D
      db_crashtest.py: remove need for shell · 67510eef
      Daniel Black 提交于
      Summary:
      Before:
      $ ps -ef
      
      build     1713    16  0 Jul11 ?        00:00:00 make crash_test
      build     3437  1713  0 Jul11 ?        00:00:00 python -u tools/db_crashtest.py --simple blackbox
      build     3438  3437  0 Jul11 ?        00:00:00 [sh] <defunct>
      build     3440     1 99 Jul11 ?        5-03:01:25 ./db_stress --max_background_compactions=1 --max_write_buffer_number=3 --sync=0 --reopen=20 --write_buffer_size=33554432 --delpercent=5 --block_size=16384 --allow_concurrent_me
      
      After:
      
      build      1706     16  0 02:52 ?        00:00:01 make crash_test
      build      3432   1706  0 02:55 ?        00:00:00 python -u tools/db_crashtest.py --simple blackbox
      build      4452   3432 99 04:35 ?        00:01:42 ./db_stress --max_background_compactions=1 --max_write_buffer_number=3 --sync=0 --reopen=20 --write_buffer_size=33554432 --delpercent=5 --block_size=16384 --allow_concurr
      Closes https://github.com/facebook/rocksdb/pull/2571
      
      Differential Revision: D5421580
      
      Pulled By: maysamyabandeh
      
      fbshipit-source-id: d6c3970c38ea0fa23da653f4385e8e25d83f5c9f
      67510eef
  2. 14 7月, 2017 5 次提交
  3. 13 7月, 2017 5 次提交
  4. 12 7月, 2017 4 次提交
    • A
      Improve the design and native object management of Stats in RocksJava · 000bf0af
      Adam Retter 提交于
      Summary: Closes https://github.com/facebook/rocksdb/pull/2551
      
      Differential Revision: D5399288
      
      Pulled By: sagar0
      
      fbshipit-source-id: dd3df2ed6cc5ae612db0998ea746cc29fccf568e
      000bf0af
    • I
      Bump version to 5.7 · 269d383d
      Islam AbdelRahman 提交于
      Summary:
      Bump version to 5.7
      Closes https://github.com/facebook/rocksdb/pull/2566
      
      Differential Revision: D5400043
      
      Pulled By: IslamAbdelRahman
      
      fbshipit-source-id: 74aae4ff143d370d7d89807e5be08a6ab827da40
      269d383d
    • Y
      Fixes db_bench with blob db · c32f2722
      Yi Wu 提交于
      Summary:
      * Create info log before db open to make blob db able to log to LOG file.
      * Properly destroy blob db.
      Closes https://github.com/facebook/rocksdb/pull/2567
      
      Differential Revision: D5400034
      
      Pulled By: yiwu-arbug
      
      fbshipit-source-id: a49cfaf4b5c67d42d4cbb872bd5a9441828c17ce
      c32f2722
    • D
      db_bench_tool: fix buffer size · fcd99d27
      Daniel Black 提交于
      Summary:
      Found by gcc warning:
      
      x86_64-pc-linux-gnu-g++ --version
      x86_64-pc-linux-gnu-g++ (GCC) 7.1.1 20170710
      
      tools/db_bench_tool.cc: In member function 'void rocksdb::Benchmark::RandomWithVerify(rocksdb::ThreadState*)':
      tools/db_bench_tool.cc:4430:8: error: '%lu' directive output may be truncated writing between 1 and 19 bytes into a region of size between 0 and 66 [-Werror=format-truncation=]
         void RandomWithVerify(ThreadState* thread) {
              ^~~~~~~~~~~~~~~~
      tools/db_bench_tool.cc:4430:8: note: directive argument in the range [0, 9223372036854775807]
      tools/db_bench_tool.cc:4492:13: note: 'snprintf' output between 37 and 128 bytes into a destination of size 100
           snprintf(msg, sizeof(msg),
           ~~~~~~~~^~~~~~~~~~~~~~~~~~
                    "( get:%" PRIu64 " put:%" PRIu64 " del:%" PRIu64 " total:%" \
                    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
                    PRIu64 " found:%" PRIu64 ")",
                    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
                    gets_done, puts_done, deletes_done, readwrites_, found);
                    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
      cc1plus: all warnings being treated as errors
      Makefile:1707: recipe for target 'tools/db_bench_tool.o' failed
      Closes https://github.com/facebook/rocksdb/pull/2558
      
      Differential Revision: D5398703
      
      Pulled By: siying
      
      fbshipit-source-id: 6ffa552bbd8b59cfc2c36289f86ff9b9acca8ca6
      fcd99d27
  5. 11 7月, 2017 3 次提交
  6. 08 7月, 2017 3 次提交
  7. 07 7月, 2017 2 次提交
  8. 06 7月, 2017 5 次提交
  9. 03 7月, 2017 1 次提交
  10. 01 7月, 2017 4 次提交
  11. 30 6月, 2017 4 次提交
    • A
      Regression test for empty dedicated range deletion file · d310e0f3
      Andrew Kryczka 提交于
      Summary:
      Issue: #2478
      Fix: #2503
      
      The bug happened when all of these conditions were satisfied:
      
      - A subcompaction generates no keys
      - `RangeDelAggregator::ShouldAddTombstones()` returns true because there's at least one non-obsoleted range deletion in its map
      - None of the non-obsolete tombstones overlap with the subcompaction key-range
      
      Under those conditions, we were creating a dedicated file for range deletions which was left empty, thus causing an error in VersionEdit.
      
      I verified this test case fails before the #2503 fix and passes after.
      Closes https://github.com/facebook/rocksdb/pull/2521
      
      Differential Revision: D5352568
      
      Pulled By: ajkr
      
      fbshipit-source-id: f619cae39984ce9bb9b7a4e7a9ac0f2bb2ce43e9
      d310e0f3
    • M
      Add a fetch_add variation to AddDBStats · e9f91a51
      Maysam Yabandeh 提交于
      Summary:
      AddDBStats is in two steps of load and store, which is more efficient than fetch_add. This is however not thread-safe. Currently we have to protect concurrent access to AddDBStats with a mutex which is less efficient that fetch_add.
      
      This patch adds the option to do fetch_add when AddDBStats. The results for my 2pc benchmark on sysbench is:
      - vanilla: 68618 tps
      - removing mutex on AddDBStats (unsafe): 69767 tps
      - fetch_add for all AddDBStats: 69200 tps
      - fetch_add only for concurrently access AddDBStats (this patch): 69579 tps
      Closes https://github.com/facebook/rocksdb/pull/2505
      
      Differential Revision: D5330656
      
      Pulled By: maysamyabandeh
      
      fbshipit-source-id: af64d7bee135b0e86b4fac323a4f9d9113eaa383
      e9f91a51
    • Z
      skip generating empty sst · c1b375e9
      zhangjinpeng1987 提交于
      Summary:
      When a compaction job output nothing, there is no necessary to generate a empty sst file which will cause `VersionEdit::EncodeTo` failed.
      ref https://github.com/facebook/rocksdb/issues/2478
      Closes https://github.com/facebook/rocksdb/pull/2503
      
      Differential Revision: D5350799
      
      Pulled By: ajkr
      
      fbshipit-source-id: df0b4fcf3507fe1c3c435208b762e75478e00143
      c1b375e9
    • Y
      fix format compatible test · 67b417d6
      Yi Wu 提交于
      Summary:
      The comma "," is not a valid separator for bash arrays.
      Closes https://github.com/facebook/rocksdb/pull/2516
      
      Differential Revision: D5348101
      
      Pulled By: yiwu-arbug
      
      fbshipit-source-id: 8f0afdac368e21076eb7366b7df7dbaaf158cf96
      67b417d6
  12. 29 6月, 2017 1 次提交