Uploaded image for project: 'Flink'
  1. Flink
  2. FLINK-7289 Memory allocation of RocksDB can be problematic in container environments
  3. FLINK-14877

Enable FRocksDB to optimize performance via options when bounding total memory

    XMLWordPrintableJSON

Details

    • Sub-task
    • Status: Resolved
    • Major
    • Resolution: Implemented
    • None
    • 1.10.0
    • None
    • None

    Description

      Currently, as we plan to keep base on RocksDB-5.17.2 and cherry-pick write-buffer manager related rocksJava commits to FRocksDB. Options below are recommend when bounding the total memory.

      tableConfig.setCacheIndexAndFilterBlocksWithHighPriority(true);
      tableConfig.setPinTopLevelIndexAndFilter(true);
      

      However, RocksDB-5.17.2 does not include these java API. But luckily, these features are already implemented on RocksDB's native code. Cherry pick https://github.com/facebook/rocksdb/pull/4589 with only updates on java side should be enough.

      Attachments

        Issue Links

          Activity

            People

              yunta Yun Tang
              yunta Yun Tang
              Votes:
              0 Vote for this issue
              Watchers:
              5 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: