Uploaded image for project: 'Hadoop HDFS'
  1. Hadoop HDFS
  2. HDFS-9198

Coalesce IBR processing in the NN

    XMLWordPrintableJSON

Details

    • Improvement
    • Status: Closed
    • Major
    • Resolution: Fixed
    • 2.0.0-alpha
    • 2.8.0, 2.7.3, 3.0.0-alpha1
    • namenode
    • None
    • Reviewed

    Description

      IBRs from thousands of DNs under load will degrade NN performance due to excessive write-lock contention from multiple IPC handler threads. The IBR processing is quick, so the lock contention may be reduced by coalescing multiple IBRs into a single write-lock transaction. The handlers will also be freed up faster for other operations.

      Attachments

        1. HDFS-9198-branch-2.7.patch
          41 kB
          Vinayakumar B
        2. HDFS-9198-Branch-2.8-withamend.diff
          40 kB
          Uma Maheswara Rao G
        3. HDFS-9198-Branch-2-withamend.diff
          40 kB
          Uma Maheswara Rao G
        4. HDFS-9198-trunk.patch
          38 kB
          Daryn Sharp
        5. HDFS-9198-trunk.patch
          38 kB
          Daryn Sharp
        6. HDFS-9198-trunk.patch
          36 kB
          Daryn Sharp
        7. HDFS-9198-trunk.patch
          36 kB
          Daryn Sharp
        8. HDFS-9198-trunk.patch
          31 kB
          Daryn Sharp
        9. HDFS-9198-branch2.patch
          32 kB
          Daryn Sharp

        Issue Links

          breaks

          Sub-task - The sub-task of the issue HDFS-10641 TestBlockManager#testBlockReportQueueing fails intermittently

          • Major - Major loss of function.
          • Resolved
          is related to

          Bug - A problem which impairs or prevents the functions of the product. HDFS-10563 Block reports could be silently dropped by NN

          • Critical - Crashes, loss of data, severe memory leak.
          • Resolved

          Improvement - An improvement or enhancement to an existing feature or task. HDFS-12619 Do not catch and throw unchecked exceptions if IBRs fail to process

          • Minor - Minor loss of function, or other problem where easy workaround is present.
          • Resolved
          relates to

          Improvement - An improvement or enhancement to an existing feature or task. HDFS-14186 blockreport storm slow down namenode restart seriously in large cluster

          • Major - Major loss of function.
          • Reopened

          Test - A new unit, integration or system test. HDFS-13554 TestDatanodeRegistration#testForcedRegistration does not shut down cluster

          • Minor - Minor loss of function, or other problem where easy workaround is present.
          • Resolved

          Activity

            People

              daryn Daryn Sharp
              daryn Daryn Sharp
              Votes:
              0 Vote for this issue
              Watchers:
              35 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: