Uploaded image for project: 'Hadoop HDFS'
  1. Hadoop HDFS
  2. HDFS-11180

Intermittent deadlock in NameNode when failover happens.

    XMLWordPrintableJSON

Details

    Description

      It is happening due to metrics getting updated at the same time when failover is happening. Please find attached jstack at that point of time.

      Attachments

        1. HDFS-11180.00.patch
          3 kB
          Akira Ajisaka
        2. HDFS-11180.01.patch
          3 kB
          Akira Ajisaka
        3. HDFS-11180.02.patch
          4 kB
          Akira Ajisaka
        4. HDFS-11180.03.patch
          9 kB
          Akira Ajisaka
        5. HDFS-11180.04.patch
          16 kB
          Akira Ajisaka
        6. HDFS-11180-branch-2.01.patch
          15 kB
          Akira Ajisaka
        7. HDFS-11180-branch-2.6.01.patch
          13 kB
          Akira Ajisaka
        8. HDFS-11180-branch-2.7.01.patch
          13 kB
          Akira Ajisaka
        9. HDFS-11180-branch-2.8.01.patch
          15 kB
          Akira Ajisaka
        10. jstack.log
          107 kB
          Abhishek Modi

        Issue Links

          relates to

          Bug - A problem which impairs or prevents the functions of the product. HDFS-11352 Potential deadlock in NN when failing over

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

          Test - A new unit, integration or system test. HDFS-11290 TestFSNameSystemMBean should wait until JMX cache is cleared

          • Major - Major loss of function.
          • Resolved

          Activity

            People

              aajisaka Akira Ajisaka
              abmodi Abhishek Modi
              Votes:
              2 Vote for this issue
              Watchers:
              24 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: