Details

    Description

      I'm encountering an issue with the Mapreduce HistoryServer processing the history files for large jobs. This has come up several times with for jobs with around 60000 total tasks. When the HistoryServer loads the .jhist file from HDFS for a job of that size (which is usually around 500 Mb), the HistoryServer's CPU usage spiked and the UI became unresponsive. After about 10 minutes I restarted the HistoryServer and it was behaving normally again.

      The cluster is running CDH 5.3 (2.5.0-cdh5.3.0). I've attached the output of jstack from a time this was occurring.

      Attachments

        1. MAPREDUCE-6222.009.patch
          20 kB
          Ray Chiang
        2. Screen Shot 2015-05-20 at 11.16.25 AM.png
          267 kB
          Ray Chiang
        3. MAPREDUCE-6222.008.patch
          18 kB
          Ray Chiang
        4. MAPREDUCE-6222.007.patch
          18 kB
          Ray Chiang
        5. MAPREDUCE-6222.006.patch
          12 kB
          Ray Chiang
        6. MAPREDUCE-6222.005.patch
          12 kB
          Ray Chiang
        7. MAPREDUCE-6222.003.patch
          12 kB
          Ray Chiang
        8. MAPREDUCE-6222.002.patch
          11 kB
          Ray Chiang
        9. JHS New Display Top.png
          126 kB
          Ray Chiang
        10. JHS Original Display Top.png
          295 kB
          Ray Chiang
        11. MAPREDUCE-6222.001.patch
          9 kB
          Ray Chiang
        12. head.jhist
          9.09 MB
          Andrew Johnson
        13. historyserver_jstack.txt
          1.30 MB
          Andrew Johnson

        Issue Links

          Activity

            People

              rchiang Ray Chiang
              ajsquared Andrew Johnson
              Votes:
              0 Vote for this issue
              Watchers:
              12 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: