Uploaded image for project: 'Jira Data Center'
  1. Jira Data Center
  2. JRASERVER-64255

jvm crash G1ParScanThreadState::copy_to_survivor_space

    XMLWordPrintable

Details

    Description

      NOTE: This bug report is for JIRA Server. Using JIRA Cloud? See the corresponding bug report.

      Summary

      This is not a JIRA bug, but the placeholder ticket for JVM problem. Main goal is to track the problem on our side.
      Periodically JVM crashes with following error:

      #  SIGSEGV (0xb) at pc=0x00007f36c1177ae2, pid=26874, tid=0x00007f36a79da700
      #
      # JRE version: Java(TM) SE Runtime Environment (8.0_101-b13) (build 1.8.0_101-b13)
      # Java VM: Java HotSpot(TM) 64-Bit Server VM (25.101-b13 mixed mode linux-amd64 compressed oops)
      # Problematic frame:
      # V  [libjvm.so+0x5c3ae2]  G1ParScanThreadState::copy_to_survivor_space(InCSetState, oopDesc*, markOopDesc*)+0x1d2
      
      ...
      Stack: [0x00007f36a78da000,0x00007f36a79db000],  sp=0x00007f36a79d9410,  free space=1021k
      Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, C=native code)
      V  [libjvm.so+0x5c3ae2]  G1ParScanThreadState::copy_to_survivor_space(InCSetState, oopDesc*, markOopDesc*)+0x1d2
      V  [libjvm.so+0x5c450b]  G1ParScanThreadState::trim_queue()+0x4ab
      V  [libjvm.so+0x59ed17]  G1ParEvacuateFollowersClosure::do_void()+0x27
      V  [libjvm.so+0x5aa923]  G1ParTask::work(unsigned int)+0x463
      V  [libjvm.so+0xae5a6f]  GangWorker::loop()+0xcf
      V  [libjvm.so+0x924698]  java_start(Thread*)+0x108
      

      This is internal JVM error in G1ParScanThreadState code.

      Environment

      • Oracle 8u101
        • Java HotSpot(TM) 64-Bit Server VM (25.101-b13) for linux-amd64 JRE (1.8.0_101-b13)

      Notes

      There are some bugs open for this in Java bug tracker:

      But unfortunately, they are closed as Cannot Reproduce.
      Commend from Oracle team:

      This can be caused by any bug that corrupts heap memory. It could be an issue with GC, with the compiler, with bad native code...

       
      We have had success in preventing this bug from occurring in previous support cases by determining the source of frequent Full GCs and fixing the underlying problem, reducing the frequency of Full GCs

      Workaround

      Restart JVM/Jira

      Attachments

        Issue Links

          Activity

            People

              Unassigned Unassigned
              ayakovlev@atlassian.com Andriy Yakovlev [Atlassian]
              Votes:
              6 Vote for this issue
              Watchers:
              24 Start watching this issue

              Dates

                Created:
                Updated: