I have noticed it many times in the past. I think they popped up when I launched a test that crashed the tserver with an exception. When I was doing development this happened often enough that I made a script to do grep and kill them. Under normal circumstances they die as normal.
On Wed, Aug 31, 2016 at 8:41 AM, Josh Elser <[email protected]> wrote: > Tried running the ITs overnight for 1.8.0rc3 > > {noformat} > Running org.apache.accumulo.test.replication.UnorderedWorkAssignerRe > plicationIT > OpenJDK 64-Bit Server VM warning: INFO: os::commit_memory(0x00000000f1900000, > 157810688, 0) failed; error='Cannot allocate memory' (errno=12) > # > # There is insufficient memory for the Java Runtime Environment to > continue. > # Native memory allocation (malloc) failed to allocate 157810688 bytes for > committing reserved memory. > # An error report file with more information is saved as: > # /home/elserj/accumulo-180-rc3/accumulo-1.8.0-src/test/hs_err_pid5342.log > Running org.apache.accumulo.test.replication.KerberosReplicationIT > Running org.apache.accumulo.test.replication.ReplicationRandomWalkIT > Running org.apache.accumulo.test.replication.SequentialWorkAssignerIT > Running org.apache.accumulo.test.replication.GarbageCollectorCommuni > catesWithTServersIT > Running org.apache.accumulo.test.replication.WorkMakerIT > zsh: killed mvn verify -fn > {noformat} > > What I'm seeing now is that I have 28 orphaned Java processes, all of them > appear to be from MAC on these specific test classes. > > Has anyone else noticed that we're leaking these processes? Are we not > launching them correctly such that when the PPID dies, they continue to > live? >
