[
https://issues.apache.org/jira/browse/MAPREDUCE-1869?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Allen Wittenauer updated MAPREDUCE-1869:
----------------------------------------
Summary: TaskTracker hangs due to fork() being MT-unsafe (was: TaskTracker
hung in org.apache.hadoop.fs.LocalDirAllocator)
After talking with various folks, the current thinking is that this is related
to http://bugs.sun.com/view_bug.do?bug_id=5049299 and friends. My stack trace
matches other traces where fork() has blown up.
> TaskTracker hangs due to fork() being MT-unsafe
> -----------------------------------------------
>
> Key: MAPREDUCE-1869
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-1869
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Components: tasktracker
> Affects Versions: 0.20.2
> Environment: Solaris 10 Update 7
> Java 1.6.0_14
> Reporter: Allen Wittenauer
> Priority: Critical
> Attachments: jstack.txt, pstack.txt
>
>
> A TaskTracker process on our grid appears to be locked up and not sending
> heartbeats to the namenode. Attaching jstack and pstack output. Even tho
> the hangs appear to be in LocalDirAllocator, the local file system seems to
> be a-ok.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
