[
https://issues.apache.org/jira/browse/HADOOP-5392?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12679100#action_12679100
]
Amar Kamat commented on HADOOP-5392:
------------------------------------
HADOOP-4638 added this piece of code
{code}
synchronized (trackerToJobsToCleanup) {
Set<JobID> jobs = trackerToJobsToCleanup.get(trackerName);
jobs.add(taskId.getJobID());
}
{code}
Here jobs can be null. A null check should be made. This issue was detected
during HADOOP-5392 testing
> JobTracker crashes during recovery if job files are garbled
> -----------------------------------------------------------
>
> Key: HADOOP-5392
> URL: https://issues.apache.org/jira/browse/HADOOP-5392
> Project: Hadoop Core
> Issue Type: Bug
> Components: mapred
> Reporter: Amar Kamat
> Assignee: Amar Kamat
> Priority: Blocker
> Fix For: 0.20.0
>
>
> Jobtracker crashed in the recovery stage for a job with 0 byte job.xml.
> Ideally one would expect the jobtracker to try and recover as many jobs as
> possible.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
