[ https://issues.apache.org/jira/browse/MAPREDUCE-1703?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Vinod K V resolved MAPREDUCE-1703. ---------------------------------- Resolution: Duplicate Going to track this trivial fix as part of MAPREDUCE-1707 that is making changes to the same class. > TaskRunner would crash in finally block if taskDistributedCacheManager is null > ------------------------------------------------------------------------------ > > Key: MAPREDUCE-1703 > URL: https://issues.apache.org/jira/browse/MAPREDUCE-1703 > Project: Hadoop Map/Reduce > Issue Type: Bug > Components: tasktracker > Affects Versions: 0.21.0 > Reporter: Amareshwari Sriramadasu > Fix For: 0.22.0 > > > If TaskRunner throws an Exception before initializing > taskDistributedCacheManager, it would crash in finally block at > taskDistributedCacheManager.release(). TaskRunner would crash without doing > tip.reportTaskFinished() thus not failing the task. Task will be marked > FAILED after "mapred.task.timeout" because there is no report from the task. > We should add a not null check for taskDistributedCacheManager in finally > block. -- This message is automatically generated by JIRA. - You can reply to this email to add a comment to the issue online.