[
https://issues.apache.org/jira/browse/TEZ-301?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13733382#comment-13733382
]
joeyli edited comment on TEZ-301 at 8/8/13 11:19 AM:
-----------------------------------------------------
Getting the #runningtask is a little tricky because of container reuse. Only
TaskAttemptListenerImpTezDag#getTask() know #tasks assigned to containers, but
it doesnt know when those tasks complete. so I use a map to track
#assignedtasks against each vertex. Then in VertexImpl#getVertexProgress(), I
can get #runningtasks by #assignedtasks-#completedtask.
The interestingthing is this still work when multi-Dag running in a single AM
because the vertexid is unique.I create a patch, please take a look, thanks.
was (Author: fiberlijun):
Getting the #runningtask is a little tricky because of container reuse.
Only TaskAttemptListenerImpTezDag#getTask() know #tasks assigned to containers,
but it doesnt know when those tasks complete. so I use a map to track
#assignedtasks against each vertex. Then in VertexImpl#getVertexProgress(), I
can get #runningtasks by #assignedtasks-#completedtask.
The interting thing is this still work when multi-Dag running in a single AM
because the vertexid is unique.I create a patch, please take a look, thanks.
> Running task count in DAG Progress is always 0 ( even as tasks complete ).
> --------------------------------------------------------------------------
>
> Key: TEZ-301
> URL: https://issues.apache.org/jira/browse/TEZ-301
> Project: Apache Tez
> Issue Type: Bug
> Reporter: Hitesh Shah
> Labels: TEZ-0.2.0
>
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
