[
https://issues.apache.org/jira/browse/HADOOP-630?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12470421
]
Nigel Daley commented on HADOOP-630:
------------------------------------
This bug makes it difficult to debug jobs after the JT has been taken down.
Starting up the JT to analyse jobs through the JT web UI is only possible if
you also bring up the HDFS used by the job. Given all the tests I run, with
different HDFS instances, etc, I find this restriction very inconvenient.
> The web/ui for the JobTracker doesn't come up until after the namenode is out
> of safe mode
> ------------------------------------------------------------------------------------------
>
> Key: HADOOP-630
> URL: https://issues.apache.org/jira/browse/HADOOP-630
> Project: Hadoop
> Issue Type: Bug
> Components: mapred
> Affects Versions: 0.7.2
> Reporter: Owen O'Malley
> Assigned To: Owen O'Malley
>
> The http status server for the job tracker isn't initialized until after the
> working directory is cleared, meaning that it can't happen until after the
> name node leaves safe mode.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
