[
https://issues.apache.org/jira/browse/HADOOP-5388?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12678703#action_12678703
]
Sreekanth Ramakrishnan commented on HADOOP-5388:
------------------------------------------------
I am trying to monitor the task failures. Suppose if we want a script which is
run periodically on a cluster which looks at the running jobs in the cluster
and then check for task failures in the same and outputs the
jobid|attemptid|failure reason.
In this case, currently I would have to hard code per page level.
True, in long run we should be doing this REST/XML way but as a stop gap
measure we should be doing something so that these can be done. We should
proably expose services from JobTracker which provides JSON/XML data to
different clients which can make used of those information. But then I think
arriving there would take quite a bit of time.
> Add ids to tables in JSP pages to ease scraping of the data.
> ------------------------------------------------------------
>
> Key: HADOOP-5388
> URL: https://issues.apache.org/jira/browse/HADOOP-5388
> Project: Hadoop Core
> Issue Type: Improvement
> Reporter: Sreekanth Ramakrishnan
> Priority: Minor
> Attachments: HADOOP-5388.patch
>
>
> Currently, the tables which are generated by the JSP pages are lacking id's.
> If the tables had id then it would ease the pain of writing scrapping
> utilities.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
