hive-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Joydeep Sen Sarma (JIRA)" <>
Subject [jira] Commented: (HIVE-247) Need access to job_id from session state
Date Fri, 23 Jan 2009 21:29:59 GMT


Joydeep Sen Sarma commented on HIVE-247:

take a look at hive-176.

suresh has added a hivehistory object per session and this history object contains a list
of queries and the jobs spawned by them and stats for these jobs.

we can use this perhaps for this purpose?

> Need access to job_id from session state
> ----------------------------------------
>                 Key: HIVE-247
>                 URL:
>             Project: Hadoop Hive
>          Issue Type: Improvement
>          Components: Query Processor
>         Environment: Hadoop 0.19.0, Hive trunk
>            Reporter: Edward Capriolo
>            Priority: Minor
> When using the CLI the tracking URL and kill command are printed to the console:
> {quote}
> Starting Job = job_200901231237_0001, Tracking URL = http://hadoop1.jointhegrid.local:50030/jobdetails.jsp?jobid=job_200901231237_0001
> Kill Command = /opt/hadoop/hadoop-0.19.0/bin/../bin/hadoop job  -Dmapred.job.tracker=hadoop1.jointhegrid.local:54311
-kill job_200901231237_0001
> {quote}
> The ExecDriver has a static Map ExecDriver.runningJobKillURIs. The HiveSessionManager
used in the HWI is spawning multiple SessionState instances inside one JVM. There is no way
to determine the jobid from the SessionState. As a result, I do not have access to the job
id and can not kill it.
> Preliminary suggestions:
> * Add another static accessible structure to ExecDriver that can relate job_id to a session.
> * Have exec driver set the current running job_id inside HiveConf so the CLISession is
aware of the job id.

This message is automatically generated by JIRA.
You can reply to this email to add a comment to the issue online.

View raw message