Help needed to kill a "rogue" server job

RamonaReed · Post by **RamonaReed** » Mon Apr 12, 2004 10:23 am

Please I need help ASAP. I cannot stop a server job that is running. I have had the DBA kill the oracle session, and I have submitted a stop from the DataStage Director, but nothing is stopping it. Our server administrator even re-booted our DataStage server, and it's still running. How do you stop something like this?????

thanks in advance
Mona

chulett · Post by **chulett** » Mon Apr 12, 2004 10:33 am

There's a big difference between a job that says it is running versus one that actually is. If you've rebooted the DataStage server than your job, she ain't running.

It is just a flag and it says it is 'Running' because it never heard back that the job ended.

Use the 'Clear Status File' command from Director to clear this condition.

RamonaReed · Post by **RamonaReed** » Mon Apr 12, 2004 10:42 am

I've tried that, and it still indicates that it's running, Also when I right click on the job and select Monitor, the elapsed time is still increasing, so is it running or isn't it? HELP !!!!!

kcbland · Post by **kcbland** » Mon Apr 12, 2004 11:07 am

Try recompiling the job, if you rebooted there can't be any runaway threads. You have a "software" lock, basically just a status in a table is invalid. Recompile and that should fix it.

RamonaReed · Post by **RamonaReed** » Mon Apr 12, 2004 11:17 am

Thank you Kenneth !!!!! :D That cleared the status, now I only have to be able to clear the log file. Thank you again.

Mona

ray.wurlod · Post by **ray.wurlod** » Mon Apr 12, 2004 3:33 pm

What you see in the Director is not the actual status of the job. Rather, it's the most recent entry that the job has been able to write to its status file, which is what Director reads.
If the job has aborted suddenly (or been killed), then it never gets the opportunity to update its status records, so the Director appears to show it as still running.
This is why re-compiling, or clearing the status file, has the apparent effect of "stopping" the job. It's not really stopping it; rather it's updating the record that indicates that the job is running.