Job hangs at midnight

SteveG · Post by **SteveG** » 19 Sep 2010 2:36

We have seen 2 different jobs that hang in Yellow mode starting at midnight. These jobs are cyclical and run every 5 minutes, 24x7. But they do not hang every night. Each job has hung just once.

We are running on Unix 6.4. We also saw this on the previous version 6.2(?).

The jobs call a script that calls a SQL Agent job.
Any thoughts?

Thanks

Echelon · Post by **Echelon** » 21 Sep 2010 8:07

First check the agent logs. Increase the debug level on the agent side so you can get more info.

I assume that the problem will be on the SQL side as you had the same problem on 6.2 and now on 64.
So the best it is to run the SQL script manually to find out how long it takes to finish and if any errors show.

gbyrnes · Post by **gbyrnes** » 24 Sep 2010 7:56

Hi Steve,

This is something to do with the SQL Agent.
To be sure that it is not Control-M you could add another 5-minute Cyclic Job to run next to the SQL Agent Job that just does a sleep or at least something to hit the OS (i.e. hostname, etc.) - do NOT use a Dummy Job as this does NOT go to the OS.
If the SQL Agent Job hangs without the "test" Job hanging then it is a SQL Agent issue - no need to burden the Control-M environment with debugging.
(If you do feel the need to debug the Control-M/Agent then I would recommend only doing it for say 20 minutes around midnight instead of 24x7 - depending on the Control-M version you can toggle the Control-M/Agent debugging via command-line (read a Control-M Job)).

Good luck...
Cheers, Graeme.