KBA Job Control and code=-14

sportclimber · Post by **sportclimber** » Fri Feb 17, 2012 1:46 pm

Hi everyone:

We encountered the following problem with DataStage and particularly with KBA Job Control and we are trying to determine root cause so that we can prevent this from happening again.

KBA Job Control batches saw jobs as completed normal when DataStage hadn't even run the job. Thus we had a situation where dependent jobs were then kicked off when their parent jobs hadn't been run.

In other words, it seemed as if KBA told DataStage, "run job x". DataStage didn't/couldn't run it and sent back something to KBA Job Control. KBA Job control interpreted that something as "Job complete" instead of "Job couldn't be run".

We think that something was going on in DataStage at the time this happened. Some of our DataStage Sequences got errors (eg, "Error calling DSRunJob(jobName1), code=-14 [Timed out while waiting for an event]"). And, we got the dreaded "ACCESS DENIED: Project is currently locked. The project is currently locked by the Administrator." Please note, that soon after getting this message, we could log on to the projects again without a reboot. At the time we had a large number of processes running and some of them were very system intensive.

We are guessing that DataStage reached its server resource limit (based on a search of the -14 code here in the forums) and couldn't kick off any kind of process (be it a job or a login).

If this is the case, does KBA Job Control not correctly trap the -14 code? Or perhaps something else is going on and DataStage sent a different code to KBA Job Control causing KBA Job Control to think the job had run successfully?

Thanks for any input on this issue,
Pam

ketfos · Post by **ketfos** » Sat Feb 18, 2012 7:23 pm

What is KBA Job Control ?

ray.wurlod · Post by **ray.wurlod** » Sun Feb 19, 2012 2:29 am

A suite of utilities from Ken Bland & Associates. You can find more (including their web address) by searching DSXchange. The web address is part of Ken's signature block.

chulett · Post by **chulett** » Sun Feb 19, 2012 8:49 am

Pam, the first question that comes to mind is: have there been any changes made to the code that I left there? I don't recall ever seeing the behaviour you've noted but have seen something almost the opposite where it goes to start a job and never hears back that it started and thus you get a 'living dead' running hang.

However, if the code is the same then I can poke around in the area that controls all that and see if anything looks... promising. I would imagine that you are correct that the system stress you were seeing manifest as those -14 errors played a role here as well.

sportclimber · Post by **sportclimber** » Sun Feb 19, 2012 9:35 am

Hi Craig:

Nothing has changed in the code base.

I'd definitely prefer to have the behavior you've seen then what we saw!

Thanks,
Pam