DSXchange

I have a problem whereby the job was aborted. I checked the Log and it indicate the following error message:

Batch::<Batchname>.JobControl(fatal error from DSRunJob): Job control fatal error (-14)
(DSRunJob) Job <jobname> apprears not to have started after 60 seconds.

Kindly advice.

Do an exact search on "Job control fatal error (-14)". This itself will return you 9 posts. Have a head start there. Trust me it will save you a lot of time.

Looks like I will need to Set the time for Inactivity in Inactivity Timeout box.

Thanks.

That won't help. That is client inactivity. You are simply overloading your server. Stop doing that, or get a more powerful server.

I'm not sure why... cause we have a very powerful server Dual processor and 2 GB memory and there are no issues with the performance and resource of the machine.

In addition to that, some of the error message includes - "Abnormal termination of stage..."

It's not always about how Big and Beefy your server is, or the number of jobs running at the same time - it can be about the number of jobs you try to start at the same time.

Your 'abnormal termination' errors are a separate issue. Suggest you start a new thread and include as much detail as you can on it, including what you get when you Reset the job and find a 'From previous run...' entry in the log.

May I know how may number of jobs been called parallely by the batch job?
Is it possible for you to serialize it?

As Ray mentioned this is usually due to the fact that you are overloading your system. Means you are starting too many processes at the same time. If a process could not be scheduled 60s after being initiated, DS terminates it.

You should try to more serialize your sequence flow.

You may also try to ask IBM (Ascential) support if they can provide you a patch that extends the timeout period. I know this has been provided for some UNIX systems, not sure for Windows....

Best regards

Klaus

Hello again Klaus,

are you certain about that patch?

I am at a site where they had this problem and I was told here that there was no workaround for this issue; there were some configuration changes done to allow more processes in total but that the 60 second limit was hardcoded in several places and couldn't be changed. This is an AIX site. It would be great if this truly did exist.

Additional Suggestions: Disable any screen saver on the Windows server. Verify that the firewall and/or anti-virus software is not imposing too great an impact on startup times for executables. Verify that the Windows server is not being used as either a primary or backup domain controller. Monitor resource consumption with Task Manager or Performance Monitor to "spot the hogs".

Hello Arnd, hope you're doing well

Yes, I've got that patch for a SOLARIS and an AIX customer, both on release 7.51A.

I think they don't want to provide that patch in general, but on insisting specific customer request it was provided and helped resolve their issue.

Greetings from good old Bavaria...

Klaus

Klaus,

thanks for that information & gruesse right back! I'll pass on the information here to get the customer to request that patch; they are suffering from that issue and their workarounds are not permanent solutions.

Danke,

Thanks for the pointers Klaus. We have jobs running in parallel. Around 6-7 jobs at a time. That might be one of the problem, as you mentioned earlier. Just curios, is there a limit whereby Datastage could handle an "X" amount of parallel jobs running or is it based on the 60 seconds window time frame?

How long is a piece of string? The answer is "it depends". More jobs can be handled at once if they don't do much compared to fewer jobs that each do a lot. If you make extensive use of common hashed files, you might investigate sharing these (see dsdskche.pdf in your manuals). Ultimately try to design jobs so that they do the minimum amount of processing and I/O to achieve the desired result.

Iam not at PX right no, but I am sure, there is a Evnironmental variable, where you can try setting it for time outs.