Page 1 of 1
Posted: Sun Nov 16, 2008 9:33 pm
by chulett
I don't recall this coming up before, so I built a quick harness. Ran a test with a Sequence job running an existing Server job and it behaved as one would expect - run it under Sequence control and it showed that when interogated, run standalone before or after that and it returned an empty value.
Just realized you had one Sequence running another, something different than I tried. Really shouldn't behave differently, but who knows? I'll try that and post back.
ps. DataStage 7.5.1A on HP-UX.
Posted: Sun Nov 16, 2008 10:00 pm
by chulett
Ok, I'm seeing the same behaviour as you are, Sequence running Sequence. After the first stand-alone run post abort/reset of my "B", it still shows the original controller when interogated. However, if I run it stand-alone after that again, then it comes back empty.
It's almost as if it is reporting your reset / rerun as originally started. Do you have checkpointing enabled in the Sequence jobs? I did and now I'm working how much of an effect that has on this.
Posted: Mon Nov 17, 2008 8:41 pm
by aworth
Thanks for your input.
I'm not really sure what you meant by "checkpointing enabled in the Sequence jobs", so I guess the answer would be no.
The simplest form to show this would be the follow...
Batch::B Code
Code: Select all
ControllingBatch = DSGetJobInfo(DSJ.ME, DSJ.JOBCONTROLLER)
Call DSLogFatal("The main program has begun: ":ControllingBatch,"JobControl")
Code
Code: Select all
* Setup Batch::A, run it, wait for it to finish, and test for success
hJob1 = DSAttachJob("Batch::B", DSJ.ERRFATAL)
If NOT(hJob1) Then
Call DSLogFatal("Job Attach Failed: Batch::B", "JobControl")
End
ErrCode = DSSetDisableProjectHandler(hJob1, @FALSE)
ErrCode = DSSetDisableJobHandler(hJob1, @FALSE)
ErrCode = DSRunJob(hJob1, DSJ.RUNNORMAL)
ErrCode = DSWaitForJob(hJob1)
Status = DSGetJobInfo(hJob1, DSJ.JOBSTATUS)
If Status = DSJS.RUNFAILED Or Status = DSJS.CRASHED Then
* Fatal Error - No Return
Call DSLogFatal("Job Failed: Batch::B", "JobControl")
End
Thanks and Regards,
Allan Worthington...
aworth@imb.com.au
Posted: Tue Nov 18, 2008 12:23 am
by chulett
So, when you said "Batch" you literally meant "Batch colon colon" as in hand-coded job control code maintained via the Director? Huh. No wonder the checkpoints question didn't make sense, I was specifically referring to a Sequence job.
As noted (and regardless) I saw the same behaviour you did. Now, as to whether that's a bug or a feature we'd probably need to ask IBM.
Posted: Tue Nov 25, 2008 11:04 pm
by aworth
FYI. I had a response from IBM Support
Subject: Re: Fw: *** PMR 76203,999,616*** DataStage Support - DSGetJobInfo and JOBCONTROLLER
Hi Allan,
We have successfully reproduced your issue on our testing environment. It seems to us that status files are left over even though job has been reset from the Director client.
As a possible solution, you can do the following
1/ Director client -> Highlight the job (e.g. Batch::A)
2/ Reset the job as you did before
3/ Click Job -> Clear Status File
4/ Re-run the job
In the meantime, I have escalated your issue with our engineering team to see if there is any existing fix / solution available. I will keep you informed if there is any update from their end.
Posted: Wed Nov 26, 2008 12:54 am
by chulett
Thanks... please let us know if they come back with an update.