DSXchange

I have a sequence job which calls several jobs. My children jobs finish fast but after each job, that is, after this line in the Director:

Job Xyz has finished, status = 1 (Finished OK)

there is a line like:
Parent_Seq_Job..JobControl(@Xyz_Label): Report on job: Xyz

************************************************
---
---
Job run stats


Producing the info takes a very long time. I already tried APT_NO_JOBMON and APT_PERFORMANCE_DATA

How do I prevent it to speed up my job significantly?

I guess It is waiting for the job to finish running, tests the status, and then
Run another one...

Try this...and see whether it works or not

In your Job control page remove or Comment out the Code which is testing the status..
Normally that will be similar like this More info can be found on the Designer guide..Job Control Routines section

All my jobs are running fine, nothing is aborting. The job is finishing very quick. Its just that it is taking a very long time ending one job and starting another.

Report on job is the fourth compilation option for a sequence, to be found in its job properties. Uncheck this box and re-compile the sequence.

Thanks Ray but it turns out that that was not cause. The biggest delay is between the line:

.........Job Xyz has finished, status = 1 (Finished OK)

and the next line:

..........Job run requested (...)

I am assuming that, eventhough it says (Finished OK) for the job, the actual work of the job finishes right when the line 'Job run requested' is logged for the next job. Can you confirm?

I don't believe so. The "job finished" message should be the last thing that the job does.

You will need some more exact form of tracing, possibly with the aid of your official support provider, to determine what's happening between the end of one job and the start of the next. For example you might try setting APT_STARTUP_STATUS.

If you have 8.1.x, where are you carrying the logs - in the traditional Project location or in the XMETA database? Are your logs... large? I'd also be curious if your &PH& directory has a large number of files in it and if this happens with all jobs or if this is specific to this one sequence.

Thanks Craig.

1) I think my logs are in the Xmeta database. What do you mean traditional project locaton? RT* folders?

2)The log in the Director for this job isn't large at all.

3)I cleared &PH&. It didn't help.

4)I haven't had this with another job, then again, I don't have another job sequence where the child jobs are this data intensive.

OK, then it sounds like you'll need to pursue the course that Ray outlined. And yes, by 'traditional' I meant the 'Universe' repository we've always had and the RT_LOGnnnn dynamic hashed files. You can confirm by checking (off the top of my head) the setting of RTLogging / ORLogging in your environment.

RTLogging/ORLogging is set on a per-project basis, in the DSParams file.

Tip: When installing, edit the DSParams file in the Template directory, so that subsequently-created projects already have these values set how you want them.

1)I set APT_STARTUP_STATUS on all children jobs and the sequence. All children jobs are finishing very fast. I don't see any additional info in the log of the sequence job. What should I be expecting?

2)I also tried changing the values of ORLogging/RTLogging. Do I have to stop and start Datastage services for this to take affect?

The "report on job" is a compilation option in the sequence's Job Properties.

I tried both checking and unchecking that option. With APT_STARTUP_STATUS=True, I don't see any additional information. The problem remains - 80% of the time is taken between:

.........Job Xyz has finished, status = 1 (Finished OK)

and the next line:

..........Job run requested (...)