What would cause a job to abort in the Prelude Scheduler?

[DonQuixote]

I have a program that extracts Customer Data and puts it into a flatfile.  It is scheduled to run at 9:45pm nightly.  Lately, it works on some nights and aborts on others; not the same night every week.  But, it always runs if you manually set it off.  What would I be looking for to solve this problem?

[Tom Pellitieri]

Any number of issues can abort a job - usually a request for user input is the biggest offender (e.g., a call to SB.DISP).

Did you check the most recent _PH_ log for the aborted job?  If you are in Scheduler Inquiry, you can press F7-Last _PH_ to see this.  You can also search the _PH_ file directly from TCL.  Usually, at least the most recent three jobs are present.

Assuming you're not on a very old version, you can find the records at TCL using the following SELECT statement:

SELECT _PH_ WITH F3 = "[!xxx in SCHEDULER ]"

Where "xxx" is the Scheduler Job Number (Sched#).  Note also that the word SCHEDULER is all upper case. 

You can then edit the _PH_ file and look for errors.  I recommend making sure the editor shows unprintable characters (in ED/AE, enter the up-arrow (^) as a command before listing).  Any errors should be near the bottom of the file.

[Colin Alfke]

Ditto on Tom's note to check the _PH_ directory for the output from the job. Most times the reason is noted. We have a daily overnight job. There are two main culprits that cause it to fail:
1. The server is being backed up and the file is locked at the O/S level causing a write failure
2. The server is rebooted (windows update, maintenance, power issue....)

hth
Colin

[Thomas]

This could also be cause by a UNIX permissions issue.  If the scheduler gets started by different users, and your users have different permissions, that could cause your job to abort.  But as previously stated, the Job log will give you some hints, as to why this job crashes.