Job control process (pid XXXXXX) has failed

[caola]

I have a sequence job that invokes a korn shell.
The sequence is like:
job to create a flat file---->invoke a korn shell that use the flat file as input

In the korn shell it reads from the flat file which has a list of items and then does some db2 database operation for each item in that file.

There may be up to 30000 items in the flat file, so it may take 1 hour for the ksh to complete.

The invoke of ksh is not shown in director, but i can find the process by ps -ef in aix. After nearly one hour's wait, the sequence aborts.
The warning is just "Job control process (pid XXXXXX) has failed".
It seems the sequence lost the track of the ksh status.

My question is:
In the ksh it prints each item that processed to the screen, is there a limit on the length of the total output to screen?
Is there a limit of wait time for the ksh to finish?
If there is too much output to the screen, will it lead to above error?
Is it a known ds bug? as version 7 is out of support, there's no way to find the info of fix pack..

[chulett]

How exactly are you executing the korn shell from the Sequence job? If you are using the standard Execute Command, it won't be logged until it completes so it can tell you the status and whatever the 'command' writes to standard out. There is definitely a limit on how much a single log entry can hold but I don't recall anything failing because of that, instead it just gets truncated.

I think you need to debug this from the shell script... write debug messages (what I call breadcrumbs) to another file as you work, you can even tee off standard out/error as well. This so you can see how far it gets before it fails and if anything evil is going on.

[qt_ky]

It may also help to debug your script by redirecting its output to a new log file. Then you can check in the log file for clues.