Concurrent Managers Crashing and FNDLIBR processes going defunct

Hi,

We observed an issue in one of our R12 instances where the concurrent managers were crashing and going defunct. There were no notable error messages in Internal Concurrent Manager log. The log indicated that all the concurrent queues were started up correctly as per the workshift definition.

However, the logs of the custom concurrent managers which were crashing had the following error:

APP-FND-00362: Routine afpbep cannot execute request &REQUEST for program &PROGRAM, because the environment variable &BASEPATH is not set for the application to which the concurrent program executable &EXECUTABLE belongs.

I had no reason to suspect any relinking issues even if we had patches applied to this system prior to this issue because the relink log did not show any failures. Also, there were other managers which use the FNDLIBR executable that were working fine without issues.

Further investigation revealed that in the log of the concurrent managers which were crashing, ahead of the aforesaid error, there was one concurrent request that was picked up by the manager for execution. Each of the crashed managers had tried to execute the request before they crashed.

I cancelled the identified custom concurrent request and looked up for similar requests and cancelled one more of such a request which was scheduled.

Then bring up the concurrent managers - they work like charm!

Further investigation revealed that the concurrent program definition was incorrect. Rather, the custom top under which this program was registered was not defined correctly. At the shell level, the custom top was defined as $XXCUST_TOP where as in
the custom application registration within apps, it was defined as CUST_TOP (Base Path).

I think the straight forward way of fixing this issue with the custom top is to update the base path in application registration to XXCUST_TOP.

- Aravind Kamath Posral