LLAWP process is buggy after executed kill -9 command

Document ID : KB000047273
Last Modified Date : 14/02/2018
Show Technical Document Details

Issue:

In R12.52 SP1 CR05, we encountered 500 error after shutting down LLAWP process by using "kill -9". 

BUT, in R12.0, not encounterd 500 error, because LLAWP restarted properly. 

We believe that this 500 error is related to shared semaphore process. 

When I confirmed by ipcs command, semaphore processes are existing after kill command. 

But, in R12.0, semaphore processes are cleared after kill command. 

We think that R12.0 is better than R12.52. Because in R12.0, if LLAWP stop unexpectedly, LLAWP is recovered by using existing semaphores and stating up automatically. 

If in R12.52, if LLAWP stop unexpectedly,LLAWP is not recovered and customer has to wait until semaphores are cleared for a bit long time. 

 

Environment:

Apache WebAgent: R12.52 SP1 CR05

 

Cause: 

This issue is product bug.

Normally when LLAWP goes down while Apache is running  , it is restarted during subsequent requests and DoManagement call.

This was the behavior observed in 12.0.

The current behavior in 12.52 is due to a bug which seems to have got introduced in 12.52SP1. Due to this bug if the LLAWP process gets killed without a clean shutdown, its process id remains in memory and hence the new LLAWP process fails to get spawned.

 

Workaround:

This issue will be fixed in R12.52 SP1 CR07