ASAP Cannot Be Stopped In A Veritas Cluster When The Fork_agent Is Killed
(Doc ID 1342280.1)
Last updated on MAY 11, 2021
Oracle Communications ASAP - Version 5.2.4 and later Information in this document applies to any platform.
When abruptly stopping the ASAP fork_agent process (using kill –9 ) ASAP could not be successfully stopped on the current node.
This prevented the ASAP resource group being restarted either on the current node, or on the failover node.
This problem was tracked down to an issue in the ASAP.ksh script. In the stop_asap function, which is called from the Veritas clean script, the script was getting stuck in a loop waiting for the ASAP critical processes to stop. The script first calls out to the stop_asap_sys script to gracefully stop all ASAP processes, it then loops waiting for the critical processes to stop, and then finally kills any remaining java processes if they are still running. The fork_agent is responsible for shuting down the ASAP java processes, and since it was stopped as a part of the test, the java processes remain running after the execution of the stop_asap_sys script. These java processes then interfere with the loop that checks that the ASAP critical processes have been stopped. The script believes the CTRL server is still running, however, the script is erroneously picking up the java processes and not the CTRL server process. This issue was fixed by further constraining the list of processes checked by the script to remove the java processes.
To view full details, sign in with your My Oracle Support account.
Don't have a My Oracle Support account? Click to get started!