Pages

Tuesday, March 29, 2016

Manager Status Shows System Hold, Fix Manager before resetting counters

Manager Status Shows System Hold, Fix Manager before resetting counters

Issue started after bouncing all the servers.

Cause:

An orderly shutdown of the concurrent manager was not performed. The CP Diagnostic Request Analyzer" shows that a CP Shutdown was initiated at on 16-OCT-12 at 22:22:41 and then cancelled.

Yet, the database alert log file shows that the database was shutdown within four seconds of the concurrent managers being shutdown/cancelled, not providing enough time to perform an orderly shutdown of the concurrent processing server.

Solution:

1.Shutdown the Concurrent Processing (CP) server in an orderly manner and verify that the shutdown is complete before invoking a shutdown of the database and/or the database listener. Check the last CP process to exit/terminate, the Internal Manager, with the Unix command: ps -ef | grep FNDLIBR

If a manual shutdown is being performed, use the same sequence found in the adstpall.sh – "Stop All Oracle E-Business Suite Enabled Services" script: first the CP Server--once down--shutdown the App Listener, the Forms Server, the Web (Apache) Server. Lastly, if necessary, shut down the database listener and/or database.

Note 1:The start up is the same, but in reverse order. In this case, a complete bounce of all servers (apps/db tier) restored a normal manager status/state.

2.In the event that the CP server was not shutdown properly, please run the Concurrent Manager Recovery feature to address any Concurrent Manager / Concurrent Processing issues within the Oracle Application Manager.. In some cases, it may be necessary to bounce the database.


Note 2:Shutting down the database or the listeners without an orderly shutdown of the Concurrent Processing server is the equivalent of pulling the power cord on a Unix box. 999 times out of 1000, the Unix system will recover without any incident; however, at some point file/disk corruption or an odd state will occur; similarly, the same applies to the CP server, manager queue and process state information may not be current or correct.

Note 3:The Concurrent Manager Recovery was created by CP Development. It is a helpful for re-initializing the managers, particularly when a manager's normal operation may have been compromised by manual intervention by a DBA or System Administrator (by manual killing of OS / RDBMS processes) or due to unplanned system outages (an accidental closing of the database without doing an orderly shutdown of the concurrent managers, power / network outages, etc).

OR

1.Go to the Concurrent Manager Administer screen >.

2.Please click the "Fixed" button and restart the concurrent manger.

NOTE:The button will be enabled only when the manager is stuck in: "System Hold, Fix Manager". Then bring the mouse control to the specific manager to get the button enabled.

Reference metalink Doc ID 1507580.1

No comments:

Post a Comment