For et par måneder siden oplevede vi på et par af vores IAS 10G instanser, at listener prosessen hang, og det blev den ved med indtil man genstartede den. Det er jo en anelse kritisk på produktionen, så vores W/A var en gang om ugen, at genstarte disse prosseser.

Jeg sad så et øjeblik og surfede, og stødte på en Indisk Oracle DBA, der havde samme problem, samt en gut der havde prøvet det før. Det løste faktisk vores problem.

"I am an Oracle DBA working with an organization in India. We are running Oracle 10.1.0.3 on Sun OS 5.9. We have been facing serious problems for the last month. Sometimes our listener stops responding. I recevied complaints from our clients that the listener was unable to find a handler for the service. If I check the status of the listener, it shows successfully. But if I reload or start the listener again, it does not show the respective handler for my particular server. Then, to make my database normal I need to shut it down and start it again. After restarting, the listener starts working again. One thing I noticed is that my CPU utilization goes very high. Due to lack of resources it is not possible to add more processor. Please suggest what steps need to be taken."

I have seen mention of this problem under the exact circumstances you are describing. This problem with the listener hanging and high CPU utilization can happen when both Oracle 10g database and application server are installed on the same server. If you have 10giAS installed on the database server, you can try renaming the file $ORACLE_HOME/opmn/conf/ons.config and then restarting the listener. This file defines ports used by Oracle Notification Service (ONS), which may conflict with the listener and cause it to hang. If this does not resolve the problem, I would recommend opening a TAR with Oracle Support. For more information on this bug, you can review Doc ID # 284602.1 on Oracle's Metalink Web site.