Hello all,I actually saw that a lot on 10.2 RAC databases on Linux at least: the status of the service as recorded in the OCR didn't match what was in v$active_services. One thing to remember is that dbms_service.stop_service only stops the service on the active node-if it's a TAF service I'd expect it to be up and running on more nodes-again the views should tell you where. You should connect to the instances that are reportedly running the service and stop it via dbms_service. Then try dropping the service. I wouln't use dbms_service to do so, AFAIK it doesn't propagate the fact the service was dropped to the Clusterware layer.
Also worth noting that dbms_service is pretty much deprecated in the current release-have a look at the PL/SQL API reference if you're interested.
Hope this helps, Martin Martin Bach https://martincarstenbach.wordpress.com http://uk.linkedin.com/in/martincarstenbach On 04/08/2011 15:13, Jeremy Schneider wrote:
I haven't come across this before, but have a few more ideas for things to check...
.... -- //www.freelists.org/webpage/oracle-l