Intelligent Agent Fails to Restart

We are running 9.2.0.4, two instance (and two node) RAC on Win2k3 Server EE.

I have a TAR open for another issue (RMAN jobs scheduled through OMS are completing per "jobout" log files on target node, but OMS never receives a "job complete" message from the IA and OEM reports the job as "started" forever, until deleted and resubmitted).

Anyway, one of the suggested solutions involved modifying snmp_rw.ora (bind IA to single NIC) and restarting the IA on the target node. The problem is that the IA will not restart. It fails with error 259 (in Windows Services Manager)

The last entry in the nmiconf.log (for each attempted started) indicates that it is "Discovering services on SBLA8100..." After each start attempt, I have to manually kill the process tree for that service (DBSNMP continues to run) before I can try again.

I have been down this road before (we actually had problems with the IA starting upon a reboot when we first put the DB into production; that was cleared up by disabling the SNMP-based HP Insight monitoring agents on the server, putting the IA on manual start and scripting its startup), so I'm very familiar with the IA "clean start" process. I've attempted it many times and it seems that only a reboot is going to clear this up.

I've heard some comments that this type of thing is typical of the IA, especially at the 9.2.0.4 level. Has anyone else experienced this problem? If so, has anyone seen it go away by going to 9.2.0.6? We were holding off on the 9.2.0.6 patchset as we've been pretty solid at 9.2.0.4 and we're going to 10g in November, but I'm open-minded on that plan. Any thoughts/suggestions? (and yes, I've already mentioned this in the TAR; all I've found in metalink is the "clean start" suggestion amongst some other odds and ends, which I've tried)

Thanks.

Jesse Received on Fri Jul 15 2005 - 14:06:04 CDT