OEM GC 10gR3 jobs and UDMs

Hey all,

I've installed OEM GC 10.2.0.3.0 on OEL 4, after giving up trying to follow Oracle's instructions for moving 10gR2->R3 on Windows. Things are mostly working, but I've seen a few important glitches.

First, is there any method for notification of OEM job failures? I can't find one here or in the 10gR1 DB Control that this is replacing. Kind of a bummer when our RMAN backups fail because of a full flash recovery area and there's no emails generated to let me know about it. And I can't find any sort of mention of this in Metalink or any of the online docs (local to OEM or at tahiti.oracle.com).

Second, I've created a user-defined metric to monitor the size of our top-n tables. It worked great when testing it at five minute intervals. As soon as I altered the interval to the production once-a-day, OEM cleared all the alerts with the message "CLEARED: Collection disabled". But the collection is enabled and is still running once-a-day, as scheduled, according to the Collection Timestamp on the metric.

I won't mention how adding the SQL Server plugin apparently broke my test Agent. <sigh>

Seeing as my average resolution time is >4weeks for an SR, I thought I'd try here first...

Anyone?
TIA!
Rich

--
http://www.freelists.org/webpage/oracle-l