This content has been marked as final. Show 6 replies
With or without Data Guard?
The behavior you are describing is called "shooting itself in the head" and is caused by an instance determining that it is no longer a member of the cluster. This can be caused by a cache fusion memory interconnect failure, a storage access failure, or other problem. Potentially enough different causes that you should open an SR at metalink immediately. Be sure you run an RDA first and attach it to the SR.
But one quick question ... why Veritas anything? It is totally unnecessary with 10gR2 RAC and is not just adding cost and complexity it is eating CPU like a piranha.
We are not using dataguard.
Veritas Storage Foundation is used for creating the replication for the DR sites.
One of the RAC nodes is rebooting every 3 days and all the HP-UX log files were showing some panic GAB and related to Veritas Cluster. We have esclated to Symantec also. But they said reboot could be due to Oracle 10g RAC. So we are struggling to find the cause.
Does this Veritas Storage Foundation consume too much memory also? TOP command give only 100 MB free during peak.
I have raised SR.
I see no value in an Oracle RAC cluster in using any Veritas tool.
Data Guard does better replication and it is free.
ASM does better storage management and it is free.
If you want to solve the problem I would seriously suggest dumping Veritas and going with a best-of-breed Oracle solution.
Why? Because it works. And is fully supported.
If you can not then create an RDA and open an SR at metalink. That is the best we can help you with.