AIX

Hi,

I am setting up a hacmp oracle cluster with 2 resource groups - one on each node.
Binaries are local - so is configuration but its the same on both nodes - so are the startscripts.

If I start the cluster, each DB comes up nicely on its homenode.

Now I have the problem that when I failover my resourcegroups, IP and filesystems come up properly but the DBs only sometimes. In most cases I can start the DBs with their corresponding startscripts, sometimes only the listener comes up.
Same result when I fail the DBs back to their own homenode - they sometimes come up, sometimes only the listener, sometimes nothing.

Manually executed on both nodes both startscripts are working just fine.

We run AIX 6.1 ML10 + hacmp 6.1.6

Anyone any idea ?

Regards
zxmaus

It could be a timing issue in the order that the resources start.

When you've started up manually then you know that IP addresses etc are all active in your resource group. That isn't the case with all HACMP resource groups.

There is a flag in the TCP/IP settings to make sure that the IP Address is on before the application is started. That might help.

Otherwise its time to trawl through the Application and HACMP logs.