On Fri, 2006-02-10 at 09:33 -0500, Greg Forte wrote: > I am slowly building something similar, though with 10g and ES and not > Itanium (EM64T, though), and the outcome so far is a qualified "maybe". > Oracle is mostly OK with failing over from node to node, using just init > scripts I've crafted and the standard CS stuff - the database engine and > the listener work fine, as long as I temporarily change the hostname > during startup (and shutdown) of the services to the one it expects to > see (which is an alias associated with a floating IP address that always > gets assigned to the node running Oracle, anyway), but Enterprise > Manager is not so happy - parts of it work, parts of it don't, > specifically backup scheduling, patch grabbing/application, and > communication with the listener. Other services that use Oracle have > been set up and are happy as clams, though. I'm still waiting for some > help from Oracle on what the issue might be with EM. I'll post further > info to the list as it comes ... if anyone else has been here before, > please do pipe up! Here's a RHCS4-ized Oracle iAS (infrastructure) script. I think it lives in CVS head. An older version of this script was used for CFC certification on RHEL 2.1 - so things may have changed since then, but it's probably a reasonable start. I don't recall having to have the tweak the hostname; that sounds like a new thing to me. Also, I don't know all the possible ways to get Oracle running -- chances are excellent that this won't work except for an "iAS Infrastructure" installation. -- Lon
Attachment:
oracleas
Description: application/shellscript
-- Linux-cluster@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/linux-cluster
