I have a client who is running a 24/7 operation. They're on Windows 2012 Server with the physical boxes located within a datacentre. They're running 10.2B08 Enterprise (Replication licenses are bought but it's not yet implemented).
There's a slight conflict developing at the moment because the datacentre want to keep the OS patched and up to date for security reasons, but the client doesn't want to give them the downtime necessary for rebooting.
So the question they've come to me with is in regard to windows clustering of some sort. From my research I gether clustering is supported for Windows 2008 R2, but not other OSs? It's something I have no experience about and was wondering what solutions others have used to get around this conundrum.
You will still have a downtime as cluster failover will bring the db down on an active node and will start it on secondary node. Ie, the users will be kicked out. I am just wondering whether it is worth the hassle, as the reboot is usually quick nowadays, so it would be a matter of application/db being down 10 mins (reboot) as opposite to 2 mins (fail-over) give or take. I don't think that 10.2B procluster is 2012 compatible, but assuming databases are managed by adminserver, you can cluster the AdminServer service, or just add _mprosrv.exe manually in cluster manager. Or if this is virtual, use the failover on the hypervisor level.
Thanks Libor. That's really useful to know. I don't know anything about Windows clustering at all. I'll try and dissuade them as it makes life a lot easier!