I'm trying to put together a plan to recover from potential OVM failures, but I'm unclear about a number of points. Any suggestions and clarifications would be most welcome.
We're installing a backup agent on all the VM's (Red Hat and Oracle Linux). If the VM was somehow completely lost, we plan to re-create it, do a fresh o/s install, and restore the filesystems from backup.
It runs as VM. To recover from a problem, we plan to reinstall the VM Manager using the same UUID, rather than backing it up. If we were to back it up, are there critical configuration files we should back up?
This is exported and backed up regularly.
Oracle VM Server:
I'm not sure about this one. If we somehow lost one server (we have a 3-node cluster), if we just reinstall Oracle VM Server with the same ip and hostname would all be well after I rediscovered it? As long as the other two nodes were up and running, would the rebuilt server be added to the ocfs2 cluster when it's rediscovered?
Also, what information is in the server pool filesystem and the database? How do you all back up the server pool filesystem (or do you?) ? We can't install our backup agent on the Oracle VM Servers.
there is no need to backup the Oracel VM Server .
if something wrong with the OVS,not the ocfs2,you may wipe the db for the OVS,and the restart the agent then reboot the OVS.The ovs would be fresh as you just installed it