I'm trying to put together a design for a client that isn't full HA/clustered for cost reasons, but would like to take advantage of instant recovery. The backup server would not be as powerful or have the storage of the main server, so it would be a stop gap measure to stream the VM's from the backup repository (NAS) - until the main server was restored in event of a hardware failure.
I have done some testing in a lab environment with two servers, both running hyper-v (one just standing by) - and a copy of Veeam installed on each Hyper-V base, and got it kind of working.
I say kind of - because I had the backup server running a trial copy of Veeam Backup and Replication, as well as the main server - but in production, one would only have one license available.
I then simulated a failure and opened up Veeam on the backup server, and thought I would just import the config - but that failed because it the config had different ip addresses and hostnames from the other server.
Okay...so I just did a fresh discovery of the NAS repository, imported existing backups, hit instant recover and that worked great.
Problem is, in a production environment if Veeam is installed in Hyper-V - it is gone when the server dies, so...would I be able to just apply the license key from the Veeam install on the failed main server to the backup server or will it spit the dummy at me and leave me without a way to quickly recover the VM's?
Is the only real way to do this to have another physical backup server with Windows installed on it instead of a NAS and then manage the Hyper-V environment from this backup server rather than installing directly on the Hyper-V hosts?
Hope that makes sense....if that's the case...sure would be nice to have a linux iso/appliance that just installed on bare metal or kvm or something so I didn't need to buy extra windows/windows server licenses.