hardware wise i'm thinking skip raid 1, use a "enterprise grade" ssd intel s3500 and apply every 15 min incremental replication to a backup box running hyper-v. i known it's not the same as raid 1 redundancy, but truthfully my entry level server is more likely to have a power supply crap out on me than the ssd i think. i can live with firing up a 15 min old replica of my server on a backup box, then non-urgently fix whatever part i need to on the main server, ssd or powersupply or entire replacement if need be. thoughts to this strategy?
I would do the opposite for sure. I would even use two consumer grade drives in RAID 1 instead of the S3500 alone. Server Power Supplies tend to be very reliable. There's also redundant power supplies but those are quite expensive. $400-500 for Redundant PSUs. They are also very loud. I only use them when I'm running two or more VMs on a server. At that point, it becomes necessary.
If you are going to virtualize, it's better to do it from the start. P2V migrations of domain controllers don't always work well. The best thing about Hyper V's backup is that it is VSS aware making restores much easier compared to third party solutions. I mean there are good third party ones like BackupExec, but those can be quite pricey.
In terms of live failover, replication is the best option. Restoring your backup from last night would be at least a day behind. For most people, that's fine, but if you want swift restoration with virtually no downtime, that's the way to go.
One good server and one equivalent desktop is a good combo for replication. Hyper V has the replication built into it so there's no need for 3rd party software. Use a gigabit link between the servers on separate NICs. 10 Gigabit is popping up on server boards too nowadays.
.png "ACUF Donor I")
