We run all of our dozen+ SQL servers as virtuals under Hyper-V. Our servers are configured with the Master, MSDB and Model system databases on E: and F: (the same as out user database) and Temps on G:. Hyper-V allows snap-shots of drives.
It seems to me like it would be better to put Master, Model and MSDB on the C: (system) drive. If I ever have to do complete server restore I can just use the last snapshot of the C:. Boot it. Then restore all of my user databases.
Does anyone see a problem with my thinking?
In the old days when SQL server lived on actual hardware the C: drive dying was a real risk and it made sense to put the system databases on the data and log drives. In the virtual world this doesn't make as much sense to me. If I install a service pack that blows up a server I can always fall back to the last good server snapshot. I do SQL backups Master, Model and MSDB... but they rarely change. We don't add/remover users often or change security. All the jobs in MSDB are generic across all of our servers and do not change. We do a complete set of Complete + Differential + log backup so we can restore our user databases to any point in time down to 10 minutes. Taking a snapshot of our Data (E) and Logs (F) drives seems redundant, not very useful in a disaster recover situation and takes up a bunch of backup media.
We did a disaster recover drill. The old SQL server with the system databases on the C: drive was a lot faster to get up and running than the new SQL servers with the system databases on E: and F:.
What are your thoughts?