I'm running servers and I can't afford the "downtime".
Here on the Server forum, we all run servers
Well, there is a way to suspend your VMs to backup them. But, the time difference between "suspend" and "shutdown" is in equal the time which the guest OS needs to be correctly shut down (one minute) and to start (also one minute).
Think about it: shut down means downtime. Suspend means downtime, too.
Is there a way to "suspend" or "pause" the VM's?
Sure. But I would really suggest you:
never suspend any production machine if you want to keep your correct data. Let me describe, why.
Do you know something like suspending in the real server world? No? Me too. And your servers don't know this as well.
Now, imagine, your server software works with transactions. Nowadays, it does
every server OS and every server application, such as MSSQL, Exchange, Oracle, MySQL...
Let's say, we run some DB on a (real) server, which computes a bigger transaction right now. And it looses the power. What will the DB software do after you brought the server back? It will see the not committed transaction and will roll back it. So, the data is consistent.
And now let's play this game on a virtualized server. It accomplishes our transaction, and in the middle of it, our backup script / software / user suspends the machine. After restoring, the DB software will think: "Hey, I had a crash, but my transaction runs further. No way!" and will produce a rollback. And it means, your transaction is either deleted without logging (best case) or is executed until the suspending. I have also seen that the same transaction was accomplished twice...
