Comparing HA-Lizard and XenServer HA
-
Years of uptime, as awesome as it sounds doesn't really work.
It sets the VM to a suspend state while it build the backup file. (At least this is what NAUBackup does).
So technically speaking your Uptime counter should still continue, as far as the VM knows.
-
@DustinB3403 said:
Years of uptime, as awesome as it sounds doesn't really work.
It sets the VM to a suspend state while it build the backup file. (At least this is what NAUBackup does).
So technically speaking your Uptime counter should still continue, as far as the VM knows.
I take it you can't use that on things like SQL and Exchange then?
-
I haven't had anything to try yet that is heavily used. We have our Asset management system which is built on a SQL Database, and that gets backed up weekly without issue.
But there are only 2-4 people accessing the database at a time, at the most.
-
I was under the impression that there was only a blip in the machine's activity as the snapshot was taken (this usually doesn't take but a second or two, right?).
-
Just testing now, it takes maybe 3-5 seconds to run per VM in our environment.
-
And that is only the Snapshot portion, migrating it off host obviously takes a bit of time.
-
@DustinB3403 said:
And that is only the Snapshot portion, migrating it off host obviously takes a bit of time.
Sure, but other than using some resources shouldn't really be noticed by the VMs themselves.
-
@DustinB3403 After the snapshots are complete, the VMs should go back to running.
-
@dafyre Correct they do.
Immediately after those first few seconds. The data transfer off host takes the longest amount of time to complete.
I have 4 VM's that we backup (care about) that use 696GB of data, so the transfer time takes a bit.
-
@DustinB3403 How much does that affect the performance of the VMs that are running while your backups are being copied off?
-
I've never monitored them.
The work is being done by Dom0, not the VM.
And with the resources being statically assigned I can't imagine that there is much of a hit to the performance of the VM's them self.