Failover Best Practice

[lukeup]

Hello,

I am testing failover and failback and a few questions have arisen about the best way to do it.

Firstly, the production vm is shut down/broken and then failed over to the replica, which keeps the same IP and everything is working. Great.
But once the production VM is ready for fail back there are problems:

1) Both VMs need to be on for failback but have the same IP.

2) If the production VM is back online for failback with the same IP, users could mistakenly use the server before it is ready, and there would be a lot of confusion.

3) Obviously, the replica is being used as production whilst failback occurs, so when do you stop using that server and move back users.

It would be better if you could failback and then boot up the production VM when ready. That way, access could be controlled.

It seems easier to delete the original VM and replicate back the other way, then failover to it when ready.

What is the best practice for failback to production server?

Thanks

[foggy]

1) Both VMs need to be on for failback but have the same IP.

Actually, this is not true. The original VM needs to be powered off to perform failback (and if it is powered on, it is powered off automatically when failback is started).

[lukeup]

Thanks for reply, am just trying failback again

So when the failback completes does Veeam handover to the original VM and shut down the replica?

Do users carry on working on the replica right up until the servers swap back?

Is there any noticable downtime?

If it isn't automatic, how do I know when to shut down the replica and move back to the original?

Is there a window where data is lost?

None of this is explained in the documentation, I am not sure what failback actually does

[lukeup]

Also, calculating original signature hard disk is taking ages. 19% in 30mins.

It was only failed over for a few minutes and no data changed.

[foggy]

So when the failback completes does Veeam handover to the original VM and shut down the replica?

You can select whether to power on the original VM on failback completion or not while going through the Failback wizard. The replica VM is stopped automatically.

Please review the detailed description of the failback process given in the product user guide, on p.50. Should answer all your questions.

[foggy]

Also, calculating original signature hard disk is taking ages. 19% in 30mins.

It was only failed over for a few minutes and no data changed.

Please see the existing topic for explanation.

[lukeup]

Ok thanks