Testing replica DR Site a failure what went wrong.

[cag]

This is the procedure I used to test our DR Site.

Veeam replication max 2 replica's.

to start up the replica I did the following.

1 take snapshop of the VM
2. set the networking to a seperate VLAN
3. Start the VM. Started OK.

Seams to start OK.
4. Shut down the VM
5. Delete all Snapshots
6. Replicate the server again.

Do all the same steps as above, except now the VM Blue screens on startup.

Tried to restore to the previous replica state. System hung on updating disk.

Now the VM Doesn't work at all. Complains about parent disk.

This was a test. Is there a problem in my procedure ?

[Vitaliy S.]

5. Delete all Snapshots

You need to revert to snapshot state (Choose "Go to"), instead of deleting it (doing this commits snapshot contents into VMDK). If you choose to delete the snapshot, subsequent incremental job runs will apply data to wrong replica state, producing corrupted resulting image. BSOD is to be expected in that case.

For proper testing replica procedure, please refer to this post.

Thanks.

[cag]

That is the procedure I followed. The post speaks about discarding the snapshop and never used the term goto, and the GUI pictures did not show this important step.

But I understand what you mean. I will have to re-do all my replica's and start again.

[tsightler]

Well, in your steps above I don't see that you stopped the replica job while you were testing. That's pretty important if you decide to perform your failover testing outside of Veeam because if a replication cycle happens while you have a snapshot of the replica then everything will be completely messed up. Why not just use the Veeam built in failover testing facility? It will handle everything for you correctly.

[cag]

Doesn't Using the failover wizzard destore the ability to continue replication after the process completes.

Also I know I did not indicate the step to stop the replication but there was no need to since it only runs once a day and I was testing outside of the replication schedule.

[Vitaliy S.]

Doesn't Using the failover wizzard destore the ability to continue replication after the process completes.

No, it doesn't. Failover wizard is doing exactly the same things as you've tried to perform manually.

[tsightler]

Right, the Failover wizard will handle everything for you, and then you can "undo failover" to revert back. It's a pretty nice feature. However, you problems were caused by deleting the snapshot rather than reverting and then deleting the snapshot. I can see how the previous instructions were unclear on that point.