Manually testing replica failover (including DC)

[nebojsa]

Hi, our DR procedure is based on Veeam replicas and we need to perform a test of the DR plan manually, i.e. without Veeam technologies such as SureBackup, Failover Plans etc since Veeam BR server will be unavailable during the test. So basically I need to manually emulate the SureReplica process.

I have two questions in this regard:

1. in order not to corrupt replica jobs, is it enough to just revert to the latest replica snapshot (restore point) before powering on the replica and reverting to it again after powering off the replica (when the test is done), or is it necessary to create an additional helper snapshot as described in post10941.html?p10941 ?

2. as a part of the test, we first need to recover our two domain controllers replicas, one of which (DC1) holds all FSMO roles. DC replicas are created from backups which have Application Aware Processing enabled. I was planning on:
- powering on DC1
- waiting for it to reboot twice (non-authoritative restore)
- performing authoritative restore as described in https://www.veeam.com/kb2119 section "Restoring entire AD infrastructure (AKA “all DC’s are lost”)"
- powering on DC2
- waiting for it to reboot twice (non-authoritative restore) and start replicating before I proceed with the rest of the VMs

My questions are:

2a. is this plan OK or should I power on DC1 and DC2 at the same time as described in KB 2119? SureReplica jobs which we use power them on one at a time and pass without a problem.

2b. Will DC1 and DC2 after manually powering them through vSphere client, automatically perform a non-authoritative restore, or I need to go through additional hoops, eg. post54554.html#p54554 ?

Thanks a lot!

[DGrinev]

Hi,

First of all, I'd like to ask is there any reason why you manage your replicas from the production site?
You can dedicate replication management to the separate Veeam B&R instance at the DR site. This allows you to failover in case of disaster at production.
Also, you will be able to utilize all features available in SureReplica which makes testing much easier.

In case of manual testing you should shut down production VMs or isolate them from the replicas, otherwise, your DCs will go crazy. There are so many things you should not forget.
The manual approach is possible, but all the things already automated in SureReplica specifically to avoid any mistakes you can make accidentally. Thanks!

[nebojsa]

Hi Dmitry, thanks for your reply.

I am aware of this, but the customer's request is to perform testing in this way and implementing a new VBR instance is currently not an option because of limited time. DR site will be completely isolated from the production site and VMs during the testing.

[DGrinev]

1. All changes will be discarded unless you initiate the merge process. You can review this thread for additional information.
2. You can fire up both DCs without delay.

The major part of the manual testing described in the KB2119 that you've mentioned. Thanks!