Backup Host for Replication Failover

[Karl Kunze]

Hi,

at time we run a single ESXi 6.0-host, on which Veeam B&R is replicatiing to a NAS.

Because this host is a little aged we added an additional server as second ESXi 6.0-host and plan to prepare this new machine as backup-machine to be powered on in case the old one dies.

I plan to place a copy of our Veeam B&R-vm on this new host. So in case the backup-machine gets powered on Veeam B&R is at hand fast and available for doing the failover to the NAS-stored replicas gracefully via the Veeam B&R-Gui (I had to do disaster-recovery from the replicas manually without Veeam B&R once, and that was not fun).

From that back-in-business-point I would be able to plan and commence the failback (or would you call it restore) to the datastore of the backup-machine.

Could this be a working strategy?

Regards

Karl

[foggy]

Hi Karl, not sure I totally understand your scenario - are you doing replication or backup to NAS now? Looks like you're doing backups and plan to restore VMs from the backup to the second (normally turned off) host in case where the original host dies, am I right? As opposed to replication, where you have a second host always turned on and holding up-to-date ready-to-run replica VMs to failover to.

[Karl Kunze]

HI Alexander,

Veeam B&R runs on the old host and we backup to disk and to tape on the old host. Additionally we replicate from the old host to the NAS. The scenario I am thinking about starts with the old host gone with dead hardware. In that case we have only the replicas on the NAS at hand, but not a running Veeam B&R-machine to initiate the failover.

My idea is to put a shutdown copy of the Veeam B&R on the new host, link the replicas on the new host, so that in case its needed the new host can be started and the failover-process can be initiated with the Veeam B&R-Gui and with the option, to do the failback towards the datastore of the new host.

Do you get the idea?

Regards

Karl

[DGrinev]

Hi Karl,

It's recommended approach that you are hosting another instance of Veeam B&R in the DR site responsible for replication jobs.
So, in case of disaster you can immediately initiate failover. Thanks!

[Karl Kunze]

Hm, ok, I will ask with a different approach:

What I did test so far is replicating the Veeam B&R-vm not only to the NAS but direct to the datastore of the new backup host. This is done by a second replication-job scheduled as follow-up to the first replication-job (first one for replicating all the production-vms). So I have a Veeam B&R-machine with the latest entries of the replicated production-vms in its database.

My problem is, that I need the second host being up and running all the time for this to work.

My intention is to bring this second host to life only when the first host has died. Then I would bring up the second host manually and start the replicated Veeam B&R-vm manually. What I see now as a problem after testing is the fact, that this manually brought up Veeam B&R-vm does not have the latest snapshot-/replica-/restore-point-data of the production-vms in its database.

Is there a way to bring up that replicated Veeam B&R-vm manually and then sync its replica-information-database with the replicas found on the NAS?

Regards

Karl

[foggy]

If NAS is connected to the source host as a datastore and holds replica VMs, you could connect it to the target host in case of DR and import replicas there for further failover.

[Karl Kunze]

Ok, connect NAS to target host, start the backup/restored Veeam B&R-vm on the target host, but how would I import the replicas there to have access to different restore-points i.e.?

Regards

Karl

[foggy]

You need to register replica VMs on the new host and then map your job to them - after that you will be able to failover using Veeam B&R.

[Karl Kunze]

Ok, registered one replica on the new host resulting in an entry without "_replica"-suffix. Is this correct? I mean, it is with is settings like datastore etc. linked to the replica on the NAS?

Then I struggle with your word "map". Do you mean the mapping-option getting availabe when checking the "low-connection-bandwith" in the replication-job-settings?

Regards.

Karl

[foggy]

Yes, I mean this option. Please note, that you need to perform at least one job run after mapping is done (which will consolidate all older restore points).

[Karl Kunze]

As we are talking about DR with lost original-vms and initiating the "failover to replica"-process including access to different restore-points, how can I perform at least one replication-job-run including consolidation.

Regards

Karl

[foggy]

Then I still don't get how do you replicate VMs to NAS. Anyway, to perform failover, you need a replication job with the original VM as a source and replica VM, to which you're going to fail over, as a target. You cannot fail over just to any VM on whatever host.

[Karl Kunze]

I do not want the backup-host be permanentely online. We do not need that. I want the backup-host be prepared to be switched on in case the old host goes down. Switch it on and then manually start an also prepared Veeam-backup-vm on the backup-host. I need to be able to first initiate the failover to the replicas on the NAS - in this scenario the orginal-vms are gone - and then failback to production with redirecting the restore to the datastore of the backup-host.

I got this all including the mapping, but I do not understand what you mean with "Please note, that you need to perform at least one job run after mapping is done (which will consolidate all older restore points)". Or shall I do this mapping before DR while everything is fine as part of preparing the Veeam-backup-.vm on the backup-host?

Regards

Karl

[Karl Kunze]

Hi,

I am still a little lost here. Any further help on finishing my preparations are highly welcome.

Regards

Karl

[foggy]

I do not want the backup-host be permanentely online. We do not need that. I want the backup-host be prepared to be switched on in case the old host goes down. Switch it on and then manually start an also prepared Veeam-backup-vm on the backup-host. I need to be able to first initiate the failover to the replicas on the NAS - in this scenario the orginal-vms are gone - and then failback to production with redirecting the restore to the datastore of the backup-host.

To be able to failover within Veeam B&R, you need a replication job, which cannot be performed to an offline host. I just don't get how you're going to store your replicas on NAS. Or do you mean backups, not replicas?