Veeam Replication to Datastores

[jimmycartrette]

I'm setting up a replication, I was hoping that I would be able to replicate to a datastore cluster, but it looks like that's not possible (unless there is a way you know of)

So the original plan was to use tags (Bronze, Silver, Gold) for RPOs/Replication job schedules, but now it starts to make more sense to organize those RPOs by datastore so the job always goes from the right datastore to the right datastore. Is that what everyone else does?

Any downside to having 11 Replication jobs instead of 1 for Bronze?

[Vitaliy S.]

Hi,

Yes, it is not possible to select the datastore cluster itself. If I were you, I would suggest using tags as source for replication jobs (in this case all new VMs will be picked up automatically) and create multiple jobs (VMs should be grouped by their RPO and VM role) for better job management.

Thanks!

[jimmycartrette]

I wanted to use tags, but I cannot see how anything but datastore sourcing works since we must target individual datastores.

It has to do with sizing. So if I have 11 LUNs, all of the same replication category, and I have 11 LUNs at the destination, then to balance, I have two choices.
1. Use tags so for this bronze tier we have one job, but then go into the veeam job and go through all of the vms and choose (mapping 1:1 is easiest) a target datastore in the target window.
2. Use 11 jobs, one for each datastore. The target for LUN0 will be DRLUN0 so the balancing of datastores is done on the front end.

I do have one question - now that I have replicated my VMs, for VMs with existing replicas, if I move them to a different job that has a different target datastore, will it respect the existing datastore or will it move the replica to the new target datatore? If it respects it, I see how that now I've seeded the replicas I could create one job and just target a mostly empty datastore for new VMs that get picked up.

Any other way, it just seems best to replicate based on datastores (associate RPOs and jobs with datastores), instead of tags.

Am I wrong here?

[Vitaliy S.]

It has to do with sizing. So if I have 11 LUNs, all of the same replication category, and I have 11 LUNs at the destination, then to balance, I have two choices.
1. Use tags so for this bronze tier we have one job, but then go into the veeam job and go through all of the vms and choose (mapping 1:1 is easiest) a target datastore in the target window.
2. Use 11 jobs, one for each datastore. The target for LUN0 will be DRLUN0 so the balancing of datastores is done on the front end.

Hmm... creating jobs per datastore seems to be easier.

I do have one question - now that I have replicated my VMs, for VMs with existing replicas, if I move them to a different job that has a different target datastore, will it respect the existing datastore or will it move the replica to the new target datatore? If it respects it, I see how that now I've seeded the replicas I could create one job and just target a mostly empty datastore for new VMs that get picked up.

If you move a VM to another job with a different destination, the replication job will create a new VM, since it will not find any existing VM replicas in the target location. If you want to continue using VM replicas as a target for your source VMs, you will need to use replica mapping functionality.