VMware Cloud Community
wayneoakley-msi
Enthusiast
Enthusiast

how to move the replication target to a new datastore?

i have 50+ replicated virtual machines on a datastore on a san

a new san is coming online and i need to relocate the replication target to that datastore.

cannot seem to find any way of moving the datastore content other than visiting each replication job, reconfiguring to use the new datastore and wait while it does a full sync again.

does anybody happen to be aware of an easier and more efficient way of relocating the replcation target and its content?

thanks

Reply
0 Kudos
4 Replies
mvalkanov
VMware Employee
VMware Employee

Hi,

Basically you can:

1. Create a copy of the base vmdks (without the hbrsrvdisk vmdks).

2. Unconfigure the old replications.

3. Configure the replications using the copied vmdks as initial copies. It will do a "full-sync", but mostly exchange checksums and not whole data blocks.

To create the copy:

A) If using SRM - perform test recovery, clone the recovered vm, unregister it from vCenter Server and re-use its disks as initial copies.

B) If not using SRM - perform recovery without syncing latest changes and without powering on the recovered VM. Migrate the VM to the new datastore, unregister it from vCenter Server and re-use its disks as initial copies.

C) Manually copy the vmdks with scp / cp.

D) Copy the vmdks with Datastore Browser download / upload. You would need to edit the new vmdks and fix the disk uuid values to match the original ones, as otherwise they won't be usable as initial copies.

Regards,

Martin

Reply
0 Kudos
sakibpavel
Enthusiast
Enthusiast

The following steps will allow you to move the VMs and to configure replication to continue with the new datastores.

1. Set up new datastores on the target hosts with the new storage.
2. Rename the VM to not include the _VZreplicate in the name
3.  Storage vMotion the VMs to the new datastores, this will relocate the  VM directory to the new datastores including all the VMware and  vReplicator files and will keep the same name as what is shown in the  inventory
4. Rename the VM back to include the _VZreplicate in the name
5. Edit your replication jobs to have vReplicator replicate the VMs to the new datastore
6. Run the jobs again this time the migrated VMs will replicate the changes to the new datastores

Reply
0 Kudos
wayneoakley-msi
Enthusiast
Enthusiast

Thanks Martin.

Kind of what I thought might be necessary, was not aware of the need to change the disk uuid if doing a datastore browser.

Given the effort with each of the steps I think it might just be easier to reconfigure each of the jobs (which needs to be done anyways) and let it do a full sync with data copy. Will take longer but less opportunity for errors.

Reply
0 Kudos
wayneoakley-msi
Enthusiast
Enthusiast

Thanks but still seems to be a lot of manual effort for a fairly basic function, seems a bit short-sighted by the vReplicator team to not facilitate migration of the target datastore to other storage.

Given fast storage it could be faster and a lot safer to just delete all the jobs and recreate them pointing to the new storage and let it full sync on its own. Far less effort on my part to be sure.

Maybe the dev folks will provide something in a future release.

Reply
0 Kudos