Replica fails after cold migration

[singy2002]

Hi,
I am trying to understand if what I am doing is wrong or if this is just a support issue.

So I have a VM on cluster 1 which is on FC storage. I need to move this VM to cluster 3 which is iSCSI storage, the two clusters cannot see each others storage so the only possible solution is a cold migrate.

I cold migrate the VM, bring it up, its still on the same VLAN, still managed by the same VCenter so i assume everything should be ok. When the next replica task runs it fails after 9kb. now i am fortunate in that I can remove this VM from the destination server and readd it to the job and let it do a full replica again (100MB LES) but i want to prevent this sort of thing in the future as we will be doing ALOT of moving of systems between old FC and New iSCSI and I am not permitted to allow the two clusters to share storage (if only)

So should this work? can i make a change somewhere to correct this? does the job need to be recreated or atleast this VM removed and readded into the job and from the destination host.

Hope this makes sense.
Regards
Andy

[foggy]

Andrew, Veeam B&R tracks the unique moref ID of a VM (assigned by vCenter). This ID is not preserved during cold migration and the VM is no longer recognized as the same VM by the product. To continue the replication, you should re-add the VM to the backup job and map it to the existing replica to avoid full job run (no need to remove the VM from the destination server). Thanks.

[singy2002]

how do i "map it to the existing replica"

[foggy]

Described in user guide, p.112.

[singy2002]

Sorry i should have mentioned this, we are still on V5, does this still apply?

[foggy]

Well, I thought you were talking about v6's Quick Migration.. And replica mapping is also available in v6 only.

How did you migrate the VM? ID change might not be the case if you initiated migration using the corresponding vSphere Client command as vCenter should preserve the ID.

[singy2002]

Yes i shut the VM down and did a migrate VM and datastore to a different cluster/storage.

[foggy]

Then I'd suggest to proceed with support team on this. Logs review is required here.