Hi,
Just to provide an update to this issue incase anyone else attempts the same process:
VMware support reported:
********************************************************************************************************************************************
It seems Networker following this process:
- saves a copy of the VMX file during backup (instead of obtaining the ConfigInfo for the VM via the API)
- during restore, it copies the VMX file back to the datastore (instead of creating/reconfiguring the VM based on the saved ConfigInfo).
- register the VMX
- Reconfigure the registered VMX to create disks
This will work if the VM is being restored back to the same datastore and has access to the original disks. If the VM is being restored to a different location where is does not have access to the original disks, they are creating the disks but the VMX still has replication configuration entries that say that these disks are being replicated. When HBR checks the reconfigure request, it sees that the disk operation is "edit". In this case, it checks if the old disk length and the new disk length is the same and fails the reconfigure request if the sizes do not match. But since the old disks are not longer available, HBR cannot figure out the old disk size and it may be showing up as 0 in the ConfigInfo. So it fails the reconfigure operation.
If they had followed the supported method for restoring the VM by using the APIs to create it, they would not run into this issue. Though, I think even in that case there might still be some issues in some scenarios where only some disks are restored (eg, datastore of one of the disks
is lost) but as long as the backup partners use the APIs, the workaround for these more uncommon cases would be to restore the entire VM.
********************************************************************************************************************************************
I raised this with EMC who reported:
********************************************************************************************************************************************
Networker has not been qualified for SRM + Vshpere replication( which has been confirmed by ENG_team ). Once eng-team approves then it will be open for support. In that case we can have support call with VMware to resolve the issue. Since it is not been tested with Networker, We will not able to go further to discuss with VMware or troubleshoot.
Yes you are right in saying that, if networker talks to VMware there may be solutions to the problem you have. But unfortunately for now, it is not seems to be possible; as networker code has not been prepared to support SRM + Vsphere Replication.
********************************************************************************************************************************************
EMC then referenced the following link:
http://solutions.emc.com/emcsolutionview.asp?id=esg130882
So it seems that for the foreseeable future Networker will not support the restore of VADP image backups if the VM is also protected using VMware SRM.