Replica Destination Thick VMDK

[douglas_carson]

Hi Guys,
Just new to veeam backup and replication. I have been testing the replication piece with a few clients in mind. I do however have a couple of questions.
In the test lab i have the below:
Main site
ESX4.1i Host
1 x windows 2003 VM
1 x iSCSI HP Lefthand P4000 SAN
The ESX host has one vmfs volume on the SAN with the windows 2003 VM on the VMFS volume

DR Site
ESX4.1i Host
1 x iSCSI HP Lefthand P4000 SAN
The ESX host has one vmfs volume on the SAN which is where the replica will be placed

I noticed that the replica on the DR site was showing the vmdk was thick. Is there any way to have this thin on the DR site?
Also in the job properties i have chose SAN with failover mode is this correct for this scenario?

Thanks
Dougie

[Gostev]

Hi Douglas,

The upcoming version 5 will replicate thin disks as thin.
As for the second question, where you have your Veeam Backup server installed? Main or DR site? Is it physical or virtual?

Thanks!

[douglas_carson]

Hi Gostev
The current backup server is physical and is located in the main site.
The server is not connected to the iSCSI SAN.
The second question i have is are there any detailed docs as to what is happening at the VM level when you perform a failover?
I see that fact i takes a snap etc but would like to more fully what was happening.
The other question i have is if the primary site is down, and you have the backup server in the DR site i am guessing you can still initiate a failover without issue. (I am coming from an SRM mind set form this one)

Thanks
Dougie

[Gostev]

For physical backup server in the main site, I recommend connecting it to the iSCSI SAN and sticking with SAN+failover mode. Right now, since it is not connected to SAN, it always works in "failover" mode over regular network (which is not optimal performance-wise, and loads production network).

Failover to latest restore point simply creates snapshot (to protect replica VMDK from modification, in case you will want to continue replication later, or failover to different restore point) and starts up the VM. You can do the same manually with vSphere Client (even skipping snapshot part if it is not needed), Veeam Backup is not necessary.

Failover to different restore point also creates snapshot (again to protect replica VMDK), then modifies VM disks to take their state back to require restore point, and starts up the VM. Failover to restore point other than latest requires Veeam Backup server present, and cannot be performed manually.

Thanks!

[Bunce]

Hi Douglas,

The upcoming version 5 will replicate thin disks as thin.
As for the second question, where you have your Veeam Backup server installed? Main or DR site? Is it physical or virtual?

Thanks!

Hi Gostev,

What will occur on existing replica's that have been replicated as thick, once we upgrade to V5? Will they 'shrink' back to thin at the destination or will we have to recreate the job / re-seed / etc?

Cheers,
A

[Gostev]

Unfortunately such transition is impossible, so it will be required to recreate the job.