New User - Restore Questions

[Neurobit]

Hello,

We just implemented Veeam this week, and I am still learning the ins and outs of the product.

First of all, I am very impressed with the performance!

Our (brand new) vSphere environment consists of a cluster of three ESX 4.1 boxes, and a Compellent SAN.
We are backing up to a Unistar NAS over Ethernet (two bonded Gig NICs on the NAS), and we are getting pretty good throughput (compared with BackupExec).
Our Mail server (248 GB VM) backed up in 3 hours on Wednesday, and yesterday in 2 Hours! (Fluctuated between 190 - 35 MB/s).

Question 1
Now I am testing restores, and I have a test 2003 box that I created and on which I deleted some system files to make it unbootable.
My question is, when I restore the full VM, will it replace the current one, or do I need to delete it before the restore process?

Question 2
A few of my servers have LUN volumes that reside on the SAN.
I have redundant controllers and FC switches, so hypothetically, were I to lose the SAN, how can I restore my VMs that will still be pointing to the SAN even though all the data is being backed up with Veeam?

Thanks in advance for your answers,
Paul

[TMassa]

1) If you change the name during the restore, it will make a new virtual machine and register it in VMWare for you. If you leave the name the same, it will overwrite the files (after warning you, of course)

2)You should be able to restore the individual VMDK (if they are indeed being backed up as "local disk") and register the disk on the VM before powering up the restore...disable the iSCSI connection to the SAN and change drive letter/path of the additional disk to match your old LUN volume.

[Neurobit]

Tony,

Thanks for the reply.

1) The restore worked great. Restored the 25 GB VM in 3.5 minutes.

2) Veeam is backing up the VMs to our NAS, not a local disk (mapped as a backup store in ESX). Is this what you're referring to?

Thanks again,
Paul

[TMassa]

As for #2, I assume that the data that your LUN volumes are being backed up like a normal VMDK, correct? You're using iSCSI initator?
If you SAN is down, you can restore your VM (and LUN volumes) to a new storage location, but the VM will still think it has a LUN connection to your SAN when it boots. You will have to break this in case of an emergency and restore the "drive" as a VMDK.

If the SAN is restored and you restore the VM, you shouldn't need to do anything. There may be an "extra" drive on the restored VM because Veeam will only see it as an attached disk (VMDK).

I'm really only speculating on point #2.

[Neurobit]

Thanks again.

Yes, we are backing up the LUN volumes as VMDKs (no iSCSI, we have a FC fabric).
I am just thinking worst case. We have redundant SAN controllers & FC switches and the three ESX hosts connected to both, but I'd still like to know that if something catastrophic happened to the SAN, that I could still bring up my crucial boxes up until the SAN is back in service.

Paul