VMWARE and VDR troubles.

Rangler · 13 Dec 2011 at 10:17

HI All,

Is anyone using VDR at all.?
I have it at a client but i'm not sure its working at all the storage targets it backs up to are nearly 98% full and i don't believe that VDR is reclaiming space.
Indexing is also extremely slow on the VDR as well?

Should it automatically reclaim space or does it purge base on backup sets to hold in the job itself.?

Also would i be right in saying that VDR is NOT a DR solution it only good for restoring on a current live environment.?
If not a DR solution then i need to be looking at alternatives like VEEAM.

Thanks in advance for any info.

sidethink · 13 Dec 2011 at 14:04

My understanding is that VDR is 'just' a backup/recovery tool and not a DR solution. We have quite a bit of VMWare in our infrastructure but dont use VDR for backups, we use Netbackup.

Like most backup tools it probably needs told to purge/overwrite old jobs based on some policy - sorry I cant help you more on that.

For DR purposes, the VMWare answer is Site Recovery Manager.

droyden · 13 Dec 2011 at 14:21

I used VDR (altho it was v1), IIRC it will purge data after the defined retention period, how long is your retention currently?

Rangler · 13 Dec 2011 at 16:53

Was 14 days but due to it eating up space on the VDR san ive reduced it to 5.

Going to clear down the SAN and backup stores then reattatch them to the VDR and see how it goes as its allways hanging when either cataloging or indexing targets.

Gerbil · 14 Dec 2011 at 13:29

Used v1 of VDR and found it next to useless. Started OK, but then quickly spent most of its time verifying the integrity of the datastore it was backing up to and when that's in progress you can neither backup or restore from it. If it gor hung up it also used to leave VMs with lingering snapshots that needed to be serached for and deleted.

My understanding is that the v2 VDR appliance released with vSphere 5 is a lot better and can be deployed on (and backup) ESX/ESXi 4 servers. I've only installed the v2 VDR appliance long enough to re-familiarize myself with it before tackling my VCP5, but it does look like some improvements have been made. In particular if an integrity check is interrupted it does now appear to carry on from the point it got interrupted rather than starting again. If you want dedup then just make sure that you keep the destination backup disk to less than 1TB (or 500 GB for a CIFS store). VDR allows you to backup to a bigger disk, but VMWare quite openly say that it won't reliably dedupe under those circumstances.

I certainly wouldn't call it a DR solution. More a quick way to restore trashed VMs or the odd deleted file with the file level recovery agent. The trouble is the datastore you backup to is inevitably located on a server or SAN storage in the same datacenter as the servers you're backing up so is likely to be consumed by the same distaster.