You can do that but how would you restore if your storage unit fails?
We have found that a physical Veeam server gives you quicker recovery. With the instant recovery option.
Hi @tjanse227,
yes your idea is possible. In that case your are running an all in one deployment, except the repository role.
You could probably deploy some separate VMs as a proxy for the HotAdd job, to run these tasks explicit on them, not on your VBR itself.
When it comes to restores you will be also good to go, but remember the 3-2-1 rule oder better 3-2-1-1-0 rule.
As @Vpaulcb mentioned, think about „what if my prod SAN fails?“ (with only one SAN, you will lose your VBR as well). Then you should have a plan, how to quickly get your VBR configuration back running: Configuration backup outside your prod.. maybe in/on a separate Site.
Instant recovery option will be also available, but you need your VBR intelligence for that.
Best, Markus
I also would not bother with direct storage access in such a small environment. You’d get a very good backup performance but restore with direct storage access has some caveats (thin/thick etc) and you might end up with a hot add proxy in a vm here anyways.
But as @Vpaulcb states, have a look at a small physical all in one VBR server. You might even use just network (NBD) as proxy mode. Advantage is, you're fully independent from the VI stack and could use infant recovery in case of a storage failure.
If you're not satisfied with NBD performance, which has gotten way better in recent years btw, you might always add in a virtual proxy in VM. Could even be Linux if licenses are an issue.
Thanks everyone, much appreciated - I will go with the physical server performing all roles, including repository then.
Hello,
what you have described, is method to use physical servers for example with C:\ drive local and application disks are SAN disks. So you don’t need to backup with agent only the entire physical machine, you can backup SAN disk via snapshots of SAN, then you have consistent backup of application disks.
But i would go for VBR as suggestions above….
Thanks everyone, much appreciated - I will go with the physical server performing all roles, including repository then.
The current Veeam Best Practice (within VMCE) deployes Veeam within a VM and make the VM as available as possible (replication to another target etc.).
I try to avoid having a simple deployment scenario but that depends on the customer and the size of the environment. If that fits your needs all is set. :-)
