Instant Recovery Question

[HFX902]

Hi All,

Currently my team is working on the best option for disaster recovery. Including the fastest way to be up and running after an outage.
We are using a 10gb switch with 10gb pipe from Veeam server which is on a dell PE320 with backups backing up to a dell PE4000.

My question is if we preform instant recovery on vm will changes made to this vm reflect when when it is safe to restore the vm back to its original DS. (Or will the vm just be back at square 1?)

Also any other suggestions on speedy recovery methods using veeam? we have about 150 VMs with 40 High Priority.

[Gostev]

Hi!

Yes, changes will persist when instantly recovered VM is migrated back to production storage.

Since Instant VM Recovery is normally good for a few VMs only, I recommend that you use replication for those 40 high priority VMs (in addition to backup). This way, in case of disaster you will be able to power any amount of these VMs on, and they will be running at full speed immediately - not constraining by backup storage and vPower NFS data processing performance.

Thanks!

[HFX902]

What do you recommend for the remaining VMs? A full entire restore?

[foggy]

If you have DR infrastructure capable of running all VMs and would need the entire environment running while production is completely lost, you can use replication for the rest of VMs as well and configure failover plans in accordance with VMs priority.

[HFX902]

What would my limit VMs be using Instant Recovery? Also have you used Quick Migration?

[foggy]

The limit depends on the underline storage performance, here are a couple of posts with good explanation: one and two.

[HFX902]

Hi Foggy thanks for your help

This is our SAN http://www.dell.com/us/business/p/equallogic-ps4110x/pd

Please let me know your opinion on this SAN

Also I see Instant recovery is not a true DR, enough to get you by.


Can you recommend veeams best solution for DR?

[Gostev]

Equally big factor is actual workload running in a VM. Many applications are very light on disk I/O (DC, DNS, DHCP) and you can potentially have 10 or even more of this kind of VMs running through instant recovery from a single backup repository at acceptable performance. There are also applications that do very heavy disk I/O (like busy Exchange or SQL Server), and of course you will not be able to run many of those simultaneously with instant VM recovery.

[HFX902]

I have spoke with Veeam and they are fully confident with our current backup infrastructure we will have no issues running 40 high priority vms using Instant recovery. Followed by quick migration when the time is right.

[Gostev]

I assume you have talked to a Veeam sales person?
Here, you are speaking with Veeam R&D.

[chrisdearden]

I suppose if you were to spread the restores over a number of separate repositories, It *might* be possible, but it's certainly not a config I'd be happy to recommend.

[Gostev]

It really depends mostly on how heavy on disk I/O applications running in those VMs are... "fully confident" blanket statement covering every possible workload is impossible.

[HFX902]

Thank you all for your feedback!

[Delo123]

40 instant recoveries from one repository is a very very big risk. Not only will things be awefull slow but the risk of losing data (and all changes since the launch of the instant recovery) is enormous...
Instant recovery of multiple is NO alternative to replicas (been there, done that, lost 2 days worth of exchange items)...