Instant recoery VS restore full VM

[hasayeretFMG]

Good morning,

Is there any documentation which compares pros and cons of instant recovery and full VM restore?
It almost feels to me that 99.9% of the cases instant recovery will be the way to go since the machine is always accessible to the users and you can still perform a migration.
In fact, I almost do not understand why the restore full VM option is still available and the only reason I can think about is that it might be a faster way to restore since the machine is down and no changes are being made on it (like in an IR scenario)

Thoughts?

[skrause]

Instant VM recovery's performance is dependent on a large number of factors. The primary one that can have a major impact on running production vms this way is the type and speed of your backup storage.

Instant VM recovery is great as an option for some scenarios, but in others it can be almost useless. A common example would be for enterprises that use FiberChannel for their production SAN and do not have adequate Ethernet infrastructure to run storage traffic for production VMs as well as normal networking.

A full restore will always "work" to restore your data to some kind of production storage, instant VM recovery is situational.

[hasayeretFMG]

Instant VM recovery's performance is dependent on a large number of factors. The primary one that can have a major impact on running production vms this way is the type and speed of your backup storage.

Instant VM recovery is great as an option for some scenarios, but in others it can be almost useless. A common example would be for enterprises that use FiberChannel for their production SAN and do not have adequate Ethernet infrastructure to run storage traffic for production VMs as well as normal networking.

A full restore will always "work" to restore your data to some kind of production storage, instant VM recovery is situational.

Thank you so much for your answer!
In our case, we are running our Veeam on the same server where the Veeam data is stored: A windows server with just tons of disks while our production environment is really a SAN-based environment. Sounds like in our environment Instant Recovery would be the way to go? Or did I miss something

[skrause]

Again, it is situational and will vary between every instance you need to recover your data.

If you have one VM that crashes and want to get it back up? Instant VM recovery is probably fine.

Every VM you have gets whacked with a crypto-locker and you need to restore them all? Doing a series of Full restores to your SAN would likely be a better option for most VMs.

Instant recovered VMs are added to your VM environment as new VMs with new UIDs so you need to take that into account when you use Instant Recovery+storage migration to get them back on your production SAN.

[hasayeretFMG]

Thank you!
We have a 1.1TB Exchange server which needs to be moved from 1 host to another and we realized that do that through Veeam would be the fastest way.
I assume that in a situation like this when we have our emails available all the time, instant recovery would be the best way since the machine is online all the time. Correct?

[channdeep]

Just to share my observation for which is faster between Instant_VM_Recovery (migration to production) and Full_VM_Restore - so the short answer is that the Instant_VM_Recovery (migration to production) is faster.

Full_VM_Restore took 5 hours 42 minutes to recover a 5TB VM with the VM being unavailable all the time.

Instant_VM_Recovery (migration to production) took 4 hours 18 minutes to recover the same 5TB VM with the VM being available all the time.

Both restores were run at almost an idle time of the backup server with Veeam9.5 on HyperV platform.

Hence my observation that Instant_VM_Recovery (migration to production) is faster.

The key difference noticed in the restore process is that Full_VM_Restore runs the restore process on all of the VHD files at the same time, so it's full speed gets divided between its different internal sessions.

Whereas, the Instant_VM_Recovery (migration to production) runs the restore process on a single VHD file at a time. Guess that single internal session gets full resources then. Once one VHD file restore is complete, the process moves automatically to the next VHD file and so on. So, this is the difference noticed i.e. breaking one big restore in smaller chunks.

Hope it helps someone.

Thanks and regards.

[Thommo]

Great insight Channdeep, thanks for sharing this info. We are about to restore a whole network of vms (about 50) onto a brand new VMware network in a new location, I was hoping to see what method might be faster as we have a tight 3 day window to bring the whole network offline

[Mildur]

If this is for a migration scenario, why not use replica jobs?
Prior to the maintenance window, you can replicate all vms. On the 3 days window, do the failover. Only changed blocks since the last replication are replicated when you do the failover. Veeam even can use re-ip feature for windows machines if you have other subnets at the new location.

[Andreas Neufert]

Planned Failover is what you need.
VM replication will replicate VMs on specific selected schedule.
Planned Failover (you can use Failoverplans to sort start order) will shutdown the VM, replicate the last changes and boot up the VM on the other side.
https://helpcenter.veeam.com/docs/backu ... ml?ver=110
https://helpcenter.veeam.com/docs/backu ... ml?ver=110

[vmtech123]

I have many SANs in my environment. Some are NVME, some are hybrid, and some are just disk. If an only disk SAN is my repository, an instant recovery is going to boot those VM's up on the Veeam SAN/Repo and present to production. This could be really good for a single server or even a few in a pinch while I migrate it back to the production SAN. The more VM's, the more you will push that storage. I would create some test VM's and see how it works for you. Perhaps a job with several to see the performance. Try and have things running on them too.

The pros of instant recovery is you just boot your backup rather than restore it to the other storage so it comes up quick (if the storage is decent). You can copy it back while users are utilizing the VM.
The cons are you may need better storage for Veeam.

It's the opposite for a VM restore. you need to wait up front, but then it boots up on it's normal storage.