Replication - amount of CPU

[Akson]

Hello!

Is it possible to change the amount of CPU, which starts with replica? In SureBackup there is an option that changes the amount of RAM: "Amount of memory that is allacate VM".
At that moment, if server where you run replica has fewer processors than the source server, the action fails with the error.

[veremin]

Is it possible to change the amount of CPU, which starts with replica?

No, it is not. Thanks.

[Akson]

Is a possibility, that the option to change the number of processors was added in the next update Veeam?

[veremin]

Nope, as you might be the first one to request this functionality. Thanks.

[Vitaliy S.]

At that moment, if server where you run replica has fewer processors than the source server, the action fails with the error.

During failover you can change the number of vCPUs manually via vSphere Client and then start the VMs.

[Akson]

Yes, I can change number of CPUs via vSphere Client but then I can't use Veeam to migrate to the production environment. I have to do the migration with VMware.

[Vitaliy S.]

Not sure I fully understand why you can't use replication jobs to move VMs to the production environment. Can you please clarify? Will it be a permanent failover?

[Akson]

Not sure I fully understand why you can't use replication jobs to move VMs to the production environment. Can you please clarify? Will it be a permanent failover?

In the standard case, when we use replication, we can migrate to the production environment - Veeam is responsible for it. When I manually run the VM from a replica, it is no longer work Veeam mechanisms. We migrate the VM, but only from VMware.

[veremin]

As mentioned, it's not possible to change amount of CPU, using VB&R GUI. However, during failover (or once it's finished) you can access the failovered VM and change those settings via vSphere Client. Thanks.

[dellock6]

you can at least add additional vCPU if you first configure the VM to support hotadd (not to be confused with VADP hotadd aka virtual appliance mode).
There are some limits based on the involved guest OS, but with it you can add in real time new CPU and memory. Not remove as far as I remember.

Luca.

[willn]

OK, so I am putting together my replication plan and going to be replicating our production VMs to the DR site. My main question is if there is a way to reconfigure the production VMs to have less RAM than the production (we will run in "crippled" mode in DR).

My only thought about how to do this is to replicate the VMs initially to get them to the DR site, manually change the RAM requirements on the replica VM and then map those replicas on the original replication job, that way it won't overwrite the settings with the original. Hopefully that makes sense.

Does this sound like the best way to accomplish this?

[veremin]

There is no functionality to achieve what you're after in automatic fashion.

Speaking about manual tweaking memory value, I haven't tested that myself and not sure whether that re-configuration would persist after subsequent replication cycle.

Thanks.

[SteelContainer]

We have had a prospective client ask about the reduction of specs for cloud replicas. Count me down for a feature request.

[veremin]

We've discussed this at service provider subforum once and I doubt that I get the idea behind making replica VM less provisioned in terms of resources (besides price, of course)?

In the way I see it replica by definition is an exact copy of production VM, so, if production VM hosting mission-critical application requires the defined amount of resources to run successfully, why to make it crawl during disaster when you need it most? If VM can live ok with less resources, why to overprovision it in production in the first place?

Thanks.

[SteelContainer]

That's a good point. Yes it is simply that the client wants to pay less for the replica.

[veremin]

Then, it might be worth asking your customer the very same questions to make sure he or she understands the consequences this functionality might potentially bring. Anyway, we got the idea. Thanks.

[mcrape]

I recently freed up some older servers and decided to use Veeam to replicate VMs to them. Given that we already have pretty good backups (on and offsite), this is just another layer of protection. I figured we would be better off with replicas on older hardware vs. not having them at all.

These older servers don't have as much RAM as our production servers, and I was thinking it would be nice to have a feature as part of the replication job to alter the VM's settings. For example, what might be an 8 GB machine in production we would scale down to a 4 GB machine if it was running on the replica servers. Yes, we would likely see a performance hit, but in our environment, I would rather see a performance hit vs. having downtime to do a restore.

Thoughts?

[PTide]

Hi,

That sounds like a valid request. For now with current version of VB&R you can utilize post-job scripts in order to edit your replica .vmx file automatically after every sync interval.

Thank you.

[mcrape]

Thanks - I'll poke around and see.

Would I need to create a new script for each different customization?

[PTide]

Would I need to create a new script for each different customization?

There should be a separate script for each replica job. If you have all VMs in one replication job then it's ok to have one script for all VMs customization.

[relionbrian]

Currently when you failover to a replica, the backup server needs to have the same amount of memory as the production server. There is no where to configure an option to boot with less memory similar to how Surebackup allows you to do this. Modifying the memory setting in Hyper-V Manager simply gets overwritten by Veeam during the failover.

We often re-purpose old servers to act as Veeam Backup servers and they usually have less memory compared to the production server.

I know it's ideal to upgrade the memory on the backup server but it sure would be nice to have the option to boot with less memory