Off-site Replication and resource allocation

[xefil]

Hello!

It's possible to limit the resources of the mirrored VM in the DR-Site?
I mean, in the Production Site the VMs has more resources (memory, cores) and the DR-Site has not so much resources available. Due the DR-Site should run only the necessary time the Production Site is back, the VMs could be allocated with minor resources.
Is it possible to configure it automatically on Veeam?

Thank you!

Simon

[foggy]

Hello, Simon. There's no such capability currently in VBR. Thanks for the feedback.

[xefil]

Hello!

Thank you for the answer!
Maybe a feature for the future? why not

Simon

[lewish]

Hi.

We run a DR site with hosts that only have 8 CPU's available per host. If/when we upgrade to ESX5 we want to increase the number of CPU's for 1 server in the Production site to 12....does this mean it won't replicate to a host with fewer CPU's available.

Cheers Lewis

[Vitaliy S.]

Lewis, no it doesn't mean that. You do not need to have identically the same host in the DR and Production to be able to replicate VMs.

[lewish]

Thanks Vitaliy for that.

[xefil]

Hello!

I need another help on this issue. In case I've a DR Site with same capabilities as Production Site all works fine; I mean in case of switch over on DR Site we can maintain the VMs running without problems.
In case the DR Site has minor resources, if we need to switch on DR Site I've to change manually the resources (CPU and RAM) before the startup, right?
Until here ok.
But what happen if I would like to failback to Production Site? What happen to the VMs that on the Production Sites would have more resources than in the DR Site?

Thank's for the help!

Simon

[foggy]

Simon, the production site VMs will keep their CPU and RAM resources after failback (configuration will not be overwritten).

[xefil]

Thank you for the answer!

Have a nice day!

simon

[johnlong]

Is it possible to change the resource allocation of replica VM's? I have a server in a DR site that has less ram than the production server. Unfortunately I cannot add more ram, the board is maxed out. Replacing the server at this point in time is not an option either. I know I can modify the ram allocation before failover or before turning the VM on but it would be nice if I could set the amount of RAM allocated to the replica in Veeam.

I tried changing the amount of ram on the replica but it changed back after replication.

Thanks!

[dellock6]

I do not think is possible, since replica is copying also the vmx configuration file, where the amount of ram is defined. If you modify the replica version, at the next replica job the file will be owerwritten again since from the Veeam point of view, it's not equal to the source version.
None the less, it could be a nice feature for a future release.

Luca.

[loda117]

Hi Guys,
Starting a new venture with Veeam and this time it is replicating.
We have deployed a new server with 2 VMs (SBS2011, Sever 2012) on a 2012 Hyper-v host
Their old server is about 4 years old and can support 2012 with Hyper-v
I want to replicate main Hyper-v to the old server which now has 2012 with hyper-v enabled
I have couple of questions before I go with it
1) Old server has 8GB of RAM total will this be a problem since the new server has just 16GB RAM just allocated for SBS2011 and 8GB for server 2012?
2) Old Server is also lacking in total storage space (-200GB) will this be a problem? (Storage overall is plenty to host both of those server and then some just wondering)
3) Different processors? Will this a problem?

Thanks.

[Vitaliy S.]

1-3. All these do not matter, so feel free to re-use your other server as replication target. Thanks!

[robg]

I uncovered a problem today that I hope can be improved upon in Veeam.

I had to resort to a replica on a VM that lives on a host that is different than its replication target.

The main VM host has 24 cpus, the target just 4. The VM was configured to use all 24 of the CPUs. However, the failover process failed because the 2nd host could not start it with the 24 setting. Veeam thought the replica was running when it was actually not, and I had to "undo failover" - which probably does nothing in this case.

So I had to change the CPU setting to 4 on the replica target and start it manually.. And now in order to fail back, I have to replicate it back to the main host manually as well.

This may be a feature request, is there any way out of this situation? Maybe a check can be made to make sure that the replica is actually capable of running in the target host - or more flexibility in the failover process (like a way to adjust the replica settings in advance).

[foggy]

Rob, thanks for the feedback. The correct way of failing over in such case implies manually adjusting replica VM resources to be able to start on the target host. In this case failback should work just fine.

[robg]

It doesn't work.. I tried adjusting the replica manually to reduce the CPU count and it still fails once the failover is initiated. I think it's trying to use the properties of the source VM.

[robg]

Another thing is that when you merge threads like this (I'm sure to keep the forum more organized) - I do not get the email notification anymore because I was not subscribed to this thread. I have to go into my post history to find responses.

[foggy]

It doesn't work.. I tried adjusting the replica manually to reduce the CPU count and it still fails once the failover is initiated.

Right, this does not work so for the snapshot replica Veeam B&R is currently using. I based my reply on one of the posts above, which referred to the old replica format, where configuration was not overwritten during failover. Sorry for that.

[robg]

There's no way around it? Cause I bet the best solution is to design a way to alter the target VM, like another page of settings in the job. Seems important enough in cases where the primary and secondary hosts are not matched.

[veremin]

But if the hosts are not matched in terms of resources, how can you guarantee that VMs running mission-critical applications wouldn't crawl, once they are failovered to less-provisioned host? If those VMs can consume less resources, why to overprovision them?

I'm wondering because because we consider replicas being the exact copies of production VMs, thus, it's not possible to tweak the resources provided.

Thanks.

[dirkdigs]

hello i created a replica job however my DR hyper-v server is low on ram. My source VM is suing 24 GB ram. I want the replica to use less .. says 12 GB. where can i configure this?

[foggy]

Jason, please review this thread for the answer. Basically, this is not configurable.

[Isaac Rainsford]

feature request!