sockets to instances switch - feedback

[mcz]

Dear PM's,

we've recently switched from the socket-based license model to the instances because it made a bit more sense to us. We did some calculations about how many instances we needed and then made the order. Installing the license on the server was quickly done, also on veeam one side - but then I got complaints by veeam one that we had exceeded the license.

Asked our seller and finally created a ticket where I've got told that replicas do consume an instance... Man...
When we do calculations about the protected vm's, you don't expect to need more licenses for the monitoring side. Of course, somehow it makes sense that you need an instance per monitored vm, but that's totally different compared to the sockets model...

For me it just doesn't make sense when I need say 30 instances for my workloads that I'd have to pay for 50 instances for the monitoring. Of course, you can exclude them, but it's not the same as 'back then'. Sometimes I've got the feeling that the whole licensing model (especially the switch) has got a bit too complicated allthough it looks quite simple at first sight.

Please let me know your thoughts - thanks!

[Vitaliy S.]

Hi Michael,

Hmm... that's a good point. I would agree with you and say that there should be an exception in licensing for replicated VMs if they are powered off and the source VM still exists. If one of the rules does not apply, then Veeam ONE should be consuming the license key. Let me discuss it with the RnD team and I will come back to you.

Thanks!

[Gostev]

Yeah this looks like an oversight on our part.

[mcz]

Thanks for the positive feedback, Vitaliy & Anton! Very very interesting to me that nobody else has mentioned it before. Probably others had enough instances...

[wishr]

Hi Michael,

Thank you for bringing it up! I guess customers were excluding replicas from monitoring in such cases.

Anyway, we are looking into this scenario and will update this post once we have something to share.

By the way, are you testing your replicas? If so, how are you doing that (failover, SureReplica), and how much time do the replicas usually spend in the "powered on" state during such tests?
Also, what monitoring and reporting capabilities of VONE you are using for replicas? Most of the time, replicas are powered off, hence the question.

Thanks

[mcz]

By the way, are you testing your replicas? If so, how are you doing that (failover, SureReplica), and how much time do the replicas usually spend in the "powered on" state during such tests?

Yes, we are testing it via SureReplica and (planned) Failover. So the duration of the power-on also depends on the amount of vm's in the application groups and could also last an hour or so.

Also, what monitoring and reporting capabilities of VONE you are using for replicas? Most of the time, replicas are powered off, hence the question.

As a general answer I'd say that I don't care about OS-specific issues like no space on the disk because that would happen on the source-VM as well. But I'd love to get events related to the hypervisor, like when there's a snapshot which is 300 days old... AND if you failover to a replica, you're interested about the OS-events as well, so it always depends on the current usage of the replica if you wanna know everything, nothing or just a bit.

In general I'd love to be able to monitor replicas without limitations, they are the longer arm of the source-vm and very important... If your replica fails during a failover, you'd ask why you ever had a monitoring tool when it didn't warn you long enough. You get the point, I'm sure.

Thanks!

[mcz]

Just wannted to update this topic: Obviously the vm's created by the sureBackup-Jobs are also counted - which is not a big surprise but there it's not possible to list them in the exclusion due to the random name they get...

[Vitaliy S.]

Michael, yes, we are aware of that and are trying to come up with a solution that would address not only replicated VMs but others that are not always powered on.

[rsomby2]

Same issue here, underlicensed when it comes to monitoring. Not fair
I already made a exclusion for _replica but its still exceeded. So i had a look at what vms i have and see we have alot of old powered off vms (going to be archived/deleted, but takes time)
If there was an option to choose a criteria to exclude powered off vms, that would be nice.

[Vitaliy S.]

Yes, we will adjust our licensing logic in the next major update to address this situation (powered off VMs will not be consuming a licence key). Thanks for your feedback!