[ovirt-users] Cluster Processor Change

[Michal Skrivanek]

> On 7 Apr 2018, at 14:27, Clint Boggio <cboggio at inlinenetworks.com> wrote:
> 
> Forgive me if this is a duplicate, as I had spam filter problems previously when I sent this inquiry.
> 
> Environment Rundown:
> 
> OVirt 4.2
> 6 CentOS 7.4 Compute Nodes Intel Xeon
> 1 CentOS 7.4 Dedicated Engine Node Intel Xeon 
> 1 Datacenter 
> 1 Storage Domain
> 1 Cluster
> 10Gig-E iSCSI Storage 
> 10Gig-E NFS Export Domain
> 20 VM’s of various OS’s and uses 
> 
> The current cluster is using the Nehalem architecture.
> 
> I’ve got the deploy two new VMs that the current system will not allow me to configure with the Nehalem based cluster, so I’ve got to bump up the architecture of the cluster to accommodate them.

why not? You are the one selecting teh VM’s CPU…why not use Nehalem too?

> 
> Before i shut down all the current VMs to upgrade the cluster, I have some questions about the effect this is going to have on the environment.
> 
> 1. Will all of the current VM’s use the legacy processor architecture or will I have to change them ?

running VMs will change on their power cycle. Unless you used per-VM CPU Type override (I suppose you did not)

> 
> 2. Can I elevate the cluster processor functionality higher than the underlying hardware architecture  ?

if all your hosts support it

> 
> 3. In regards to the new cluster processor, will all of the processor architectures below the one I choose be an option for the existing and future VMs ?

not sure I understand the question. Normally you can use any CPU as long as it is supported by all hosts in the cluster. You can use per-VM override to bypass that, but that’s usually not a good idea

Thanks,
michal
> 
> I apologize for the long post and I hope that I haven’t left out any vital information.
> _______________________________________________
> Users mailing list
> Users at ovirt.org
> http://lists.ovirt.org/mailman/listinfo/users