Hello. I've been through the Performance Best Practices and Google, but can't seem to find a straight answer.
We are designing VMs for migrating our Mission Critical databases. Normally the databases can run find with 2vCPU, but about 3 times a year a special job is run that going to need up to and possibly above 16vCPU. Minimizing the number of times the VM is powered off is critical.
Does anyone know what the recommended practice would be for sizing this VM? Does anyone have any recommendations themselves or any previous experience with this kind of situation? Thanks.
I have no personal experience but the following suggests "Add the maximum VCPUs that might be needed ..."
vNUMA is disabled if VCPU hotplug is enabled (2040375)
Note: You did not mention the guest OS and version.