We have citrix cluster where multiple vms are running by esxtop script we are seeing
Numa % local
While it should be 100% for all VMs
Any suggestion appreciated.
Not certain I follow what you're saying. Can you provide more info?
Ideally, Numa % local should be 100% for all VMs
Do you also have a number of wide VMs on this host (or ping ponging between it and others) which cross NUMA boundaries? If so, this could be the reason.
It is necessary to know the dimensions of the NUMA node and the VMs configuration.
In addition, I have long researched the behavior of the NUMA balancer and found out that it does not work as it normally does.
this host have high configuration vm ,
you mean do we required check ping request between vm, or vm to host