Common hypervisors (Proxmox, Vmware, KVM..) do not prevent the sum of allocated virtual resources (such as vCPU) to exceed the number of physical ones, even if some of them (OpenStack) do limit the overcommittment ratio (virtual_resources/physical_resources) to a configurable value.
I'm confused about this paragraph from this article
XenServer previously allowed overcommitting vCPUs. However, this behavior is blocked as it can lead to undesirable effects when running a VM a vCPU that cannot be matched to a pCPU
I can both read it as "overcommittment is blocked in Xen" or "one VM cannot exceed the number of physical cores but sum of all VM vCPUs can"
Is it (still) possible to overcommit cpu/memory with Xen?