Can't create guest with more than 240 VCPUs
Issue
-
Appreciate this may be somewhere in the documentation, but are the numbers of CPUs and amount of RAM reserved for the hypervisor auto-generated based on node's capacity, or are they values we can tune?
-
I have a 128 core (with SMT enabled) hypervisor and I can't seem to deploy a virtual flavour with more than about 240 CPUs. 16 (or 8 physical) seems a lot to be reserving for the hypervisor, and it also seems to be reserving a large amount of 1TB RAM.
-
Can we adjust that? We are not looking to overcommit CPUs on these nodes, just deploy a single huge VM with as many resources as possible.
Environment
- Red Hat OpenStack Platform 16.2 (RHOSP)
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.