Comment 1 for bug 1879878

This looks like the resize just hit a race condition.
when using cpu pinning cpus are not claimed on the compute host numa toplogy blob
so its perfectly feasible for a compute node to claim to fail even if schduler and placement claims
passed.

marking this as incompleted as we dont know what release of nova this is being reported against.
if its pre train this is invalid as we expect this to race. if its after train then we should see why we are not doing the cliam
as part of the resize verify step. we should be in which case this maybe vaild.

in general however we cannot fully resolve this untill we have numa in placement as there will always be a race between the schduler and compute node resouce tracker but there should not be one between resize_verify and the perodic task even today
unless the vm raced with another vm for the resouces on the host.