Comment 3 for bug 1834694

I've done debugging to figure out the root cause of this and those notes are here:

https://review.opendev.org/#/c/668252/1/nova/scheduler/host_manager.py@626

tl;dr we fail to save free_disk_gb to the compute node, but later create the resource provider with inventory, so the scheduler gets the provider from placement but the node doesn't have free_disk_gb so the HostState isn't updated and gets run through the filters/weighers anyway and blows up in CPUWeigher.