resource tracker for bare metal nodes tries to subdivide resource

[Robert Collins]

after deploying a small instance on big hardware:

2013-05-09 08:48:30,085.085 19736 AUDIT nova.compute.resource_tracker [-] Auditing locally available compute resources
2013-05-09 08:48:30,208.208 19736 AUDIT nova.compute.resource_tracker [-] Free ram (MB): 97792
2013-05-09 08:48:30,208.208 19736 AUDIT nova.compute.resource_tracker [-] Free disk (GB): 2038
2013-05-09 08:48:30,209.209 19736 AUDIT nova.compute.resource_tracker [-] Free VCPUS: 23
2013-05-09 08:48:30,308.308 19736 INFO nova.compute.resource_tracker [-] Compute_service record updated for ubuntu:96deccd5-0ad9-4bb5-979b-009bebac52fc

This should show 0, 0 and 0 : the size of the instance is not the amount to subtract :).

I don't know if this is just cosmetic.

[aeva black (tenbrae)]

This suggests that BaremetalHostManager isn't doing what it's supposed to, which is ensure that a node's resources are consumed all-or-nothing.

[Sean Dague (sdague)]

Extremely old bm bug, marking as won't fix