Log unknown providers during quota calculation

When determining quota usage for known nodes log cases where the
provider config isn't known. One theory is that this can happen when
launchers are restarted and that it is otherwise normal but without more
information it is hard to know for sure. Log the node data when we hit
this sutation so that we can dig in more and debug this.

Change-Id: I466e64a999ebf867ada9462a0c00a90de4b70125
This commit is contained in:
Clark Boylan 2017-12-20 10:14:07 -08:00
parent ded8fd80d8
commit 310ea2fe9d

View File

@ -221,8 +221,16 @@ class OpenStackProvider(Provider):
if node.provider == self.provider.name:
if pool and not node.pool == pool.name:
continue
provider_pool = self.provider.pools.get(node.pool)
if not provider_pool:
self.log.warning(
"Cannot find provider pool for node %s" % node)
# This node is in a funny state we log it for debugging
# but move on and don't account it as we can't properly
# calculate its cost without pool info.
continue
node_resources = self.quotaNeededByNodeType(
node.type, self.provider.pools.get(node.pool))
node.type, provider_pool)
used_quota.add(node_resources)
return used_quota