openstack/ironic
Code Issues Proposed changes

Update resources only for specific node during deletion.

Browse Source 
At the moment during Nova instance termination resources are
updated for all Ironic nodes. As result with high number of nodes,
removing N instances initiates N x M requests, where M is the total
number of nodes. This commit allows to update resources per specific
node on termination.

Depends-On: Iffad4a6ab1aaa5fea591c19f6d330dc861c5675d

Change-Id: Ib54741c8564e18f23c4f50bed190dfc66db7ba76
Closes-Bug: #1552446
This commit is contained in:
vsaienko 2016-03-03 01:07:33 +02:00 committed by Vasyl Saienko
parent 982c9a97ab
commit 16d8819891
2 changed files with 12 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
ironic/nova/compute/manager.py
View File

@ -81,13 +81,15 @@ class ClusteredComputeManager(manager.ComputeManager):
pass
@lockutils.synchronized(CCM_SEMAPHORE, 'ironic-')
def _update_resources(self):
"""Update our resources
def _update_resources(self, node_uuid):
"""Update the specified node resource
Updates the resources while protecting against a race on
Updates the resources for instance while protecting against a race on
self._resource_tracker_dict.
:param node_uuid: UUID of the Ironic node to update resources for.
"""
self.update_available_resource(nova.context.get_admin_context())
self.update_available_resource_for_node(
nova.context.get_admin_context(), node_uuid)
def terminate_instance(self, context, instance, bdms, reservations):
"""Terminate an instance on a node.
@ -100,4 +102,4 @@ class ClusteredComputeManager(manager.ComputeManager):
instance,
bdms,
reservations)
self._update_resources()
self._update_resources(instance.node)

5
releasenotes/notes/node-deletion-update-resources-53862e48ab658f77.yaml Normal file
View File

@ -0,0 +1,5 @@
---
fixes:
- Fixed performance issue for 'ironic.nova.compute.ClusteredComputeManager'
when during Nova instance termination resources were updated for all
Nova hypervisors.