Quantum Manager VM launch failure
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
OpenStack Compute (nova) |
Fix Released
|
High
|
dan wendlandt | ||
neutron |
Fix Released
|
Critical
|
dan wendlandt |
Bug Description
If the first VM to be spun with DHCP-enabled (default on) is quickly followed by a second VM (e.g., do to an automated script launching VMs), you will reliably hit a race condition that causes the second VM to fail to successfully launch.
This occurs with Quantum Manager, and the race condition is related to the check that determines if a quantum port for the DHCP server has been created yet. The solution is a simple one-liner: to add @synchronized for the enable_dhcp method so that it cannot be run simultaneously for two VMs.
diff --git a/nova/
index b433652..4fc6f37 100644
--- a/nova/
+++ b/nova/
@@ -364,6 +364,7 @@ class QuantumManager(
+ @utils.
def enable_dhcp(self, context, quantum_net_id, network_ref, vif_rec,
Changed in nova: | |
status: | New → In Progress |
assignee: | nobody → dan wendlandt (danwent) |
Changed in quantum: | |
status: | New → In Progress |
importance: | Undecided → Critical |
milestone: | none → essex-3 |
Changed in nova: | |
importance: | Undecided → High |
milestone: | none → essex-3 |
Changed in quantum: | |
status: | In Progress → Fix Committed |
assignee: | nobody → dan wendlandt (danwent) |
Changed in quantum: | |
status: | Fix Committed → Fix Released |
Changed in nova: | |
milestone: | essex-3 → 2012.1 |
Changed in quantum: | |
milestone: | essex-3 → 2012.1 |
Reviewed: https:/ /review. openstack. org/3447 github. com/openstack/ nova/commit/ 6142230ccf25556 50dbb902a5c342a 342e9b2582
Committed: http://
Submitter: Jenkins
Branch: master
commit 6142230ccf25556 50dbb902a5c342a 342e9b2582
Author: Dan Wendlandt <email address hidden>
Date: Wed Jan 25 21:36:22 2012 -0800
bug 921931: fix Quantum Manager VM launch race condition
Change-Id: I1f3a0f001477ca a5f595c93c49f6c 9e03a50149a