Openstack network selection is not passed from the controller to the models
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Canonical Juju |
Fix Released
|
Critical
|
Tim Penhey |
Bug Description
When working on an Openstack cloud with multiple networks the right network to use can be specified on bootstrap with the argument '--config network=foo' to the juju bootstrap command. After doing this the controller comes up fine, adding a new model also works fine but upon adding applications to the new model allocation of Nova VMs fails.
A juju status shows the machines in this state.
machines:
"0":
juju-status:
current: down
message: agent is not communicating with the server
since: 29 Sep 2016 15:23:03Z
instance-id: pending
machine-status:
current: pending
since: 29 Sep 2016 15:21:15Z
series: trusty
Checking the log on the controller you see errors like (slightly santized to avoid a bunch of hex user data):
4da59b22-
caused by: request (http://
tags: | added: canonical-is |
tags: | added: ateam |
Changed in juju: | |
status: | In Progress → Fix Committed |
Changed in juju: | |
status: | Fix Committed → Fix Released |
The controller is bootstrapped with the network config property to select the network. That is not stored on the controller so future deploy commands come back with the error from Nova:
"Multiple possible networks found, use a Network ID to be more specific."
I think that the juju controller-config should be outputting the stored config of the network specified for the cloud.