os-collect-config unable to restart neutron-l3-agent
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
neutron |
Expired
|
Undecided
|
Unassigned | ||
tripleo |
Invalid
|
High
|
Unassigned |
Bug Description
The command 'os-collect-config --force --one' fails on overcloud-
+ service neutron-l3-agent restart
neutron-l3-agent stop/waiting
start: Job failed to start
[2014-01-28 23:17:33,155] (os-refresh-config) [ERROR] during post-configure phase. [Command '['dib-run-parts', '/opt/stack/
And in /var/log/
Stderr: '/usr/bin/
2014-01-28 23:17:33.073 13463 ERROR neutron.
2014-01-28 23:17:33.133 13463 CRITICAL neutron [req-5f2c30e1-
Command: ['sudo', '/usr/bin/
Exit code: 99
Stdout: ''
Stderr: '/usr/bin/
Was able to fix by adding filter to /etc/neutron/
kill_l3_agent: KillFilter, root, /opt/stack/
description: | updated |
Changed in tripleo: | |
status: | New → Confirmed |
Changed in tripleo: | |
importance: | Undecided → High |
status: | Confirmed → Triaged |
This appears to be caused by neutron- ns-metadata- proxy being run out of a venv, rather than root. The rootwrap to kill this service is set to /usr/bin/python, but needs to reference the venv python in order to work.