Ask Your Question
0

Ocata - Failed to start OpenStack Neutron Linux Bridge Agent

asked 2017-04-19 17:11:08 -0500

sledge gravatar image

updated 2017-04-21 11:22:50 -0500

Hi,

I have followed that guide on centos 7 https://docs.openstack.org/ocata/inst...

Everything is running fine except the linux bridge agent fail to start, it starts for a couple of second and fail right after on both my compute and controller node ( exact same hardware and os version )

The logs seem the same than here: https://bugs.launchpad.net/neutron/+b...

My logs:

/var/log/neutron/linuxbridge-agent.log

2017-04-19 17:19:25.902 10638 ERROR neutron Traceback (most recent call last):
2017-04-19 17:19:25.902 10638 ERROR neutron   File "/usr/bin/neutron-linuxbridge-agent", line 10, in <module>
2017-04-19 17:19:25.902 10638 ERROR neutron     sys.exit(main())
2017-04-19 17:19:25.902 10638 ERROR neutron   File "/usr/lib/python2.7/site-packages/neutron/cmd/eventlet/plugins/linuxbridge_neutron_agent.py", line 21, in main
2017-04-19 17:19:25.902 10638 ERROR neutron     agent_main.main()
2017-04-19 17:19:25.902 10638 ERROR neutron   File "/usr/lib/python2.7/site-packages/neutron/plugins/ml2/drivers/linuxbridge/agent/linuxbridge_neutron_agent.py", line 932, in main
2017-04-19 17:19:25.902 10638 ERROR neutron     launcher.wait()
2017-04-19 17:19:25.902 10638 ERROR neutron   File "/usr/lib/python2.7/site-packages/oslo_service/service.py", line 329, in wait
2017-04-19 17:19:25.902 10638 ERROR neutron     status, signo = self._wait_for_exit_or_signal()
2017-04-19 17:19:25.902 10638 ERROR neutron   File "/usr/lib/python2.7/site-packages/oslo_service/service.py", line 314, in _wait_for_exit_or_signal
2017-04-19 17:19:25.902 10638 ERROR neutron     self.stop()
2017-04-19 17:19:25.902 10638 ERROR neutron   File "/usr/lib/python2.7/site-packages/oslo_service/service.py", line 228, in stop
2017-04-19 17:19:25.902 10638 ERROR neutron     self.services.stop()
2017-04-19 17:19:25.902 10638 ERROR neutron   File "/usr/lib/python2.7/site-packages/oslo_service/service.py", line 688, in stop
2017-04-19 17:19:25.902 10638 ERROR neutron     service.stop()
2017-04-19 17:19:25.902 10638 ERROR neutron   File "/usr/lib/python2.7/site-packages/osprofiler/profiler.py", line 153, in wrapper
2017-04-19 17:19:25.902 10638 ERROR neutron     return f(*args, **kwargs)
2017-04-19 17:19:25.902 10638 ERROR neutron   File "/usr/lib/python2.7/site-packages/neutron/plugins/ml2/drivers/agent/_common_agent.py", line 119, in stop
2017-04-19 17:19:25.902 10638 ERROR neutron     self.set_rpc_timeout(self.quitting_rpc_timeout)
2017-04-19 17:19:25.902 10638 ERROR neutron   File "/usr/lib/python2.7/site-packages/osprofiler/profiler.py", line 153, in wrapper
2017-04-19 17:19:25.902 10638 ERROR neutron     return f(*args, **kwargs)
2017-04-19 17:19:25.902 10638 ERROR neutron   File "/usr/lib/python2.7/site-packages/neutron/plugins/ml2/drivers/agent/_common_agent.py", line 471, in set_rpc_timeout
2017-04-19 17:19:25.902 10638 ERROR neutron     self.state_rpc):
2017-04-19 17:19:25.902 10638 ERROR neutron AttributeError: 'CommonAgentLoop' object has no attribute 'state_rpc'
2017-04-19 17:19:25.902 10638 ERROR neutron 
2017-04-19 17:19:25.903 10638 INFO oslo_rootwrap.client [-] Stopping rootwrap daemon process with pid=10649

I don't get the proposed workaround in the bug description :

fix was to just ...
(more)
edit retag flag offensive close merge delete

Comments

"fix was to just loop until we get an interface with a name (mac)" means that they created a loop in the source code until a device with MAC address is found.

I.e. this is nothing you can do as a deployer or user, but it's what the developers did.

Bernd Bausch gravatar imageBernd Bausch ( 2017-04-19 21:48:15 -0500 )edit

The easiest way to fix your deployment is applying the few changes (https://review.openstack.org/#/c/439766/) directly to /usr/lib/python2.7/site-packages/ neutron/plugins/ml2/drivers/linuxbridge/agent/linuxbridge_neutron_agent.py. Literally 5 lines. Don't forget to restart the agent.

Bernd Bausch gravatar imageBernd Bausch ( 2017-04-19 21:52:05 -0500 )edit

That's the loop that was added. Before, they expected the MAC address to be on the first device:

    for device in devices:
        mac = ip_lib.get_device_mac(device.name)
        if mac:
            break
Bernd Bausch gravatar imageBernd Bausch ( 2017-04-19 22:12:23 -0500 )edit

I did it and restarted the agent with the same results and error logs. It does not fix my issue. Is that possible that the cause is not the same but the error log is identical ?

sledge gravatar imagesledge ( 2017-04-20 14:41:44 -0500 )edit

Sorry I put you on the wrong path. Certainly possible. When I see that an attribute is accessed that doesn't exist in an object, my first guess is a mismatch of several software components.

Bernd Bausch gravatar imageBernd Bausch ( 2017-04-21 06:42:20 -0500 )edit

2 answers

Sort by ยป oldest newest most voted
0

answered 2017-04-22 10:11:39 -0500

sledge gravatar image

updated 2017-04-22 10:12:14 -0500

After a couple of tests I have found that the problem only happens with the OVH kernell. There is a lot of network devices that appears in the ovh default centos 7 installation that are not there with the original centos 7 kernell that might cause issue.

Thanks for your help.

edit flag offensive delete link more
0

answered 2017-04-19 21:11:36 -0500

imacdonn gravatar image

There is no workaround in the bug - only a fix. Since the fix it not released yet, you could apply the patch manually - download https://git.openstack.org/cgit/openstack/neutron/patch/?id=42631e311779b9a779ad8e7bd6654a58c6124fca (https://git.openstack.org/cgit/openst...) and copy to the affected node(s), and run something like:

cat patchfile | (cd /usr/lib/python*/site-packages && patch -p1)

and skip the "tests" part of the patch.

edit flag offensive delete link more

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower

Stats

Asked: 2017-04-19 17:11:08 -0500

Seen: 292 times

Last updated: Apr 22