Xen ovs-agent-plugin polling manager is reported as not active

[huan]

My environment is XenServer + Neutron with ML2 plugin, OVS as driver, VLan type.
This is a single box environment installed by devstack.
When it began to run, I found q-agt always had error logs like this:

2015-09-09 05:15:23.653 ERROR neutron.agent.linux.ovsdb_monitor [req-2dfc6939-277b-4d54-83d5-d0aecbfdc07c None None] Interface monitor is not active

Dig deep into the code, I found the callstack is trace from OVSNeutronAgent.rpc_loop() to neutron/agent/linux/utils.py remove_abs_path(). So I temporarily added debug log and found

2015-09-09 05:15:23.653 DEBUG neutron.agent.linux.utils [req-2dfc6939-277b-4d54-83d5-d0aecbfdc07c None None] cmd_matches_expected, cmd:['/usr/bin/python', '/usr/local/bin/neutron-rootwrap-xen-dom0', '/etc/neutron/rootwrap.conf', 'ovsdb-client', 'monitor', 'Interface', 'name,ofport,external_ids', '--format=json'], expect:['ovsdb-client', 'monitor', 'Interface', 'name,ofport,external_ids', '--format=json'] from (pid=11595) cmd_matches_expected /opt/stack/neutron/neutron/agent/linux/utils.py:303

So, it's clear that after remove absolute path, the command still cannot match, so it will lead to the ERROR log of "Interface monitor is not active"

[Rossella Sblendido (rossella-o)]

It seems a problem with cmd_matches_expected ....this is used all over the place, is the polling manager the only one that it's reported as not active?

[huan (huan-xie)]

Hi,
So sorry for my late reply.
Yes, I also think the problem is with cmd_matches_exepcted. And polling manager is the only one that reported not active as I saw in the log.
I think the problem is whith xenserver as hypervisor, q-agt which runs in compute node and will talk to dom0, we will set q-agt's root_helper as
"root_helper = sudo /usr/local/bin/neutron-rootwrap-xen-dom0 /etc/neutron/rootwrap.conf"
And this is different from the one q-domu that runs in domU and thus make this problem happen.

BR//Huan

-----Original Message-----
From: <email address hidden> [mailto:<email address hidden>] On Behalf Of Rossella Sblendido
Sent: Wednesday, September 09, 2015 3:50 PM
To: Huan Xie
Subject: [Bug 1493662] Re: Xen ovs-agent-plugin polling manager is reported as not active

It seems a problem with cmd_matches_expected ....this is used all over the place, is the polling manager the only one that it's reported as not active?

** Summary changed:

- ovs-agent-plugin polling manager fails to monitor interface table
+ Xen ovs-agent-plugin polling manager is reported as not active

--
You received this bug notification because you are subscribed to the bug report.
https://bugs.launchpad.net/bugs/1493662

Title:
  Xen ovs-agent-plugin polling manager is reported as not active

Status in neutron:
  New

Bug description:
  My environment is XenServer + Neutron with ML2 plugin, OVS as driver, VLan type.
  This is a single box environment installed by devstack.
  When it began to run, I found q-agt always had error logs like this:

  2015-09-09 05:15:23.653 ERROR neutron.agent.linux.ovsdb_monitor [req-
  2dfc6939-277b-4d54-83d5-d0aecbfdc07c None None] Interface monitor is
  not active

  Dig deep into the code, I found the callstack is trace from
  OVSNeutronAgent.rpc_loop() to neutron/agent/linux/utils.py
  remove_abs_path(). So I temporarily added debug log and found

  2015-09-09 05:15:23.653 DEBUG neutron.agent.linux.utils [req-2dfc6939-277b-4d54-83d5-d0aecbfdc07c None None] cmd_matches_expected, cmd:['/usr/bin/python', '/usr/local/bin/neutron-rootwrap-xen-dom0', '/etc/neutron/rootwrap.conf', 'ovsdb-client', 'monitor', 'Interface', 'name,ofport,external_ids', '--format=json'], expect:['ovsdb-client', 'monitor', 'Interface', 'name,ofport,external_ids', '--format=json'] from (pid=11595) cmd_matches_expected /opt/stack/neutron/neutron/agent/linux/utils.py:303

  So, it's clear that after remove absolute path, the command still cannot match, so it will lead to the ERROR log of "Interface monitor is not active"

To manage notifications about this bug go to:
https://bugs.launchpad.net/neutron/+bug/1493662/+subscriptions

[Bob Ball (bob-ball)]

See also bug 1495423

[Armando Migliaccio (armando-migliaccio)]

This bug is > 240 days without activity. We are unsetting assignee and milestone and setting status to Incomplete in order to allow its expiry in 60 days.

If the bug is still valid, then update the bug status.

[Launchpad Janitor (janitor)]

[Expired for neutron because there has been no activity for 60 days.]

[huan (huan-xie)]

I think we don't need to fix this bug as it bug I reported is caused by my configuration problems, so we can close this bug