yahoo-eng-team team mailing list archive

[Jian Wen <wenjianhn@xxxxxxxxx>]

The fix works.

** Changed in: neutron
       Status: Incomplete => Fix Released

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to neutron.
https://bugs.launchpad.net/bugs/1255850

Title:
  L3_agent: if openvswitch fails to restart, _rpc_loop will block
  forever.

Status in OpenStack Neutron (virtual network service):
  Fix Released

Bug description:
  Related to bug 890180

  How to reproduce the bug in a devstack environment.

  1. To simulate that the ovs daemon crashs, and we are not able to successfully restart it.
     $ sudo /etc/init.d/openvswitch-switch stop

  2. Remove an interface from a router. Because of the ovs daemon is down, ovs-vsctl will
     waits forever. As a result, _rpc_loop hangs waiting for _process_routers with holding
     the lock.
     $ neutron router-interface-remove router2 0a33e6c2-9176-4d88-840e-3879cbee5fd2

  3. Try to assign a floating IP to an instance
     $ neutron floatingip-associate c6b3ee4a-1eac-4438-a561-9aaf57e1b373 17e11c92-8707-4941-bc2d-455a1b15707d

     Expected results:
        The floating IP has been added to the qg-* device.
        It's OK for the moment if we can't remove the interface from the router.
     Actual results:
        The floating IP has not been added to the qg-* device.
        _rpc_loop hangs without any explicit log messages about the blocking.

To manage notifications about this bug go to:
https://bugs.launchpad.net/neutron/+bug/1255850/+subscriptions