yahoo-eng-team team mailing list archive

[Launchpad Bug Tracker <1430999@xxxxxxxxxxxxxxxxxx>]

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

** Changed in: neutron
       Status: Incomplete => Expired

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to neutron.
https://bugs.launchpad.net/bugs/1430999

Title:
  get_devices_details_list times out in a scaled up environment

Status in neutron:
  Expired

Bug description:
  The rpc call get_devices_details_list made from within the
  neutron_openvswitch_agent can timeout as an environment scales up and
  the port list grows.  This prevents the neutron_openvswitch_agent from
  starting successfully.  Increasing the rpc_response_timeout is one
  solution but doesn't seem to be ideal.  Another solution is to
  repeatedly call get_devices_details_list with a fixed subset size of
  devices.  The subset size could be set so that the rpc call always
  completes within the default rpc_response_timeout and a timeout
  failure would be independent of scale.

  When update_device_list(devices_up=[600 ports]) called by agent, it is getting timeout. I see response comping for this RPC after 2 minutes(Default RPC timeout is 1 minute).
  As there is timeout, agent is again trying to process these 600 ports.
  As agent is busy processing these ports all time, it is ignoring wiring up of new port added by nova and hence nova is not receiving network-vif-plugged notification, and failing building of instance with "Timeout waiting for vif plugging callback for instance".

To manage notifications about this bug go to:
https://bugs.launchpad.net/neutron/+bug/1430999/+subscriptions