yahoo-eng-team team mailing list archive

[Giulio Fidente <1403921@xxxxxxxxxxxxxxxxxx>]

Public bug reported:

when working with a topology with multiple, balanced, neutron-server
nodes, the neutron l3 agent failover facility
(allow_automatic_l3agent_failover) works as expected if the cluster is
stable (aka, the agent dies but server remains stable)

it does not instead when the entire node is brought down and the server
it was connected to initially becomes unavailable, the remaining server
nodes will recognize agent is died, but relocation of routers never
happens

** Affects: neutron
     Importance: Undecided
         Status: New

** Description changed:

- when working in a topology with multiple, balanced, neutron-server
- nodes, the neutron l3 agent failover (allow_automatic_l3agent_failover)
- works as expected if the cluster is stable (aka, the agent dies but
- server remains stable); it fails to relocate the routers instead when
- the entire node is brought down (despite the remaining server nodes
- noticing the agent hosting a router is now down)
+ when working with a topology with multiple, balanced, neutron-server
+ nodes, the neutron l3 agent failover facility
+ (allow_automatic_l3agent_failover) works as expected if the cluster is
+ stable (aka, the agent dies but server remains stable)
+ 
+ it does not instead when the entire node is brought down and the server
+ it was connected to initially becomes unavailable, the remaining server
+ nodes will recognize agent is died, but relocation of routers never
+ happens

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to neutron.
https://bugs.launchpad.net/bugs/1403921

Title:
  allow_automatic_l3agent_failover fails if neutron-server goes down as
  well

Status in OpenStack Neutron (virtual network service):
  New

Bug description:
  when working with a topology with multiple, balanced, neutron-server
  nodes, the neutron l3 agent failover facility
  (allow_automatic_l3agent_failover) works as expected if the cluster is
  stable (aka, the agent dies but server remains stable)

  it does not instead when the entire node is brought down and the
  server it was connected to initially becomes unavailable, the
  remaining server nodes will recognize agent is died, but relocation of
  routers never happens

To manage notifications about this bug go to:
https://bugs.launchpad.net/neutron/+bug/1403921/+subscriptions