yahoo-eng-team team mailing list archive

[Slawek Kaplonski <1771841@xxxxxxxxxxxxxxxxxx>]

Public bug reported:

When backup L3 agent in HA deployment is restarted, there is 20-30 seconds of data plane connectivity break to FloatingIP.
It looks that during start of L3 agent, it first enable IPv6 forwarding: https://github.com/openstack/neutron/blob/master/neutron/agent/l3/namespaces.py#L106 and then disable it for non master node: https://github.com/openstack/neutron/blob/master/neutron/agent/l3/ha_router.py#L419
That cause send of MLDv2 packet:

08:38:29.815180 Out fa:16:3e:9e:de:0e ethertype IPv6 (0x86dd), length
112: :: > ff02::16: HBH ICMP6, multicast listener report v2, 2 group
record(s), length 48

and that breaks somehow connectivity as packets are sent to backup node
instead of master one.

** Affects: neutron
     Importance: Undecided
     Assignee: Slawek Kaplonski (slaweq)
         Status: Confirmed


** Tags: l3-ha

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to neutron.
https://bugs.launchpad.net/bugs/1771841

Title:
  packet loss during backup L3 HA agent restart

Status in neutron:
  Confirmed

Bug description:
  When backup L3 agent in HA deployment is restarted, there is 20-30 seconds of data plane connectivity break to FloatingIP.
  It looks that during start of L3 agent, it first enable IPv6 forwarding: https://github.com/openstack/neutron/blob/master/neutron/agent/l3/namespaces.py#L106 and then disable it for non master node: https://github.com/openstack/neutron/blob/master/neutron/agent/l3/ha_router.py#L419
  That cause send of MLDv2 packet:

  08:38:29.815180 Out fa:16:3e:9e:de:0e ethertype IPv6 (0x86dd), length
  112: :: > ff02::16: HBH ICMP6, multicast listener report v2, 2 group
  record(s), length 48

  and that breaks somehow connectivity as packets are sent to backup
  node instead of master one.

To manage notifications about this bug go to:
https://bugs.launchpad.net/neutron/+bug/1771841/+subscriptions