yahoo-eng-team team mailing list archive

[Dongcan Ye <1991791@xxxxxxxxxxxxxxxxxx>]

Public bug reported:

OVN version: 21.12.1
OVS version: 2.15.6
OpenStack version: Yoga

In my test environment, two nodes as HV Chassis and enabled these chassis to host gateways.
1. Add a network and subnet.
2. Add a router, then attach subnet to router.
3. Set router's external gateway.
4. Binding a floatingip to an instance.

Then if we use 'ip link set dev *** down' to make the physical tunnel
interface down. The floatingip can't access.

# ovn-nbctl list Logical_Router_Port lrp-95c640e8-7249-41ee-a407-98a2bc8741ca
_uuid               : 2b3f9e79-33ee-45aa-9a23-ad9e9e460455
enabled             : []
external_ids        : {"neutron:network_name"=neutron-d1d2dcab-e701-4cc2-8c64-c7d81d5f33e5, "neutron:revision_number"="47", "neutron:router_name"="801a8e21-e368-4b9f-8119-830d7f743322", "neutron:subnet_ids"="a3bbe66c-f963-4337-ae8e-bb5fd3e7bde9"}
gateway_chassis     : [11f482cd-18fc-42d4-a3a6-e036ae428371, dc861c0a-e8d6-4b98-ae7f-de0e418f3163]
ha_chassis_group    : []
ipv6_prefix         : []
ipv6_ra_configs     : {}
mac                 : "fa:16:3e:38:5b:50"
name                : lrp-95c640e8-7249-41ee-a407-98a2bc8741ca
networks            : ["192.168.135.248/20"]
options             : {}
peer                : []


We can't see the gateway_chassis priority changes.

# ovn-nbctl list gateway_chassis
_uuid               : 11f482cd-18fc-42d4-a3a6-e036ae428371
chassis_name        : "d2692315-6389-4e95-b211-a3dd54ec4582"
external_ids        : {}
name                : lrp-95c640e8-7249-41ee-a407-98a2bc8741ca_d2692315-6389-4e95-b211-a3dd54ec4582
options             : {}
priority            : 2

_uuid               : dc861c0a-e8d6-4b98-ae7f-de0e418f3163
chassis_name        : "00b6dd97-6360-4ae5-b1f4-a68cf2e8c70c"
external_ids        : {}
name                : lrp-95c640e8-7249-41ee-a407-98a2bc8741ca_00b6dd97-6360-4ae5-b1f4-a68cf2e8c70c
options             : {}
priority            : 1


After read the HA guide in OVN[1], it said we can set cpath_down config.
But there is no effective.

ovn-nbctl --wait=hv set NB_Global . options:"bfd-cpath-down"=true
ovs-vsctl set interface ovn-d26923-0 bfd:cpath_down=true


In guide[2], we test three situations, it works for us.
(1) The gateway chass is shutdown.
(2) The ovs-vswitchd is stopped.
(3) The ovn-controller is stopped.


[1] https://docs.ovn.org/en/latest/topics/high-availability.html
[2] https://docs.openstack.org/neutron/latest/admin/ovn/routing.html

** Affects: neutron
     Importance: Undecided
         Status: New


** Tags: ovn

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to neutron.
https://bugs.launchpad.net/bugs/1991791

Title:
  [ovn][l3-ha] BFD monitoring not working for network down

Status in neutron:
  New

Bug description:
  OVN version: 21.12.1
  OVS version: 2.15.6
  OpenStack version: Yoga

  In my test environment, two nodes as HV Chassis and enabled these chassis to host gateways.
  1. Add a network and subnet.
  2. Add a router, then attach subnet to router.
  3. Set router's external gateway.
  4. Binding a floatingip to an instance.

  Then if we use 'ip link set dev *** down' to make the physical tunnel
  interface down. The floatingip can't access.

  # ovn-nbctl list Logical_Router_Port lrp-95c640e8-7249-41ee-a407-98a2bc8741ca
  _uuid               : 2b3f9e79-33ee-45aa-9a23-ad9e9e460455
  enabled             : []
  external_ids        : {"neutron:network_name"=neutron-d1d2dcab-e701-4cc2-8c64-c7d81d5f33e5, "neutron:revision_number"="47", "neutron:router_name"="801a8e21-e368-4b9f-8119-830d7f743322", "neutron:subnet_ids"="a3bbe66c-f963-4337-ae8e-bb5fd3e7bde9"}
  gateway_chassis     : [11f482cd-18fc-42d4-a3a6-e036ae428371, dc861c0a-e8d6-4b98-ae7f-de0e418f3163]
  ha_chassis_group    : []
  ipv6_prefix         : []
  ipv6_ra_configs     : {}
  mac                 : "fa:16:3e:38:5b:50"
  name                : lrp-95c640e8-7249-41ee-a407-98a2bc8741ca
  networks            : ["192.168.135.248/20"]
  options             : {}
  peer                : []


  We can't see the gateway_chassis priority changes.

  # ovn-nbctl list gateway_chassis
  _uuid               : 11f482cd-18fc-42d4-a3a6-e036ae428371
  chassis_name        : "d2692315-6389-4e95-b211-a3dd54ec4582"
  external_ids        : {}
  name                : lrp-95c640e8-7249-41ee-a407-98a2bc8741ca_d2692315-6389-4e95-b211-a3dd54ec4582
  options             : {}
  priority            : 2

  _uuid               : dc861c0a-e8d6-4b98-ae7f-de0e418f3163
  chassis_name        : "00b6dd97-6360-4ae5-b1f4-a68cf2e8c70c"
  external_ids        : {}
  name                : lrp-95c640e8-7249-41ee-a407-98a2bc8741ca_00b6dd97-6360-4ae5-b1f4-a68cf2e8c70c
  options             : {}
  priority            : 1

  
  After read the HA guide in OVN[1], it said we can set cpath_down config.
  But there is no effective.

  ovn-nbctl --wait=hv set NB_Global . options:"bfd-cpath-down"=true
  ovs-vsctl set interface ovn-d26923-0 bfd:cpath_down=true

  
  In guide[2], we test three situations, it works for us.
  (1) The gateway chass is shutdown.
  (2) The ovs-vswitchd is stopped.
  (3) The ovn-controller is stopped.

  
  [1] https://docs.ovn.org/en/latest/topics/high-availability.html
  [2] https://docs.openstack.org/neutron/latest/admin/ovn/routing.html

To manage notifications about this bug go to:
https://bugs.launchpad.net/neutron/+bug/1991791/+subscriptions