[Bug 1813291] Re: instances created before an upgrade from queens to rocky lost their networking information after the ugprade

[Corey Bryant]

I wonder if a relation is needed between nova-cloud-controller and
memcached to recreate. That was needed to recreate the other bug for
LP#1812672.

-- 
You received this bug notification because you are a member of Ubuntu
OpenStack, which is subscribed to Ubuntu Cloud Archive.
https://bugs.launchpad.net/bugs/1813291

Title:
  instances created before an upgrade from queens to rocky lost their
  networking information after the ugprade

Status in Ubuntu Cloud Archive:
  Incomplete

Bug description:
  Hi,

  Using bionic with 18.11 charms. I recently upgraded an openstack
  cluster from queens to rocky. Instances created before the upgrade
  stopped showing network information in "openstack server list" - see
  https://pastebin.canonical.com/p/tJ4ZcDDDx5/.

  Trying to understand what was going on, I realized that their
  "network_info" column in the "instance_info_caches" table ("nova"
  database) in mysql was empty :
  https://pastebin.canonical.com/p/qhGzh2WF5F/

  I tried 2 different things to fix this :

  1. delete their row in the "instance_info_caches" table. The row comes
  back when the same empty network_info

  2. openstack server stop/start the instance : doesn't work. IIRC it
  does nothing. I tried to kill the qemu process, reset-state and
  restart, but the instance comes up with no networking

  3. delete the neutron port, and manually recreate and re-attach it :
  this appears to bring back networking after the instance is
  stopped/started. This is the instance that has network info in the
  "openstack server list" above.

  Thanks

To manage notifications about this bug go to:
https://bugs.launchpad.net/cloud-archive/+bug/1813291/+subscriptions