Comment 8 for bug 1909004

Hi Jake. As this is (possibly?) in a lab setting, it may be helpful to grab a juju-crashdump of the model which would grab all the logs from all the units? https://github.com/juju/juju-crashdump -- there are/have been some ovn issues fixed during the most recent dev cycle, so it also may be worth (in a lab setting only) trying it with the ~openstack-charmers-next charms.