charm gets stuck on config-changed hook and holds juju machine-lock while trying to create br-ex

This issue was faced by a customer on their first small bionic-stein cloud
(9 nodes, 7 nova-compute + 2 neutron-gateway).

They did not face this issue in any of their larger bionic-stein clouds
(39 nodes, 37 nova-compute + 2 neutron-gateways).

I was able to reproduce this in an OrangeBox (10 node - all nova-compute).

The bundle I used to reproduce this issue - https://paste.ubuntu.com/p/SWP5ydqhH9/

The juju status (problem is on the octavia LXD containers): https://paste.ubuntu.com/p/K2Yc8yp454/

Output of ps aux | grep juju on one of the units shows that there is a python3 process running the config-changed hook of the neutron-ovs : https://paste.ubuntu.com/p/HbnkgttwGf/

juju machine-lock.log: https://paste.ubuntu.com/p/DxH3264bv9/

Also I saw that the ovs-vswitchd service fails to run.

I was led to believe that this issue may be due to this bug: https://bugs.launchpad.net/charm-neutron-openvswitch/+bug/1906280 and so I applied the patch that adds
limits.kernel.memlock: unlimited to neutron-ovs's lxd-profile.yaml.

However after applying the patch and upgrading the charm and recreating the containers, the result is still the same so I think it is a diffrent issue altogether.

Here are the crashdumps before the patch:https://drive.google.com/file/d/11gbHuSIOG47BMG4AP0kbgE1pphKj8vaA/view?usp=sharing

and after the patch: https://drive.google.com/file/d/1GY0tJH2PcNfuUu4UatoSS88PPXoKRRfk/view?usp=sharing