Comment 10 for bug 2023668

We just upgraded our pre-production cloud from xena->yoga and this is happening mid upgrade, preventing completion and creating an indeterminate state.
Original deploy was 20.04 w/ Xena which worked fine a few hours ago and now does not (with those symptoms all over the place) after upgrading to yoga stable.

@Athos Ribeiro: 3 "control nodes" with 15 nova/zun/ovn nodes
Image is now https://quay.io/repository/openstack.kolla/ubuntu-source-rabbitmq/manifest/sha256:40e788bb4122fd9546a7f9c8ebee48c8a4de20cdd2673cc9c7767f720261018e

Resetting queues in rabbitmq does not help, nor does enabling om_enable_rabbitmq_high_availability and going through the reconfiguration procedure. Looks like an actual bug in the container.