Hi all,

We are looking for the best way to process system upgrades on our controller nodes and especially the best way to restart them (we have 3 controllers nodes with routers ha).

If we just restart them one by one (waiting long time between each), we have already seen rabbitmq cluster errors (needing a complete restart of rabbitmq).

With a kolla stop, a restart and a kolla deploy, some docker containers are restarted on others nodes.

The administration guides describes how to add or remove a node but not this simple case.

How do you proceed securely?

Thanks for all,

Yves