Open Stack

Some of our clusters are heavily used, and in those clusters we get complaints when we reboot a controller (or sometimes when we deploy and containers restart). Is that normal, or does it mean that we have something configured wrong?

The symptoms are intermittent 504 from endpoints, and VM creation/deletion failing or partially completing, for example the VM is created but without DNS records.

We are not following the "Removing existing controllers" procedure [1] before rebooting the controller; is that necessary to avoid these issues?

1. https://docs.openstack.org/kolla-ansible/latest/user/adding-and-removing-hosts.html