The Hypershift in-place upgrade controller currently uses a single pod for each Machine Config Daemon (MCD) instance it needs to run. This is because the original design called for the MCD to be run as a one-shot Pod. To make the controller more robust, we should monitor the pods it creates (see: https://issues.redhat.com/browse/MCO-259).

As a precursor to that work, we should modify the controller to use a DaemonSet to create and manage these pods on all nodes so we can address all the pods simultaneously or individually. The MCD itself contains a state machine which coordinates config rollouts across the nodes, so it should just work as-is.

Done When:

• The Hypershift in-place upgrade controller uses a DaemonSet to manage MCD pods.
• Appropriate unit and e2e tests are written.
• Note: This does not include MCD pod monitoring work as that will be covered by https://issues.redhat.com/browse/MCO-259.