Destroy the bootstrap node in an attempt to test backup and restoration.

[Ebaneck]

Component:

'tests', 'backup', 'restore' 'ci'

Why this is needed:

While performing a bootstrap node recovery, we do not have any real need for the old bootstrap node and as such we should destroy it completely.

What should be done:

Add a CI step to destroy the old bootstrap node.

Implementation proposal (strongly recommended):

Seems to me that there is no real need to destroy the entire open stack multi-node infrastructure. We could destroy the bootstrap node with a kubectl delete node bootstrap

It looks like some extra cleanup is needed such as

• Maybe cleaning up Persistent Data if any exist on the host (prometheus data for example)
• For any hosts that had ran an etcd service, there could be a couple of items that may need to be cleaned up (e.g Remove the etcd backups by running rm -r /var/lib/etcd/member/*.)

Test plan:

[alexandre-allard]

Closing this issue as it has already been done in #2061 (#1687)