Boot server errors when Consul cluster is degraded

[nshalman]

Issue Description

When Consul cluster is degraded, the boot server spews errors that look like:
error=requested ip not assigned to this mac
which are probably the result of KV being unable to write to Consul.

When writes to KV fail, we may want to try to bubble that error up to the administrator.

Steps to Reproduce (if applicable):

1. Bootstrap a cluster
2. Boot a second node
3. Destroy second node
4. Attempt to boot a new second node

Acceptance Criteria (how do we know when this is done?)

Related Tickets

Additional Info

Output of uname -a:

Environment (physical, VM, cloud provider, etc.):

Other Relevant Info: