Sometimes for external applications it is important to identify the
cloud instance of the nodes. Until this patch there was no contract
that the node name returned by List was also the unique identifier of
the cloud instance. This new API ensures that an external application
can reliably retrieve the relevant instance id of the nodes.
Signed-off-by: Federico Simoncelli <fsimonce@redhat.com>
--sync_nodes=false gives user flexibility to add/remove nodes in the
cluster using REST api/kubectl cli and at the same time can use
cloud provider for other resources like persistent disks, etc.
During startup, if nodes have already been defined the controller
does not need to create them, just needs to keep them up to date.
When defining a static list, the controller would loop and fail
repeatedly if the node existed.
I don't know if we need to update the node during registration,
as long as the sync loop checks it later.
