GetInstanceStatusByName doesn't return an error when > 1 instance if found

[EmilienM]

/kind bug

What steps did you take and what happened:

GetInstanceStatusByName doesn't return an error when > 1 instance if found and this can lead to cluster problems if we have two instances with the same name. Could be a leftover or a staled resource, that needs to be removed before creating a new one.

What did you expect to happen:

An error should be returned.

Environment:

• Cluster API Provider OpenStack version (Or git rev-parse HEAD if manually built): latest

[EmilienM]

/assign EmilienM

[EmilienM]

@mdbooth so in the current state of things, if two instances have the same name, cleanup will never happen. This will likely change once we have the OpenStackServer controller that I'm working on at the moment; so I wonder if we want to provide a "fix" for this into 0.10 or if we're fine with that behaviour now and we'll improve with the new controller.

[mdbooth]

I think if we have 2 instances with the same name and can't distinguish between them we're just going to have to report an error to the user. I'm not sure we can safely do anything about it, now or with OpenStackServer.

[EmilienM]

ok so i'll submit a PR that just reports an error for now, gotcha.