(POOLER-140) Ensure a VM is alive at checkout

[mattkirby]

This commit duplicates the vm_ready? check to the API layer to allow for API to validate that a VM is alive at checkout. Without this change API relies upon the checks in pool_manager validating pools. This change should allow for additional insight into whether a machine is in a ready state and resopnding at checkout time.

[kevpl]

seems fine to me.

my one question reading through is whether or not we can eliminate all the duplication of this line by putting it in a before :each or let statement:

allow_any_instance_of(Vmpooler::API::Helpers).to receive(:open_socket).and_return(socket)

[mattkirby]

@kevpl I found that when the allow statement into a before(:each) block that it caused additional failures. I think this was because cases that were supposed to fail would show vm_ready? as true. I think it would be possible to reduce duplication by restructuring the tests to confine the before statements to where the addition is needed. It seemed like more work than what I ended up with though, so unless your preference is strong I think specifying the addition for each test works ok. Feel free to point out if I'm overcomplicating it though.

[mattkirby]

I could not find a way to use allow to stub the return value of the open_socket method without using allow_any_instance_of. I suspect there's something I don't understand with where sinatra is loading the helpers, and that if I were to understand that better there is likely another way to get at this. I'm all for a cleaner solution, but after several days of banging my head against this I did not find one.