Rely on 10 minute starting_timeout instead of a heartbeating thread

[jrafanie]

The server process allows "starting" workers up to 10 minutes to start
before marking them as not responding. We can use this starting_timeout
instead of heartbeating in a thread while another thread runs the
ansible setup playbook (configure).

cc @Fryguy @gtanzillo @carbonin

I manually tested this, I'm open to suggestions on how to unit test this.

[miq-bot]

Checked commit jrafanie@14ee97b with ruby 2.2.6, rubocop 0.47.1, and haml-lint 0.20.0
1 file checked, 0 offenses detected
Everything looks good. ⭐

[Fryguy]

I'm confused...this doesn't use the do_before_work_loop mechanism at all.

[jrafanie]

I'm confused...this doesn't use the do_before_work_loop mechanism at all.

@Fryguy Yeah, we can't without creating a new thread. The point was to remove the heartbeating thread. We can't block the server so we'd have to create another thread to do the setup_ansible in the do_before_work_loop. This PR moves all of the logic into the do_work_loop in a single thread.

[carbonin]

I'm for this, we've gotten burned by this extra thread at least once already.

[jrafanie]

@Fryguy can you review/merge? @carbonin keeps hitting this issue when the extra thread spins up but doesn't complete.

[carbonin]

I think @Fryguy is suggesting that instead of moving things into the thread, we should move the thread into EmbeddedAnsibleWorker#start_runner which would map to where we call fork for the other workers.