(maint) change maint-mode state of status implementation to :starting

[wkalt]

This changes our status implementation to return :starting while in maint mode,
and bumps to the status-service version that enables the change.

[puppetlabs-jenkins]

Refer to this link for build results (access rights to CI server needed):
http://jenkins-enterprise.delivery.puppetlabs.net/job/enterprise_puppetdb_init-multijob_githubpr-master/278/
Test PASSed.

[kbarber]

How do I artificially create this maintenance mode so I can manually test? :-). Sorry, I presume this is HA stuff I'm not across :-).

Incidentally, while I was starting and just seeing what it did while I kept curling, I get an 'unknown' state: https://gist.github.com/kbarber/e36fcd54373cb8c75247. Incidentally, this is a pre-existing condition it would seem.

[wkalt]

Nope, not HA stuff. The easiest thing to do is probably query

curl -X GET http://localhost:8080/status/v1/services

in a bash loop during PDB startup, and observe that you get a few "starting" statuses.

[wkalt]

To your second point that's a bug with the mq-depth component. I'll add some handling there.

[kbarber]

@wkalt okay, well in that case here are my results: https://gist.github.com/kbarber/490684628d9acb69521c

I don't see the transition, I'm guessing its hidden behind these exceptions. Notice however, there is another problem with db-spec? I think we have at least 2 cases of failure in this mechanism today it seems.

[kbarber]

@wkalt looks good doing manual testing now: https://gist.github.com/kbarber/9ca6f28ff1745b866a27

[puppetlabs-jenkins]

Refer to this link for build results (access rights to CI server needed):
http://jenkins-enterprise.delivery.puppetlabs.net/job/enterprise_puppetdb_init-multijob_githubpr-master/289/
Test PASSed.