For RHOS: IP of compute node was lost in orchestrator logs

[Anastasia Palkina]

ISO #91
"release": "3.2.1",
"nailgun_sha": "71bb3cbc6aeda57d6676f5a8e87eb08b80f1ea11",
"ostf_sha": "c54f0ff0e5765f0c89497100eb34386829b00bbd",
"astute_sha": "cb4868f0481d6c7d2d213ceb065ae3c40ab699ce",
"fuellib_sha": "2d45962a302b311c3aa8b7688e80f2ca1d81b2ed"

1. Create new environment (RHOS, simple mode)
2. Add controller, compute and cinder
3. Start deployment. It was successful
4. You can see in orchestrator logs that IP of compute node was lost

[13352] e7365e5d-893f-4c7a-b89c-966bb8d7c92a: MC agent 'puppetd', method 'runonce', results: {:sender=>"16", :statuscode=>0, :statusmsg=>"OK", :data=>{:status=>"stopped", :output=>"Called /usr/bin/puppet agent --onetime --ignorecache --logdest syslog --trace, Currently stopped; last completed run 1386164427 seconds ago", :idling=>0, :running=>0, :stopped=>1, :runtime=>1386164427, :lastrun=>0, :enabled=>1}}

And no one message about this node later.
Only in the end of deployment:
[13352] e7365e5d-893f-4c7a-b89c-966bb8d7c92a: Spent 193.790575792 seconds on puppet run for following nodes(uids): 16,17

[Matthew Mosesohn (raytrac3r)]

Doesn't impact Health Checks but shows as error

[Matthew Mosesohn (raytrac3r)]

Andrey came to me and explained that this is an astute problem with checking status of compute + cinder nodes deploying simultaneously. Astute tracks Cinder node deployment correctly, but doesn't track Compute. Compute node is still deploying when Fuel Web reports that all nodes are ready for deployment because astute doesn't track its progress.

Passing back to Vladimir Sharshov as Andrey requested.

[Mike Scherbakov (mihgen)]

This could be a duplicate of bug with failure of second role deployment