Astute fails deployment if primary controller deployment failed even only 1 time

[Vladimir Kuklin]

This bug: https://bugs.launchpad.net/fuel/+bug/1348171 discovered an issue. With only one puppet failure on the primary controller we have deployment marked as failed.

[Vladimir Sharshov (vsharshov)]

We lost access via mcollective with node '2' (controller) and marked this one as error. If any node in a cluster has error, we do not uploaded cirros image. I think no bug here.

We can show in the end of deploy before post-hook processing final state for all nodes (success/error). It will save time for such case to easy understand what failed without web-interface. What do you think?

Log:

2014-08-04T11:20:33 debug: [375] Retry #1 to run mcollective agent on nodes: '2'
2014-08-04T11:21:15 debug: [375] Retry #2 to run mcollective agent on nodes: '2'
2014-08-04T11:21:58 debug: [375] Retry #3 to run mcollective agent on nodes: '2'
2014-08-04T11:22:41 debug: [375] Retry #4 to run mcollective agent on nodes: '2'
2014-08-04T11:23:23 debug: [375] Retry #5 to run mcollective agent on nodes: '2'
2014-08-04T11:24:05 debug: [375] Data received by DeploymentProxyReporter to report it up: {"nodes"=>[{"uid"=>"2", "status"=>"error", "error_type"=>"deploy", "role"=>"controller"}]}
2014-08-04T11:24:05 debug: [375] Data send by DeploymentProxyReporter to report it up: {"nodes"=>[{"uid"=>"2", "status"=>"error", "error_type"=>"deploy", "role"=>"controller"}]}
2014-08-04T11:24:05 debug: [375] Data send by DeploymentProxyReporter to report it up: {"nodes"=>[{"uid"=>"2", "status"=>"error", "error_type"=>"deploy", "role"=>"controller"}]}
2014-08-04T11:24:05 debug: [375] Nodes statuses: {"succeed"=>[], "error"=>[], "running"=>["5"]}
2014-08-04T11:24:05 info: [375] Casting message to Nailgun: {"method"=>"deploy_resp", "args"=>{"task_uuid"=>"8ae326ed-deb0-48ce-8175-53673d972785", "nodes"=>[{"uid"=>"2", "status"=>"error", "error_type"=>"deploy", "role"=>"controller"}]}}