Deploy from task till the end finished with puppet errors

[Egor Kotko]

Steps to reproduce:
1. Create and provision cluster: 1Controller, 1 Compute, 1 Cinder;

2. On the controller execute tasks till the 'virtual_ips'
/api/clusters/1/deployment_tasks?end=virtual_ips
On the compute and cinder execute tasks till the 'hosts'
/api/clusters/1/deployment_tasks?end=hosts

3. Execute on all nodes task started from 'conntrackd'
/api/clusters/1/deployment_tasks?start=conntrackd

Expected result deployed cluster

Actual: deployment failed with errors in puppet log on controller.
http://paste.openstack.org/show/482430/

[Matthew Mosesohn (raytrac3r)]

Possibly a duplicate of https://bugs.launchpad.net/fuel/+bug/1519366

[Kyrylo Galanov (kgalanov)]

It's not a bug. The issue caused by incorrect input data.
Automatic deployment puppet calls: http://paste.openstack.org/show/482578/
Puppet calls after executing provided commands: http://paste.openstack.org/show/482596/