deploy ha with nova vlan failed on attempts to start keystone

[Tatyanka]

{"build_id": "2014-07-30_00-03-17", "mirantis": "yes", "build_number": "518", "ostf_sha": "09b6bccf7d476771ac859bb3c76c9ebec9da9e1f", "nailgun_sha": "9112643217a8e591da78e554c3d499b2835943a3", "production": "docker", "api": "1.0", "fuelmain_sha": "327981b9846b79cb9fc99b03650d382b4e1b3b21", "astute_sha": "6db5f5031b74e67b92fcac1f7998eaa296d68025", "release": "5.0.1", "fuellib_sha": "6e6bf6cba3dd8a105f7e7f52f8de68c5b4bc32fe"}

Steps to Reproduce:
1. Deploy centos HA with nova vlan
    3 controllers 2 computes

Expected result:
Cluster is ready

Actual:
cluster in error state
Puppet can not start keystone on second controller with next error
http://paste.openstack.org/show/89138/

btw - when I try to execute the same command mannually - keystone starts without problems

[Tatyanka (tatyana-leontovich)]

Also re-deploy of this environment finish with success, so seems there is some floating problem may be caused of low performance environments

[Bogdan Dobrelya (bogdando)]

#518 ISO doesn't contain expected patches, hence setting to invalid