Comment 16 for bug 1533197

The problem was reproduced on build #562 on scale env 14 at even smaller scale (on order of magnitude!).

Environment:
    * 3 controllers
        - Sahara, Murano activated
        - Ceph is used for everything
    * 2 LMA nodes
    * 5 computes

I've reproduced the steps above on different amount of booting VMs:
  * During boot of 15 instances (3 per compute) I've got following instances failed:
    - e960617e-041f-4348-8d3c-1a38b39cec11
  * During boot of 20 instances (4 per compute) I've got following instances failed:
    - 04e0746b-5ba6-4622-bd8a-acf68f6c77e1
    - 528a9bc1-5374-494b-b405-3ac35d33e3a4

Please find diagnostic snapshot here:
https://drive.google.com/a/mirantis.com/file/d/0B9tzODpFABxkdk9JeDNEZ0l1Y0k/view?usp=sharing