Invalid estimation of cluster deployment progress

[Anastasiia Naboikina]

Steps to reproduce:
1. Create new cluster.
2. Add controller and compute.
3. Start cluster deployment.
4. Wait approximately 1-2 minutes.

Expected result:
cluster deployment progress bar shows less than 50%, something in between 10% to 30% and
is growing gradually as deployment continues.

Actual result:
cluster deployment progress bar shows 50% right away and doesn`t change until operation system is installed on nodes.

Reproduced on iso 4.0-94

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/60835

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/60836

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/60835
Committed: http://github.com/stackforge/fuel-web/commit/3424388fac56e38d39b8b91d9648f42aebe6cc46
Submitter: Jenkins
Branch: master

commit 3424388fac56e38d39b8b91d9648f42aebe6cc46
Author: Evgeniy L <email address hidden>
Date: Fri Dec 6 18:46:00 2013 +0400

    Fix progress bar calculation when provisioning is running

    Separated deployment and provisioning progress bar
    calculation because orchestrator now sends progress
    for provisioning task, not for deployment task
    as it was before.

    Closes-bug: #1257342
    Change-Id: Ie0c84ee92ec6f027b483e6f9116755f21489ec8d

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/60836
Committed: http://github.com/stackforge/fuel-astute/commit/3237ca81ef8ee9b238004089e282ebef6bf96ea0
Submitter: Jenkins
Branch: master

commit 3237ca81ef8ee9b238004089e282ebef6bf96ea0
Author: Evgeniy L <email address hidden>
Date: Mon Dec 9 15:17:27 2013 +0400

    Fix provisioning progress calculation

    * don't send ready status in provision handler
      because nodes are not provisioned yet in
      this handler
    * fix undefined variable in provision handler
    * fix nodes calculations statuses in
      watch_provision_progress method. Firstly get
      all nodes in error status, then retrieve nodes
      which ready and add their status to result
      variable.
    * fix typo in logs

    Closes-bug: #1257342
    Closes-bug: #1249994
    Change-Id: I56e06ade671795fc8c87b7e84e159df7abdba34b