Slaves don't boot via PXE after deployment stop on provisioning

[Artem Panchenko]

api: '1.0'
astute_sha: a3432e6e31ffd6f1c56386b2eb54afeacb74750b
build_id: 2014-05-20_01-10-31
build_number: '213'
fuellib_sha: 3d92142a5643af82596f0450e39282550a45e5db
fuelmain_sha: 68c62519bc788fd8ff27e4576a6cdf7e7fac14c0
mirantis: 'yes'
nailgun_sha: ab7f7dfddadfe0e08a39693c6d33aa0250f20142
ostf_sha: 353f918197ec53a00127fd28b9151f248a2a2d30
production: docker
release: '5.0'

Steps to reproduce:

 1. Create new environment (simple + Centos/Ubuntu + NovaDHCP)
 2. Add 1 controller node
 3. Start deployment
 4. Stop deployment when provisioning is in progress

Expected result:

 - slave node is rebooted and boots to bootstrap

Actual:

 - after reboot slave node still tries to boot from hard drive (see attached screenshot)

After some investigation I found that bootsector on HDD isn't cleared before restarting of slave node. It should be done by functions described here: https://github.com/stackforge/fuel-astute/blob/master/lib/astute/ssh_actions/ssh_erase_nodes.rb , but
it seems that they work incorrectly.

To make work erasing of hard drives during stopping of deployment on provisioning I've changed /usr/lib64/ruby/gems/2.1.0/gems/astute-0.0.2/lib/astute/ssh_actions/ssh_erase_nodes.rb script on master and restarted 'astute'. You can find the file I used to implement this workaround in attachments.