MBR is not cleaned after node/cluster deletion

[Artem Panchenko]

Fuel version info (9.0 #250): http://paste.openstack.org/show/496473/

After node or whole cluster removal it still can boot from disk, so sometimes (when 1st PXE try fails due to STP forward delay) baremetal servers are marked as discovered, but actually they are booted into old target system. I added debug logging with `dd` command to the erase_node mcollective agent and got this:

"Running command 'dd if=/dev/zero of=sda bs=446 count=1 oflag=direct'"

(full output: http://paste.openstack.org/show/496472/)

As you can see value for `of=` arg is pasted incorrectly, '/dev/' is missed.

Steps to reproduce:

1. Deploy cluster
2. Remove cluster
3. Try to boot deleted node from HDD

Expected result:

node does't boot from HDD, MBR is clean

Actual result:

node boots from hard drive

Reproducibility:

always

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/314187

[OpenStack Infra (hudson-openstack)]

Change abandoned by Artem Panchenko (<email address hidden>) on branch: master
Review: https://review.openstack.org/314187
Reason: since https://review.openstack.org/#/c/309959/ is merged this patch is not needed anymore

[Artem Panchenko (apanchenko-8)]

Fixed by https://review.openstack.org/#/c/309959/

[Dmitry Pyzhov (dpyzhov)]

Backport is here: https://review.openstack.org/#/c/316704/

[Maksym Strukov (unbelll)]

Verified as fixed in 9.0-404