We need to cleanup list of nodes after we will add non-ubuntu compute nodes manually

[Timur Nurlygayanov]

Steps To Reproduce:
1. Deploy OpenStack cloud with fuel
2. Deploy additional non-ubuntu compute nodes
3. Check "fuel nodes" output on fuel master node:

fule nodes

Expected Result:
we will see only actual information about OpenStack nodes, and we will not see information about failed / offline compute nodes.

Observed Result:
we can see many failed nodes:
[root@nailgun ~]# fuel nodes
id | status | name | cluster | ip | mac | roles | pending_roles | online | group_id
---|----------|---------------------|---------|-------------|-------------------|------------|---------------|--------|---------
3 | ready | slave-03_controller | 1 | 10.109.5.5 | 64:68:70:a2:cc:d3 | controller | | True | 1
9 | discover | Untitled (d1:df) | None | 10.109.5.10 | 64:b1:3b:4d:d1:df | | | True | None
5 | ready | slave-04_compute | 1 | 10.109.5.6 | 64:b7:54:dc:82:68 | compute | | False | 1
4 | ready | slave-01_controller | 1 | 10.109.5.3 | 64:89:3f:c6:bf:ed | controller | | True | 1
2 | discover | Untitled (b6:65) | None | 10.109.5.7 | 64:8b:46:9b:b6:65 | | | False | None
1 | ready | slave-02_controller | 1 | 10.109.5.4 | 64:d8:2c:aa:39:0a | controller | | True | 1
7 | discover | Untitled (41:c5) | None | 10.109.5.8 | 64:8d:19:3a:41:c5 | | | True | None
8 | discover | Untitled (88:c7) | None | 10.109.5.11 | 64:a7:e4:47:88:c7 | | | True | None
6 | discover | Untitled (f6:b6) | None | 10.109.5.9 | 64:e8:12:4d:f6:b6 | | | True | None

[Ivan Berezovskiy (iberezovskiy)]

If we are speaking about our tests for non-ubuntu computes, we add this ubuntu compute node to be able to run OSTF tests later, because non-ubuntu computes aren't connected to Fuel, so Fuel thinks that compute nodes amount is 0. For real environments we don't need to use this workaround.