[fuel-ci] fuel-library stable/5.0 doesn't pass ubuntu deployment test

[Aleksandra Fedorova]

https://fuel-jenkins.mirantis.com/job/5.0.fuel-library.ubuntu.deploy_ha_vlan/84/

Test has been triggered on merged code and failed at OSTF run:

{u'status': u'error', u'testset': u'smoke', u'name': u'Create volume and attach it to instance', u'duration': u'350 s.', u'step': None, u'taken': 0.00475192070007324, u'message': u'', u'id': u'fuel_health.tests.smoke.test_create_volume.VolumesTest.test_volume_create', u'description': u' Target component: Compute\n\n Scenario:\n 1. Create a new small-size volume.\n 2. Wait for volume status to become "available".\n 3. Check volume has correct name.\n 4. Create new instance.\n 5. Wait for "Active" status\n 6. Attach volume to an instance.\n 7. Check volume status is "in use".\n 8. Get information on the created volume by its id.\n 9. Detach volume from the instance.\n 10. Check volume has "available" status.\n 11. Delete volume.\n 12. Verify that volume deleted\n 13. Delete server.\n '},

The issue seems to be floating as seen from build history:
https://fuel-jenkins.mirantis.com/job/5.0.fuel-library.ubuntu.deploy_ha_vlan/

[Ryan Moe (rmoe)]

With access to a failed environment I found that compute nodes seem to be dying after/while launching VMs. I manually brought all nodes online and attempted to run OSTF smoke tests [0]. Test 9 failed and at almost the exact same time the compute node where those VMs were scheduled went away. There was no network connectivity and I wasn't able to access the host over VNC. The compute node VM was still running according to libvirt. I have also attached a diagnostic snapshot.

[0] http://paste.openstack.org/show/204907/

[Alexey Stupnikov (astupnikov)]

I have set bug's status to Won't fix for 5.0.x milestone since this bug has duplicate nominations to 5.0.x milestone