The provision stage fails due to '#<RuntimeError: Could not find any hosts in discovery data provided>'

[Sergey Novikov]

Detailed bug description:
the issue was found by
https://product-ci.infra.mirantis.net/job/9.x.system_test.ubuntu.repetitive_restart/130/testReport/(root)/ceph_partitions_repetitive_cold_restart/ceph_partitions_repetitive_cold_restart/

https://product-ci.infra.mirantis.net/job/9.x.system_test.ubuntu.error_node_reinstallation/131/testReport/(root)/reinstall_failed_compute_deployment/reinstall_failed_compute_deployment/

Steps to reproduce:
            1. Enable kernel v4.4 http://paste.openstack.org/show/590002/
            2. Create a cluster
            3. Add 3 nodes with controller and mongo roles
            4. Add a node with compute and cinder roles
            5. Deploy
Expected results: The deployment passes

Actual result: The deployment fails on provisioning with "Error running provisioning: #<RuntimeError: Could not find any hosts in discovery data provided>"

more details from astute's log http://paste.openstack.org/show/590001/

Reproducibility:
rare

Description of the environment:
snapshot #537

[Vladimir Kuklin (vkuklin)]

It seems that the issue happened due to the fact that nodes went offline because they did not have necessary kernel modules installed. This could happen due to an incorrect list of packages being specified for the provisioning phase with the script provided in the bug description. I would suggest firstly to check if provisioning of at least one node succeeds and also provide the logs of the node rebooted properly and was able to ping the master node. Moreover, the test cases provided have been passing through provisioning stage for over a week. I am marking this bug as incomplete until there is an environment to debug.

[Oleksiy Molchanov (omolchanov)]

Marking as Invalid, because of no activity for more than a month.