[system tests] eth1 interface with public network was down after all cluster nodes restart

[Andrey Sledzinskiy]

http://jenkins-product.srt.mirantis.net:8080/view/5.1_swarm/job/5.1_fuelmain.system_test.ubuntu.thread_3/26/testReport/junit/%28root%29/ceph_multinode_restart/ceph_multinode_restart/

Steps:
1. Create cluster - Ubuntu, simple, flat nova-network, 1 controller+ceph, 1 compute, 2 cinder+ceph
2. Deploy cluster
3. Restart all nodes

Actual - eth1 is down on compute node after restart (node-3)

Logs are attached

[Vasilios Tzanoudakis (vtzanoudakis)]

If you restart the node the interface is still down?

[Andrey Sledzinskiy (asledzinskiy)]

Didn't try it, issue is floating and doesn't occur very often

[Dmitry Borodaenko (angdraug)]

We don't know enough about this issue to mark it high priority and target it for a maintenance release. Marked as Incomplete and moved to 6.0.

[Andrey Sledzinskiy (asledzinskiy)]

Issue was again reproduced on 5.1.1 CI
Logs are attached
After reboot node-4 on which eth1 was down interface became Up

[Nastya Urlapova (aurlapova)]

Andrey, are you really sure that is qa issue?

[Nastya Urlapova (aurlapova)]

May be we have to ask our developers to help us with investigation.

[Andrey Sledzinskiy (asledzinskiy)]

It's likely not a qa issue but it's floating and is rarely reproduced. I'll ping someone to help with investigation next time I'll see it.