Gate jobs are failing in different ways related to network issues

[Gabriele Cerami]

gate queue keeps resetting for problems not directly related to tripleo, that appear to be random but related with the network.

reset at 12 UTC caused by

http://logs.openstack.org/82/562282/8/gate/tripleo-ci-centos-7-containers-multinode/07ab872/job-output.txt.gz#_2018-04-25_09_28_03_833396

queue is now going to reset with failures in various patches, on different job for a different reason than before,

http://logs.openstack.org/43/563843/3/gate/tripleo-ci-centos-7-undercloud-containers/462364e/job-output.txt.gz#_2018-04-25_13_09_18_725041

http://logs.openstack.org/96/564096/1/gate/tripleo-ci-centos-7-containers-multinode/4e4f6f9/job-output.txt.gz#_2018-04-25_12_41_08_799021

Same error appeared in two different providers too

This is blocking also an important patch that is needed to fix promotion.

[Quique Llorente (quiquell)]

Comparing vxlan_networking log in both kind of jobs, they don't appear to do the same

The one with vxlan_network ssh problem:
http://logs.openstack.org/43/563843/3/gate/tripleo-ci-centos-7-undercloud-containers/462364e/logs/undercloud/home/zuul/vxlan_networking.sh.log.txt.gz

The one with vxlan_networking passing:
http://logs.openstack.org/82/562282/8/gate/tripleo-ci-centos-7-containers-
multinode/07ab872/logs/undercloud/home/zuul/vxlan_networking.sh.log.txt.gz

Also in the last one there is a
Failed to get unit file state for neutron-openvswitch-agent.service: No such file or directory
2018-04-25 13:08:14 | + '[' '' = enabled ']' doesn't look good.

[Gabriele Cerami (gcerami)]

Randomly arrived. Randomly left. Zuul queue has been under 7 hours for the past 2 days. Marking as invalid