Tempest DVR HA multimode tests fails due to no FIP connectivity

[Slawek Kaplonski]

I checked results of neutron-tempest-dvr-ha-multinode-full for about 50 patches and I found that quite common reason of failures is problem with SSH to instance via Floating IP address.

Example of such failed tests are:

http://logs.openstack.org/65/552065/1/check/neutron-tempest-dvr-ha-multinode-full/cac8ec8/logs/testr_results.html.gz

http://logs.openstack.org/97/552097/1/check/neutron-tempest-dvr-ha-multinode-full/47364b0/logs/testr_results.html.gz

http://logs.openstack.org/51/414251/57/check/neutron-tempest-dvr-ha-multinode-full/e7ff316/logs/testr_results.html.gz

http://logs.openstack.org/78/552678/1/check/neutron-tempest-dvr-ha-multinode-full/5ed008a/logs/testr_results.html.gz

http://logs.openstack.org/81/551981/3/check/neutron-tempest-dvr-ha-multinode-full/09af070/logs/testr_results.html.gz

[Brian Haley (brian-haley)]

In the first log there I see this in the console log:

=== network info ===
if-info: lo,up,127.0.0.1,8,::1
if-info: eth0,up,,8,fe80::f816:3eff:fe89:fad4

So the instance never got an IP on eth0.

Doesn't really narrow it down since it could be a DVR flow issue, or the DHCP agent, or ???

[Brian Haley (brian-haley)]

Checked the other logs, all fail DHCP as above or:

Starting network...
udhcpc (v1.20.1) started
Sending discover...
Sending discover...
Sending discover...
Usage: /sbin/cirros-dhcpc <up|down>
No lease, failing