Fix connectivity issue during external bridge recreation test

In the faults test where we are testing if neutron-openvswitch-agent
can properly recover after deletion and recreation of the external
bridges, we are deleting all bridges which are in bridge-mappings for
the neutron-openvswitch-agent.
As we are doing that usually for an agent which runs on controller, we
may lost connectivity to the OpenStack api as well. Better to save ip
address to ping before the test to not to query API when it might be
not accessible.

Change-Id: Ieb9a59223639e3bbd997c6af748593d21cf8cba7
This commit is contained in:
Alex Katz 2020-11-04 18:13:56 +02:00 committed by Federico Ressi
parent a4354c5274
commit ad0881a25d
1 changed files with 4 additions and 3 deletions

View File

@ -75,7 +75,8 @@ class OpenvswitchTest(testtools.TestCase):
@undercloud.skip_if_missing_undercloud
def test_recreate_physical_bridge(self):
# Check if vm is reachable before test
ping.ping_until_received(self.stack.ip_address).assert_replied()
ip_add = self.stack.ip_address
ping.ping_until_received(ip_add).assert_replied()
network_l3_agents = neutron.list_l3_agent_hosting_routers(
self.router_id)
@ -86,6 +87,6 @@ class OpenvswitchTest(testtools.TestCase):
agent['host'],
ovs_agent['configurations']['bridge_mappings'].values())
ping.ping_until_unreceived(self.stack.ip_address).assert_not_replied()
ping.ping_until_unreceived(ip_add).assert_not_replied()
self._create_bridges()
ping.ping_until_received(self.stack.ip_address).assert_replied()
ping.ping_until_received(ip_add).assert_replied()