openstack
/
neutron
Code Issues Proposed changes

[Fullstack] SG test - first wait until all vms can ping each other 

In fullstack security group test, after fake VMs are created there
check if connectivity to some custom port is working fine with port
security disabled.
After that there is called "block_until_ping" method for each vms.

This patch changes that to first wait if we can ping vms and later do
netcat tests.
Even if that will not solve problems with failures of this test, we
may know more if the issue is caused by netcat or it's just no
ICMP connectivity between VMs at all.

Change-Id: Ie9e2170c761c9a10f3daa991c3fb77f304bb07e2
Related-Bug: #1742401
(cherry picked from commit 1e9a3bffd2)
This commit is contained in:
Slawek Kaplonski 2020-06-24 12:32:14 +02:00
parent e99d5b66ef
commit a4514641e9
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
neutron/tests/fullstack/test_securitygroup.py
View File

@ -133,15 +133,15 @@ class TestSecurityGroupsSameNetwork(BaseSecurityGroupsSameNetworkTest):
tenant_uuid, subnet_cidr)
# 0. check that traffic is allowed when port security is disabled
vms[0].block_until_ping(vms[1].ip)
vms[0].block_until_ping(vms[2].ip)
vms[1].block_until_ping(vms[2].ip)
self.assert_connection(
vms[1].namespace, vms[0].namespace, vms[0].ip, 3333,
net_helpers.NetcatTester.TCP)
self.assert_connection(
vms[2].namespace, vms[0].namespace, vms[0].ip, 3333,
net_helpers.NetcatTester.TCP)
vms[0].block_until_ping(vms[1].ip)
vms[0].block_until_ping(vms[2].ip)
vms[1].block_until_ping(vms[2].ip)
# Apply security groups to the ports
for port, sg in zip(ports, self.index_to_sg):