Comment 19 for bug 1840465

While performing validation, I was unable to reproduce the issue in bionic-rocky deployment. Inspecting the code, the fix was not present. I added logging, and confirmed that neutron is now including the key "security_group_rules" in the response, therefore the error is no longer triggered. I added logging to confirm that it was indeed the case, and got:

[Tue Nov 12 22:32:51.766942 2019] [wsgi:error] [pid 26153:tid 139930429884160] [remote 10.5.0.43:40358] sg: {'id': '40ca8baa-a3a5-412d-b955-7f9fb3e86e26', 'name': 'test3', 'tenant_id': 'a3ef2c1a2a8e4b628f592e0f87ef3c11', 'description': '', 'security_group_rules': [], 'tags': [], 'created_at': '2019-11-12T22:11:56Z', 'updated_at': '2019-11-12T22:12:03Z', 'revision_number': 3, 'project_id': 'a3ef2c1a2a8e4b628f592e0f87ef3c11'}

I also re-tested this with a limited user account, same results.

I proceeded to upgrade the package, confirmed the code has been changed and includes that fix. Ran the test case and saw that it does not cause any problems. Added additional logging to confirm the new condition (the fix) is not invoked as it is not necessary and indeed it was not being invoked.

I briefly checked neutron, python-neutronclient and horizon code, but I haven't been able to track down the change that caused the key 'security_group_rules' to no longer be missing.

I proceeded to reproduce the problem in xenial-queens to re-confirm the test case, and I couldn't.

Therefore my conclusion is that the test case needs to be improved, as it seems to be missing details that are required to reproduce the problem.