Fix failing registration test #17346
Conversation
shubhamsg199
added
No-CherryPick
| @@ -221,12 +221,13 @@ def test_positive_force_register_twice(module_ak_with_cv, module_org, rhel_conte | |||
| assert result.status == 0 | |||
| assert rhel_contenthost.subscribed | |||
| assert f'Unregistering from: {target_sat.hostname}' in str(result.stdout) | |||
| assert f'The registered system name is: {rhel_contenthost.hostname}' in str(result.stdout) | |||
| hostname = rhel_contenthost.execute('hostname').stdout.strip() | |||
There was a problem hiding this comment.
It looks like the unregistering of the host is not done properly and RHEL10 is only impacted here. I am guessing it an issue of RHEL10 rather than automation.
There was a problem hiding this comment.
Yeah, that was my initial guess but looking at the logs, it says unregistered.
There was a problem hiding this comment.
@shubhamsg199 @shweta83 Looking at the test failure in CI, when hostnamectl set-hostname {name} it never worked on < RHEL 10 since we use containers for this test and especially for RHEL10 we don't have any container images created yet it and it default checkout a VM where set-hostname works and this started to fail.
2025-01-16 04:41:22 - gw0 - broker - DEBUG - a33bb7c2198e executing command: hostnamectl set-hostname ZugqfcdVhq.example.com
2025-01-16 04:41:22 - gw0 - broker - DEBUG - a33bb7c2198e command result:
stdout:
stderr:
System has not been booted with systemd as init system (PID 1). Can't operate.
Failed to connect to bus: Host is down
status: 1
IMO this is expected behaviour, so if we're considering to fix it this way then we should consider using no_containers for this test, but before that I'd like to understand the rational behind changing the hostname step if is that really required for this test?
There was a problem hiding this comment.
Thanks for the explanation @Gauravtalreja1 , I think we should update the test and use no_containers
The reason behind changing the hostname step is because of the BZ attached which is the test requirement
There was a problem hiding this comment.
The test scenario of the bug was testing case-sensitiveness of the hostname. But as @Gauravtalreja1 said, the scenario was never tested due to containers.
The problem here is we are asserting rhel_contenthost.hostname with hostname which will always fail. We need to fix it and run the test on VM.
| reg_id_new = re.search(reg_id_pattern, result.stdout).group(1) | ||
| assert f'The system has been registered with ID: {reg_id_new}' in str(result.stdout) | ||
| assert reg_id_new != reg_id_old | ||
| assert ( | ||
| target_sat.cli.Host.info({'name': rhel_contenthost.hostname}, output_format='json')[ | ||
| target_sat.cli.Host.info({'name': hostname.lower()}, output_format='json')[ |
There was a problem hiding this comment.
If hostname is in uppercase before registration, then after registration does it get changed to lowercase ?
There was a problem hiding this comment.
then we don't need .lower()
There was a problem hiding this comment.
Out of curiousity, have we evaluated whether this is still a valid test? BZ 1361309 is very old, and the code that was changed to fix it [1] is no longer present in the same module upstream [2]. I was unable to find it elsewhere in the Katello codebase either.
[1] https://github.com/Katello/katello/pull/6214/files
[2] https://github.com/Katello/katello/blob/master/app/models/katello/host/subscription_facet.rb
Problem Statement
test_positive_force_register_twiceis failing as we are assertingrhel_contenthost.hostnamebut we are changing the hostname in the testSolution
Update the test to assert the updated hostname