[6.14.z] Use RHEL8 for container contenthost

[Satellite-QE]

Cherrypick of PR: #15179

Problem Statement

Some of our tests are failing to pull docker images with OCI artifacts using obsolete version of docker client provided by RHEL7 extras repo.
(see BZ#2262793)

Solution

Let's move forward to RHEL8 and newer version of podman/docker.
Proposed changes in this PR:

1. Use RHEL8 with newer podman/docker client.
2. Rescope container_contenthost fixture to module. This speeds up current tests from ~45 minutes to ~15 minutes. The downside is we need to ensure podman logout is performed on teardown (via request.addfinalizer).

Related Issues

With the new podman new issue was identified (which has not been hit with the old docker client): https://issues.redhat.com/browse/SAT-25328
Since the fix PR is up already, let's test it with packit.

PRT test Cases example

trigger: test-robottelo
pytest: tests/foreman/cli/test_container_management.py
Katello:
    katello: 11006

[Satellite-QE]

trigger: test-robottelo
pytest: tests/foreman/cli/test_container_management.py
nailgun: 1151
Katello:
    katello: 11006

[Satellite-QE]

PRT Result

Build Number: 7138
Build Status: UNSTABLE
PRT Comment: pytest tests/foreman/cli/test_container_management.py --external-logging
Test Result : ============ 1 failed, 3 passed, 54 warnings in 1441.37s (0:24:01) =============

[shweta83]

@vsedmik Please check the failure and rebase the PR.

[vsedmik]

The PRT failure seems to be directly linked to the search issue mentioned in the PR description. So I wonder if the packit setup failed for z-stream (while it worked well in master where the packit/PRT passed) since the Katello fix PR raised against Katello master might fail to be used at z-stream Satellite? Maybe @Griffin-Sullivan could say if it's possible.

Nevertheless, the failing test is expected to keep failing until the fix is merged, backported and delivered downstream. It is not directly related to this PR (apart from the fact that this PR revealed it).