Add some randomness to CV publish break down #14782
Conversation
vsedmik
added
CherryPick
|
trigger: test-robottelo |
| ).create() | ||
| try: | ||
| publish_task = cv.publish(synchronous=False) | ||
| sleep_time = random.randint(0, 60) # publish takes ~70s in 6.15 and SatLab VM | ||
| sleep(sleep_time) | ||
| target_sat.power_control(state='reboot', ensure=True) |
There was a problem hiding this comment.
I wonder if mere f-m service restart wouldn't do the same job here (w/o need to spin AAP WF).
There was a problem hiding this comment.
Decided to test that out with parametrization (second commit).
|
PRT Result |
Satellite-QE
added
the
PRT-Passed
| ).create() | ||
| try: | ||
| publish_task = cv.publish(synchronous=False) | ||
| sleep_time = random.randint(0, 60) # publish takes ~70s in 6.15 and SatLab VM | ||
| sleep(sleep_time) |
There was a problem hiding this comment.
Are we able to use wait_for task here? so if publish takes longer than normal we can still pass through?
There was a problem hiding this comment.
Actually there is nothing specific we would wait for here. We just need to break the publish task in a random time while it's still running and check it succeeds later (after resume if needed). So we just need "wait for random time" which is done by that sleep.
There was a problem hiding this comment.
Looks good, Ack 👍
Pending comment from @ColeHiggins2's review.
|
trigger: test-robottelo |
|
PRT Result |
Satellite-QE
removed
the
PRT-Passed
|
trigger: test-robottelo |
|
PRT Result |
Satellite-QE
added
the
PRT-Failed
|
PRT Result |
| def test_positive_reboot_recover_cv_publish(target_sat, function_entitlement_manifest_org): | ||
| @pytest.mark.parametrize('reboot', [True, False], ids=['vm_reboot', 'fm_restart']) | ||
| def test_positive_reboot_recover_cv_publish( | ||
| target_sat, module_sca_manifest_org, module_big_repos, reboot |
There was a problem hiding this comment.
| target_sat, module_sca_manifest_org, module_big_repos, reboot | |
| module_target_sat, module_sca_manifest_org, module_big_repos, reboot |
module_big_repos is using module_target_sat and the test is using target_sat, seems like a mistake.
Same suggestion for the rest of the target_sat occurrences.
There was a problem hiding this comment.
Yeah, but I guess it won't help since we are a bit destructive here, so we get new Satellite for each test run here, which means we need new setup for each parametrized test which makes me cry and doubt if I really want to do the setup twice just to check the f-m restart.
There was a problem hiding this comment.
Okay so it seems it did help, PRT passed.
|
trigger: test-robottelo |
|
PRT Result |
Satellite-QE
added
PRT-Passed
* Add some randomness to CV publish break down * Parametrize for f-m service restart too (cherry picked from commit 869694e)
Add some randomness to CV publish break down (#14782) * Add some randomness to CV publish break down * Parametrize for f-m service restart too (cherry picked from commit 869694e) Co-authored-by: vsedmik <[email protected]>
* Add some randomness to CV publish break down * Parametrize for f-m service restart too
Problem Statement
In CV eval it was suggested to add some randomness into the CV publish interruption so that we cover different phases of the publish task and go more real-life.
Solution
Add some sleep and few other adjustments.
PRT test Cases example
trigger: test-robottelo
pytest: tests/foreman/destructive/test_contentview.py