Project

General

Profile

Actions

action #134672

closed

[qe-core] test fails in consoletest_finish

Added by mgrifalconi 8 months ago. Updated 6 months ago.

Status:
Resolved
Priority:
High
Assignee:
Category:
Bugs in existing tests
Target version:
-
Start date:
2023-08-28
Due date:
% Done:

100%

Estimated time:
Difficulty:

Description

Observation

failing since 14d, reschedule of test did not help

openQA test in scenario sle-15-SP1-Server-DVD-Incidents-Minimal-s390x-qam-minimal-full@s390x-zVM-vswitch-l3 fails in
consoletest_finish

Test suite description

Testsuite maintained at https://gitlab.suse.de/qa-maintenance/qam-openqa-yml. minimal = base pattern, minimal (enhanced base) pattern are additional convenience paclkages

Reproducible

Fails since (at least) Build :30250:ca-certificates-mozilla

Expected result

Last good: :30232:dtb-aarch64 (or more recent)

Further details

Always latest result in this scenario: latest

Actions #1

Updated by szarate 8 months ago

The flow of the test is a bit strange,

it should not kill the Xserver all the time, only if it is detected, but we're restarting it at the same time, so we're def running into a race condition, IMO we should first:

Stop services, mask packagekit, restart the services but I'm starting to wonder if the purpose of this module is just to mask packagekit and nothing else... if so, we could figure out a way to reduce the waste a bit further even...

Actions #2

Updated by dzedro 8 months ago

consoletest_finish behaved always weird, I changed it lately to do reboot on 15-SP2+
To make it completely robust I could do reboot on all SLE15 versions, I will add also 15-SP1
https://github.com/os-autoinst/os-autoinst-distri-opensuse/pull/17495/commits/f05895d5c5c73995af31aee8596e4f3690b801ff

Actions #3

Updated by dzedro 8 months ago

  • Status changed from New to In Progress
  • Assignee set to dzedro
Actions #4

Updated by dzedro 8 months ago

szarate wrote in #note-1:

The flow of the test is a bit strange,

it should not kill the Xserver all the time, only if it is detected, but we're restarting it at the same time, so we're def running into a race condition, IMO we should first:

Stop services, mask packagekit, restart the services but I'm starting to wonder if the purpose of this module is just to mask packagekit and nothing else... if so, we could figure out a way to reduce the waste a bit further even...

It worked until something unrelated changed, I don't understand why some unrelated changes mess with the behavior.
IMO it got worse with never versions and systemd and and, can't just kill vnc anymore, it will leave processes behind and who know what will happen.
I think rebooting the system is cleanest approach.

Actions #5

Updated by dzedro 8 months ago

  • Status changed from In Progress to Feedback
  • % Done changed from 0 to 100
Actions #6

Updated by dzedro 6 months ago

  • Status changed from Feedback to Resolved
Actions

Also available in: Atom PDF