action #33517
closed[functional][y][yast][sporadic][medium] yast2_control_center: possible conflict with PackageKit doing background refresh
0%
Description
Observation¶
openQA test in scenario opensuse-Tumbleweed-DVD-x86_64-yast2_gui@64bit fails in
yast2_control_center
Reproducible¶
Fails since (at least) Build 20180314
Expected result¶
Last good: 20180313 (or more recent)
Suggestions¶
Call pkcon_quit within test module.
As an alternative, we can explore if pressing yes is stable enough to fix in the test. (First option is still preferable)
Further details¶
Always latest result in this scenario: latest
either we explicitly terminate PK before hand or we ensure the pk refresh is completed (the test sometimes passes, most of the times fails).
Updated by okurz over 6 years ago
- Subject changed from yast2_control_center: possible conflict with PackageKit doing background refresh to [functional][y][yast]yast2_control_center: possible conflict with PackageKit doing background refresh
- Target version set to Milestone 16
Updated by okurz over 6 years ago
- Due date set to 2018-04-10
- Priority changed from Normal to Urgent
- Target version changed from Milestone 16 to Milestone 15
failing reproducibly every time since the mentioned build
Updated by dimstar over 6 years ago
failing reproducibly every time since the mentioned build
not really - whenever the PK conflict does not show up (and I see it) I can create one or two more needles. The failure always happens to be in the same module, but it is not always the exact same error
e.g. https://openqa.opensuse.org/tests/643136#step/yast2_control_center/20 failed on missing needles for the printer config
Updated by riafarov over 6 years ago
- Subject changed from [functional][y][yast]yast2_control_center: possible conflict with PackageKit doing background refresh to [functional][y][yast][sporadic] yast2_control_center: possible conflict with PackageKit doing background refresh
- Description updated (diff)
- Status changed from New to Workable
I see, in cases like these feel free to add the subject keyword "[sporadic]" so that no one states: "One job did not fail, is it solved?" ;)
Updated by cwh over 6 years ago
- Assignee set to cwh
So unconditionally calling pkcon quit
before starting sw_single should maybe solve that problem, shouldn't it?
Updated by okurz over 6 years ago
- Subject changed from [functional][y][yast][sporadic] yast2_control_center: possible conflict with PackageKit doing background refresh to [functional][y][yast][sporadic][medium] yast2_control_center: possible conflict with PackageKit doing background refresh
yes
Updated by cwh over 6 years ago
My first try to solve this by a simple "pkcon quit" was not successful:
http://boltzmann.suse.de/tests/43#step/yast2_control_center/2
Using this way we don't know how long PackageKit will take until it actually quits. So maybe it is better to go for the button-clicking solution.
Updated by okurz over 6 years ago
yes, I also do not have any better plan now. Should be ok to have a multi-tag assert_screen with if match_has_tag
. https://github.com/os-autoinst/os-autoinst-distri-opensuse/blob/master/tests/x11/kontact.pm can provide an example.
Updated by cwh over 6 years ago
- Status changed from In Progress to Feedback
The clicking approach works – using a wisely selected needle even with all the repeated dialogs that ask to try to quit again.
http://boltzmann.suse.de/tests/55
Here the PR for the code:
https://github.com/os-autoinst/os-autoinst-distri-opensuse/pull/4775
and the needle:
https://github.com/os-autoinst/os-autoinst-needles-opensuse/pull/349
Updated by cwh over 6 years ago
While continuing to update needles for that test I found two more cases when PackageKit can sporadically block that test:
- online-update
- online-repositories
I keep fixing them as they go. I guess I got them all, now.
http://boltzmann.suse.de/tests/65
Updated by cwh over 6 years ago
- Status changed from Resolved to Feedback
One more PR to fully fix this.
https://github.com/os-autoinst/os-autoinst-distri-opensuse/pull/4818
Updated by okurz over 6 years ago
- Due date changed from 2018-04-10 to 2018-05-08
- Target version changed from Milestone 15 to Milestone 16
The latest test run seems to be fine: https://openqa.opensuse.org/tests/662672
@cwh do you think we should resolve this one or do you think you want to gather more statistics?