Actions
action #162272
closed[qe-tools][RPi3] test fails in prepare_firstboot - use retries with exponential back-off in os-autoinst size:S
Status:
Resolved
Priority:
High
Assignee:
Category:
Feature requests
Target version:
Start date:
2024-06-14
Due date:
% Done:
0%
Estimated time:
Tags:
Description
Observation¶
I did a manual installation on my RPi 3 without any issue. It looks like a setup issue on openQA.
openQA test in scenario sle-15-SP6-JeOS-for-RaspberryPi-aarch64-jeos-realhw-RPi@RPi3B fails in
prepare_firstboot
Test suite description¶
Run a test on real RPi hardware. To be used eg. with RPi4 machine.
Setup: https://confluence.suse.com/pages/viewpage.action?spaceKey=~dheidler&title=Hardware+Automation
Reproducible¶
Fails since (at least) Build 6.75 (current job)
Expected result¶
Last good: 6.74 (or more recent)
Suggestions¶
- Look into what was found in #158404 which was about the last time this test failed ~1 month ago
- This looks to be failing sporadically only? Try and reproduce it - take care to only run first_reboot to avoid reflashing the disk
- Meanwhile the scenario is ok again but we want to improve. The call
select_console('root-ssh')
establishes an ssh connection within the os-autoinst backend and as visible in https://openqa.suse.de/tests/14266260/logfile?filename=autoinst-log.txt there is already retrying but seemingly only 5 times up to 1m, not longer. We should ensure that there are longer retries with exponential backoff up to let's say 30m - Add exponential retry in https://github.com/os-autoinst/os-autoinst/blob/master/backend/baseclass.pm#L1211
- Consider using a higher value than 5 for the specific test scenario or machine related to RPi
Further details¶
Always latest result in this scenario: latest
Updated by okurz 5 months ago
- Tags changed from bugbusters, reactive work, infra, rpi to bugbusters, reactive work, rpi
- Subject changed from [qe-tools][RPi3] test fails in prepare_firstboot size:S to [qe-tools][RPi3] test fails in prepare_firstboot - use retries with exponential back-off in os-autoinst size:S
- Category changed from Regressions/Crashes to Feature requests
Updated by livdywan 5 months ago
- Status changed from Workable to In Progress
- Assignee set to livdywan
- Meanwhile the scenario is ok again but we want to improve. The call
select_console('root-ssh')
establishes an ssh connection within the os-autoinst backend and as visible in https://openqa.suse.de/tests/14266260/logfile?filename=autoinst-log.txt there is already retrying but seemingly only 5 times up to 1m, not longer. We should ensure that there are longer retries with exponential backoff up to let's say 30m- Add exponential retry in https://github.com/os-autoinst/os-autoinst/blob/master/backend/baseclass.pm#L1211
- Consider using a higher value than 5 for the specific test scenario or machine related to RPi
I'm looking into the exponential retry on the backend side.
Updated by livdywan 5 months ago
- Consider using a higher value than 5 for the specific test scenario or machine related to RPi
https://gitlab.suse.de/qac/qac-openqa-yaml/-/merge_requests/1703
Actions