action #170323
closed[openQA][infra][http(s)] Many baremetal SUT machine fails to get autoyast profile from worker machine via HTTP(s) Server returned code 404
0%
Description
Observation¶
Many test runs failed with new Build42.3 due to
An error occurred while fetching the profile: Cannot find URL 'http://worker33.oqa.prg2.suse.org:20163/Dsf7Y9CYz27DGa1t/files/bare-metal1.oqa.prg2.suse.orgvirt_autotest/host_unattended_installation_files/autoyast/dev_host_15.xml' via protocol HTTP(S). Server returned code 404
Please refer to screenshot as below:
and
https://openqa.suse.de/tests/16024606#step/installation/4
https://openqa.suse.de/tests/16024607#step/installation/4
https://openqa.suse.de/tests/16022849#step/installation/4
https://openqa.suse.de/tests/16022859#step/installation/4
https://openqa.suse.de/tests/16022854#step/installation/4
My very preliminary observation is worker33 and worker34 have such issue. Worker35 might not have such issue.
Steps to reproduce¶
- Prepare ipxe script for baremetal host installation
- Reboot and boot from ipxe
- Installation starts and tries to fetch autoyast from worker machine
Impact¶
All tests replying on ipxe host installation can not pass this step
Problem¶
- HTTP service on worker machine
- Worker process on worker machine
- Merge request https://gitlab.suse.de/openqa/salt-pillars-openqa/-/merge_requests/937/diffs
- Communication between baremetal SUT and worker machines
Suggestions¶
- Check worker machine state
- Check service state on worker machine
- Check worker process on worker machine
- Check communication between baremetal SUT and worker machines
- Check merge request https://gitlab.suse.de/openqa/salt-pillars-openqa/-/merge_requests/937/diffs
Workaround¶
n/a
Files
Updated by waynechen55 15 days ago
- Subject changed from [openQA][infra][http(s)] Baremetal SUT machine fails to get autoyast profile from worker machine via HTTP(s) Server returned code 404 to [openQA][infra][http(s)] Many baremetal SUT machine fails to get autoyast profile from worker machine via HTTP(s) Server returned code 404
Updated by dheidler 14 days ago · Edited
- File clipboard-202411271356-5ctwk.png clipboard-202411271356-5ctwk.png added
- File clipboard-202411271357-zpumh.png clipboard-202411271357-zpumh.png added
This is also seems to be the same issue as jobs like this one: https://openqa.suse.de/tests/16026151
Even though the screenshots look like as if the wrong PXE menu was starting up
the video shows that the chainloaded ipxe script from baremetal-support.qe.prg2.suse.org can't load the kernel.
The blue default ipxe menu is then shown because the firmware fails over to the 2nd NIC.
Updated by dheidler 13 days ago · Edited
- Status changed from New to In Progress
- Assignee set to dheidler
This was most likely caused due to infra not actually deploying https://gitlab.suse.de/OPS-Service/salt/-/merge_requests/5789 when merging it despite stating otherwiese.
See https://suse.slack.com/archives/C02AJ1E568M/p1732709040320289
This was causing old configs being serverd from baremetal-support server in nue2 instead of the instance in prg2.
This should be fixed now as I manually ran salt-call state.apply on suttner1 and suttner2.