action #130141
closed
test fails in podman_netavark on SLEM on Azure
Added by ph03nix about 1 year ago.
Updated about 1 year ago.
Description
Observation¶
openQA test in scenario sle-micro-5.4-BYOS-Azure-x86_64-slem_containers@64bit fails in
podman_netavark
openQA test in scenario sle-micro-5.4-BYOS-Azure-ARM-x86_64-slem_containers@64bit fails in
podman_netavark
The issue is present on Azure and Azure-ARM, but not on GCE or AWS.
Test suite description¶
The base test suite is used for job templates defined in YAML documents. It has no settings of its own.
Reproducible¶
Fails since (at least) Build 0114 (current job)
Expected result¶
Last good: 0113 (or more recent)
Further details¶
Always latest result in this scenario: latest
Files
- Related to action #124631: Usage of netavark and aardvark in podman networks added
- Description updated (diff)
https://openqa.suse.de/tests/11210668#step/podman_netavark/82 looks like an issue with establishing the ssh tunnel (password required) but I'm not sure.
I could not reproduce any issues with a test instance after installing netavark and aardvark-dns in one transaction:
zoidberg:~ # transactional-update pkg install netavark aardvark-dns
...
Please reboot your machine to activate the changes and avoid data loss.
New default snapshot is #4 (/.snapshots/4/snapshot).
transactional-update finished
zoidberg:~ # reboot
Instance works just as expected.
For some reason in the failing step the user bernhard does not exist anymore. The ssh config assumes the user to be present.
This is weird. The user bernhard is created and exists after the slem_prepare step:
Compare this to the screenshot of podman-netavark, where bernhard does not exist anymore:
There is something fishy going on here.
jlausuch wrote:
Azure and its own peculiarities... what a surprise...
I smell something fishy here, but will need some time to investigate. Today with all the meetings it's impossible.
The problem is that toolbox removes the testapi user bernhard after running, even if other tests (e.g. PC test runs) still need it. This is a test issue.
- % Done changed from 0 to 70
Restarted all affected job runs:
- Status changed from In Progress to Resolved
- % Done changed from 70 to 100
Also available in: Atom
PDF