QA » openQA Project » openQA Tests

https://bugzilla.suse.com/show_bug.cgi?id=1062771 is a ship stopper bug for SLE15 and we want to help with investigation.

We want to test if the current KOTD which should contain a fix (see bug comments) will prevent this problem.

Currently trying to add the kernel repo as an "addon" during installation of an openQA job:

openqa_clone_job_osd 1589683 _GROUP=0 TEST=okurz_bsc1062771_btrfs ADDONURL=kernel ADDONURL_KERNEL=http://download.suse.de/ibs/Devel:/Kernel:/SLE15/standard/

Created job #1591961: sle-15-Installer-DVD-s390x-Build543.1-btrfs@s390x-zVM-ctc -> https://openqa.suse.de/t1591961

If this test shows that the updated kernel was used at least in the installed system (not in the installer, of course) then we should reconduct more to hit the sporadic bug or ensure that we can't hit it anymore.

If the above does not work we can try the following alternatives:

• Create a DUD including the kernel and potentially more needed modules
• Create a new ISO - was it "mkiso-something"?

So
http://lord.arch.suse.de/tests/744/file/autoinst-log.txt
showed that an updated kernel was used and the original problem was reproduced.

For the test to work I needed two needles "addon_products_sle-addon-products-kernel-20180406" and "import-untrusted-gpg-key-devel-kernel-20180406" which I also pushed to master.

I commented on the bug, let's say what jeffm says.

I retriggered on production anyway because it might be helpful to show it there as well: https://openqa.suse.de/tests/1596405#live

So somehow the test result was not exactly clear. My test showed that the problem would not be fixed but a newer build with a newer kernel showed no problems anymore anyway. Maybe the older kernel in the installer did something to the filesystem before rebooting into the newer kernel? In any way, I commented back on the bug and suggested to have it fixed.