[sle] test fails in installation: Grub installation failed
|Category:||Bugs in existing tests|
- Assignee set to leli
Actually all sle11sp4 host installation are failed the same, see https://openqa.nue.suse.com/tests/2493463.
According to history jobs, there were successful run seven days ago, Feb 19. I checked code commit history, following two commits are related:¶
Author: Michal Nowak email@example.com
Date: Wed Feb 20 10:23:19 2019 +0100
Replace wait_boot_on_local_disk Regression from a42ed584bd400b0a37b6cb6df01b124cbb222edc. Fails here: * https://openqa.suse.de/tests/2478650#step/boot_to_desktop/3 * https://openqa.suse.de/tests/overview?distri=sle&version=15-SP1&build=173.1&groupid=111 Validation run: http://nilgiri.suse.cz/tests/300
Author: lemon.li firstname.lastname@example.org
Date: Wed Feb 20 10:55:13 2019 +0800
Boot from local disk on installation on aarch64
Since michal already fixed one regression by a42ed584bd400b0a37b6cb6df01b124cbb222edc, root cause is still there.
Lemon, would you please help to fix this blocking issue for virtualization beta4 milestone test on sle11sp4 host autoyast installation?
- Assignee changed from leli to xlai
From the log:
[2019-02-22T13:58:48.711 CET] [debug] >>> testapi::_handle_found_needle: found reboot-after-installation-by-autoyast-sle11sp4-20170724, similarity 1.00 @ 121/353
[2019-02-22T13:58:48.734 CET] [debug] IPMI:
[2019-02-22T13:58:48.735 CET] [debug] /var/lib/openqa/cache/openqa.suse.de/tests/sle/tests/autoyast/installation.pm:252 called opensusebasetest::wait_boot
[2019-02-22T13:58:48.735 CET] [debug] <<< testapi::select_console(testapi_console='sol', await_console=0)
'hostname' => 'localhost',
'ikvm' => 0,
'port' => 44770
XIO: fatal IO error 11 (Resource temporarily unavailable) on X server ":48227"
after 15738 requests (15581 known processed) with 0 events remaining.
xterm: fatal IO error 11 (Resource temporarily unavailable) or KillClient on X server ":48227"
[2019-02-22T13:58:49.750 CET] [debug] Driver backend collected unknown process with pid 199683 and exit status: 1
[2019-02-22T13:58:49.751 CET] [debug] Driver backend collected unknown process with pid 199685 and exit status: 84
[2019-02-22T13:58:49.751 CET] [debug] Driver backend collected unknown process with pid 199695 and exit status: 0
[2019-02-22T13:58:49.878 CET] [debug] Connected to Xvnc - PID 205345
icewm PID is 205374
I compared the failed log with previous successful log. The failed job failed for it waste time on the IO error, so it can't catch the needle 'reboot-after-installation' any more.
What I think:
1. I can't find where show it is related with my push.
2. First to check the IO error.
3. I guess there is something not stable for the test flow, it expected to catch the second 'reboot-after-installation' (stage 2) in very short time after catch the first 'reboot-after-installation' (stage 1). ( From video, we can see the system will go from needle 'reboot-after-installation' to 'grub' very quickly).
I checked code, again. This should be not related to lemon, sorry :(.
From git log tests/autoyast/installation.pm, seems introduced by :
Author: Jan Baier email@example.com
Date: Mon Feb 18 19:07:25 2019 +0100
Utilize autoyast for baremetal testing Speed up installation over IPMI by utilizing autoyast installation procedure with QAM profiles. This should also fix issue with QR images over svirt backend for s390x architecture.
I submit a PR to fix it.
but my local verification has not finished. I'll update the status.