action #167383
closedcoordination #169270: [SLE-Micro][epic] Establish ppc64 test runs for SLEM6
test fails in disk_boot for toolbox
Added by mdati 3 months ago. Updated about 2 months ago.
100%
Description
Observation¶
openQA test in scenario sle-micro-6.1-Container-Image-Updates-x86_64-sle_micro_toolbox_image@64bit repeatedly failed in
disk_boot, see previuos runs.
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 6.1_3.6
Expected result¶
Last good: 6.1_3.5 (or more recent)
Further details¶
Always latest result in this scenario: latest
Note¶
On boot menu, key enter, to select first item, is missing.
Files
| 3-boot_ERRORc.png (12.3 KB) 3-boot_ERRORc.png | Error device missing /dev/disk/* | mdati, 2024-10-03 08:28 |
Updated by mdati 3 months ago
- Tags changed from firstboot to firstboot, toolbox
- Project changed from openQA Tests (public) to Containers and images
- Category deleted (
Bugs in existing tests)
Note that, I paused the test for debugging in VNC, manual key return entered, then test run resumed: the test continued normally and passed: https://openqa.suse.de/tests/15527898.
But on next rerun it will fail, missing the key enter selection automation.
Updated by mdati 3 months ago · Edited
- Tags changed from firstboot, toolbox to toolbox, boot
- Subject changed from test fails in disk_boot for toolbox x86 to test fails in disk_boot for toolbox
In group SLE Micro Toolbox updates build 6.1_3.7,
the test on x86_64 then passed next run, a clone, paused and manually selected the right key;
on ppc64le still fail on [disk_boot][(https://openqa.suse.de/tests/15547875#step/disk_boot/2).
Updated by mdati 3 months ago
- Related to action #162689: [sle-micro 6.1] add ppc64le image into testing queue added
Updated by mdati 3 months ago · Edited
- File 3-boot_ERRORc.png 3-boot_ERRORc.png added
After some investigation, it resulted that, with KEEP_GRUB_TIMEOUT=1, in [disk_boot] we skip grub checks and on boot menu it remains waiting for (manual) key-ret; but setting =0 the boot proceeds, on grub too.
MR 1844 of jlausuch, to change it in the template, merged.
But, another issue appeared during boot phase (VR 15581119), a device missing: 
Due to that, the system returns to the first boot menu, in loop: so, we still shall fix it, to let module pass.
Updated by mdati 3 months ago
- Related to deleted (action #162689: [sle-micro 6.1] add ppc64le image into testing queue)
Updated by mdati 3 months ago
- Blocks action #162689: [sle-micro 6.1] add ppc64le image into testing queue added
Updated by mdati 3 months ago · Edited
The problem may be is in the SLEM 6.1 ppc64le [HDD_1 creation phase] or in the raw image itself.
All other architectures HDD_1 are ok, tests pass in SLE-M Toolbox group
Updated by mdati 3 months ago · Edited
- Status changed from Workable to In Progress
- Assignee set to mdati
This Issue in Toolbox sle-m 6.1 ppc64le was noted and multi-hands managed since last Sep-17; threads:
https://suse.slack.com/archives/C02CGKBCGT1/p1726580157444859
https://suse.slack.com/archives/C02CGKBCGT1/p1726658367118569
https://suse.slack.com/archives/C02CGKBCGT1/p1727358030039419
https://suse.slack.com/archives/C02CGKBCGT1/p1727863595800619
Here last code updates (merge requests):
https://gitlab.suse.de/qac/qac-openqa-yaml/-/merge_requests/1821
https://gitlab.suse.de/qac/qac-openqa-yaml/-/merge_requests/1823
https://github.com/os-autoinst/os-autoinst-distri-opensuse/pull/20230
https://gitlab.suse.de/qac/qac-openqa-yaml/-/merge_requests/1824
https://gitlab.suse.de/qac/aceto-magico/-/merge_requests/19
https://gitlab.suse.de/qac/qac-openqa-yaml/-/merge_requests/1825
https://gitlab.suse.de/qac/qac-openqa-yaml/-/merge_requests/1826
https://gitlab.suse.de/qac/qac-openqa-yaml/-/merge_requests/1839
https://gitlab.suse.de/qac/qac-openqa-yaml/-/merge_requests/1844
To date the issue status is:
- Create bootable a sle-m 6.1 ppc64le
qcow2image is published ok inhdd/, (raw.xzimages instead, created in assets,but don't moved in hdd), BUT - in SLE Micro Toolbox updates, fails always for
ppc64lebecause that published image contains some config. error preventing it from booting. - Therefore Container Host flavors
-ppc-,podmantests fail (see note 4. - The
dockertests pass instead, havingno toolboxto run.
Updated by ph03nix 3 months ago
I see that this is still failing in https://openqa.suse.de/tests/15616178#step/disk_boot/13 with "no suitable video mode found". This test run uses the SL-Micro.ppc64le-6.1-Default-Updated.qcow2 HDD image produced in https://openqa.suse.de/tests/15618478, which works fine. I wonder were the culprit is?
@mdati do you know more?
Updated by mdati 3 months ago · Edited
kind reply to https://progress.opensuse.org/issues/167383#note-10:
No, https://openqa.suse.de/tests/15618478, does NOT work fine: the HDD is only published ok.
As already explained in https://progress.opensuse.org/issues/167383#note-4, when that new HDD is used in Container-Image-Updates-ppc64le it fails in this step, that is , but the error no suitable video mode found is "minor", not the blocking one, and self recovered: in fact the expected Grub menu NEWLY reappears and stop there, waiting for a next key-enter, never more occurring.
That minor error I saw occurred also in images that booted ok.
It is a loop:
- test start with GRUB Menu,
- auto key-enter,
- boot phase proceeds, but then... Error
device missinghappens, - system fall back to GRUB menu, the minor error occurrs,
- but then back GRUB Menu appears.
If you here re-click enter, that loop rerun.
Updated by mdati 3 months ago · Edited
Now, next step:
After I executed many test in gr. 451 on sle-m 6.1 ppc64le, publishing as qcow2 as raw.xz or using other images in test, but always a unbootable image published.
In order try to resume testing podman in Container Hosts group 513 and unblock poo #162689, as W.A. I'm (temporary) skipping that new qcow2 hdd creation and working directly on the ppc64 raw.xz images in the HDD_1, https://openqa.suse.de/tests/15647678 , but with proper pre-updates:
(a) I added CONTAINER_IMAGE_TO_TEST 6.1 to update the toolbox default 5.5, then
(b) I triggered the install_updates, (adding "-QR" to the flavor name), to add missiing certificates for pull registry.
But then a new issues appeared in image_podman https://openqa.suse.de/tests/15641792#step/image_podman/114 validation and I am testing some code changes, in https://openqa.suse.de/tests/15647678.
Updated by mdati 2 months ago · Edited
See note in resumed poo#162689: https://progress.opensuse.org/issues/162689#note-28 WIP; to update certificates in item (b), a different solution used than note-12 above.
Updated by ph03nix 2 months ago
This ticket can IMHO be closed in favor of https://bugzilla.suse.com/show_bug.cgi?id=1227509.
Updated by mdati 2 months ago · Edited
This issue was produced by some settings in HDD creation of tests gr.377, that passed, but made the ppc image no more bootable in test of gr.451.
Felix found the problematic settings is HDDSIZEGB and created MR https://gitlab.suse.de/qac/qac-openqa-yaml/-/merge_requests/1862 to remove it; Merged ok.
Also VR were done and passed ok, confirming parameters are valid: https://openqa.suse.de/tests/15702362
Updated by mdati 2 months ago · Edited
Last SLE Micro Toolbox updates build 6.1_3.16 failed due to problematic settings.
But Felix MR https://gitlab.suse.de/qac/qac-openqa-yaml/-/merge_requests/1874 fixed the issue HDDSIZEGB_1=''.
Actually toolbox tests pass ok for all the 4 achitectures
Updated by mdati about 2 months ago
- Status changed from Feedback to Resolved
- % Done changed from 0 to 100