Project

General

Profile

Actions

action #31702

closed

[sle][migration][sle15][aarch64] test fails in reboot_gnome - match=boot-firmware timed out after 30

Added by qmsu about 6 years ago. Updated about 6 years ago.

Status:
Resolved
Priority:
Normal
Assignee:
Category:
Bugs in existing tests
Target version:
-
Start date:
2018-02-13
Due date:
% Done:

100%

Estimated time:
Difficulty:

Description

Observation

openQA test in scenario sle-15-Installer-DVD-aarch64-media_upgrade_sles12sp3@aarch64 fails in
reboot_gnome

Reproducible

Fails since (at least) Build 414.3

Expected result

Last good: (unknown) (or more recent)

Further details

Always latest result in this scenario: latest


Related issues 1 (0 open1 closed)

Related to openQA Tests - action #34999: [sle][migration][sle15][aarch64] test fails in grub_test_snapshot - need workaround to make uefi boot into harddiskResolvedqmsu2018-04-16

Actions
Actions #1

Updated by leli about 6 years ago

There is no needle with tag of 'boot-firmware'. We should add the needle to fix this issue.

linux-zyts:/var/lib/openqa/share/tests/sle/products/sle/needles # grep -Rni 'boot-firmware'
linux-zyts:/var/lib/openqa/share/tests/sle/products/sle/needles #

Actions #2

Updated by okurz about 6 years ago

This needle tag has been included with

commit 4631bd4e
Author: Jozef Pupava <jpupava@suse.com>
Date:   Thu Jan 26 09:06:13 2017 +0100

    Workaround boot HDD on aarch64

    It will still not boot because it's broken but at least it will be more obvious
    https://bugzilla.suse.com/show_bug.cgi?id=1022064

to detect the bug situation and then record a soft failure. However, the bug is marked as VERIFIED WONTFIX so I think the code block should not be there at all. At least the record_soft_failure should not be there. Maybe it should be replaced by a die-message with a helpful content pointing to the right direction, e.g. wrong test settings.

Compare to a standard aarch64-installation and test job: https://openqa.suse.de/tests/1558430#settings . This job differs because there is no setting BOOTFROM=d nor BOOT_HDD_IMAGE=1 as there is no system installed it boots from the DVD. Maybe it would help to include the same steps as for the image based tests, e.g. extra_tests_on_gnome

Actions #3

Updated by dehai about 6 years ago

  • Priority changed from Normal to High
Actions #4

Updated by leli about 6 years ago

  • Assignee set to leli
  • Priority changed from High to Normal

I will try to reproduce it on my own X86 machine and find the root cause.

Actions #5

Updated by leli about 6 years ago

In the function boot_local_disk:

if (check_var('ARCH', 'aarch64') and get_var('UEFI')) {
assert_screen 'boot-firmware';
}

I checked these failed aarch64 testcases, all with 'UEFI' = 1. So it failed for entered wrong process. I will set the 'UEFI' = 0 after upgrade to SLES15 finished and verify it on my own machine.

Actions #6

Updated by okurz about 6 years ago

that's weird because aarch64 by default always sets the flag UEFI=1. Also see the machine definition on https://openqa.suse.de/admin/machines . But it is a good idea to try out if the other branch works better by setting the flag to off once if you can.

Actions #7

Updated by okurz about 6 years ago

This is an autogenerated message for openQA integration by the openqa_review script:

This bug is still referenced in a failing openQA test: media_upgrade_sles12sp3+alladdons_aarch64
https://openqa.suse.de/tests/1570085

Actions #8

Updated by mitiao about 6 years ago

Actions #9

Updated by mitiao about 6 years ago

log said:
[debug] <<< testapi::assert_screen(mustmatch='boot-firmware', timeout=30)

See: https://github.com/os-autoinst/os-autoinst-distri-opensuse/blob/master/lib/bootloader_setup.pm#L76
we don't ever create this needle.

Actually what we need is to send 'enter' after move up to 'Boot from Hard Disk', and before sending the key, we have to satisfy the 'boot-firmware' needle.

And I don't know why the name of the tag is firmware since i didn't see firmware screenshot.

Actions #10

Updated by leli about 6 years ago

  • Status changed from New to In Progress
  • % Done changed from 0 to 60

This issue happened for aarch64 boot from disk flow depends on BOOT_HDD_IMAGE=1 and this variable changed to 0 on reboot_to_upgrade. We need to keep BOOT_HDD_IMAGE=1 for aarch64.

Actions #11

Updated by qmsu about 6 years ago

  • Related to action #34999: [sle][migration][sle15][aarch64] test fails in grub_test_snapshot - need workaround to make uefi boot into harddisk added
Actions #12

Updated by okurz about 6 years ago

could be that this code looks like copied from somewhere else and applied incorrectly but leli is right about the test variable. Though I suggest to not look for any needle "boot-firmware" because I assume it is time critical.

Actions #13

Updated by leli about 6 years ago

  • Status changed from In Progress to Resolved
  • % Done changed from 60 to 100

PR merged and verified on build 567.1.
https://openqa.suse.de/tests/1622314

Actions

Also available in: Atom PDF