[Bug 1049304] New: [Build 20170717] openQA - Live images broken
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304 Bug ID: 1049304 Summary: [Build 20170717] openQA - Live images broken Classification: openSUSE Product: openSUSE Tumbleweed Version: Current Hardware: Other URL: http://openqa.opensuse.org/tests/450944/modules/bootlo ader/steps/4 OS: Other Status: NEW Severity: Normal Priority: P5 - None Component: Live Medium Assignee: dimstar@opensuse.org Reporter: dimstar@opensuse.org QA Contact: qa-bugs@suse.de Found By: --- Blocker: --- ## Observation openQA fails in all live image based tests, failing to find the correct bootloader entries [bootloader](http://openqa.opensuse.org/tests/450944/modules/bootloader/steps/4) ## Reproducible Fails since (at least) Build [20170717](http://openqa.opensuse.org/tests/450944) (current job) ## Expected result Last good: [20170712](http://openqa.opensuse.org/tests/446982) (or more recent) ## Further details Always latest result in this scenario: [latest](http://openqa.opensuse.org/tests/latest?distri=opensuse&flavor=GNOME-Live&arch=x86_64&test=gnome-live&version=Tumbleweed&machine=64bit) ## More information After the switch to kiwi-ng (9.95), the boot media no longer show the entries for: * mediacheck * memtest -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304
Dominique Leuenberger
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304
Dominique Leuenberger
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304#c1
--- Comment #1 from David Cassany
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304#c2
--- Comment #2 from Dominique Leuenberger
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304#c3
Oliver Kurz
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304#c4
--- Comment #4 from Oliver Kurz
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304#c5
David Cassany
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304#c6
Dominique Leuenberger
Dominique, I could not find out which is the OBS project that is used for these qa tests. Could you handle this update? Or point me out which is the project I should SR to.
The images are build in openSUSE:Factory:Live / livecd-openSUSE (contains three kiwi files, with the .kiwi.in being the leading one) -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304#c7
--- Comment #7 from Fabian Vogt
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304#c8
--- Comment #8 from David Cassany
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304#c9
--- Comment #9 from Fabian Vogt
Fabian,
If choosing the mediacheck entry the image should boot normaly, it just performs 'mediacheck /dev/cdrom' and then after that, if succeeded, it boots normally. You should be capable to see a line in /var/log/boot.kiwi with 'ISO check passed' after booting. I have just retested in my env and it worked as expected. Could you verify you didn't miss the validation check? If it really does nothing please point me to an specific iso build so I can have a look at it.
I can verify that that line is present in the log. However, that is not what the user would expect from the option. It should behave just like the old mediacheck, with a visible output on the screen: https://openqa.opensuse.org/tests/446989#step/mediacheck/7
Thanks
-- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304#c10
--- Comment #10 from David Cassany
I can verify that that line is present in the log. However, that is not what the user would expect from the option. It should behave just like the old mediacheck, with a visible output on the screen: https://openqa.opensuse.org/tests/446989#step/mediacheck/7
Yes, agree, this is the expected output and this is what I saw in my dev environment building a JeOS. Strange, could you point me to the exacte project you are testing in OBS? Maybe it needs to handle plymouth in some specific way, the messages might displayed behind the scenes. I agree the user needs to see it, otherwise it doesn't have much sense. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304#c11
--- Comment #11 from Fabian Vogt
I can verify that that line is present in the log. However, that is not what the user would expect from the option. It should behave just like the old mediacheck, with a visible output on the screen: https://openqa.opensuse.org/tests/446989#step/mediacheck/7
Yes, agree, this is the expected output and this is what I saw in my dev environment building a JeOS. Strange, could you point me to the exacte project you are testing in OBS? Maybe it needs to handle plymouth in some specific way, the messages might displayed behind the scenes. I agree the user needs to see it, otherwise it doesn't have much sense.
I'm testing the i586 media from here: https://build.opensuse.org/package/binaries/home:favogt:lc/livecd-x11?reposi... -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304#c12
--- Comment #12 from Marcus Schaefer
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304#c13
Marcus Schaefer
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304#c16
Marcus Schaefer
a) how does it behave performance wise? if we lose to much time at every boot, > it's bad (our live images are already slow)
you can check the time yourself on the built iso. In my tests I got time checkmedia LimeJeOS-Leap-42.3.x86_64-1.42.3.iso app: 0x5a743c1f media: CD1 size: 260666 kB hybrid: 15360 kB pad: 300 kB check: 100% check: md5sum ok md5: 84226446a1bc69382a8fbf4211eb0c8a real 0m0.969s user 0m0.937s sys 0m0.032s with ls -lh LimeJeOS-Leap-42.3.x86_64-1.42.3.iso -rw-r--r-- 1 root root 255M Sep 8 16:44 LimeJeOS-Leap-42.3.x86_64-1.42.3.iso Since we are on dracut the bootup time is much faster than before. It would be good to see the check time numbers from your iso, which I think is bigger than mine
b) What if it fails? does plymouth disappear and/or at least give some indication, why boot does not continue? If we just remain on the splash screen, > that makes for a very bad user experience
if it fails the code calls: die "ISO check failed" The die method is from the dracut lib (/usr/lib/dracut/modules.d/99base/dracut-lib.sh) and should make sure no graphics screen hides a failed boot information You have a point in boot up time which is faster without a check. Let me know what you think -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304#c17
Dominique Leuenberger
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304#c18
--- Comment #18 from Marcus Schaefer
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304#c19
--- Comment #19 from Marcus Schaefer
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304#c20
Marcus Schaefer
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304
http://bugzilla.opensuse.org/show_bug.cgi?id=1049304#c21
--- Comment #21 from Oliver Kurz
participants (1)
-
bugzilla_noreply@novell.com