Bugzilla – Bug 728849
Garbage when enabling YaST2 firstboot
Last modified: 2011-11-09 16:53:31 UTC
1. Go to http://staging.susestudio.com/appliance/create_new. 2. Create a new 12.1 JeOS appliance. 3. In the "Configuration" tab set the "Region" setting to "Ask on first boot". 4. Install yast2-firstboot and yast2-ncurses packages. 5. Build the appliance. 6. Testdrive the appliance. 7. Observe garbage on the YaST screen similar to this: http://imagebin.suse.de/1136 I am not sure where the bug lies so based on Stano's suggestions, I am trying YaST first. Ping me if you need access to Studio's staging system (we don't have 12.1 enabled on production yet). I have a testing appliance created by applying the steps above that exhibits the problem: http://staging.susestudio.com/appliance/edit/2876#/build I did not try to reproduce this outside of Studio yet.
Making this P1 because the bug is important for 12.1 support in Studio and there is not much time to solve it.
Is it same when you have different firstboot step as first one (language)? What happens when you click next? Can you attach YaST log files? Is it specific to Testdrive or is it the same in dowloaded appliance? Where can I download the appliance anyway? (I do not know the credentials for staging.susestudio.com) (BTW, I'm sick at home, so my reaction time is long)
+ what version of yast2-installation do you have? there were some changes regarding starting firstboot service lately (in 2.21.28)
Well, my current idea is that it is caused by some incorrect order of starting services, i.e. the new systemd way. Frederic, any idea? Is it possible to turn systemd off and see if anything changes?
it would be interesting to get dmesg output when booting with systemd.log_level=debug systemd.log_target=kmsg, when yast second stage was started, to see what happened exactly.
Just a note: it's firstboot stage, not second stage, so some rules are different there.
yes, sorry, I guess firstboot was less tested with systemd than second stage..
(In reply to comment #2) > Is it same when you have different firstboot step as first one (language)? I's the same. > What happens when you click next? The garbage get cleared up by a redraw but the YaST interface is immediately covered by another messages (something with "interrupts" -- it was too fast to read the whole message). > Can you attach YaST log files? Will attach that in a minute. > Is it specific to Testdrive or is it the same in dowloaded appliance? The problems is still thre when I download the appliance and run it in qemu-kvm. > Where can I download the appliance anyway? (I do not know the credentials for > staging.susestudio.com) I published the appliance on SUSE Gallery on staging: http://staging.susestudio.com/a/LSpMjD/yast2-firstboot-jeos Access from the SUSE network should work without any credentials. (In reply to comment #3) > + what version of yast2-installation do you have? there were some changes > regarding starting firstboot service lately (in 2.21.28) The one in 12.1 repos: 2.21.25-2.1.1 (In reply to comment #4) > Well, my current idea is that it is caused by some incorrect order of starting > services, i.e. the new systemd way. Frederic, any idea? > > Is it possible to turn systemd off and see if anything changes? I tried to add the "sysinit" package and select "System V" it in the GRUB menu. The problem disappeared, which supports your systemd theory. (In reply to comment #5) > it would be interesting to get dmesg output when booting with > systemd.log_level=debug systemd.log_target=kmsg, when yast second stage was > started, to see what happened exactly. Will attach that in a minute.
Created attachment 461182 [details] /var/log/YaST2/mkinitrd.log /var/log/YaST2/mkinitrd.log from an OVF appliance exhibiting the problem.
Created attachment 461183 [details] /var/log/YaST2/y2log /var/log/YaST2/y2log from an OVF appliance exhibiting the problem.
Created attachment 461184 [details] /var/log/YaST2/y2start.log /var/log/YaST2/y2start.log from an OVF appliance exhibiting the problem.
Created attachment 461186 [details] "dmesg" output "dmesg" output taken at the point when the garbage was displayed. (I just switched to another console and did "dmesg > dmesg.txt".)
Frederic, can you see anything interesting from these? Would it help with yast2-installation containing your patch for bug #726823? David, could you try debug commands from comment 5?
(In reply to comment #13) > David, could you try debug commands from comment 5? The dmesg output is captured with these boot parameters specified, I just forgot to mention it. (The YaST2 logs are captured without them.)
ok, just tested with YaST2-Firstboot.service from yast-installation-2.21.28 and it fixes the gargabe output from network service. so, it should be fixed with goldmaster
Good, than it's duplicate of the other bug. *** This bug has been marked as a duplicate of bug 726823 ***