Bug 17839

Summary: Failures to boot from PLASMA Live iso 2016-02-28
Product: Mageia Reporter: Len Lawrence <tarazed25>
Component: Release (media or process)Assignee: Thomas Backlund <tmb>
Status: RESOLVED OLD QA Contact:
Severity: major    
Priority: Normal CC: davidwhodgins, eeeemail, ennael1, marja11, sysadmin-bugs, wilcal.int
Version: CauldronKeywords: 6dev1, NEEDINFO
Target Milestone: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Source RPM: CVE:
Status comment:
Attachments: Diagnostics file from CSM boot of PLASMA Live

Description Len Lawrence 2016-02-28 22:20:32 CET
Description of problem:
UEFI boot
After a run of messages stops at a dracut session prompt.  Impossible to use
the keyboard so no way to dump diagnostics.  Not possible to proceed.
CSM boot
A run of messages pausing at "Starting job for LSB - waiting for hotplugged network to be up" (this machine wifi only at that point) and an estimate of
5 minutes.  Runs in the background while other messages continue - all OK
until "Starting Hold until boot process finishes up..." only it never does.  Forced a reboot at that point and ended up at the dracut session prompt (note the different behaviour on the second run).  This allowed a dump of rdsosreport.txt to a usb drive (report attached).


Version-Release number of selected component (if applicable):
Mageia-6-dev1-LiveDVD-PLASMA5-x86_64-DVD

How reproducible:
On one machine the UEFI boot always proceeds the same way.
On a CSM laptop the behaviour is slightly variable but always ends in failure.

Steps to Reproduce:
1. Boot from an iso on a usb drive
2. Esc from the splash screen to view messages
3. Wait for an opportunity to dump diagnostics or until it hangs
Comment 1 Len Lawrence 2016-02-28 22:34:55 CET
Created attachment 7503 [details]
Diagnostics file from CSM boot of PLASMA Live
Len Lawrence 2016-02-28 22:35:49 CET

Whiteboard: (none) => 6dev1

Marja Van Waes 2016-02-29 08:46:30 CET

CC: (none) => davidwhodgins, eeeemail, ennael1, marja11, wilcal.int
Assignee: bugsquad => tmb

Comment 2 Marja Van Waes 2016-04-05 19:22:59 CEST
@ Len

Is this bug still valid with the publicly released iso?

Whiteboard: 6dev1 => (none)
Keywords: (none) => 6dev1, NEEDINFO

Comment 3 Len Lawrence 2016-04-05 22:22:05 CEST
@ Marja

As said in bug 17698 c#5 testing again to check symptoms.
Comment 4 Len Lawrence 2016-04-05 22:29:15 CEST
UEFI test again.  Failure confirmed.
Plasma Live iso from 2016-03-25

Dropped straight to the command line with
"Sorry, but there has been a problem starting your graphical display.....

$ drakx11
Chose nvidia series 420 and later
Plug'n'play monitor
Resolution Automatic (24bpp)
No test
Ignored request to reboot because of live session
$ systemctl default
Goes back to tty login
Login as live
$ startx start.kde
X server crash, as described before.
$ systemctl status sddm.service
disabled ... inactive(dead)
$ systemctl start sddm.service
Screen blanks
Tried tty1 and tty2
Text login prompt flashes up periodically - looks like there is no
future in it.

Alt Ctl Del to reboot
Command line again
Logged in as root and checked sddm status
sddm disabled
$ drakx11
Chose Xorg vesa, plug'n'play, 1280x1024 (not sure what the limit is for vesa
but this is a 4K screen and that resolution is too low to be usable)
No test
$ systemctl default
Straight to command line, logged in as live
$ startx start.kde
A low resolution mouse pointer appears in the centre of the "busy" symbol then
seconds later X server crash.

It looks like bug 17698 describes similar issues.
Shall follow up with the CSM report.
Comment 5 Len Lawrence 2016-04-05 23:09:55 CEST
Test of the Plasma5 iso on a CSM laptop with nvidia GeForce series 8100 graphics ran into the same problems as the in the UEFI case.  Choosing nvidia resulted in the Xorg nouveau driver being installed (apparently) and an X server crash.

So, this bug is still valid.  I am not sure of the status of 17698 though.  Since it applied to an out-dated iso and later isos reproduce its most serious issues it should perhaps be retired.
Comment 6 Marja Van Waes 2021-03-06 14:11:49 CET
Last comment about someone hitting this issue dates from almost 5 years ago.

I'm sure that there'll be a newer bug report if this still happens with recent ISOs, so closing this report as OLD

Resolution: (none) => OLD
Status: NEW => RESOLVED