reproduced it 3 times in a row. 
The last time I looked at journalctl -b again, but it was way too short.

var/log/Xorg.0.log had the same errors as those in the attached journalctl -b output:


Jun 06 17:14:07 localhost gdm-Xorg-:0[1651]: (EE) systemd-logind: failed to get session: PID 1651 does not belong to any known session


Jun 06 17:14:13 localhost gdm-Xorg-:0[1651]: (EE) AIGLX error: dlopen of /usr/lib/dri/i915_dri.so failed (/usr/lib/dri/i915_dri.so: undefined symbol: _glapi_tls_Dispatch)
Jun 06 17:14:13 localhost gdm-Xorg-:0[1651]: (EE) AIGLX: reverting to software rendering
Jun 06 17:14:13 localhost gdm-Xorg-:0[1651]: (EE) AIGLX error: dlopen of /usr/lib/dri/swrast_dri.so failed (/usr/lib/dri/swrast_dri.so: undefined symbol: _glapi_tls_Dispatch)
Jun 06 17:14:13 localhost gdm-Xorg-:0[1651]: (EE) GLX: could not load software renderer

Tested the same GnomeLiveCD (on a USB stick) on a different laptop
https://wiki.mageia.org/en/User:Marja/QA/Hardware#Lenovo_ThinkPad_SL510
with and without changing language in boot screen and/or language choice screen. GDM comes up without failwhale.
There are no errors in Xorg.0.log either (except the systemd-logind error, which can be ignored, as said yesterday by tmb and in a different bug report by coling)

For the record, tried again on the first laptop (that has this issue), with all localization (language, timezone) settings default, except for the keyboard (Nl)
 
Neither "Alt f4" nor "systemctl restart prefdm.service" nor 
"systemctl stop prefdm.service" followed by "systemctl start prefdm.service" helped.

This is not an installer issue

CC: (none) => sysadmin-bugs, thierry.vignaud
Component: Installer => Release (media or process)

I ran into this during the Live Mode of Gnome (9-Jun) version.  System installed fine for me.

CC: (none) => brtians1

(In reply to Brian Rockwell from comment #5)
> I ran into this during the Live Mode of Gnome (9-Jun) version.  System
> installed fine for me.

Anthlon x2-3800
Video Card:  Identification
Vendor: âNVIDIA Corporation

Description: âC51 [GeForce 6150 LE]

(In reply to Brian Rockwell from comment #6)
> (In reply to Brian Rockwell from comment #5)
> > I ran into this during the Live Mode of Gnome (9-Jun) version.  System
> > installed fine for me.
> 
> Anthlon x2-3800
> Video Card:  Identification
> Vendor: âNVIDIA Corporation
> 
> Description: âC51 [GeForce 6150 LE]

There are several bugs with a "oh no" gnome error, with various causes. This one is about the i915 driver, as far as I know. Are you sure you picked the right bug report?

Not sure at all, but looked closest to the same issue.  

Last round didn't have the issue.

Still valid on same hw, round 8 live GnomeCD 32bit

The errors are the same

tbh, I don't care, as long as KDE still copes with that old hw :-)

Looking forward to Mga6 and the liveCDs replaced with a light DE (and maybe all locales?)

Summary: Round 6 Mageia-5-LiveCD-GNOME-en-i586-CD: Oh no! Something has gone wrong (/usr/lib/dri/i915_dri.so: undefined symbol: _glapi_tls_Dispatch) => Round 8 Mageia-5-LiveCD-GNOME-en-i586-CD: Oh no! Something has gone wrong (/usr/lib/dri/i915_dri.so: undefined symbol: _glapi_tls_Dispatch)

btw, Alt + F4 didn't help here

Just for the record: tried whether installing from the boot menu would work better.
It didn't, the only difference is that a frozen MageiaWelcome starts after seeing the "Oh no!"
Alt F4 doesn't help, nor does systemctl restart prefdm.service

BTW one of the things to try on this error is to run ldconfig and then restart whatever crashed.

14-June round worked fine on Nvidia 6150 LE video card.  "Oh No" is gone for this equipment.

This is a duplicate of bug 16033. Lewis and I only saw it when booting from DVD, but if I read Marja's comments correctly, it can also happen when booting from a USB stick.

(In reply to Christiaan Welvaart from comment #12)
> BTW one of the things to try on this error is to run ldconfig and then
> restart whatever crashed.

This worked for me:

 - boot to "Oh no" screen
 - Ctrl-Alt-F2 to tty login
 - log in as root
 - type
     ldconfig
     systemctl restart dm.service

The shortcut of using Ctrl-Alt-Backspace to restart the X server (instead of systemctl) doesn't work.

CC: (none) => mageia

(In reply to Martin Whitaker from comment #14)
> This is a duplicate of bug 16033. Lewis and I only saw it when booting from
> DVD, but if I read Marja's comments correctly, it can also happen when
> booting from a USB stick.

Well, it only happens since round 8, in round 6/7 (the LiveCDs didn't change for round 7, IIRC) I could boot fine into Gnome... so I thought it wiser to file a separate bug report.

I'll try the ldconfig etc. workaround, though, and then report back

(In reply to Marja van Waes from comment #15)

> 
> Well, it only happens since round 8, in round 6/7 (the LiveCDs didn't change
> for round 7, IIRC) I could boot fine into Gnome... so I thought it wiser to
> file a separate bug report.
> 
I'm wrong about the round numbers, there, but the idea is the same: it didn't occur with the round before the one used when filing this bug

(In reply to Marja van Waes from comment #15)
> Well, it only happens since round 8, in round 6/7 (the LiveCDs didn't change
> for round 7, IIRC) I could boot fine into Gnome... so I thought it wiser to
> file a separate bug report.
> 
It wasn't meant as a criticism, Marja! The GNOME "Oh no" screen has caused a lot of confusion, so I think it's better not to assume you are seeing the same bug as other people until proven otherwise. I'm not sure at what stage bug 16033 really first appeared - it was mixed in with lots of other bugs.

I've just tested the workaround on my laptop with Intel (actually Optimus) graphics, and it works on that too.

(In reply to Martin Whitaker from comment #17)
> (In reply to Marja van Waes from comment #15)
> > Well, it only happens since round 8, in round 6/7 (the LiveCDs didn't change
> > for round 7, IIRC) I could boot fine into Gnome... so I thought it wiser to
> > file a separate bug report.
> > 
> It wasn't meant as a criticism, Marja! 

Don't worry, Martin :-)

If it did feel as criticism, then it was nothing compared to how critical I can sometimes be of myself ;-)

> The GNOME "Oh no" screen has caused a
> lot of confusion, so I think it's better not to assume you are seeing the
> same bug as other people until proven otherwise. I'm not sure at what stage
> bug 16033 really first appeared - it was mixed in with lots of other bugs.
> 
> I've just tested the workaround on my laptop with Intel (actually Optimus)
> graphics, and it works on that too.

Thanks for having mentioning that workaround and mentioning it again.... it just made me find out that, for me, this bug occurs less often

I hadn't tested a LiveCD for round 9 pre-release final, so grabbed a Mga5 GnomeLiveCD from a mirror now.

It booted fine from USB (tried twice from a "cold" machine). I had nearly closed this bug, but decided to try a reboot, too. That resulted in the Oh no! again.

running ldconfig and then restarting dm.service does indeed solve the problem.

Note that the screen seemed frozen for a long time (the cursor was frozen, too) before the desktop suddenly nicely appeared.

Closing as dup of 16033 (the ldconfig workaround working, convinced me it is very likely the same bug, after all)

*** This bug has been marked as a duplicate of bug 16033 ***

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