Mageia forum

Hi.

When installing VirtualBox (it had disappeared by itself and mysteriously) my kernels have also been updated. But to boot I need to do it with version 5.1.14, as the rest are blocked at startup, showing this error.



How can I solve it?

Are you sure that this is blocking the boot, are you sure it's not just a problem with graphics driver or X server not starting fully ?

Apart from that, have a look at e.g. https://unix.stackexchange.com/question ... nse-buffer or https://forum.manjaro.org/t/firmware-bu ... 1-00/52124
and try disabling TPM in UEFI/BIOS.

Hi, thanks for your response.

I only know that when those two error phrases appear, the boot stops, so I could not assure you that it is blocking the boot. How can I find out if it's a graphics driver or X server problem?

If it really stops that early during boot, you can't. You could start failsafe/recovery mode though, to see if that boots - simply select it from bootloader in the advanced menu.

Also, did you try to disable the TPM as mentioned ?

doktor5000 wrote:If it really stops that early during boot, you can't. You could start failsafe/recovery mode though, to see if that boots - simply select it from bootloader in the advanced menu.

Also, did you try to disable the TPM as mentioned ?

Hi.

First I have tried to disable the TPM.

(Inhabilitado=Disabled)

But the only change has been that, the screen that crashes on startup, two more sentences are displayed that it did not show before.



Then I have started Mageia in recovery mode, and it has started in command line, but once I have set the administrator password I do not know what I must do to return to normal boot mode with Plasma.

Linares wrote:Then I have started Mageia in recovery mode, and it has started in command line, but once I have set the administrator password I do not know what I must do to return to normal boot mode with Plasma.

Try as root with

Code: Select all

systemctl isolate graphical.target

When I do systemctl isolate graphical.target, this other bunch of lines appears, until the system is locked again.

https://gyazo.com/d69fa005b55e9b92d61c8d406b5e7fc9

After you ran that, can you switch to another tty via e.g. Ctrl+Alt+F2 ?

Yes, I can. With Ctrl+Alt+F2 I'm able to switch another tty, and the scenario is not much different. In the other tty I type systemctl isolate graphical.target, and the following messages appear.

https://gyazo.com/1cc61dbbaf3e28a40c6584aaa3d65334

After that, I can keep entering commands, but always in tty. The graphical environment does not start.

If you can switch to another tty, then after entering systemctl isolate graphical.target in the first one, switch to another, and in that one write the journal logs to a file as root

Code: Select all

journalctl -a > /tmp/journal.log

and then later, upload /tmp/journal.log here as an attachment.

I have done as you have told me, but when I enter the tmp folder, there is no file.

You did it as your regular user, and as the message says, you don't have permission to view the system logs. You need to run that as root, as mentioned previously.

doktor5000 wrote:If you can switch to another tty, then after entering systemctl isolate graphical.target in the first one, switch to another, and in that one write the journal logs to a file as root

Code: Select all

journalctl -a > /tmp/journal.log

and then later, upload /tmp/journal.log here as an attachment.

Hello again.

I already have the journal.log file, but we have a problem. This file occupies 199 Mb and I cannot attach it in this forum. I have compressed it with gzip until leaving it at 7 Mb, but the maximum allowed is 5 Mb.

Sorry, forgot a letter there, my bad
journalctl -a will output all the journal logs, but we only need the one from the most recent boot. Try again with

Code: Select all

journalctl -ab > /tmp/journal.log

Hi. Here I have the journal.log file.

Seems you're missing the kernel-devel package for 5.6.6:

may 17 19:19:24 Mageia dkms-autorebuild.sh[1137]: nvidia-current (430.64-8.mga7.nonfree): Installing module.
may 17 19:19:24 Mageia dkms-autorebuild.sh[1137]: Kernel devel files for 5.6.6-desktop-1.mga7 not installed. Cannot install this module.

Apart from that, your display manager fails to start (probably due to the issue with the graphics driver), and then core-dumps:

Code: Select all

may 17 19:19:26 Mageia sddm[4087]: Starting...
may 17 19:19:26 Mageia sddm[4087]: Logind interface found
may 17 19:19:26 Mageia sddm[4087]: Adding new display on vt 1 ...
may 17 19:19:26 Mageia sddm[4087]: Loading theme configuration from ""
may 17 19:19:26 Mageia sddm[4087]: Display server starting...
may 17 19:19:26 Mageia sddm[4087]: Running: /usr/bin/X -nolisten tcp -auth /var/run/sddm/{8c476d62-cc1d-4525-a932-4e2902167467} -background none -noreset -displayfd 17 -seat seat0 vt1
[...]
may 17 19:19:26 Mageia sddm[4087]: Failed to read display number from pipe
may 17 19:19:26 Mageia sddm[4087]: Display server failed to start. Exiting
may 17 19:19:26 Mageia audit[1]: SERVICE_STOP pid=1 uid=0 auid=4294967295 ses=4294967295 msg='unit=sddm comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=failed'
may 17 19:19:26 Mageia systemd[1]: sddm.service: Main process exited, code=dumped, status=6/ABRT
may 17 19:19:26 Mageia systemd[1]: sddm.service: Failed with result 'core-dump'.

In the meantime you could try a different graphics driver, like nouveau or vesa via failsafe mode => drakx11, or try a different display manager.

If there are two incidents, I think it will be best to go in parts.

If the kernel-devel package for 5.6.6 has disappeared, how can I recover it?

Well, so far it's one issue - graphics driver not working properly or wrong one chosen. Hence X server is not starting.

And how you can "recover" it - simply install it.

Code: Select all

urpmi kernel-desktop-devel-latest

But I have a problem with kernels and by doing this I could definitely lose all access to Mageia.

At startup I had 4 kernels. Two working properly. The other two do not, with these problems that we are seeing. (Go to viewtopic.php?f=7&t=13456&p=79017#p78820)

In another forum, other users recommended me to update quietly, that nothing would happen to me. I did it and I only got one more unusable kernel installed, leaving three broken, and only one useful one.

If I do this, I can find four broken kernels, none useful, and goodbye Mageia.

Linares wrote:But I have a problem with kernels and by doing this I could definitely lose all access to Mageia.

Well, you basically have 2 choices: either keep it the way it is, and stay with your working kernel 5.1.14 and never install any kernel update again, without getting any of the security fixes.
Or, you try to fix the installation. It's your choice.

Linares wrote:At startup I had 4 kernels. Two working properly. The other two do not, with these problems that we are seeing. (Go to viewtopic.php?f=7&t=13456&p=79017#p78820)

This is the first time that you mention this.

And nobody asked to update your kernel, but you need to install the kernel headers for the kernel that you want to boot, if you use dkms packages. Or you install the respective precompiled nvidia module for that kernel, or you switch to a different graphics driver, as suggested.

Please provide the output of

Code: Select all

rpm -qa | grep kernel | sort

doktor5000 wrote:Please provide the output of

Code: Select all

rpm -qa | grep kernel | sort

This is the output:



The first four are the ones that appear in the boot list, but only 5.1.14 works. The kernel that I must keep at all costs, for being the only one that is keeping me in Mageia.

So if you want to use the dkms nvidia driver for kernel 5.6.6, then you would need to install the package kernel-desktop-devel-5.6.6
Or switch to a different graphics driver, as already mentioned.

urpmi kernel-desktop-devel-latest has not worked.



"Falló la instalación, algunos archivos están ausentes" is "Installation failed, some files are missing". :-/

Well, it could not download any packages. That is either because your network connection is not up, your repo hdlists are not up-to-date or your mirror is not in sync.

Try again after

Code: Select all

systemctl start network.service
urpmi.update -a

Also, as you skipped a kernel update, looking at your screenshot - you would install kernel-desktop-devel-5.6.8, and you don't have that kernel installed so that wouldn't help at all.
So better download directly kernel-desktop-devel-5.6.6

systemctl start network.service
urpmi.update -a
urpmi kernel-desktop-devel-latest

That sequence has worked without incident but on reboot I choose kernel 5.6.6 and it freezes again on this screen: viewtopic.php?f=7&t=13456&p=79038#p78820

The rpm -qa | grep kernel | sort returns

Code: Select all

kernel-desktop-5.1.14-1.mga7-1-1.mga7
kernel-desktop-5.3.11-1.mga7-1-1.mga7
kernel-desktop-5.5.15-3.mga7-1-1.mga7
kernel-desktop-5.6.6-1.mga7-1-1.mga7
kernel-desktop-devel-5.3.11-1.mga7-1-1.mga7
kernel-desktop-devel-5.6.13-5.mga7-1-1.mga7
kernel-desktop-devel-latest-5.6.13-5.mga7
kernel-desktop-latest-5.5.15-3.mga7
kernel-firmware-20190603-1.mga7
kernel-firmware-nonfree-20200316-1.mga7.nonfree
kernel-userspace-headers-5.5.15-3.mga7
virtualbox-kernel-5.3.11-desktop-1.mga7-6.0.14-6.mga7
virtualbox-kernel-5.5.15-desktop-3.mga7-6.0.20-1.mga7
virtualbox-kernel-5.6.6-desktop-1.mga7-6.0.20-2.mga7
virtualbox-kernel-desktop-latest-6.0.20-2.mga7

PS: After this awesome help, I owe you a few beers at the Beer Festival.