6.5.11-desktop-5.mga9 hangs on boot

[isadora]

Hi
A week ago, I think, Mageia 9 was updatet with in the Mageia control center with "Update your system". I have first booted the system now.

The problem is that 6.5.11-desktop-5.mga9 hangs on boot.

There is this error on the screen before it hang:

Code: Select all

Failed to start mdmonitor.service.
See 'systemctl status mdmonitor.service' for details.

I then booted 6.5.11-desktop-5.mga9 in recorvery mode
As root I typed 'systemctl status mdmonitor.service'
The result is:

Code: Select all

mdmonitor.service - Software RAID monitoring and management
  Loaded: Loaded (/usr/lib/systemd/system/mdmonitor.service: enabled; presset: enabled)
  Active: inactive (dead)

I don't know if the above is the problem.

Or is this the problem:

Code: Select all

[root@localhost log]# tail -50 Xorg.0.log.old
[    26.275]    compiled for 1.21.1.7, module version = 0.1.1
[    26.275]    ABI class: X.Org Video Driver, version 25.2
[    26.275] (II) LoadModule: "glx"
[    26.275] (II) Loading /usr/lib64/xorg/modules/extensions/libglx.so
[    26.276] (II) Module glx: vendor="X.Org Foundation"
[    26.276]    compiled for 1.21.1.8, module version = 1.0.0
[    26.276]    ABI class: X.Org Server Extension, version 10.0
[    26.276] (II) LoadModule: "nvidia"
[    26.276] (II) Loading /usr/lib64/xorg/modules/drivers/nvidia_drv.so
[    26.276] (II) Module nvidia: vendor="NVIDIA Corporation"
[    26.276]    compiled for 1.6.99.901, module version = 1.0.0
[    26.276]    Module class: X.Org Video Driver
[    26.276] (II) v4l driver for Video4Linux overlay mode (V4L2)
[    26.276] (II) NVIDIA dlloader X Driver  535.129.03  Thu Oct 19 18:44:59 UTC 2023
[    26.276] (II) NVIDIA Unified Driver for all Supported NVIDIA GPUs
[    26.276] (WW) Falling back to old probe method for v4l
[    26.276] (II) systemd-logind: releasing fd for 226:0
[    26.277] (II) Loading sub module "fb"
[    26.277] (II) LoadModule: "fb"
[    26.277] (II) Module "fb" already built-in
[    26.277] (II) Loading sub module "wfb"
[    26.277] (II) LoadModule: "wfb"
[    26.277] (II) Loading /usr/lib64/xorg/modules/libwfb.so
[    26.277] (II) Module wfb: vendor="X.Org Foundation"
[    26.277]    compiled for 1.21.1.8, module version = 1.0.0
[    26.277]    ABI class: X.Org ANSI C Emulation, version 0.4
[    26.277] (EE) NVIDIA: Failed to initialize the NVIDIA kernel module. Please see the
[    26.277] (EE) NVIDIA:     system's kernel log for additional error messages and
[    26.277] (EE) NVIDIA:     consult the NVIDIA README for details.
[    26.277] (EE) NVIDIA: Failed to initialize the NVIDIA kernel module. Please see the
[    26.277] (EE) NVIDIA:     system's kernel log for additional error messages and
[    26.277] (EE) NVIDIA:     consult the NVIDIA README for details.
[    26.277] (WW) Falling back to old probe method for v4l
[    26.277] (EE) NVIDIA: Failed to initialize the NVIDIA kernel module. Please see the
[    26.277] (EE) NVIDIA:     system's kernel log for additional error messages and
[    26.277] (EE) NVIDIA:     consult the NVIDIA README for details.
[    26.277] (EE) NVIDIA: Failed to initialize the NVIDIA kernel module. Please see the
[    26.277] (EE) NVIDIA:     system's kernel log for additional error messages and
[    26.277] (EE) NVIDIA:     consult the NVIDIA README for details.
[    26.277] (EE) No devices detected.
[    26.277] (EE)
Fatal server error:
[    26.277] (EE) no screens found(EE)
[    26.277] (EE)
Please consult the Mageia support
    at https://bugs.mageia.org
 for help.
[    26.277] (EE) Please also check the log file at "/var/log/Xorg.0.log" for additional information.
[    26.278] (EE)
[    26.279] (EE) Server terminated with error (1). Closing log file.


Is there a way to get some more logs ?

Is this a good ide ?
urpmi.update -a
urpmi --auto-select

I can boot with 6.4.16-desktop-3.mga9

From here the system is up to date.

Best regards
Anders

[morgano]

Raid failing is normal if you do not have raid... ( I do not have and i also get that message - it apparently always tries to look for RAID)

It apparently fail to initialise nvidia kernel module.

Yes make sure the system is updated.

Do

Code: Select all

dkms status

say it is installed for the kernel?

Do you have dkms-nvidia-current installed? It is needed to be installed/updated at the same time as nvidia-current or kernel is installed/updated for the kernel module to be built. I have got hit by this and system do not warn.

Check dkms-nvidia-current is installed, and then probably easiest for you is to uninstall kernel 6.5.11-desktop-5, and then istall it again - to get the module built.
While doing this, you can have a terminal open and watch the progress by viewing the journal: issue

Code: Select all

journalctl- f

in a terminal and leave it running before installing the new kernel.

If it fail despite kernel module seem OK, you could try kernel-linus-6.5.11-2.mga9, which is a less pathed kernel. (For me it works slightly better for nvidia -(resume from suspend need power cycling the monitor when i use desktop kernel))

[anp66]

Hi Morgano

Thanks for your input.

This command:

Code: Select all

dkms status

gives the following result:

Code: Select all

nvidia-current,535.129.03-1.mga9.nonfree, 6.4.16-desktop-3.mga9, x86_64: installed
nvidia-current,535.86.05-1.mga9.nonfree, 6.4.9-desktop-4.mga9, x86_64: installed
nvidia-current,535.86.05-1.mga9.nonfree, 6.5.11-desktop-5.mga9, x86_64: installed
nvidia-current,535.86.05-1.mga9.nonfree, 6.4.16-desktop-3.mga9, x86_64: built
nvidia-current,535.86.05-1.mga9.nonfree, 5.15.126-desktop-1.mga8, x86_64: installed


You're right !.
The only kernel built is the one I can log into, graphically.

Is there a recovery disk I can make now, so that I have the option of logging in graphically if I destroy more ?

I think your strategy of uninstall kernel 6.5.11-desktop-5, and then install it again, is the right one!.

I would actually like to uninstall all 4 cores that have not been built.

Do you know how to do it from a terminal ?

In my panic, I tried the following:

Code: Select all

urpmi.update -a

It did not work. It looked like there was no connection to the repository.

Best regards
Anders

[anp66]

Hi again Morgano

A little more info.
When I search for dkms-nvidia-current packages in the graphical environment, the result is 2 hits:

Code: Select all

dkms-nvidia-current 535.129.03 1.mga9.nonfree x86_64
NVIDIA kernel module for GeForce 745 and later cards
dkms-nvidia-current 535.86.05 1.mga9.nonfree x86_64
NVIDIA kernel module for GeForce 745 and later cards


Should I uninstall version 535.129.03 ?
(I really want to do that.)

Best regards
Anders

[morgano]

Most important is to make the updates and install to work.
I also believe your problem originates in failed update.
Please see https://wiki.mageia.org/en/Mageia_9_Err ... g_software

During boot with a kernel that have not the correct module built, it is supposed to build it automatically (have patience then, it takes a couple minutes on a fast machine).
- provided you have the kernel-something-devel- prm installed for that kernel version, and the dkms-nvidia of the nvidia version.

[anp66]

Now it went wrong, have a system I can't boot graphically. Have no nvidia-current in status build, they are all in status: installed.

No matter which urpmi command is done from the command line, the error will be:

Code: Select all

Retrieving failed: aria2 failed: exited with 19

I think the basic problem is a lack of connection to a medium ?

Is there anyone who can show this command

Code: Select all

urpmi.addmedia --distrib

to be used when
version = mga9
arch = x86_64

Best regards
Anders

[Germ]

Code: Select all

# urpmi.removemedia -a

Then:

Code: Select all

# urpmi.addmedia --distrib --mirrorlist 'http://mirrors.mageia.org/api/mageia.9.x86_64.list'

Or, open MCC and do it graphically:

MCC > Configure media sources for install and update

[anp66]

Hi Germ

Thanks for your reply !

Havn problems with the addmedia command.
Get this error message :

Code: Select all

mirror list not fund
Could not find a mirror from mirror list http://mirrors.mageia.org/api/mageia.9.x86_64.list

Traid a lot of combinationd without any luck 😳.

Can someone see what I do wrong ?

Best regards
Anders

[doktor5000]

Are you sure that your network is actually started? The previous aria2 failures and your previous remark does indicated otherwise.
Try to ping 8.8.8.8 to see if you do have a network connection.

[anp66]

Your are right. I don't have any network.
Due to problems with the Nvidia packages, in status installed and not in build, I am running as root in recovery mode.

I don't know how to start the network from recovery mode.

Can I get the network by shifting to runlevel 3:

Code: Select all

init 3

or should I do something like :

Code: Select all

systemctl start systemd-metworkd
# systemctl start systemd-resolved

Best regards
Anders

[doktor5000]

init 3 should still work, which should translate to

Code: Select all

systemctl isolate multi-user.target

[morgano]

For some reason when i tried earlier this year i needed to use "telinit 3" instead of "init 3".

[anp66]

Thanks doktor5000 !
Now, from recovery, into text mode with network.

Then the following was done with sucess:

Code: Select all

urpmi.update - a
urpmi --auto-select

It seamt som nvidia packages was build.

Then, optimistic, booting into the graphically enviroment.
In the top of the screen it says:

Code: Select all

Building nvidia-current driver...
Installing nvidia-current driver...

But the boot process hang in a black screen.

Then with root in text mode I did:

Code: Select all

dkms status

The resultat was not good:

Code: Select all

nvidia-current, 535.129.03-1.mga9.nonfree, 6.5.11-desktop-5.mga9, x86_64: installed
nvidia-current, 535.129.03-1.mga9.nonfree, 5.15.126-desktop-1.mga8, x86_64: installed
nvidia-current, 535.129.03-1.mga9.nonfree, 5.15.122-desktop-1.mga8, x86_64: installed


I would like the nvidia packages to be build.

I have, without luck, been looking in log files, but did not find 'the smoking gun'.

Something is mixed up with this nvidia stuff and I don't know how to get it in ordrer.

What do I do now ?

Best regards
Anders

[doktor5000]

From your output boot with 6.5.11 kernel should work, at least for the nvidia driver as it has been built.

Check what kernel packages you have installed.

Code: Select all

rpm -qa | grep kernel | sort

For every installed kernel you need to respective -devel package with the same version.
Maybe you're simply missing kernel-desktop-devel-latest package which takes care about that automatically on every kernel update.

As an example, with 3 installed kernels (6.4.9, 6.4.16, 6.5.11) this is what it should look like

Code: Select all

[doktor5000@Mageia9]─[00:59:56]─[~] rpm -qa | grep kernel | sort
kernel-desktop-6.4.16-3.mga9
kernel-desktop-6.4.9-4.mga9
kernel-desktop-6.5.11-5.mga9
kernel-desktop-devel-6.4.16-3.mga9
kernel-desktop-devel-6.4.9-4.mga9
kernel-desktop-devel-6.5.11-5.mga9
kernel-desktop-devel-latest-6.5.11-5.mga9
kernel-desktop-latest-6.5.11-5.mga9
kernel-firmware-20230625-1.mga9
kernel-firmware-nonfree-20230814-1.mga9.nonfree
kernel-userspace-headers-6.5.11-5.mga9
remove-old-kernels-0.99.21.7-1.mga9

[morgano]

Maybe whatever problem i hit on this system, is similar to what you have hit it too:

I have noticed that when *switching between nvidia drivers*, that next boot with same kernel that was in use ends with black screen, but i have not had time to dig into it.
And yes despite dkms reporting modules are in place.
I opened https://bugs.mageia.org/show_bug.cgi?id=32579

One thing to check is if your kernel boot command line contain 'nokmsboot', it should.
To see when choosing boot: when in grub boot menu, highlight the correct choice and press key [e].
To see what was used in current boot: issue

Code: Select all

$ journalctl -b | grep 'Command\ line'

And for the previous boo exchange "-b" to "-b-1", the one before that : "-b-2" etc.

On topic of nvidia issues you could also try to install kernel-desktop-6.5.13-2.mga9 in backports testing which fixes another nvidia issue https://bugs.mageia.org/show_bug.cgi?id=32541#c19. Install that and at the same time the corresponding -devel- package, and the kernel modules should be built. It is very possible that then booting to this kernel works if the cause is similar to what I experience where it is only problem with the kernel that it i was in use when running drakx11...

Anyway I seem to get out of the problem by either uninstall then reinstall the problematic kernel,
And/or boot a working kernel choice, use drakx11 to change to nouveau or xorg kernel mode setting, reboot, uninstall all nvidia packages stuff, reboot, then use drakx11 to install nvidia driver. (probably not all steps are needed, but that is what I did.)

[anp66]

Hello again..

A 1000 thanks for your input and time!

There are a lot of kernels in the system, so the following has been created:

Code: Select all

rpm -qa | grep 6.5.11 | black
cpupower-6.5.11-5.mga9
kernel-desktop-6.5.11-5.mga9
kernel-desktop-devel-6.5.11-5.mga9
kernel-desktop-devel-latest-6.5.11-5.mga9
kernel-desktop-latest-6.5.11-5.mga9
kernel-userspace-headers-6.5.11-5.mga9

There is also kernel firmware.......

There is no remove-old-kernels, is that why I have so many old kernels lying around ?

Looking in GRUB boot menu 'nokmsboot' exists.

Doing journalctl is the result:

Code: Select all

journalctl -b | grep 'Command line'
Dec 10 20:28:07 localhost kernels : Command line: BOOT_IMAGE=/boot/vmlinuz-6.5.11-desktop-5.mga9 root=UUID=xxxxx ro failsafe

It doesn't say 'nokmsboot', but remember that a text environment is run, in the recovery session, as there is no graphical environment that works.

In the /boot folder, the following is done:

Code: Select all

find *6.5.11*
config-6.5.11-desktop-5.mga9
initrd-6.5.11-desktop-5.mga9.img
symvers-6.5.11-desktop-5.mga9.xz
System.map-6.5.11-desktop-5.mga9
vmlinuz-6.5.11-desktop-5.mga9

Mysterious, it seems that an image exists even though 'dkms status' says it is only 'installed'.

There are several things here.
To get a better overview, can old kernels be deleted with remove-old-kernels ? (which I must have installed).

After the cleanup, how do I reinstall kernel-desktop-6.5.11-5.mga9 with associated devel and latest packages ?

Best regards
Anders

[morgano]

can old kernels be deleted with remove-old-kernels ? (which I must have installed).

Yes, see https://wiki.mageia.org/en/Removing_pac ... ld_kernels

how do I reinstall kernel-desktop-6.5.11-5.mga9 with associated devel and latest packages ?

Just select them in drakrpm. Or use urpmi.

[anp66]

Then there is a little more info.

remove-old-kernels was installed and run.
This meant that 30 old kernels = 5925 MB were deleted.
Now I have the 3 newest kernels and understand that remove-old-kernels run via a crontab job once a week.

What remains then is the problem of reinstalling kernel-desktop-6.5.11-5.mga9 or installing kernel-desktop-6.5.13-2.mga9

Best regards
Anders

[anp66]

Yes, there is again a working graphical login.

From recovery mode as root, the following was done:

Repository 'core backports testing' was added by changing urpmi.cfg

Code: Select all

urpmi kernel-desktop-6.5.13-2.mga9

Then Mageia was rebooted. In the boot sequense this was stated:

Code: Select all

The system has to be rebooted due to display driver change.

You could then log into the graphical environment.

The only thing that now confuse me a little is the following:

Code: Select all

[root@localhost ~]# dkms status
nvidia-current, 535.129.03-1.mga9.nonfree, 6.5.11-desktop-5.mga9, x86_64: installed
nvidia-current, 535.129.03-1.mga9.nonfree, 5.15.126-desktop-1.mga8, x86_64: installed
nvidia-current, 535.129.03-1.mga9.nonfree, 5.15.122-desktop-1.mga8, x86_64: installed


In the list I had expected to see kernel-desktop-6.5.13-2.mga9 as builed ?

Thanks again for your time both of you !

The remaning for me is to sum up the case after which it comes in SOLVED.

Best regards
Anders

[morgano]

You forgot to install kernel-desktop-devel-6.5.13-2.mga9 with the kernel.
Because of that it the kenrnel module was not built.
When booting that kernel the system detected the missing module and switched from using nividia to use nouveau.

So:
Open a terminal and issue "urpmi kernel-desktop-devel-6.5.13-2.mga9". Then keep terminal open and issue "journalctl -f" and let it run so you see next step operating:
Use MCC/drakx11 to switch to nvidia-current (listed as Driver: Production, same as GeForce 745 and later)
In terminal you should see the module getting built.
Reboot.

[doktor5000]

As kernel-desktop-devel-latest is installed that should have been automatically installed.

[anp66]

Hi Morgano

Tanks for your explanation !

what do you mean with:
Use MCC/drakx11 to switch to nvidia-current (listed as Driver: Production, same as GeForce 745 and later)

In
MCC.Hardware.set_up_the_graphical_server
graphical_card: custom = Nvidia GeForce 745 series and later

In
MCC.Boot.set_up_display_manager
x GDM (GNOME Display Manager) <--- it is the chosen one (Not Harry Potter)
XDM (X Display Manager)

I can't find anything about drakx11 in MCC

Best regards
Anders

[morgano]

Ah, sorry for the in-house wording. Several names mean the same thing.

The "set_up_the_graphical_server" icon actually start drakx11 which is also the same as XFdrake - you can launch it like that from terminal without starting MCC if you like.

So, first alternative.

[anp66]

Hi again Morgano

Just to be sure, the following 2 packages were installed:

Code: Select all

urpmi kernel-desktop-devel-latest-6.5.13-2.mga9
urpmi kernel-desktop-latest-6.5.13-2.mga9


Then I runt the drakx11 and chose graphic card : GeForce 745 and later

while I was doing that, another window was used to keep an eye on what was happening using: 'journalctl -f'.
I didn't spot any errors and unfortunately no 'binds' either.

Then this was done:

Code: Select all

[root@localhost anders]# remove-old-kernels

 System: Mageia release 9 (Official) for x86_64 | Kernels in /boot/:4 | AUTO:1 | KEEP:3 
 ==> kernel-desktop
 1   : Keep  : U   : kernel-desktop-6.5.13-2.mga9-1-1.mga9.x86_64  2023-12-11T16:40:19 CET
 2   : Keep  :     : kernel-desktop-6.5.11-5.mga9.x86_64           2023-11-29T10:19:17 CET
 3   : Keep  :     : kernel-desktop-5.15.126-1.mga8-1-1.mga8.x86_64 2023-08-25T23:22:24 CEST
 4   : Remove:     : kernel-desktop-5.15.122-1.mga8-1-1.mga8.x86_64 2023-07-27T03:17:06 CEST
 ==> kernel-desktop-devel
 1   : Keep  :     : kernel-desktop-devel-6.5.13-2.mga9-1-1.mga9.x86_64 2023-12-13T08:47:30 CET
 2   : Keep  :     : kernel-desktop-devel-6.5.11-5.mga9.x86_64     2023-11-29T10:19:19 CET
 3   : Keep  :     : kernel-desktop-devel-5.15.126-1.mga8-1-1.mga8.x86_64 2023-08-25T23:22:10 CEST
 4   : Remove:     : kernel-desktop-devel-5.15.122-1.mga8-1-1.mga8.x86_64 2023-07-27T03:16:52 CEST
               U = In use now
Remove 2 (1 kernel:) ? y/N/i (i=confirm for each) y

removing kernel-desktop-devel-5.15.122-1.mga8-1-1.mga8.x86_64
removing package kernel-desktop-devel-5.15.122-1.mga8-1-1.mga8.x86_64
      1/1: removing kernel-desktop-devel-5.15.122-1.mga8-1-1.mga8.x86_64
                                 ###########################################################################################################################################################################################################
removing kernel-desktop-5.15.122-1.mga8-1-1.mga8.x86_64
removing package kernel-desktop-5.15.122-1.mga8-1-1.mga8.x86_64
      1/1: removing kernel-desktop-5.15.122-1.mga8-1-1.mga8.x86_64
                                 ###########################################################################################################################################################################################################
 Gain:278 MB - Kernels in /boot/: 3


I am a happy man because kernel-desktop-6.5.13-2.mga9-1-1.mga9.x86_64 is 'In use'.

So this was done:

Code: Select all

[root@localhost anders]# dkms status
nvidia-current, 535.129.03-1.mga9.nonfree, 6.5.11-desktop-5.mga9, x86_64: installed
nvidia-current, 535.129.03-1.mga9.nonfree, 5.15.126-desktop-1.mga8, x86_64: installed
nvidia-current, 535.129.03-1.mga9.nonfree, 6.5.13-desktop-2.mga9, x86_64: installed

[root@localhost anders]# rpm -qa | grep kernel-desktop | sort
kernel-desktop-5.15.126-1.mga8-1-1.mga8
kernel-desktop-6.5.11-5.mga9
kernel-desktop-6.5.13-2.mga9-1-1.mga9
kernel-desktop-devel-5.15.126-1.mga8-1-1.mga8
kernel-desktop-devel-6.5.11-5.mga9
kernel-desktop-devel-6.5.13-2.mga9-1-1.mga9
kernel-desktop-devel-latest-6.5.13-2.mga9
kernel-desktop-latest-6.5.13-2.mga9


6.5.13-desktop-2.mga9, x86_64: installed, is unfortunately not in status 'built'. I don't know how a 'build' can be provoked.

However, you don't have to spend any more time on the problem. You have been a great help and I have become a little wiser.

Is in a better place than when I started. Now I can log into the graphical environment again. I'm not sure which graphics driver is used, but I've learned to clean up old kernels and also know how to switch to 'init 3' from recovery mode.

So again 1000 thanks for your help and time.

Best regards
Anders

[isadora]

Would you be so kind, to mark the topic solved.
You can do so, by writing [SOLVED] leftwards of the topic-title in your first message.
Thanks ahead.