Join Log in
Score:5
Athena avatar Ubuntu

After Kernel 5.11 update, computer won't boot? Related to nVidia drivers. (Kernel 5.11.0.25.27~20.04.10)

id flag
Athena

Almost solved - Please help!

I thought earlier it was due to installed i386 architecture, when it in fact was because of the newest Ubuntu 20.04.10 Linux Kernel 5.11.0.25.27

    $ dpkg --list | grep linux-image
ii  linux-image-5.11.0-25-generic              5.11.0-25.27~20.04.1                  amd64        Signed kernel image generic
rc  linux-image-5.4.0-42-generic               5.4.0-42.46                           amd64        Signed kernel image generic
rc  linux-image-5.4.0-52-generic               5.4.0-52.57                           amd64        Signed kernel image generic
rc  linux-image-5.4.0-53-generic               5.4.0-53.59                           amd64        Signed kernel image generic
ii  linux-image-5.8.0-63-generic               5.8.0-63.71~20.04.1                   amd64        Signed kernel image generic
ii  linux-image-generic-hwe-20.04              5.11.0.25.27~20.04.10                 amd64        Generic Linux kernel image

It is the last one that apparently is causing the problem on my HP Pavilion dv7-3130eo. I was able to boot in the 5.8.0-63-generic Kernel, and now trying to find a way to remove the newest Kernel.

I tried running this:

    sudo apt-get remove "linux-image-5.11*"
Reading package lists... Done
Building dependency tree       
Reading state information... Done
Note, selecting 'linux-image-5.11.0-1009-gcp' for glob 'linux-image-5.11*'
Note, selecting 'linux-image-5.11.0-1007-azure' for glob 'linux-image-5.11*'
Note, selecting 'linux-image-5.11.0-1008-oracle' for glob 'linux-image-5.11*'
Note, selecting 'linux-image-5.11.0-25-generic' for glob 'linux-image-5.11*'
Note, selecting 'linux-image-5.11.0-1013-oracle' for glob 'linux-image-5.11*'
Note, selecting 'linux-image-5.11.0-22-lowlatency' for glob 'linux-image-5.11*'
Note, selecting 'linux-image-5.11.0-1014-aws' for glob 'linux-image-5.11*'
Note, selecting 'linux-image-5.11.0-1014-gcp' for glob 'linux-image-5.11*'
Note, selecting 'linux-image-5.11.0-1012-azure' for glob 'linux-image-5.11*'
Note, selecting 'linux-image-5.11.0-22-generic' for glob 'linux-image-5.11*'
Note, selecting 'linux-image-5.11.0-1009-aws' for glob 'linux-image-5.11*'
Note, selecting 'linux-image-5.11.0-25-lowlatency' for glob 'linux-image-5.11*'
Package 'linux-image-5.11.0-1007-azure' is not installed, so not removed
Package 'linux-image-5.11.0-1008-oracle' is not installed, so not removed
Package 'linux-image-5.11.0-1009-aws' is not installed, so not removed
Package 'linux-image-5.11.0-1009-gcp' is not installed, so not removed
Package 'linux-image-5.11.0-1012-azure' is not installed, so not removed
Package 'linux-image-5.11.0-1013-oracle' is not installed, so not removed
Package 'linux-image-5.11.0-1014-aws' is not installed, so not removed
Package 'linux-image-5.11.0-1014-gcp' is not installed, so not removed
Package 'linux-image-5.11.0-22-generic' is not installed, so not removed
Package 'linux-image-5.11.0-22-lowlatency' is not installed, so not removed
Package 'linux-image-5.11.0-25-lowlatency' is not installed, so not removed
The following packages were automatically installed and are no longer required:
  amd64-microcode intel-microcode iucode-tool linux-headers-generic-hwe-20.04 thermald
Use 'sudo apt autoremove' to remove them.
The following additional packages will be installed:
  linux-image-unsigned-5.11.0-25-generic
Suggested packages:
  fdutils linux-doc | linux-hwe-5.11-source-5.11.0 linux-hwe-5.11-tools linux-modules-extra-5.11.0-25-generic
The following packages will be REMOVED:
  linux-generic-hwe-20.04 linux-image-5.11.0-25-generic linux-image-generic-hwe-20.04 linux-modules-extra-5.11.0-25-generic
The following NEW packages will be installed:
  linux-image-unsigned-5.11.0-25-generic

Not sure if, I should press "yes".. or find another way to just completely uninstall the newest version? Please do help me with this question? How to exactly remove or deactivate the problematic Kernel and make sure that the Kernel that works is being kept and protected?

Bug report about the Ubuntu 20.04.2 LTS kernel 5.11.0-25

Similar Post: 20.04.2 won't boot after linux-image-5.11.0-22-generic:amd64 got auto installed

It seems that it has something to do with the Kernel update paired with my Nvidia drivers (that came with the Ubuntu system).
Here are some similar experiences, but these threads don't provide any clear directions how to solve it, other than removing the Kernel, but how does one do that in a proper way?:

updating kernel 5.8 to 5.11 ubuntu 20.1 wont boot

NVIDIA drivers don't work after kernel update to 5.11

OLD SOLVED: This was not the real cause of the problem! With the help of good people in here, I was guided how to enter Recovery Mode, where I decided to try to remove i386 architecture that I had installed some months ago. I ran the following command: sudo apt-get remove libc6:i386.

My computer has Nvidia Geforce, and is 64bit Ubuntu. It might be that some new updates made me unable to boot. I had previously installed i386 architecture in order to run a computer game (Enemy Territory), but it didn't work. Now I removed i386, and my computer works again. Thank you everyone!

The Old Post: Description of first experiences with the inability to boot the computer, without knowing what caused the problem:

My HP Pavilion dv7-3130eo computer suddenly would not boot. Everything loads, even the Ubuntu logo shows up. I press "ESC" to follow the terminal output. I am only able to quickly glance: GRUB failed boot detection. (confirmed by checking the boot.log1) At some point the screen becomes black. The only visible element is a:

_

In the upper left corner. The screen remains stuck like that. When I click on the power button, it shuts down by itself after a few minutes. There is also

What I tried to do: First I was able to boot from a new Ubuntu-Installation USB key, I did this many times. Here i first tried to mount the harddisk and removed some files. I quickly found out that the harddisk was far from full. It still could not start.

Then I installed the Boot-repair packages while in the USB-Boot, and tried to repair the boot. When trying to do that, am met with this message: The current session is in BIOS-compatibility mode. Please disable BIOS-compatibility/CSM/Legacy mode in your UEFI firmware.

I then restart and enter BIOS to see how to enable UEFI. The BIOS however does not give any such options, overall it seems to be a very simple or old version of BIOS.

I restart and boot again from the USB, where I use the Boot-repair packages to run a test. This is the full pastebin output: Pastebin of Boot-repair test

This is what I noticed in the pastebin:

=============================== StdErr Messages ================================

File descriptor 63 (pipe:[76976]) leaked on lvs invocation. Parent PID 22927: /bin/bash

Suggested repair: ______________________________________________________________

The default repair of the Boot-Repair utility would purge (in order to fix packages) and reinstall the grub-efi-amd64-signed of
sda5,
using the following options:        sda1/boot/efi,
Additional repair would be performed: unhide-bootmenu-10s  use-standard-efi-file    

Blockers in case of suggested repair: __________________________________________

 The current session is in BIOS-compatibility mode. Please disable BIOS-compatibility/CSM/Legacy mode in your UEFI firmware, and use this software from a live-CD (or live-USB) that is compatible with UEFI booting mode. For example, use a live-USB of Boot-Repair-Disk-64bit (www.sourceforge.net/p/boot-repair-cd), after making sure your BIOS is set up to boot USB in EFI mode. This will enable this feature.

Confirmation request before suggested repair: __________________________________

The boot of your PC is in BIOS-compatibility/CSM/Legacy mode. You may want to retry after changing it to EFI mode.
Alternatively, you may want to retry after deactivating the [Separate /usr partition:] option.
Are you sure you want to continue anyway?

Final advice in case of suggested repair: ______________________________________


Please do not forget to make your UEFI firmware boot on the Ubuntu 20.04.2 LTS entry (sda1/efi/****/shim****.efi (**** will be updated in the final message) file) !
The boot of your PC is in BIOS-compatibility/CSM/Legacy mode. You may want to retry after changing it to UEFI mode.

I search the web to find out how to enable the UEFI. As I am not able to find any immediate solution, I try to the boot-repair again, this time I find a box in the advanced options, that ignores the UEFI. I try to run the repair, and it is succesful. This is the box i unticked: seperate boot/efi partion

This is the pastebin the boot-repair repair gives me: Repair pastebin

I try to restart the computer. But nothing has changed. It still won't boot.

Do anyone have any clues or suggetions? All kind of help would be highly appreciated!

To clarify: The problem still remains the same. I am not able to succesfully load Ubuntu, half way through it stops and a black screen with a _ sign appears in the upper left corner.

In BIOS, it says I am running BIOS Version F.1A InsydeG20 Setup Utility. Rev. 3.5 The computer has a Nvidia Geforce graphic card.

Thank you for your help

Sincerely Athena

3411
1 + 19
ChanganAuto avatar
us flag
ChanganAuto
Unfortunately, like some many other before, you don't understand that UEFI isn't a mode/feature. UEFI IS the firmware that replaced the old BIOS firmware many years ago. Although many users and vendors wrongly use "BIOS" to this day, that ain't make it so. Actually if you have UEFI you don't have BIOS and vice-versa. And to complicate thing even further most UEFIs still to this day have a special "BIOS-compatibility mode" (AKA "CSM" AKA "Legacy") that sort of emulates the behavior of the old 1981 BIOS. The warning is actually telling you to **disable** that mode because you have UEFI and ...
0
Reply
ChanganAuto avatar
us flag
ChanganAuto
... the OS has been correctly installed in UEFI mode. But because CSM is enabled then most external (bootable) media can boot in either mode. How it boots is how it installs and also how Boot-Repair will or won't be able to do something as it must boot in the same mode of the installed OS(es).
0
Reply
Athena avatar
id flag
Athena
Thank you for your reply. It is very helpful! How would you suggest me to disable that mode?
0
Reply
ChanganAuto avatar
us flag
ChanganAuto
That's entirely up to you and the firmware settings. Those are different depending on brand/model and firmware version.
0
Reply
Athena avatar
id flag
Athena
@ChanganAuto Thank you. I have provided the details of my computer and the Firmware, according to what I could find out. Is there a way to find more details about my firmware? Is there a way to use the Ubuntu USB to configure the UEFI, so that I can update GRUB properly (Or whatever is necessary to make my computer boot up as normal again?) Do you have any suggestions? Thank you! Athena :-)
0
Reply
ChanganAuto avatar
us flag
ChanganAuto
Even with CSM enabled the worse that happens is having two entries for the same bootable external media, one explicitly mentioning "UEFI" and that's the one you want to boot. However, if you did the USB with some tool like Rufus (Windows) the default settings are for BIOS mode (BIOS/MBR) and should be UEFI/GPT, it only boots in mode or the other. That said, your problem is UNLIKELY to be solved with Boot-Repair. This tool isn't an universal panacea, the best it can do is to re-install Grub. Your problem may have to do with file system corruption.
0
Reply
oldfred avatar
cn flag
oldfred
Your fstab shows the mount of an ESP - efi system partition, which means system is UEFI. But you have grub installed to MBR and are using MBR partitioning. UEFI strongly suggests using gpt partitioning with UEFI. Ubuntu does install in UEFI mode to MBR, but probably should not. Was it originally BIOS or UEFI?
0
Reply
Athena avatar
id flag
Athena
Thank you for all these valuable clues! @ChanganAuto I looked a bit into it. The current USB I am booting from, seems to have a filesystem partition that has EFI System in it. Do you suggest me to make a Rufus boot? Furthermore. I did disk test and disk repair, I also ran fsck, nothing indicates it is a file system corruption, or at least it doesn't seem to be able to fix it. Even if it is corruption to the particular boot-area. What can I do to restore it?
0
Reply
Athena avatar
id flag
Athena
The harddrive itself, seems to have a /dev/sda1 Filesystem partition with the partition type of: W95 FAT32 (Bootable). It is true that there are two other partitions: /dev/sda2/ with the partition type: Extended. And /dev/sda5/ with the partition type: Linux (Ext4) I wonder why Ubuntu is able load to a certain extend, but stops and freezes at some point. Is it really the GRUB that is failing or is it something else? @oldfred Thank you!
0
Reply
oldfred avatar
cn flag
oldfred
Can you press escape key right after UEFI/BIOS screen (if in UEFI boot mode) and get grub menu. With only one install menu does not normally show. And then can you choose the recovery boot second line or in submenu.
0
Reply
Athena avatar
id flag
Athena
Thank you @oldfred. Quick update, I found the boot.log and managed to find only one "fail" message, reproduced 10 times: GRUB failed boot detection I will what you told me to do, I tried something similar earlier. When I attempted last time this happened: I click ESC at the HP screen (This might be the EUFI/BIOS?. After clicking ESC there, a number of options appear, including entering BIOS, which has no use. But I don't believe the this is the GRUB menu. The only moment the GRUB menu appears is when I boot the USB.
0
Reply
Athena avatar
id flag
Athena
F1 System Information F2 System Diagnostics F9 Boot Device Options F10 BIOS Setup F11 System Recovery F12 Nework Boot ENTER - Continue Startup These are the options I get.
0
Reply
Athena avatar
id flag
Athena
@oldfred Following your instructions to press ESC after the UEFI/BIOS (HP) screen, I was able to get GRUB and select Recovery Mode. Here I was able to perform some changes. Including installing updates and autoremove. But it still did not allow me to boot. I then tried it again. Opened the sudo terminal. And decided to try to remove the i386 architecture I had previously installed some months ago. I did this by: sudo apt-get remove libc6:i386 IT WORKS NOW! I was able to boot. Thank you for your help!
0
Reply
Athena avatar
id flag
Athena
I found out it was due to the Kernels. But I am not sure if I should just remove the newest Kernel, and how to do that correctly? Or maybe the Kernel has to be updated? I tried running some things in the terminal and I have pasted my output in the above post. @oldfred
0
Reply
oldfred avatar
cn flag
oldfred
Is that the latest UEFI version? (They still call it BIOS). How did you install nVidia driver? Post this: `dkms status` If you installed nVidia from repository it should automatically update into new kernel. If you change nVidia driver you must totally purge old version or else you have conflicts and then bigger issues. nVidia install, purge if needed. https://ubuntuforums.org/showthread.php?t=2383560&p=13735336#post13735336
0
Reply
Athena avatar
id flag
Athena
I am now running the 5.8 Kernel, and it works. Here is the output for `dkms status`:`nvidia-340, 340.108, 5.8.0-63-generic, x86_64: installed` Do you think it is necessary to purge nVidia drivers? I think the drivers came with Ubuntu, I didn't do anything to install them. I also didn't change drivers, I believe (except if the i386 architecture I also installed months ago can be counted as that?) I am not sure how to find out what UEFI version I have, and if it is the latest. I earlier wrote what I could see when I opened BIOS. @oldfred Thank you for your help!
0
Reply
ar flag
user68186
Welcome to Ask Ubuntu. This is a question answer site. Please don't put SOLVED in the question title. And don't put the answer inside the question. You are most welcome to answer your own question. Accept your answer by clicking on the gray check mark ✔️ next to your answer and turn it green ✅. This will mark your question solved and help others.
0
Reply
Athena avatar
id flag
Athena
Thank you for your response @user68186 I would like to put the answer, but I still don't have the definite answer yet. So I am still waiting for others to continue helping. I thus wrote "partly solved", since the reason for the problem was discovered, but there was not given any clear guidance how to completely solve the problem. Thank you for your guide and help and understanding :-) Athena
0
Reply
Tim avatar
jp flag
Tim
I too have this problem when I upgraded from 5.11.0-25 to 5.11.0-31. I too have the nvidia drivers. I can't change to nouveau drivers... I am booting back into 0-25 for now. I tried to reinstall grub. It seems to be related to my wierd drive setup. I have a like 4 nvme drives and 4 hard drives. one nvme has windows and one with ubuntu. the others are data. I get a black screen like intramfs isn't loading or something right after grub. very annoying have to manually boot.
0
Reply
Score:1
avatar Ubuntu
cn flag
Petrika

I think I found a solution based on some other experience:

  • apply fix for 5.11 kernel:

    1. Boot in first available kernel that is working (for me it is 5.8.0.59)
    2. Add repository: sudo add-apt-repository ppa:kelebek333/nvidia-legacy
    3. Apply fix: sudo apt install xorg-modulepath-fix

Change from Nvidia to nouveau drivers.

Try to reboot with 5.11.

For me this was working combination.

Note: Definitely there is some bug with lib32gcc1 and libgc6xx and their incompatibility with this and later kernel updates. We have to wait for this fix for nvidia drivers.

0 + 1
Athena avatar
id flag
Athena
Thank you for this valuable solution. I havn't tried it yet, but I will do and I will post my result.
0
Reply
Score:0
Alexandru Razvan Miu avatar Ubuntu
ph flag
Alexandru Razvan Miu

I've had a similar issue when upgrading distro from 20.04 to 21.04, which came with a much newer kernel (5.8 to 5.11.. work PC has been sitting idle since the pandemic). Updating the gcc package fixed for me, it's the compiler that actually tries to run and configure the nvidia-driver package, and 5.11 apparently came with a new version which the distro upgrade didn't update, or at least wasn't configured to use. So a simple sudo apt install gcc fixed it for me.

0 + 0
ธงชาติไทย
Elon Musk
I sit in a Tesla and translated this thread with Ai:

mangohost

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.