Dear Team,

Since a few weeks I have installed ubuntu 18.04 desktop on a workstation with nvidia Tesla GPU installed. However, I am stuck in a bootloop and can't login to the OS.
After adding 'nomodeset' in GRUB, I am able to boot the OS with the message below several times on startup:

PKCS#7 signature not signed with a trusted key

dmesg logs below:

dmesg -T | grep nvidia
[Fri Jan 18 17:21:11 2019] nvidia: loading out-of-tree module taints kernel.
[Fri Jan 18 17:21:11 2019] nvidia: module license 'NVIDIA' taints kernel.
[Fri Jan 18 17:21:11 2019] nvidia: module verification failed: signature and/or required key missing - tainting kernel


dmesg -T | grep -C 3 PKCS
[Fri Jan 18 17:21:11 2019] PKCS#7 signature not signed with a trusted key
[Fri Jan 18 17:21:11 2019] nvidia: loading out-of-tree module taints kernel.
[Fri Jan 18 17:21:11 2019] nvidia: module license 'NVIDIA' taints kernel.
[Fri Jan 18 17:21:11 2019] Disabling lock debugging due to kernel taint


I have seen several posts whereby a key must be created and trusted via mokutil but I am not convinced about this as it have never encountered this issue before.

Can anyone advise please?

i think the proprietary nvidia driver will always taint the kernel; maybe these messages are not the problem? they aren't even marked as errors. it even says "- tainting kernel" which means: "i am going ahead and do this now".

a bootloop?
i have never, ever heard of a bootloop on a Linux system.

Please clarify, give us a better problem description.
read the first link in my signature.

1 members found this post helpful.

Hi,

The Ubuntu startup screen was stuck on blank screen and never booted up (not a boot loop - thank you for correcting me!)

I booted into recovery mode this time, to see all the messages in the startup and I found the culprit (I don't have a screenshot as the message is no longer in syslog):

"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
nvidia-modeset xxxx

Indeed you were right. The PKCS was not the problem this whole time, it was a kernel issue with the Nvidia driver (I'm using nvidia-410)

I did three things:

1) Changed the value to 0 in /proc/sys/kernel/hung_task_timeout_secs
Noticed a server hang after a few hours of operation

2) Changed the drivers - all nvidia drivers from ppa were giving the same error message on boot

3) Changed the sysctl values as below:

$ sudo sysctl -w vm.dirty_ratio=10
$ sudo sysctl -w vm.dirty_background_ratio=5
$ sudo sysctl -p

I referred the link below:

https://www.blackmoreops.com/2014/09...conds-problem/

The server is working ok till now - I will monitor the performance to make it permanent if this is really the fix.


Anyhow, I'd like to understand the issue in depth.

I hope if is clearer now. Please let me know what I can check to gather more information.

Many thanks in advance

Hey!

Just wanted to post this in case anyone else is having the same issue.
I tried a lot of tweaking but the OS was never stable.

So I tried one thing: reverting to an older kernel and the issue was fixed!
Hope this helps someone out there

Your best bet would have been to follow the first suggestion in this tutorial ---> Automatic Install using standard Ubuntu Repository
The option you chose, #2, states: "Using graphics-drivers PPA repository allows us to install bleeding edge Nvidia beta drivers at the risk of unstable system".

Because Nvidia Tesla release was 13 years ago, the best drivers would have been the proven stable, recommended drivers available from standard Ubuntu repositories. The "bleeding edge" 410 drivers did not add any hardware components to your 13 year old graphics device and as such... did not give your old graphics device any extra functionality over the stable driver, it only succeeded in giving you what is highlighted above in red. You could be running the latest stable Ubuntu kernel for Bionic Beaver.

Hi Brains,

Thank you for reverting back and also sending the tutorial link.
I did install via the PPA, chose the recommended drivers and sadly still got the kernel instability issues.

The solution which fixed my issue in my current config, in case anyone is pulling off his/her hair in the same situation:

'Revert back to a previous kernel!'

Hi,
I'm also facing this situation now. I have spent a lot time on checking the secure boot( annoying Asus ROG motherboard) and I'm pretty sure that it's disabled in both mokutil and (UEFI)BIOS ( these two ways seem different, but I'm not sure). I'm now trying to follow your idea to install a Ubuntu 16.04.....The failure environment is Ubuntu 18.04(kernel 5.0.0-23), nvidia 440, RTX2080TI.
Hope this will help someone like me

Thank you for saying this, but if your problem continues you should really start a new thread with a dedicated problem description.
Also I'm not sure if using Ubuntu 16.04 is still good advice, or for how long.

Oh, I'm not quite familiar with the rules...Just leave it here, can I? I think they are similar failure situation, and will be searched using same keywords.
And I have solved my problem, which may be not suitable for others. It's a problem on GPU hardware where I always get "RMinitadapterfailed" when booting after installing Nvidia driver. Post the error code: "1227","1155" in ubuntu, "43" in win10. Failure occurs when start GPU after successfully installing driver

Hello tyz,

happy new year

Just to clarify - the solution that worked for me for my current config was to revert the existing Ubuntu 18.04.xx installation to an older kernel.
You can easily achieve that by choosing an older kernel from the GRUB menu or there third party tools (e.g Ukuu) that help you achieve this easily.

Hope this helps