I’ve had this one going since about January. I tried to get Nvidia Driver working back in February and gave up. In October I logged in an updated and recently my Fedora went down. I put 440.31.run on ext4 usb drive and loaded it and it worked! I had Vulkan somehow, but with only two layers, but they were from 1.1.127? I played War Thunder on Vulkan and game was very responsive and clear to play on Ground battles better than ever! I was killing. Air battles was a little more choppy and needed work. It was good for 3 days…
The I downloaded games develop bundle which gave me 9 vulkan layers but from 1.1.121 and not 127? I don’t know how those got there, but I had 'em maybe lutris. Got up to kernel 5.3.10-865 . I had Opera, just complied ffmpeg and downloaded game-develop bundle, then I updated. There were some errors. It mentioned doing something with -verify or --fix …? I do not remember, but the update was successful none the less. Upon reboot there is miscellaneous error. Close x server (5.3.10-867) Is there a way to undo the update or how do I fix? There is no Clear linux boot loader so I can’t back date the kernel. I have tried reloading and unloading the driver. and a different one … 430.64
Hi, what do you mean by
then I updated. There were some errors. It mentioned doing something with -verify or --fix …?
When do you see the error message? Was that when you update something? And what were you updating, Clear Linux or NVIDIA Driver?
There is no Clear linux boot loader
Is clr-boot-manager not present on your system?
I did a swupd bundle-add clr_boot_manager and it says not valid , skippin, failed to install 1 bundle. As far as the -verify and – fix. Those are parts of code the terminal was spitting out as I was updating just before reboot. There was an error with /var/…/…etc… like two or three. And said something about doing a swupd update -verify --fix {something along those lines I don’t remember and I didn’t do it. It said update successful.}
it may have been /var/lib/libgl64.so? thats not it but I remmeber some .so files maybe a lib or two perhaps gl? little shady on the gl part.
As far as errors go. that was all back , when gnome was still up and I had successful xserver before reboot. Now it blinks a couple of times and I get the underscore on the top right of screen. I hit control+f2 and log into console. I hit startx And I get something similar to the title of the thread. This is also when i just tried to add clr_boot_manager bundle.
Those are parts of code the terminal was spitting out as I was updating just before reboot.
Update is a transitive verb and what were you updating?
If you were updating the NVIDIA driver, how did you perform the update?
As far as errors go. that was all back
Suppose you mean black, it’s happened to me when the driver wasn’t installed correctly or when a kernel update made the driver incompatible. I suggest to uninstall the driver and see whether you are able to see the graphical desktop environment.
update as in general swupd update
I hadn’t planned on updating Nvidia driver until the next one was out, but I was planning on updating CUDA and reinstalling nvidia driver sometime in the near future.
Then just try to uninstall the NVIDIA driver.
I am unable to do this as RTX is an unkown arch to clear linux kernel. It’s weird. that is how I had it and it just loads into a console. It’s funny I can load an live session from usb stick and see the desktop, but can’t install it. it won’t let me. Not compatible. I installed this one with the old installer in January that didn’t care about things like that.
I am unable to do this as RTX is an unkown arch to clear linux kernel. It’s weird. that is how I had it and it just loads into a console. It’s funny I can load an live session from usb stick and see the desktop, but can’t install it. it won’t let me. Not compatible. I installed this one with the old installer in January that didn’t care about things like that.
The live session is 5.3.7 and i had that because i thought the one I had loaded from January was a bust. it wouldn’t let me installit . It doesn’t like the RTX 2080. Loading a Nvidia driver on the one I had going from January actually worked. Little did present me know that past me had already done all the work and set it up.
If you see unknown arch on a black screen when you boot up, it might not be related to the GPU. I’ve seen that and it’s likely solvable by modifying some boot parameters.
I’d say do this, uninstall the NVIDIA driver and use LTS kernel instead of native kernel for now.
Is lts kvm? Because it says kvm is disabled by bios. I am wondering how to enable it in bios. There is nothing about kvm in my HP bios. I know it’s the gpu as those are the only pci regions there are and it says unknown nvidia architecture. So specifically Turing because it was just fine with my pascal gtx 1050 ti a year ago using open source.
There are some errors that you can safely ignore, but I have seen this very error you mentioned, unknown arch. Actually I’m still having this error message.
What’s more important is the cause of that error message and it’d help if you provide more detailed information on the error message you see, including the context.
But just first to uninstall the driver, and use LTS kernel.
I have the nvidia driver uninstalled and rebooted ( not necessary, but did it anyway). I am confused as to about how to get on lts kernel as clr_boot_manager is invalid and I don’t have it.
I get like 6 errors all related around two problems.
none of them big deal
unable to load kvm
kvm disabled by bios
uknown nv arch
etc… basically kvm is disabled by bios and it doesn’t understand turing
Note that you never spelled it right 
It’s clr-boot-manager.
Then for those BIOS related erros, first try to update the BIOS.
It still says invalid with - replacing _
Oh so KVM is for virtual machines. Whatever don’t need that. yet…
LTS kernel can be installed via swupd.
clr-boot-manager is responsible for updating kernels and I cannot understand why this is not present in your /usr/bin.
It won’t let me add it either with swupd bundle-add command either. I added kernel-lts and got a bunch of errors about updating bios and not being able to update it to bios and wanting me to manually ad it to the bios. With iniramfs. It crashed back into the 5.3.10-867 kernel. I did flash a new rom to my hp z 620 bios went from 3.94 - 3.95. Ut was succesful. Still can’t ad clr-boot -manager though. Maybe I can wait this out. Wait for Nvidia to come out with new driver, or maybe Clear Linux will get past this with updates as before. It didn’t work back in Februaury, when I tried before this last attempt at Clear Linux Workstation. Pissed that I don’t get a normal clear linux boot manager. It’s getting late for me today…
Since Fedora crashed I have been getting all these bios error messages too. Like a qpi link is dead and a few other ones sometimes if I have to use the button to restart. i don’t know what a qpi link is. Maybe there is something deeply wrong with the mother board that Fedora caused… Clear Linux still worked though for the last 3 days from up to kernel 5.3.10-865. Fedora crashed on Saturday after updateing CUDA to 10.1 update 2 from just CUDA 10. It didn’t really crash. I entered shutdown and next time I went to start I started getting all these problems and can’t even open an emergency shell and switched over to Clear linux.
Still would like to get a boot loader going…
clr-boot-manager is provided by the bootloader bundle, which is automatically included by the kernel-native, kernel-lts, and pretty much all other kernel-* bundles. Is it really not present in /usr/bin?
It is on the cloud right? But somewhere something got messed up and it just automatically loads into latest kernel native. I’ll check it tomorrow, Friday; for that.
What do you mean by ‘on the cloud’?
What is on the cloud’? And what is the cloud you’re referring to?
The binary clr-boot-manager should be installed in /usr/bin directory in most cases. Having a Clear Linux system installed without clr-boot-manager is rather hard.