Setting up Qubes only with a CPU intergrated graphics(Urgent as GPU died.)

I originally had my GPU hidden from dom0 for passthrough and had my CPU handle the graphics.
My GPU died and when I tried to boot I was getting plymouth waiting issues. I tried to adjust the grub and even disable plymouth for what I assume is used for the Qubes graphical image on boot and still didnt proceed.

I had to ufnortunately reinstall Qubes(I do have my backups ofc) but now I get some nvme1 waiting issue.
Ive set my CPU internal graphics to be prioritzed in bios and still cant get past this issue, Ive even tried different nvmes(one at the time) in the system and saw the same issue.
Trying to the rescue option didnt help, I ofc erased the drive and tried the install again and still same error.

I need to install qubes and do so only using my CPU.
Any help would be apprciated as is urgent.

Your hardware listed at HCL?
Hardware compatibility list (HCL) | Qubes OS , check troubleshoot/remark notes.

Im using the Ryzen 9 7950X, I wouldnt see why it wouldnt be compatible for use as Ive been using it with Qubes in this configuration for nearly two years.
The primary issue is I need to install Qubes only using the CPU iGPU as previously I did so with a GPU then configured to use the iGPU.

Double check SVM in bios, does it on?

Yes

May you share photo or screenshot of ā€œnvme1 waiting issueā€? or at less quote it exactly.

Unfortunately, I think these integrated GPU are too recent for the fedora 37 in dom0, I’ve read many reports of display issue with Ryzen 7xxx X CPU :confused:

A solution could be to try installing 4.3-RC1 but it’s a release candidate, here be dragons… Or replace the GPU with a cheap one good enough for display.

1 Like

Before try 4.3.0-rc1, think that he should try installation with kernel latest.

2 Likes

Absolutely

When I had a problem like this, it was blacklisting every GPU driver module that worked. Normally it should work for all hardware.

If it is only a problem if the installer, then it can be temporary - only during installation. I think I used vesafb for months and it was OK.

Unfortunately, I don’t remember exactly what kernel command line changes are required to completely prevent loading of radeon and amdgpu modules… but someone here can surely help, if it seems useful.

Ive bought a 3050 and 3060, trying both qubes 4.3 and 4.2 and still ending up ultimately with the same errors.

The error it is always leading to:

dracut-initqueue with different warnings such as:

dracut-initqueue: stating timeoutscripts
Anaconda installer errors begin #######
It seems that the boot has failed. Possible causes include:
missing inst.stage2 or inst.repo boot parameters on the kernel cmdline. Please verify thayou have specified inst.stage2 or inst.repo,
Please also note that the ā€˜inst.’ prefix is now mandatory.

Installer errors ecountered during boot:

dracut/hooks/imitqueue/timeout/50-anaconda-error-reporting.sh:could not find
Anaconda installer error end.

This is with 4.3 and 4.2(.4?) for 4.2 after seeing cancel-multipath-wait-nvme0n1 and 4.3 has many more errors.

Ive tried adding after quiet inst.stage2=hd:LABEL=QUBES-(etc) and didnt resolve. and trying to troubleshoot with AI doesnt help much either.

I really need to use Qubes again.

Try to boot with nomodeset kernel command line option.

Did not work unfortunately

1 Like

Does it enter emergency shell at the end, where you could check/copy the boot log?
Maybe you can find errors that cause this issue there.
You can also remove quiet to see more log output.

Could you advise how to do so?
The answer is yes and I assume I will need to manually mount the usb and then copy the logs to it? Never done something like this before.

You’ll need to mount USB disk that you want to copy logs to, for example, if your USB disk partition is /dev/sdb1, then run these commands:

mkdir /mnt/usb
mount /dev/sdb1 /mnt/usb

You can find your USB disk using blkid or fdisk -l commands.
Then you can copy logs there:

mkdir /mnt/usb/logs
cp /run/initramfs/rdsosreport.txt /mnt/usb/logs
journalctl -a > /mnt/usb/logs/journalctl.txt
dmesg > /mnt/usb/logs/dmesg.txt
cp /run/initramfs/init.log /mnt/usb/logs
cp -r /tmp /mnt/usb/logs

I am not able to upload the the txt files and even trying to copy past each log as individual replies is to long

I can email the logs if this is something possible for you

Hi, you can view my logs from my proton drive. Thanks for assisting.

I don’t see anything useful in the logs.
Did you run ā€œTest media and install Qubes OSā€ GRUB menu boot entry to make sure that your media is not corrupted?
If you made sure that media is not corrupted, then try to boot Qubes OS 4.2.4 installer, edit the ā€œInstall Qubes OS R4.2.4 using kernel-latestā€ GRUB menu entry and change quiet to nomodeset at the end of the first line starting with module2.
Upload the logs if it’ll fail to boot.

Hi,

Yes Ive tried media test, Ive also done mutiple reformat of 4.2.4 and 4.3 as precaution prior.(even tried multiple different usbs)
Not sure why but removing quite with nomodeset prevents me from mounting the usb, unable to get new logs with this change.
I previously simply added nomodeset after quiet, eitherway still results in same issue.

Cant we get someone from the team to advise? Im an active user of Qubes for close to two years no, no issue before with installing since my GPU died Ive been getting this same constant error and issue with boot, bought a 3050 and 3060 and its hasnt helped.