Hello Qubes Forum!
I have been troubleshooting trying to install Qubes 4.2 on my machine, which has already been running Q4.1 for over a year. I am not doing an in-place upgrade as I still need to utilize my main install while troubleshooting. I will be attempting an in-place after I successfully get an install running on a separate drive.
I am able to proceed through the Install, all the way up till the Initial Startup where it installs the template qubes etc. After installing and rebooting, I am brought to a screen that says
LightDM.service …
start hold until boot process finishes
This brought me to this thread here https://forum.qubes-os.org/t/start-hold-until-boot-process-finishes-blackscreen-or-reboot-after-decryption-post-install/15561
Apparently for them there is a PCI device interfering with startup in the sys-usb qube, and the fix is to skip the initial startup, remove the PCI device from sys-usb, and then manually start the initial startup again.
I tried to do this, however quitting the initial startup reboots me and stays still at a console window, manually turning off the machine and rebooting does not bring up initial startup again but instead leaves me stuck at the start hold until boot process finishes like before. I am not sure if I am doing this correctly, after quitting out of the startup are you supposed to reboot to get a GUI with no qubes? or remove the PCI device via console?
I then tried editing the grub config, removing QUIET and entering qubes.skip_autostart
Doing this, it is able to proceed past the start hold until boot process finishes. The console output I get looks like this
plymouth-quit-wait.service - hold until boot process finishes
Bluetooth: hci0: Waiting for firmware download to complete
Bluetooth: hci0: Firmware loaded into 1452964 usecs
Bluetooth: hci0: Waiting for device to boot
Bluetooth: hci0: Malformed MSFT vendor eventL 0x02
Bluetooth: hci0: Device booted in 15881 usecs
Bluetooth: hci0:Found Intel DDC parameters: intel/ibt-1040-0041.ddc
Bluetooth: hci0: Applying Intel DDC parameters completed
Bluetooth: hci0: Firmware timestamp 2023.35 buildtype 1 build 70976 kauditd_printk_skb: 204 callbacks suppressed
audit: type=1131 audit(1710129487 .548:232): pid=1 uid=0 auid=4294967295 ses=4294967295 msg=‘unit=systemd-rfkill comm=“systemd” exe=“/usr/lib/systemd/systemd” hostname=? addr=? terminal=? res=success’
I am not sure where to go from here. Any help would be appreciated
Hardware
Motherboard: MSI Z790 PRO DDR4
CPU: Intel I7-13700KF
RAM: 32GB DDR4
GFX: PNY RTX 4070 TI
Other Notes
For my Qubes 4.1 install, I have to install with kernel-latest or the GUI is slow and choppy, kernel-latest works as intended. I am using the same graphics card on Qubes 4.2.
I am using Dasharo custom bios, which has improved the install actually as it fixed other problems that were prevalent before flashing it.
I have tried installing with and without kernel-latest on Q4.2
I have tried removing ALL USB devices, which did not help anything
I am yet to try a different install media / destination, but do not see this as being the issue
I am thinking it is either a graphics card issue, or a bluetooth / PCI device issue but I am not sure.
You can try to skip creating sys-usb and sys-net in Initial Setup after installation to get into dom0 and then run Initial Setup in dom0 manually using this command in dom0 terminal:
Remove all attached devices from the qube and then add the devices one by one and try to start the qube. If the start failed after you add one specific device then this device is causing the issue.
I am having the same problem but cannot get further than the Grub> prompt.
How do I edit the grub.cfg file to add qubes.skip_autostart from this Grub prompt?
Also, it is a non gui boot which is just showing:
Xen hypervisor, version 4.14.6
Xen hypervisor, version 4.14.6.config
Selecting Xen hypervisor, version 4.14.6.config brings up
Qubes, with Xen 4.14.6 config and Linux 6.6.25-1.qubes.fc37.X86_64
Qubes, with Xen 4.14.6 config and Linux 6.1.75-1.qubes.fc32.X86_64
Qubes, with Xen 4.14.6 config and Linux 6.1.62-1.qubes.fc32.X86_64
All option leads to the same
Started Command Scheduler.
Starting Light Display Manager …
Starting Hold until boot process finishes up…
Any idea how to edit grub.cfg file from the Grub> prompt?
Quite frankly because I don’t know how. I am trying to support someone thd can only go by what I have been sent. See attached pictures.
At one point I was sent a picture which looked like it would accept qubes.skip_autostart but the user was unable to type anything in (lack of knowledge?) That’s the black picture with part_msdos
Does your friend have both dedicated GPU and integrated GPU?
I’d suggest to try and connect display to another GPU.
Most probably display is connected to dGPU (e.g. Nvidia) but Qubes OS is outputting image to iGPU (e.g. Intel).
Then to type mount to see what’s mounted. I believe this is only looking at the USB drive, so nothing on his internal drive. That needs to be mounted to edit the grub.cfg and add the skip…
Maybe he’ll need this to fix the issue with missing UEFI boot entry:
First you’ll need to figure out the Qubes OS disk name.
You can use:
fdisk -l
lsblk
blkid
To figure out the available disks and partitions.
The Qubes OS disk by default should have these partitions on its disk:
part1 512 MB - EFI System - /boot/efi
part2 1024 MB - ext4 - /boot
part3 the rest of the disk size - encrypted LUKS partition with rootfs
When you know the disk name then mount the /boot and /boot/efi like this, e.g. Qubes OS disk is /dev/sda:
mkdir /mnt/boot
mount /dev/sda2 /mnt/boot
mount /dev/sda1 /mnt/boot/efi
The GRUB config should be in /mnt/boot/efi/EFI/qubes/grub.cfg or in /mnt/boot/grub2/grub.cfg.
If you want to try to start Qubes OS without qubes autostart then you can add qubes.skip_autostart option for the kernel command line (at the end of the line starting with module2).
Looks like there’s an error. To bad you’re not close to Amsterdam, I’d bring it to you to fix it cause I’m getting inpatient and tired of trying. (timotatty@hotmail.com if you know someone)
We also saw this error after we got the Xen Hypervisor menu again and tried to edit after the quiet:
Is /dev/sda your Qubes OS disk?
It seems that it was installed in legacy BIOS boot mode and not in UEFI mode. There is no EFI System partition there and disk is using MBR.
Also, there is a size mismatch, did you dd from the old Qubes OS disk to this new one? Did Qubes OS work after writing the disk image to the new disk?
Maybe you’ve made an error in editing. Can you make a photo of the place where you’ve edited the config?
Can you describe your case in more details?
Is this a new Qubes OS installation or did you use this Qubes OS successfully before but it stopped working after something happened?
When did it stop working?
