I need help /dev/mapper/pve-root does not exist

[saintofdiablo]

Needing some assistance with getting promox to boot, even when attempting to add the rootdelay=10, it doesn't seem to save. Im constantly getting "Timed out for waiting the udev queue being empty' & ALERT /dev/mapper/pve-root does not exist. This issue only occurs after running the update repositories after a fresh install.

Any suggestions?

 

[saintofdiablo]

Not sure what else i can do, I've attempt to reinstall proxmox multiple times. Issue only happens after running the update.

 

[saintofdiablo]

Timed out for waiting the udev queue being empty
Timed out for waiting the udev queue being empty
Gave up waiting for root file system device
- Boot args (cat /proc/cmdline)
- Missing modules (cat /proc/modules; ls /dev)
ALERT /dev/mapper/pve-root does not exist. Dropping to shell!

 

[gfngfn256]

Not sure why you are getting this, but look here.

 

[saintofdiablo]

Can't boot into proxmox to make that change, unless i am able to edit that in initramfs.

 

[saintofdiablo]

After reinstalling proxmox, navigated to /dev/mapper the file exists. But after running the upgrade for the repositories, the file is gone afterwards.

 

[gfngfn256]

But after running the upgrade for the repositories

How are you upgrading? What command or GUI are you using?

 

[karsten.boehm]

I got a similar situaton: fresh installation boots fine. The vg pve is available and shown in the web and the cli tools. Setting up the update sources in the web frontend (no-subscription) and updating the system (using the web or the cli) results in a updates system that cannot boot anymore. Reason also being that the vg pve cannot be found at boot time.

Tried several times with a couple of homogenious systems (Intel Modular Server) on different physical chassis. Booting was not an issue before.

This is the situation with before update:


These are the partitions (it is multipathing system, configuration worked in the past fine), boot partition is not even multipathed:


In this situation I can boot, packages are still to be updated:

After I apply these updates, the system does not boot anymore as described.

 

[karsten.boehm]

One thing that caught my attention is the fact that I got these errors in the end of the update process, see below. The server still uses old BIOS boot and no EFI if I am not mistaken so this should not be an issue. But maybe this is a problem?



Anyway I do not know, what do to about it. Try now to reboot into the new kernel to see whether it can still boot.

 

[karsten.boehm]

After a lot of testing I found out that in my case the new kernel coming with 8.2.2 was the problem. During upgrade even a dist-upgrade (from 8.1) was done on some of the machines, the new kernel was installed and led to a non booting system.
The following procedure SOLVED the problem for me:

1. Boot the system and upgrade according to the usual procedure. I used the Web-Interface for that since I got sometimes a problem with installing grub on the right disk (it did not recognize the correct one automatically, but it was possible to modify this setting in a dialogue that was opened automatically when it detected that it cannot install grub correctly).

2. Open a shell and run pve-efiboot-tool kernel list to obtain all the installed kernels. In my case there was "6.8.4-2-pve" (the offending kernel) and "6.5.13-5-pve" (the one that was used before, also 6.5.11-8 worked). Then in run pve-efiboot-tool kernel pin 6.5.13-5-pve to set the 6.5 kernel as the standard kernel to boot.

3. Make sure that the setting was pinned correctly by running pve-efiboot-tool kernel list again. Then reboot. This is how the correct setting looked like on my machines:



The problem with the missing pve volume group was gone and the system booted normally.

I hope this might be helpful for other users running in a similar problem.

 

[fiona]

Hi,
just a wild guess, but the intel_iommu setting is now enabled by default in kernel 6.8: https://bugzilla.proxmox.com/show_bug.cgi?id=5423
You could try adding intel_iommu=off to the kernel command line and see if that helps: https://pve.proxmox.com/pve-docs/chapter-sysadmin.html#sysboot_edit_kernel_cmdline

 

[karsten.boehm]

Thank you, Fiona, for the idea! I will try that in the next days when I am "physically close" to the server and report about results.

 

[karsten.boehm]

Just as a feedback, I did a quick try on one of the machines but intel_iommu=off did not help, unfortuneately.

 

[fiona]

Just as a feedback, I did a quick try on one of the machines but intel_iommu=off did not help, unfortuneately.

That's unfortunate. Looking at the original error, can you check if you installed in legacy mode but then booted in EFI mode, like was the issue in this thread: https://forum.proxmox.com/threads/efi-system-partition.137288/

 

[karsten.boehm]

I was indeed installing in Legacy Mode and one of the things that I was supposing was that I uninentionally booted in EFI-mode, but this did not seem to be the case. The machine that could boot (the old kernel) are still in legacy mode, if I am not mistaken.


For now I do have a solution (using the 6.5 kernel instead of 6.8) and I do not see major drawbacks for the moment. I am sure this will be resolved sooner or later.