[SOLVED] Efi sub loaded initrd from command line option - system cannot boot

Proxomx

New Member
Nov 25, 2021
17
1
3
35
Dear Community,

I made a simple mistake (thought I was writing in a container as root but it turned out it was proxmox host) and typed
Code:
 apt update && apt-dist upgrade
with
Code:
 apt autoremove
.

Right now after host reboot I am not able to boot, I get error ->
Code:
 efi stub loaded initrd from command line option

System is encrypted ZFS, installed on Nvme . Then my HDD storage is luks encrypted.

I am only able to boot into Debug mode from Proxmox ISO.

Current pve kernel is 5.13.19-1. and this is the only one available ( I checked with proxmox-boot-tool kernel list ).

When typed
Code:
 ls /boot/
there is missing file initrd.img-5.13.19-1-pve, the rest of flies (4) are there (system.map, config, vmlinuz, grub). I recreated initrd.img with command ->

Code:
 update-initramfs -c -k 5.13.19-1-pve

When I try command
Code:
 proxmox-boot-tool status
I get an error /etc/kernel/proxmox-boot-uuids doest not exist.

Trying to mount and init nvme ->
Code:
 proxmox-boot-tool init /dev/nvme0n1p3
but I get E: '/dev/nvme0n1p3" has wrong partition type (same error for my other disc).

Command
Code:
 efibootmgr
says EFI variables are not supported.

When I check
Code:
 ls /boot/grub/
I have only file unicode.pf2, the grub.conf is missing. So naturally update-grub gets failed to get canonical path of 'none'. I only have file /etc/default/grub.conf

Another command
Code:
 bootctl status
gives Couldn't find EFI system.... Not booted with EFI.


I tried to to reconfigure kernel ->
Code:
 dpkg-reconfigure pve-kernel-5.13.19-1-pve
and got message debconf: unable to initialize frontend: Dialog,............... Then package "pve-kernel-5.13.19-1-pve" is not installed and no information.

Would really appreciate any help / ideas how to solve this problem... How can I install fresh kernel through Debug Mode? What are my other options? This is disaster... :mad::eek:
 
Right now after host reboot I am not able to boot, I get error ->
Frist thing I would try is to boot the system without ISO - and in the boot-manager (either grub or systemd-boot - see https://pve.proxmox.com/pve-docs/chapter-sysadmin.html#sysboot_determine_bootloader_used) - edit the kernel commandline and remove the 'quiet'
Maybe this shows some messages which might show where the problem is.

Current pve kernel is 5.13.19-1. and this is the only one available ( I checked with proxmox-boot-tool kernel list ).
did you chroot into your actual system? - It sound a bit like you're running all those commands on the PVE-ISO environment and not the acutal system

check out the section on reparing a grub installation on the change wikipage:
https://pve.proxmox.com/wiki/ZFS:_S...iring_a_System_Stuck_in_the_GRUB_Rescue_Shell
(it contains the necessary commands for chrooting )

I hope this helps - else if possible share the logs (as screenshots if you cannot get them as text) - maybe we'll find a solution
 
  • Like
Reactions: Proxomx
@Stoiko Ivanov First of all thank you very much for response, much appreciated!! Honestly on Sunday morning I decided to recreated whole Proxmox host.
I couldn't find anything else to solve my problem - I stumbled upon 1st link you sent and saw also somewhere post about 'quiet' but I don't know why I didn't follow through...
2nd link I didn't see - thank you for pinpointing to it, hopefully someone else with similar problem stumbled upon here and it helps them in future :)

But since we are discussing - on the new Proxmox host I installed new kernel 5.15 but still get
Code:
 efi stub loaded initrd from command line option
when rebooting. But this time I can SSH into host (or type my password on boot screen) and everything follow as normal. Perhaps you know how should I remove this splash screen with Efi.... ?
 
But since we are discussing - on the new Proxmox host I installed new kernel 5.15 but still get
IIRC this message is normal and harmless - or do you mean that you don't see anything else after this message?

If this is the case please post:
* `cat /etc/kernel/cmdline`
* `cat /proc/cmdline`
* `proxmox-boot-tool status`
* `dmesg`
 
  • Like
Reactions: Proxomx
@Stoiko Ivanov That's exactly the problem - nothing follows after this message, it's stuck there :(

cat /etc/kernel/cmdline
Code:
root=ZFS=rpool/ROOT/pve-1 boot=zfsroot@alexpiesel
cat /proc/cmdline
Code:
initrd=\EFI\proxmox\5.15.17-1-pve\initrd.img-5.15.17-1-pve root=ZFS=rpool/ROOT/pve-1 boot=zfs
proxmox-boot-tool status
Code:
Re-executing '/usr/sbin/proxmox-boot-tool' in new private mount namespace..
System currently booted with uefi
04F3-AFD9 is configured with: uefi (versions: 5.13.19-1-pve, 5.13.19-4-pve, 5.15.17-1-pve)

And log file dmesg is in attachment :)
 

Attachments

kernel cmdline looks ok - as does the proxmox-boot-tool output (I expect that you have a single disk in this system)?

* you're sitting in front of a screen attached to the PVE node ?
* no special pci-passthrough configuration - or anything else which is not standard?

* does the issue persist if you try booting another (older) kernel? Just select the old one in the boot-menu
 
  • Like
Reactions: Proxomx
On a hunch - could you also try the following?:
* add 'simplefb' in a line of its own to /etc/initramfs-tools/modules
* run `update-intramfs -k all -u`
* run`proxmox-boot-tool refres`
* reboot

In case this renders the system unbootable - you'd need to use the ISO as described in the wiki-page to chroot into your system undo the chane and run `update-initramfs -k all -u` and `proxmox-boot-tool refresh`
 
Last edited:
  • Like
Reactions: uer and Proxomx