Upgrading from Proxmox 8 to 9 issue: GRUB error: out of memory

D

doskoi

New Member
Aug 7, 2025
5
0
1
UPDATED: This has been solved by rollback grub and kernel version to proxmox 8
https://forum.proxmox.com/threads/u...e-grub-error-out-of-memory.169391/post-790132

---

I've upgrade from proxmox 8 (kernel 6.8.12) to 9 (6.14.8-2)

after reboot.

Shows:

```
Loading Linux 6.14.8-2-pwe ...
error: out of memory.
Loading initial ramdisk
...
error: you need to load the kernel first.
Press any key to continue.
```

AND from the grub list, all the kernel 6.x got the same error, even it was able to boot before the `apt dist-upgrade`
BUT kernel 5.x (was for proxmox 7) is able to boot.

I've tried to toggle secure boot and all the way that I can google and ask AI, but not help.

I've try to boot from the live cd via Ventoy, the normal boot got the same error of "out of memory."
but grub2 mode was able to skip this screen, then stuck in the booting progress.

any help?
 
Last edited:
Hi!

If that also happens for any Live CD (e.g. Debian 12/13), then check whether there are any memory-related BIOS settings that might cause this (e.g. BAR, IGD, ...). It could also help to upgrade the BIOS firmware and/or reset BIOS settings first (if possible) and check whether the RAM is correctly reported in the BIOS.

Otherwise, where there any errors while upgrading? It could be that the bootdisk ran out of space or was corrupted during the process..
 
I've had this exact issue with MSI PRO Z690-A DDR4, where after upgrading to Proxmox 9 I was getting these exact same error messages. Even trying to boot the ISO was met with the same error while older versions worked just fine. What eventually fixed it for me was upgrading the system's bios as it was still on one from 2022.

However in my case, I was still getting the "error: you need to load the kernel first." error even after the bios upgrade, while the memory error was finally gone. However, the proxmox iso did boot just fine. After playing around with bios settigns etc, finally found out that the issue was with the second NVME drive containing nothing but a thin LVM for VMs. Without the M2 installed, the system would boot just fine without any issues. Formatting this drive and reinitializing it fixed the issue.

All by all, it was a interesting upgrade on this praticular system.
 
dakralex said:
Hi!

If that also happens for any Live CD (e.g. Debian 12/13), then check whether there are any memory-related BIOS settings that might cause this (e.g. BAR, IGD, ...). It could also help to upgrade the BIOS firmware and/or reset BIOS settings first (if possible) and check whether the RAM is correctly reported in the BIOS.

Otherwise, where there any errors while upgrading? It could be that the bootdisk ran out of space or was corrupted during the process..
Click to expand...
thanks for the reply, the point is kernel 5.x works well.
I've tried change option in BIOS (newest firmware), but didn't find out any thing help
 
AlexL said:
I've had this exact issue with MSI PRO Z690-A DDR4, where after upgrading to Proxmox 9 I was getting these exact same error messages. Even trying to boot the ISO was met with the same error while older versions worked just fine. What eventually fixed it for me was upgrading the system's bios as it was still on one from 2022.

However in my case, I was still getting the "error: you need to load the kernel first." error even after the bios upgrade, while the memory error was finally gone. However, the proxmox iso did boot just fine. After playing around with bios settigns etc, finally found out that the issue was with the second NVME drive containing nothing but a thin LVM for VMs. Without the M2 installed, the system would boot just fine without any issues. Formatting this drive and reinitializing it fixed the issue.

All by all, it was a interesting upgrade on this praticular system.
Click to expand...
thanks for the reply, I do have two ssd, one is for the OS, one for the VMs.
And my DELL is the newest BIOS ver, very interesting, when I disable the SSD0 in BIOS, the error was gone. but due to missing files or somewhat, still unable to enter the OS (stuck at the Found volume group "name" using metadata type lvm2 ..."
 
I got the error in grub CLI by run, ls:

grub> ls
(lvm/pve-root) (lvm/pve-swap) (hd0) ... (hd1) (hd1,gpt1) (hd2) (hd2,gpt3) (hd2,gpt2) (hd2,gpt1) error: out of memory.

after

grub> rmmod lvm
the OOM error is gone.
but it's only works boot from usb driver.
since the boot grub was installed on lvm, it's unable to remove from installed grub
 
Last edited:
dakralex said:
Hi!

If that also happens for any Live CD (e.g. Debian 12/13), then check whether there are any memory-related BIOS settings that might cause this (e.g. BAR, IGD, ...). It could also help to upgrade the BIOS firmware and/or reset BIOS settings first (if possible) and check whether the RAM is correctly reported in the BIOS.

Otherwise, where there any errors while upgrading? It could be that the bootdisk ran out of space or was corrupted during the process..
Click to expand...
I've spent 2 days to solved this by follow way, hope can help someone who searched on this:

1. Rollback grub to proxmox 8 version 2.06-13
The 2.12-9 no matter from installed proxmox 9 or proxmox 9 ISO, will not boot.
to debugging it, i was run command line to troubleshooting.
when `insmod lvm`, then `ls` will throw 'error: out of memory' error, but it's not the key issue.
the issue is command `linux /vmlinuz-6.x...` will throw 'error: out of memory' to block the kernel been loaded.
but 'linux /vmlinxz-5.x' is works well.

when I using the proxmox 8 ISO that get rid of this error. so when I downgrade the grub and 'update-grub', the kernel loading for 6.x is works.

2. Rollback kernel to proxmox 8 version 6.8.12-13-pve
After grub is able to load 6.x kernel.
next thing is kernel 6.14.8-2-pve (from proxmox 9) is unable to boot.
the error is:
'Check of pool pvex/barn failed (status:64). Manual repair required!'
/dev/mapper/pve-root: clean, 263837/2899968 files, 4837304/11599872 blocks
then stuck on this line, may able to enter the maintenance mode.
...

my LVM is:
PV VG Fmt Attr PSize PFree
/dev/nvme0n1p3 pve lvm2 a-- <118.74g 0 (installed boot, efi, root)
/dev/nvme1n1p1 pvex lvm2 a-- <952.87g 0

as @AlexL mentioned, remove one of SSD will get rid of the kernel loading issue,
I thought it's maybe the kernel issue, when I load the kernel from proxmox 8 (6.8.12-13-pve), the error is gone.
everything is works.

The main issue is new kernel from proxmox 9 somehow, unable to works with the exist LVM that installed on 2 disks.

@dakralex I think it's a kernel bug, but it's definitely a blocker for whom installed lvm over one disk.

---
The hardware is DELL precision serial, latested BIOS
 
Last edited:
I had a similar problem.

The following error occurred in the restart after the update. Two NVMe devices are connected, and the NVMe device that is not used for boot is removed, and the error is resolved and started.

Code: 
Loading Linux 6.14.8-2-pve ...
Loading initial ramdisk ...
error: cannot load image.

  Faild to boot both default and fallback entries.

Press any key to continue...

DeskMeet B660 Series
Intel(R) Core(TM) i5-14500
 
I'm running Proxmox on an older Dell R530 (Proxmox 5 at setup). I tried updating BIOS (as AI suggested) but this did not work out. All the 6.x Kernels weren't bootable. This problem did not occur on my Intel Nuc13 Cluster (Proxmox 8 at setup).
For me downgrading grub2 to 2.06-13+pmx7 fixed to issue and all the kernels were bootable again.

Bash: 
echo "deb https://enterprise.proxmox.com/debian/pve bookworm pve-enterprise" > /etc/apt/sources.list.d/pve8.list
apt install grub2-common=2.06-13+pmx7 grub-efi-amd64=2.06-13+pmx7 grub-pc-bin=2.06-13+pmx7 grub-efi-ia32-bin=2.06-13+pmx7 grub-efi-amd64-bin=2.06-13+pmx7 grub-common=2.06-13+pmx7
reboot

for information, my drive layout is following:
perc-raid1 on all drives:
vgs
/dev/sda3 pve
/dev/sdb1 vg-01
/dev/sdc vg-02
 
You must log in or register to reply here.
Top Bottom