[SOLVED] error: invalid cluster 0. you need to load the kernel first

Epp · 2025-04-21 17:35:15

Yes, and sadly for the second time on the same computer (the same question on this forum).

It's a dual boot laptop. Could be there is something wrong with the EFI partition.
The last days Windows is sometimes failing to complete the boot (just hangs visually). After a retry it does go through.

I think it was only days ago that I updated Arch on the laptop.

What I have tried:

-Update the Arch Iso USB
-Boot the USB with the Iso from this month
-mount /dev/nvme0n1p7 /mnt
-arch-chroot /mnt
-sudo pacman -S linux
-sudo pacman -S mkinitcpio
-sudo mkinitcpio -P
-sudo pacman -Syu

Boot partition is about 1,17gb.

Last edited by Epp (2025-04-24 20:33:17)

Lone_Wolf · 2025-04-22 10:47:54

Is https://bbs.archlinux.org/viewtopic.php?id=284755 the previous thread ?

Please post the output of lsblk -f before chrooting .

Epp · 2025-04-22 14:37:48

Yeah, correct, double stupid. Although both thanks to Dell I think.
That is that their firmware updates seriously mess up Windows and Linux (sometimes).

Sorry for acting this helpless, but updating boot files/folders on an existing install scares me.

lsblk -f

NAME        FSTYPE   FSVER LABEL       UUID                                 FSAVAIL FSUSE% MOUNTPOINTS
loop0       squashfs 4.0                                                          0   100% /run/archiso/airootfs
sda         ext4     1.0               9d8e09e5-596d-4906-82fa-8aa20a6f2872  209.5M     0% /mnt
nvme0n1                                                                                    
├─nvme0n1p1 vfat     FAT32 ESP         F028-C406                                           
├─nvme0n1p2                                                                                
├─nvme0n1p3 ntfs           OS          B83853CC385387F0                                    
├─nvme0n1p4 ntfs           WINRETOOLS  C61402F91402EBED                                    
├─nvme0n1p5 ntfs           Image       2C50C20A50C1DAAA                                    
├─nvme0n1p6 ntfs           DELLSUPPORT 1E703CF2703CD1EF                                    
└─nvme0n1p7 btrfs                      8e7e3df3-e306-4bc8-9319-077cea866b7b

boot partition folder

-rwxr-xr-x  1 root root         1 Feb 17  2023 BOOTNXT
drwxr-xr-x 41 root root      4096 Feb 17  2023 Boot
drwxr-xr-x  9 root root      4096 Apr 22 10:03 EFI
-rwxr-xr-x  1 root root  15949824 Jan  1  1980 FSCK0000.REC
-rwxr-xr-x  1 root root  12345344 Jan  1  1980 FSCK0001.REC
-rwxr-xr-x  1 root root  59453440 Jan  1  1980 FSCK0002.REC
drwxr-xr-x  2 root root      4096 Feb 17  2023 System Volume Information
-rwxr-xr-x  1 root root    442080 Feb 17  2023 bootmgr
drwxr-xr-x  6 root root      4096 Apr 29  2024 grub
-rwxr-xr-x  1 root root  59643653 Apr 18 17:49 initramfs-linux.img
-rwxr-xr-x  1 root root 127375302 Apr 18 17:49 initramfs-linux-fallback.img
-rwxr-xr-x  1 root root   7066624 Feb 17  2023 intel-ucode.img
drwxr-xr-x  3 root root      4096 Jul 22  2023 loader
-rwxr-xr-x  1 root root      1411 Apr 22 11:27 lsblk_-f
-rwxr-xr-x  1 root root       912 Apr 22 11:26 mount_folder
-rwxr-xr-x  1 root root  15368704 Apr 18 17:48 vmlinuz-linux

efi folder

drwxr-xr-x 2 root root 4096 Feb 17  2023 Boot
drwxr-xr-x 3 root root 4096 Apr 22 10:03 EFI
drwxr-xr-x 2 root root 4096 Feb 17  2023 GRUB
drwxr-xr-x 2 root root 4096 Feb 17  2023 Linux
drwxr-xr-x 4 root root 4096 Apr 21 16:42 Microsoft
drwxr-xr-x 6 root root 4096 Apr 11 13:30 dell
drwxr-xr-x 2 root root 4096 Feb 17  2023 systemd

Lone_Wolf · 2025-04-23 09:22:14

It's not clear to me (yet) how your system is setup.

What is on the nvme0n1p2 partittion ?

Are the contents of nvme0n1p1 shown in 'boot partition folder' ?

If not, what are they ?

Epp · 2025-04-23 18:09:07

Are the contents of nvme0n1p1 shown in 'boot partition folder' ?

Excuses, I was not clear. I should have left away the word folder.
Those are the files at the highest level of the partition. Not in a folder.
If I mount nvme0n1p1 to /mnt. This (from code block 2) is shown in /mnt.

-File bootx64.efi
There is a bootx64.efi file in mnt/EFI/Boot

-Folder grub
In the folder /mnt/grub there are (for example) the files x86_64-efi and grub.cfg
A file grubx64.efi is the folder /mnt/EFI/EFI
There is also a grubx64.efi file in /mnt/EFI/GRUB

-Folder Boot
The folder /mnt/Boot contains mainly locale files and Windows boot (.dll) files

-Summary for files:
/mnt/EFI/Boot/bootx64.efi
/mnt/EFI/EFI/grubx64.efi
/mnt/EFI/GRUB/grub64.efi
/mnt/grub/x86_64-efi
/mnt/grub/grub.cfg

What is on the nvme0n1p2 partittion ?

This partition is unallocated space (15MB)

The higher number of partitions is of course caused by not good cleaning.
The laptop came with Windows and BitLocker installed.
I reinstalled Windows without BitLocker, when I set up the laptop (dual boot).
To make it dual boot I had to copy all the Windows boot files to another (bigger) partion +/- 1000 instead of 100MB.
Right now Windows is installed for the third time. About 4 months ago for the last time after most likely a Dell firmware update removed the Windows-HDCP possibilty.

The higher number of efi files is most likely caused by misses at the time of the Arch install.

seth · 2025-04-24 06:50:33

You mounted sda to /mnt what looks like it's the live system/install iso?
I assume your linux root partition is nvme0n1p7 and the boot partition is actually the esp - or you might be booting from the root partition?
run an fsck on nvme0n1p1 and also "btrfs check" on nvme0n1p7 (see https://wiki.archlinux.org/title/Btrfs#btrfs_check and don't just run --repair for no reason!)

If nothing flares up, mount the root partition into /mnt, check a journal from there - https://wiki.archlinux.org/title/System … al_to_view
If it says "BOOT_IMAGE=/vmlinuz-linux" you're booting from some dedicated partition that you'll have to mount into /boot (likely the esp), "BOOT_IMAGE=/boot/vmlinuz-linux" means you're booting from the root partition.
Then re-install the kernel

Epp · 2025-04-24 17:23:49

sda was the USB device used to copy the output to my (working) Linux-computer.

Yes, nvme0n1p7 is the root partition and esp is the boot partition.

btrfs

btrfs check --readonly /dev/nvme0n1p7

Output is:

...no error found...

fsck

fsck /dev/nvme0n1p1

Output is:

There are differences between boot sector and its backup.
This is mostly harmless. Differences: (offset:original/backup)
    65:01/00
  1) Copy original to backup
  2) Copy backup to original
  3) No action

/vmlinuz-linux
Contains a free cluster (60). Assuming EOF.
/vmlinuz-linux
File size is 15368704 bytes, cluster chain lengt is 0 bytes.
Truncating file to 0 bytes.
Reclaimed 3685 unused clusters (15093760 bytes).
Dirty bit is set. Fs was not properly unmouted and some data may be corrupt.
  1) Remove dirty bit
  2) No action

Free cluster summary wrong (199829 vs. really 203514)
  1) Correct
  2) Don't correct

*** Filesystem was changed ***
The changes have not yet been written, you can still choose to leave the filesystem unmodified:
  1) Write changes
  2) Leave filesystem unchanged

I've chosen no action for all changes/corrections.

The journal says

BOOT_IMAGE=/vmlinuz-linux

So I mounted the boot partition nvme0n1p1 to /boot.
And tried to reinstall the kernel. Result of pacman -S linux was: Errors occurred, no packages were upgraded.
But sadly I can not directly find how to print the output in parts (to scroll through it).
But last lines say that files already exist in filesystem.

pacman -Syu is not working. There are 20 till 30 packages to update but error message will say up that space is not sufficient.

In short, I did not have the courage to choose "correct" for the 3 fsck findings and I see things too blurry to reinstall the kernel.
Besides the fsck output, I could think of insufficient space.
df says only 31% in use for nvme0n1p1.

Last edited by Epp (2025-04-24 20:02:28)

seth · 2025-04-24 18:58:02

Yeah, the FS on /dev/nvme0n1p1 is corrupted and the kernel is affected - it completely explains the problem.
You'll /have/ to sanitize the FS and then re-install the kernel.

Pick one of the boot sectors (your bet is as good as mine), remove the dirty bit, correct the cluster summary and write the changes.
If you're super-worried, you can dd the current partition into an image and if things go south restore that and try again.

After your fixed the FS, re-install the kernel (from the chroot), you can

pacman -S linux |& tee ~/pacman.stuff
cat ~/pacman.stuff | curl -F 'file=@-' 0x0.st

to upload the output of that if you still run into any problems.

Worst case scenario is that the boot sector is also compromised and both, sector and backup are bogus.
In that case you'll have to re-install your bootloader (grub? => grub-install …)

Epp · 2025-04-24 20:13:52

I chose to copy the backup to the original.
After that, there came only one follow-up question. I chose to remove or correct (exuses for the vagueness).

And to keep in the line of stupid. For the previous linux install action (post #7) I forgot to chroot.
That was the cause of not being able to reinstall linux.
So I've done that now, the reinstall of linux.

Result is not good. Loading Linux Linux in Grub gives (not longer invalid cluster 0 but):
error: file '/vmlinuz-linux' not found.
The file is indeed gone from nvme0n1p1. Probably I should not have copied the backup to the original.

Last edited by Epp (2025-04-24 20:20:41)

seth · 2025-04-24 20:17:27

Boot the iso and check the contents of nvme0n1p1.
Make sure to mount nvme0n1p7 into /mnt, nvme0n1p1 into /mnt/boot, arch-chroot into /mnt and then re-install the kernel.
That should™ get you /boot/vmlinuz-linux there?
exit the chroot, umount the partitions and reboot.

Epp · 2025-04-24 20:32:24

uff... it worked! Thanks very much for the help!

seth · 2025-04-24 20:41:53

As for why this might happen, see the 3rd link below. Mandatory.
Disable it (it's NOT the BIOS setting!) and reboot windows and linux twice for voodo reasons.

Epp · 2025-04-24 21:02:24

I've seen you mention it before. I think I checked it a few times in the/a Windows gui. Just saw through the link that there is a shell command. I ran it (I feel little need for fast boot or hibernation). Thanks!

Last edited by Epp (2025-04-24 21:05:09)

Arch Linux

#1 2025-04-21 17:35:15

[SOLVED] error: invalid cluster 0. you need to load the kernel first

#2 2025-04-22 10:47:54

Re: [SOLVED] error: invalid cluster 0. you need to load the kernel first

#3 2025-04-22 14:37:48

Re: [SOLVED] error: invalid cluster 0. you need to load the kernel first

#4 2025-04-23 09:22:14

Re: [SOLVED] error: invalid cluster 0. you need to load the kernel first

#5 2025-04-23 18:09:07

Re: [SOLVED] error: invalid cluster 0. you need to load the kernel first

#6 2025-04-24 06:50:33

Re: [SOLVED] error: invalid cluster 0. you need to load the kernel first

#7 2025-04-24 17:23:49

Re: [SOLVED] error: invalid cluster 0. you need to load the kernel first

#8 2025-04-24 18:58:02

Re: [SOLVED] error: invalid cluster 0. you need to load the kernel first

#9 2025-04-24 20:13:52

Re: [SOLVED] error: invalid cluster 0. you need to load the kernel first

#10 2025-04-24 20:17:27

Re: [SOLVED] error: invalid cluster 0. you need to load the kernel first

#11 2025-04-24 20:32:24

Re: [SOLVED] error: invalid cluster 0. you need to load the kernel first

#12 2025-04-24 20:41:53

Re: [SOLVED] error: invalid cluster 0. you need to load the kernel first

#13 2025-04-24 21:02:24

Re: [SOLVED] error: invalid cluster 0. you need to load the kernel first

Board footer