Wrong version of EOS is starting after Libreoffice install

I am at a loss as to what more I can do to enable my original EOS partition , the images should help explain.

Part nvme1mp2 is what I want but after installing Libre Office and a reboot I get nvmen1p1 which has none of my files.
This is my fstab:

This is the Kernel file :

I imagine there is something else I need to do to effectively ‘swap’ the partition which boots.
I don’t need access to nvmen1p1 at all.

Thanks

Don’t think this has been caused by installing Libreoffice, that would be a very strange thing if that would happen.
Can you still get to the files on the 1mp2 partition ?
Because by the looks of it you (re) installed EOS on 1mp3.

Sorry , substitute nvmen1p1 for nvme1p3 in my post.
p3 is a working but near empty EOS. All my files are on P2.
The very last thing I did yesterday was to install office using EOS Quickstart installer as I couldn’t get it using yay. Then a reboot and now this is where I am.
I managed to create a mountpoint folder on the desktop and can access my files from P2 but its the customisation and installed programs that I miss.

What bootloader, grub, systemd-boot?

efibootmgr
What is the output?

When you start the computer, you see a boot menu? If you see, you have different options to boot?

Well I have no idea what happened here , but I have noticed that your efi partition is very big (almost to the max) can’t help you there I’m afraid. But apart from that, it looks pretty normal to me , because there is a / (root) on the partition that gets started when you boot m1p3. BTW you will need m1p1 to even boot your system.

BootCurrent: 0000
Timeout: 1 seconds
BootOrder: 0000,0001,0002
Boot0000* Linux Boot Manager HD(1,GPT,9cdd74bc-719c-42d1-a50f-4a3edf9ad957,0x1000,0x200000)/\EFI\SYSTEMD\SYSTEMD-BOOTX64.EFI
Boot0001* Windows Boot Manager HD(2,GPT,f7bffd0e-95bd-4a95-a071-4d5a1095d05c,0x1000,0x30000)/\EFI\MICROSOFT\BOOT\BOOTMGFW.EFI0000424f
Boot0002 Windows Boot Manager HD(1,GPT,9cdd74bc-719c-42d1-a50f-4a3edf9ad957,0x1000,0x200000)/\EFI\MICROSOFT\BOOT\BOOTMGFW.EFI0000424f

I used to have a choice of three kernels but now its only 2.

Yes I misread the partition output. I dont want to boot the P3, I want the P2.

This is systemd-boot. I think you can look in the directory /efi/loader.
There you look at loader.conf and see what is the default conf. Maybe that has changed. Check if the correct one you want is in /efi/loader/entries.

If I boot into a live environment and simply copy the contents of P2 into P3 … that should that work ?

I did find this in efi/loader/entries :

Boot Loader Specification type#1 entry

File created by /etc/kernel/install.d/90-loaderentry.install (systemd 254.1-1-arch)

title EndeavourOS
version 6.13.1-arch2-1
machine-id 965c39e2d69749bbb85816dc0cf6d33e
sort-key endeavouros-6.13.1-arch2-1
options nvme_load=YES nowatchdog rw root=UUID=33bf60ab-76e4-48bc-b406-b919e5cef95b rw root=UUID=33bf60ab-76e4-48bc-b406-b919e5cef95b rw root=UUID=33bf60ab-76e4-48bc-b406-b919e5cef95b systemd.machine_id=965c39e2d69749bbb85816dc0cf6d33e
linux /965c39e2d69749bbb85816dc0cf6d33e/6.13.1-arch2-1/linux
initrd /965c39e2d69749bbb85816dc0cf6d33e/6.13.1-arch2-1/initrd

So I could replace the uuid which is correct for the ‘wrong’ partiton with the uuid for P2 ?

I never tried this. At least now you know that this conf is not the correct one. Maybe there is a better solution than copying the correct UUID. I think maybe better wait for other person with better knowledge on EndeavourOS systemd-boot to help.

If you still want to do, I hope you have a live usb to be able to rescue system if something goes bad.

Yes, I’ll wait a little to see what other replies come and I do have a live USB, had to use more than once last year :frowning:

1 Like

@manuel Sounds like something maybe should be looked at?

I’m assuming that the “Quick Installer” probably is a wrapper for yay. I’ve never used the installer so I’m not sure if its set to do a system update before installing. Can you post you pacman log for the time in question?

perhaps start keeping a package list
pacman -Qne --needed > Package.list (no aur packages)
pacman -Qm --needed > AURPackages.list (only aur packages)

Customizations are normally store in ~/.config a simple backup can keep these readily available (note some can just be in a dotfile outside of .config) There is also .local you may want to back up as well as those programs that do not honor the normal approach and create their own folder in your home file.

you need to run and update as the latest kernel is 6.13.5-arch1-1

Normally I believe this will work. SInce I’m not sure what actually happened All I can say is try it and see.

eos-quickstart uses pacman as far as I know.
And I assume eos-quickstart has nothing to do with the problem.
Maybe @dalto has further comments.

1 Like

EOS Quickstart only supports repo packages and runs pacman -Syu <packagelist>. So, yes, it does an update when installing packages.

The op is on an older kernel (not sure if by design) So I’m assuming that the update during his install must have gone foul somewhere or he should have had the newer kernel. Maybe the op will post the results of the pacman log so we can see better details of what happened.

One thing though I noticed that OP also has Windows installed. Now I have been reading some topics about a Windows update had messed with the bootloader of EOS. Could that have happened here?

Its Possible I wouldn’t trust M$ on a toaster but that me :grin:

One thing, you can modify the EFI w/o booting off a live disk.
I would leave the UUID alone and just edit the loader.conf and set the Default entry.
Incidentally I think efibootmgr can set this w/o doing a manual edit.

OK, I’m back.
I had to reinstall but using my 2024 live ISO failed every time, despite having used it at least twice last year. I kept seeing a message that the efi partition was full.

I had to download Mercury in Windows and I could finally install a fresh copy on my now completely empty nvme. I didn’t lose anything thanks to the tip about .local and .config. It only took 2 hours to copy those files to an old HDD :slight_smile:

Copying the partitions didn’t work because the UUID change and for some reason the efi folder was empty so I couldn’t find the old id.

As for Windows, it is on an SSD and was disconnected for the new install, Also, it is a very stripped down Win 10 and can’t self-update.

There is a pacman log here , not sure if it can shed any light on the failed installs.

Hopefully this thread might have some useful info for others in future.

Thanks to everyone.

1 Like