I’ve ran out of system space during an update and I don’t know what is using it all.
Is there a cache somewhere I need to delete?
Doing the below should cover the package caches in:
/var/cache/pacman/pkg/ and /.cache/yay/
yay -Scc && yay -Ycc
If you are running out of space in your home directory, then you may need a disk information tool like filelight or ncdu. These can help you to locate the folders/files that take up the most space, then you can decide what to do with them.
Reference; https://wiki.archlinux.org/title/List_of_applications/Utilities#Disk_usage_display
2 Likes
If you utilize btrfs you might have snapshots from before the upgrade. These can take more or less space.
1 Like
@astspikey
Don’t create duplicate topics please. Come back here and see if we can work something out.
I do have btrfs - However, this update broke my system and I can no longer see anything except “enter firmware” in my bootloader.
How would I use btrfs to revert?
Take a look here: https://discovery.endeavouros.com/applications/update-troubles-meet-timeshift/2019/11/
Scroll down to the section about not being able to boot from grub:
In case of a non-booting system, boot the device with the Live USB or another Arch-based Live USB and use Arch-chroot then follow the same instructions as above. When Timeshift is finished, you can reboot the system.
Essentially:
- Boot using the LiveUSB (your installation media)
- Use Arch-chroot
- Access Timeshift
- Revert your changes
- Boot into your installed system
Another step may be to free up some space on your device to an external device or other partition. Otherwise, you are likely to have the same issue the next time you update.
But read and follow the instructions given on the wiki.
1 Like
Time shift snapshot is about 3 months old. Are there any implications of using this? I.e. file loss?
If I can I would prefer to get the system functioning again from the live environment.
I can get into chroot.
I’ve tried pacman -Syu to finish the update.
I just don’t have any system-dboot entries.
I’ve tried mounting my efi partition. Ie. Sudo mount /dev/sdb1 /mnt/boot and looking at loader, entries etc and trying to create an entry, but it doesn’t show up in the boot menu when I reboot. Just windows 10 and reboot to firmware.
I’m using root=Partuuid and targeting my btrfs partition and the endeavour OS
I’m typing in the kernel in initramfs and .img etc.
There’s a folder in boot with my partuuid name, but it’s empty… so I don’t understand what it’s targeting. As far as I can see the kernels are located in sdb2 /boot…
When I did the pacman commands, I didn’t have sdb1 mounted, just sdb2… I e.
Sudo mount -o subvol=@ /dev/sdb2 /mnt
Feel a bit stuck 
Seems your system doesn’t have an enabled Timeshift hook. You’ll need to enable that once your system is back up and running.
Now, as you are using systemd-boot, and I’ve never used it myself, you’re gonna have to wait for someone else with experience in that area to come in and try to troubleshoot.
I assume the wiki had some instructions specific to systemd-boot, though, correct? And if so, did you follow those steps or the grub-related ones?
I’m looking at the systemd-boot stuff yeah.
My understanding is:
System stores boot info in /dev/sdb1 (EFI)
EndeavourOS system is on /dev/sdb2
/EFI should mirror /dev/sdb1…
Therefore, I’m guessing I need to mount my Btrfs endeavour partition. Somehow pass it the EFI partition by mounting it somewhere… Perhaps in /EFI.
Arch-chroot into system… Then somehow run a kernel-reinstall script?
Do you think it’s possible? Or should I just give up the goose, go for a fresh OS install and copy over my home directory? I assume that will give me access to all my files and configs? Or would it be wonky and have file permission issues?
What you said sounds like the play. But, as I said, don’t know how to do it myself.
The steps you described are pretty much the same with Grub. In fact, if you Arch-chroot in properly, you should be able to install, uninstall, or update anything via terminal.
So yeah, what you described should indeed work.
Was about to go read through it, but nah. That’s for you to do. I don’t need to learn how Systemd-boot works, since I have no intention of using it until it gets straight-forward snapshot support.
Lol!!
SUCCESS! I’m back into the system… I love linux when you actually fix something. Feels great.
Correct answer was:
Mount subvolume in /mnt/
Mount EFI partition into /mnt/boot
Run “reinstall-kernels”
Reboot. Voila. Everything was in the systemd-boot menu again, and I’m back to normal.
Phew.
2 Likes