Last Tuesday 5th May, I did a zypper dup - something like 2000 packages and a kernel update. Towards the end of the update, it flashed a message that it could not write to the “boot partition” and then that the root partion had run out of space.
I’ve never had a boot partition - there has never been a boot partition on this machine. I have the traditional UEFI, / root and home. So I don’t what it was trying to do.
I used a live disk to look at the root partition and it is indeed saturated. It is 60Gb, normally with between 20 - 25 occupied by software. So this update has found a way of filling it with 30Gb.
Further, now when I boot I get the welcome screen and can log in. I get to see the desktop taking shape and then the video crashes and I get a black screen with the the following messages: amdgpu 0000:0c:00.0 [drm] ERROR lttpr_caps phy_repeater_cut is 0x0, forcing it to 0x80 amdgpu 0000:0c:00.0 [drm] ERROR LTTR count is non-zero but invalid lane count reported. Assuming no LTTPRP present
–line repeated 7 times then– amdgpu 0000:0c:00.0 [drm] ERROR LTTPR count is nonzero but invalid lane reported. Assuming no LTTPR present. tumnbleweed 20260430 kernel 7.0.2_1 (tty1) default
and a cli login
Bearing in mind the update was also a kernel update, from the last message it’s clear that it’s trying to boot to the older setup - Grub has not been updated.
So there are 3 problems root partition bricked, video corrupted and Grub not updated.
Where do I start and what do I do?
Probably you first need to free up some space on the system-root LVM volume.
Try going into a tty session instead of logging into a graphical environment on boot by pressing “Ctrl-Alt-F4”. If you can log into a tty session as your user then try first to delete some old log files you don’t need with sudo journalctl --vacuum-time=1d. That will usually free up a bit.
If you have snapshots you can remove old ones you don’t need with sudo snapper delete <number> after running sudo snapper listto find out which ones you have. Delete all the ones you think you’ll never need.
Then you could either try to go back to the previous state before the update by doing sudo snapper rollback <number-of-snapshot-before-update>or try to complete the update with sudo zypper dup before rebooting, making sure you have enough free space on the partition with df -h.
Thanks both for clarifying things a little and it seems there is hope!
I apologise for my ignorance, but what is a “tty session” - when the message before the login prompt reads “tumbleweed 20260430 kernel 7.0.2_1 (tty1)”, does that mean I’m in a tty session already ?
I don’t understand how 2Gb of downloaded packages can inflate themselves into 20 Gb and block everything - there are snapshots but I’ve been bitten by “snapshot bloat” before and have configured things to keep a reasonable minimum.
I’ve tried using a snapshot but it does the broken video routine without showing me login screen. So likely it will do it again. When I tried a message flashed before the image started to load: "EFI warning failed to measure data for event 1 8000… (I couldn’t count the 0’s in time!)
@FrankyU2 is the “df -h” part of the “zypper dup” command ? either before or after dup I suppose… I’m sorry, I’m useless at this.
A tty session is a text only terminal session. So at the stage where you would usually log in as you into your graphical desktop environment(probably plasma or gnome, xfce etc) do a “Ctrl-Alt-F4”. This should bring you into a text only terminal where you can login as your normal user. Then your priority should be clearing whatever files you can because tux does not like to run out of space.
I would try to make sure there’s enough space to reinstall the kernel and whatever graphics drivers it needs if those got borked. Maybe run zypper dup again to try to get it to complete if there’s enough space. Then if that’s done and there is some space left over try rebooting back into your desktop normally.
Check out how much space is used for coredumps with…
sudo du -hs /var/lib/systemd/coredump
Those can be cleared with sudo systemd-tmpfiles --clean and they can take up space in a hurry if apps crash.
[If you’re using nvidia drivers there is an issue with the latest kernel so that may be a problem ].
df -h is just a program that shows you the allocated space on your partitions. You can type it at the shell prompt just like zypper dup. Doesn’t really have much to do with zypper but will tell you how much space is available on your root partition and the boot volume or whatever.
It may be that as soon as you free up some space in your root partition you’ll be able to get back into your graphical desktop without any of the other steps mentioned. To try this after you’ve freed up some space in tty press “Ctrl-Alt-F1” and login normally through the graphical interface. You might not have wireless if you need it in the tty terminal unless you can bring it up from the shell which I can’t remember how to do so this might be the best outcome. Hope this helps.
I’m guessing that this has to do with the kernel and the associated “initrd” file. That “initrd” file is generated on your computer, and it can be large.
Did the housework as advised and I’m back in ! In spite of my precautions snapper was taking up too much space. I’ll try an update again - while I’m ahead!