I was doing my weekly update today of my desktop. There were 194 updates, mostly Haskell due to Xmonad being installed. There were also new linux and linux-lts packages, and the goodies (grub config update maybe) that came with the new Endeavouros updates. The usual random updates for others.
During the -Syu update, it failed. I should have taken a screenshot, however yay bombed out, saying "nothing left to do’. It was after the kernel prepping where you see the 3-4 warnings about missing modules.
I broke out to a tty via ctrl-alt-F2 and saw multiple lines that look like:
[620497.223703] EXT4-fs error (device-sda1): __ext4_find_entry:1612: inode #14417997: comm systemd: reading directory lblock 0
Only choice I could do was power cycle. From grub, I could not boot into the latest linux kernel, as it hung in starting the ramdisk. I could boot into linux-lts. From there, after logged in, I did a pacman -S linux , which installed without incident, rebooted, and can now boot both kernels. Linux is 5.15.6-arch2-1.
Any other advice to check system and package integrity?
I thought I’d be getting the updated Endeavour grub theme that comes with the new ISO during this update. No changes there.
I’m hesitant to guess at what might have gone wrong during the update, but it appears re-installing the kernel got you up and running again. You might take a look at the troubleshooting section for pacman in the ArchWiki…
…also, if you type “yay” without arguments, it’s equivalent to “yay -Syu”.
[open a terminal and type “man yay” and it will give you full documentation. Type “q” to exit out of the man page].
Looks like a drive failure, to me. Make sure you backup all your data.
All the files I need have been backed up.
But I wonder if I should have gotten the new Endeavour grub theme I heard about with the recent release, and if other packages might be in some inconsistent state.
The pacman troubleshooting wiki section is interesting - thanks for that link.
Looking at ‘find /usr/lib -size 0’ shows nothing suspicious when comparing to another laptop that has arch - essentially a ton of kernel header files and some test data files.
I have added and removed some random packages with no issues, as the wiki states one might see database issues in such actions.
Maybe I got lucky. And perhaps pacman’s robustness covered me in whatever situation came up.
If anyone else has ideas, I’m all ears/eyes.
inxi -Faz --no-host | eos-sendlog
Thanks for looking. I don’t see anything abnormal but let me know.
You could boot with the USB installer stick and run a terminal command
sudo e2fsck /dev/sda1
Hopefully it finds nothing.
Also, is your disk nearly full? Try command
For example, your package cache might be big. Use
paccache-service-manager to find out.
This will be a hardware failure. You should check the drive with this, for example.
Backup anything you want. 2x. In computer world 1=none.
Finish updates-chroot if you have to, save up fora new HD. Buy before this one dies completely. Run it into the ground.
Yup, that’s all I got.
I went through the SMART area of the arch wiki, and all of the data for my hard drive looks OK. I’ll keep an eye on it.
There was another big forum thread from the last 36-38 hours from a user who couldn’t boot into 5.15’s kernel after an update. That sounded sort of similar to my case. From the thread, it wasn’t clear that an update failed, which is my case.
Data is backed up. I’m usually pretty good with that - it’s always a good reminder to update the files though.
Still puzzled as to what happened. I should have snapped a photo of the terminal where I was doing the update in.
Update: Potential SDD failure. I’ve had a couple more erratic crashes and loss of the file system, which is essentially death of the OS in memory. Upon researching SSD death, I believe my drive may be close. It was several years old though, so probably time.
New (larger) SSD in place, new EOS install. I’m still watching this. Super thanks for the tips and advice.