So, approximately a month ago I started getting various weird (and kinda misleading) error messages after waking Garuda from sleep. I do not reboot very often, usually only after kernel updates.
Browsers (both Chrome and FF) started to complain about low storage and asked to delete some cookies and website data.
Then Wayland Plasma sometimes refuses to start apps with error:
kf.kio.gui: Failed to launch process as service: "app-googl
e\\x2dchrome@3cfaf720bf684278b59698052f461788.service" "org.freedesktop.systemd1.NoSuchUnit" "Unit app-google\
\x2dchrome@3cfaf720bf684278b59698052f461788.service not found."
After some investigation, i found REAL reason behind this weird behaviour and error messages:
tmpfs 1,6G 1,6G 0 100% /run/user/1000
constantly monitoring jornalctl revealed this:
[ERROR glean_core::database] Failed to record metric 'extension
s.counters.browser_action_preload_result/clearAfterHover' into metrics: Error { kind: Rkv(IoError(Os { code: 2
8, kind: StorageFull, message: **"No space left on device"** })) }
Yes, rebooting system resolves this, but we are not using Windows, do we?
Solution: /etc/systemd/logind.conf
RuntimeDirectorySize=10%mem according to your needs (4G in my case).
Note: Garuda is great but… 16G RAM starting to be a tight fit. Really? Under Linux. Linux?!?
Well, as you noted, you probably need more hardware (RAM) if you want to avoid housecleaning. It’s the way of the world, more required ;0
64 GB is pretty cheap these days…
Alternatively, you could probably tool that to use REAL space vs a temporary file system.
I just noted: after clean fresh reboot:
tmpfs 4,0G 1,7G 2,4G 42% /run/user/1000
Can that be an issue? It is approximately TWO times the size of my whole HDD back in 1996.
I closed all apps (except terminal) and tmpfs still reports 1.7G tmpfs usage.
Yes, of course i CAN add more RAM. Of course i can increase tmpfs but isn’t that a little bit excessive?
Also REAL disk space vs. tmpfs. Yes I kept constantly running out of space using 120G and 250G SSD-s. Garuda snapshots and pacman cache (not the linux and apps themselves) kept using more-and-more the longer system lived. Garuda Assistant “clear cache” did not helped, actually used pacmans “deep clean” to free up some space.
Now I am on 500G SSD and almost half of it is already full of “slack”. Some time in the future it needs housekeeping.
Except some maintenance Garuda is GREAT. It is lightningly fast on this out-of-dumpster (ewaste) leftover build since 2022 (constant rolling updates).
What you are describing is normal behavior and has nothing to do with Garuda. Maintenance work is necessary everywhere. Every OS requires a certain amount of maintenance.
If you don’t delete your browser cache after closing your browser, rarely close your browser, have many tabs open, use preloading, then naturally a lot of junk will accumulate. Then you have to intervene manually from time to time.
Pacman’s cache can also be easily managed. Garuda-Rani and Garuda-Assistant can do this if you configure the number of packages to be kept, as can pacman -Scc. Alternatively, you can use pacman-cleanup-hook from the AUR. This will then be done automatically with every update or installation as you have configured.
Same for snapshots. In snapper tools you can configure how many snapshots will be kept.