I've been having this frustrating issue where my lightdm fails to start whenever my disk gets completely full. This happens a couple of times a year, and I don't notice until I can't boot into my graphical interface. I usually have to switch to a TTY and manually delete large files to get back to normal. Recently, lightdm kept trying to start in a loop, which made it really hard to log into a TTY. I'm trying to understand why running out of disk space affects lightdm this way. I used to think lightdm would just keep its info in memory or maybe limit itself to show a message about needing to delete files, but that doesn't seem to happen. Also, can anyone suggest a way to get alerts when my disk usage is above 99%? I check with `df`, but it never seems to go to 100%, even when I'm having these issues. Just last time, I was trying to create a timeshift snapshot when it failed due to lack of space, but according to `df`, I had plenty of room. It's annoying!
3 Answers
A full disk leads to unexpected issues across the board. To add, modern software is quite complex and often relies on disk access for various operations, which makes it more prone to fail when space runs low. Also, that `df` command doesn’t always show accurate free space in real-time—so you might think you have space left until you hit a wall with a task that needs it badly.
To prevent future problems, consider implementing a disk monitoring tool that alerts you when your usage hits a certain threshold, like 90% or 95%. There are tools available for Linux that can do this for you automatically, and it could save you a lot of hassle down the road!
Running out of disk space can cause issues with a lot of software, not just lightdm. Basically, if any application needs to write files or use temporary files, and it can't because the disk is full, it can lead to errors. Lightdm might be the first place you're noticing it, but it's a broader problem. Most applications write logs or caches without making it obvious when they fail, and if the disk is full, simple operations like writing can fail without any clear feedback to the user.
Exactly, and it gets trickier because when those write operations fail, the application often doesn't have a good way to deal with that. This is usually where things start going haywire!
That sounds like a great idea! Do you have any recommendations for those monitoring tools?