I'm running a script that checks the used disk space every 10 seconds using the `df` command. While it's pretty quick, I wonder if there's a more efficient method that uses fewer resources. I know I could use `stat` as in the way I listed, and it's slightly faster, but I'm curious if there's an even better approach, perhaps by accessing something directly in the `/sys/` directory. I'm also aware that rewriting the whole script in a compiled language would be overkill. If anyone has suggestions for optimizing this in shell, I'd love to hear them!
1 Answer
Honestly, `df` is already highly optimized for what you're doing. There isn't a secret file in `/sys` that gives you filesystem usage without calling home base, which would hit the kernel anyway. The differences you're looking for are just tiny tweaks in performance!
Totally right! If you run `sudo strace df /`, you’ll see how it all connects. It's interesting to see how it pulls data from `/proc/self/mountinfo` along with the stats from `statfs`. You can really see the backend magic in action!