Why Servers Crash at 3 AM

df surveys every partition; du pinpoints the one that's overflowing

3 AM. Your phone buzzes: "Disk usage at 99%." You log into the server, heart pounding. Hundreds of directories, thousands of files. Something is devouring all the space, but what?

This is where df and du become your lifeline. First, df -h to see the big picture: the /var partition is at 99%. Then du -sh /var/log/* to hunt down the culprit: a 10 GB access log that nobody ever rotated. The fix? Empty the bloated log (deleting a file a process still has open won't free the space until the process lets go) and set up logrotate so it never happens again. What could have been hours of blind searching takes just two commands and 30 seconds.

The unglamorous commands are the ones you reach for first when production is on fire.