Check Linux disk space with df -h for filesystem usage and du -sh /path for directory usage. Find large files with find / -size +100M. The team that runs these on a schedule catches full disks before services crash.
Table of contents
- df - the filesystem view
- du - directory usage
- Finding large files
- du with sorted output
- ncdu - interactive
- Free space alerts
- The ‘no space left’ trap
- FAQ
df - the filesystem view
df -hShows every mounted filesystem with size, used, available, and use%:
Filesystem Size Used Avail Use% Mounted on
/dev/sda1 100G 87G 13G 87% /
/dev/sda2 500G 200G 300G 40% /data
tmpfs 16G 1.2M 16G 1% /dev/shmThe team that runs df -h gets a quick overview. The team that uses df -h /var checks a specific mountpoint.
du - directory usage
du -sh /var/log # size of /var/log
du -sh /var/* # size of each top-level dir
du -h --max-depth=1 / # size of top-level dirsThe team that uses du -sh finds which directories are biggest.
Finding large files
# Files over 100M
find / -type f -size +100M -exec ls -lh {} \; 2>/dev/null
# Top 10 largest files in /var
find /var -type f -exec du -h {} + 2>/dev/null | sort -rh | head -10The team that uses find -size finds the offenders.
du with sorted output
du -h /var/* | sort -rh | head -10The team that sorts du output finds biggest first.
ncdu - interactive
ncdu is an interactive ncurses-based disk usage analyzer:
sudo apt install ncdu
ncdu /Navigate with arrows. The team that uses ncdu on servers with no GUI has visual disk exploration.
Free space alerts
Common cron job:
# Daily check, alert if < 10% free
df -h / | tail -1 | awk '{ if (substr($5, 1, length($5)-1) + 0 > 90) print "Disk almost full: " $0 }' | mail -s "Disk alert" [email protected]The team that has disk alerts catches full disks before they cascade.
The ‘no space left’ trap
No space left on device errors can come from:
- Filesystem full (check
df). - Inodes exhausted (check
df -i). - Reserved blocks (ext4 reserves 5% for root).
The team that checks inodes catches the ‘looks empty but isn’t’ case.
FAQ
What’s the difference between df and du?
df = filesystem-level (total/used/free for mounted volumes). du = directory/file-level (size of a specific path). The team that uses df for overview and du for drill-down has the right mental model.
Why does df show different sizes for the same disk?
Different mount points show different used/free (e.g., / vs /home). The team that runs df per mountpoint sees each filesystem’s usage.
How do I check inode usage?
df -i. The team that has ‘filesystem full’ errors but df shows free space has an inode exhaustion issue.
How do I find what filled up the disk?
du -sh /* | sort -rh | head to find the largest directory, then drill in. The team that uses this finds log files that exploded overnight.
Can I delete files but not free space?
Yes - if a process has the file open, the space isn’t freed until the process closes it. The team that restarts the process (e.g., logrotate + service restart) frees the space.
If you are sizing the infrastructure for the kind of project this post covers, the RunxBuild hosting calculator is the right place to model the line items. The compute, the memory, the storage, the bandwidth, the database - each one is a separate number, and the team’s mental model for the platform is the sum of those numbers. The RunxBuild dashboard is where the team sees the actual usage in one place.
Useful related references: