Monitoring Disk Usage and Inodes: df, du, and ncdu

Disk space running out is one of the most common emergency calls in server administration. But full disk is not the only failure mode. A filesystem can be full because of small files exhausting its inode table, even when there is plenty of free space. Understanding both df and du, plus the interactive ncdu, turns a storage panic into a quick diagnosis.

df: Filesystem and Inode Overview

The df command reports filesystem space usage. The human-readable form is:

df -h
df -h /
df -ih /

The first lists every mounted filesystem with sizes in human units. The second focuses on the root filesystem. The third shows inode usage: the -i flag swaps space columns for inode columns. If the IUse% column shows 100 percent, you are out of inodes even though df -h might show free space.

Inodes are the filesystem records that hold a file's metadata. Each file or directory consumes one. A temporary directory filled with millions of tiny files can exhaust the inode table while using very little disk space.

du: Measure Directory Sizes

The du command sums the disk usage of a directory tree:

du -sh /var/www
du -h --max-depth=1 / | sort -h
du -sh /var/log/* | sort -h | tail -20

-s summarizes, -h is human-readable. The second command lists each top-level directory and sorts by size to find what is consuming the disk. The third ranks log subdirectories by size.

To find the largest individual files:

find / -type f -size +500M -exec ls -lh {} \; 2>/dev/null
du -a / | sort -rn | head -20

The first finds files larger than 500 megabytes. The second reviews the entire tree by size. Combine these to locate the heaviest consumers in seconds.

ncdu: Interactive Disk Usage Explorer

ncdu gives an interactive, terminal-based view of disk usage with a navigable tree. Install it with your package manager:

apt install ncdu
dnf install ncdu
apk add ncdu

Then run it against a directory:

ncdu /var
ncdu -x /

The -x flag prevents crossing filesystem boundaries, which keeps the scan inside one disk. Within ncdu you navigate with arrow keys, press d to delete a file or directory, and press q to quit.

Cleaning Up Common Culprits

System logs, package caches, and old kernels are frequent offenders. Clean them safely:

journalctl --vacuum-size=100M
apt autoremove --purge
dnf clean all
find /var/log -name "*.gz" -mtime +90 -delete

journalctl --vacuum-size=100M trims the systemd journal to about 100 megabytes. apt autoremove --purge removes orphaned packages. The find line deletes compressed logs older than 90 days.

Setting Up Alerting

Prevention beats reaction. A simple cron job warns you before the disk fills:

df -h / | awk 'NR==2 { if (gsub(/%/,"",$5) > 90) print "Disk nearly full" | "mail -s alert admin@example.com" }'

On a Netbay VPS, combine disk-usage scripts with your own monitoring or a lightweight external uptime check to catch problems early.

Takeaway

Check both space and inodes with df, drill into culprits with du and ncdu, and clean up system caches and logs regularly. Monitor disk usage on a Netbay Linux VPS with high-speed SSD storage at netbayhosts.in.