Skip to main content

Disk Usage Insights

Disk pressure causes failures that look unrelated: MySQL stops, PHP writes fail, uploads break, and services crash. This guide shows a safe workflow to find what is consuming space and what you can clean up.

Quick Summary
  • Confirm space and inode usage: df -hT and df -ih.
  • Locate heavy directories: du and ncdu.
  • Clean safely: logs, caches, package caches, old backups.
  • Do not delete database files directly.

Step 1: check space and inodes​

disk-usage-check.sh
df -hT
df -ih

If inodes are near 100%, you may have millions of small files (cache, sessions, temp).

Step 2: find the largest directories​

Start at filesystem roots that commonly grow:

du-largest-directories.sh
sudo du -xhd1 / | sort -h | tail -n 20

For interactive exploration:

ncdu-root.sh
sudo ncdu /

Step 3: common safe cleanup targets​

Journal logs​

journald-usage.sh
sudo journalctl --disk-usage

You can reduce retention:

warning

Vacuuming logs deletes history. Do this only if you accept losing older troubleshooting context.

journald-vacuum.sh
sudo journalctl --vacuum-time=14d
sudo journalctl --vacuum-size=500M

Package caches​

apt-cache-cleanup.sh
sudo apt clean
sudo apt autoremove -y

Old backups​

If backups are local, ensure you have offsite copies before deleting.

find-old-backups.sh
sudo find /backups -type f -mtime +30 -name '*.tar*' -o -name '*.zip' -o -name '*.zst'

WordPress caches​

Cache directories are often safe to delete because they can be regenerated, but check the plugin first.

wordress-cache-candidates.sh
sudo du -sh /var/www/html/wp-content/cache 2>/dev/null || true
sudo du -sh /var/www/html/wp-content/*/cache 2>/dev/null || true

Things not to delete blindly​

  • /var/lib/mysql or /var/lib/mariadb (database data directory).
  • Live web roots without a restore plan.
  • Anything you are not sure about on a shared box.

Next steps​

  • If disk space is fine but the box is still slow: see [Disk I/O troubleshooting](./disk-io-troubleshooting).