Skip to content
RackStack

Disk and Filesystem Checks

Disk issues usually show up as secondary failures: services stop writing, deploys fail, temp files break, or databases go read-only.

Core commands

Section titled “Core commands”
Terminal window
df -h                            # show filesystem space in human-readable units
df -i                            # show inode usage
du -sh /var/*                    # estimate size of each path under /var
du -sh ./*                       # estimate size of each path in the current directory
mount                            # list mounted filesystems
find /var/log -type f -size +100M  # find large log files

What to check first

Section titled “What to check first”

Space usage:

Terminal window
df -h

Inode usage:

Terminal window
df -i

Big directories:

Terminal window
du -sh /var/* | sort -h

What matters

Section titled “What matters”
  • A filesystem can have free space but no free inodes.
  • Log growth is a common cause of disk pressure.
  • Temporary directories filling up can break installs, deploys, and service restarts.

Practical rule

Section titled “Practical rule”

Check both df -h and df -i. If you only check bytes, you can miss inode exhaustion completely.

Common mistake

Section titled “Common mistake”

Do not run du blindly at the root of a large filesystem unless you need to. Start in likely hot spots like /var/log, /var/lib, /tmp, and app data directories.