Find the biggest files and folders eating your disk

When your disk is suddenly full, the first question is always “what is taking all the space?”. This script answers it in seconds without installing anything. It only reads and reports; it never deletes.

The script

#!/usr/bin/env bash
set -euo pipefail

# Usage: bigfiles.sh [DIR] [COUNT] [MIN_SIZE]
DIR=${1:-.}
COUNT=${2:-20}
MIN_SIZE=${3:-10M}

[[ -d "$DIR" ]] || { echo "not a directory: $DIR" >&2; exit 1; }

echo "== Largest files over $MIN_SIZE in $DIR =="
find "$DIR" -xdev -type f -size +"$MIN_SIZE" -exec du -k {} + 2>/dev/null \
  | sort -rn \
  | head -n "$COUNT" \
  | awk -F'\t' '{ printf "%8.1f MB  %s\n", $1/1024, $2 }'

echo
echo "== Largest folders directly inside $DIR =="
find "$DIR" -mindepth 1 -maxdepth 1 -type d -exec du -sk {} + 2>/dev/null \
  | sort -rn \
  | head -n "$COUNT" \
  | awk -F'\t' '{ printf "%8.1f MB  %s\n", $1/1024, $2 }'

Usage

chmod +x bigfiles.sh
./bigfiles.sh                     # current folder, top 20, files over 10 MB
./bigfiles.sh ~ 10 500M           # home folder, top 10, files over 500 MB
./bigfiles.sh /var/log 15 1M      # log folder, files over 1 MB

Example output:

== Largest files over 10M in big ==
    25.0 MB  big/videos/clip.bin
    12.0 MB  big/my docs/report copy.bin

== Largest folders directly inside big ==
    25.0 MB  big/videos
    12.0 MB  big/my docs
     0.1 MB  big/logs

How it works

Each section is a pipeline. Reading the first one stage by stage:

1. find "$DIR" -xdev -type f -size +"$MIN_SIZE" lists regular files bigger than the limit. -xdev stays on one filesystem, so scanning / does not wander into mounted network drives.
2. -exec du -k {} + asks du for each file’s size in kilobytes. The + passes many files to one du call, which is much faster than one call per file.
3. 2>/dev/null hides “Permission denied” messages for folders you cannot read.
4. sort -rn sorts numerically, biggest first, and head keeps the top entries.
5. awk -F'\t' splits on the tab character that du puts between size and path, converts kilobytes to megabytes, and prints the path untouched, even if it contains spaces.

The second section uses du -sk (summary, kilobytes) on each direct subfolder, so you can see which folder to explore next. Run the script again on that folder to drill down.

Useful find size units

• k = kilobytes, M = megabytes, G = gigabytes, for example +1G.
• + means “more than”, - means “less than”.
• Combine with time: find . -type f -size +100M -mtime +90 finds big files not modified for 90 days, which are good candidates for archiving.

Usual suspects

• node_modules folders in old projects. Find them with find ~ -type d -name node_modules -prune.
• Docker images and build cache. Check with docker system df.
• Package manager caches, old virtual machine images, video files, and logs that are never rotated.

Once you know what to remove, delete it deliberately with rm after double-checking the path. The script stays read-only on purpose. If you want an interactive view, tools like ncdu on Linux show the same information in a browsable list.