Git Clean & GC: Housekeeping for Your Repository

In this tutorial, you'll learn about Git Clean & GC: Housekeeping for Your Repository. We cover key concepts, practical examples, and best practices.

Git clean removes untracked files from your working directory while git gc compresses Git's internal storage, keeping your repository healthy.

In this tutorial, you'll learn Git clean and garbage collection — two essential housekeeping commands that keep your repositories healthy. Untracked files accumulate over time, and Git's internal database grows with every commit. By the end, you'll maintain clean working directories and optimized repositories.

flowchart TD
  A[git clean] --> B{What to remove?}
  B --> C[Untracked files]
  B --> D[Untracked + ignored files]
  B --> E[Dry run - show what would be removed]
  
  F[git gc] --> G[Compress loose objects]
  G --> H[Remove unreachable objects]
  H --> I[Optimize packfiles]
  I --> J[Smaller, faster repository]

Git Clean: Removing Untracked Files

# Dry run - show what would be removed
git clean -n

# Remove untracked files
git clean -f

# Remove untracked files and directories
git clean -fd

# Remove untracked files AND ignored files (e.g., build artifacts)
git clean -fx

# Interactive mode - choose what to remove
git clean -i

Expected output:

Removing temp.log
Removing build/

Working with git clean Safely

Start with a dry run:

$ git clean -n
Would remove temp.pyc
Would remove logs/debug.log

Then force clean:

git clean -f

To keep certain untracked files while removing others, use .gitignore or -e:

git clean -f -e config.env

Git Garbage Collection

Git automatically runs gc periodically, but you can trigger it manually:

git gc

Expected output:

Enumerating objects: 150, done.
Counting objects: 100% (150/150), done.
Delta compression using up to 8 threads
Compressing objects: 100% (120/120), done.
Writing objects: 100% (150/150), done.
Total 150 (delta 30), reused 100 (delta 20), pack-reused 0

Aggressive Garbage Collection

For repositories that have grown very large:

git gc --aggressive

# Force garbage collection even with recent reuse
git gc --prune=now

Warning: --prune=now permanently removes loose objects. Only use it when you're sure you don't need to recover anything.

Repository Size Analysis

Check your repository's size and identify space hogs:

# Check .git folder size
du -sh .git

# Find largest objects
git rev-list --objects --all | git cat-file --batch-check='%(objecttype) %(objectname) %(objectsize) %(rest)' | awk '/^blob/ {print $0}' | sort --numeric-sort --key=3 | tail -10

Expected output:

blob a1b2c3d 10485760 large-file.zip
blob b2c3d4e  5242880 another-big-file.iso

Housekeeping Checklist

Common Errors

Practice Questions

Challenge

Create a repository with 50 commits, including some large files (create with dd). Run git clean -n then git clean -fd. Check the .git folder size. Run git gc and compare the size. Then run git gc --aggressive and compare again. Finally, find the 5 largest objects in the repository.

Real-World Task

A CI/CD pipeline is running out of disk space because the repository accumulates temporary build artifacts. Configure a post-build cleanup step in the CI pipeline that runs git clean -fdx to remove all untracked and ignored files (compiled binaries, dependencies, logs). Schedule a monthly git gc for the repository using a cron job. This optimization is standard at DodaTech for CI pipelines building Durga Antivirus Pro, where build artifacts can exceed 10 GB per pipeline run.

Built by the developers of Doda Browser, DodaZIP, and Durga Antivirus Pro.