Storing (large) binary files in Git trees is not recommended because any change to the file will result in a re-upload of the entire file when pushing, increasing the size of the Git repository. When updating a 100 MB file 100 times, your repository will be 10 GB in size!
As a remedy, there is Git LFS which stores certain files only as pointers in the Git tree, downloading the actual data from somewhere else when pulling. The repository size remains small.
GitHub applies some fees when the total amount of LFS downloads surpasses 1GB. Unfortunately, even gh-actions runs, when used with
- name: Checkout code
uses: actions/checkout@v4
with:
lfs: truecount towards that limit. To cache the LFS downloads, you can instead use this action. Simply replace the above by
- name: Cached LFS checkout
uses: rivethealth/[email protected]
# Use these to explicitly include/exclude files:
# with:
# include: "*"
# exclude: ""Check it out on the GitHub Marketplace!
- Luca Benci, Avoiding git-lfs bandwidth waste with GitHub and CircleCI
- actions/checkout issue: Cache for LFS
The scripts and documentation in this project are released under the MIT License.