Setting Up and Managing Git-Annex with Backblaze B2 Storage

Git-annex is a powerful tool for managing large files in git repositories. This guide will walk you through setting up git-annex with Backblaze B2 as remote storage, and cover common usage scenarios. B2 has a compatible API to AWS S3, offers 10GB free space, and if much cheaper than S3 for larger storage use-cases.

Initial Setup

First, initialize git-annex in your repository:

git annex init

Configuring Backblaze B2 Remote

To set up a Backblaze B2 remote, use the following command:

git annex initremote b2-upload type=S3 \
    bucket=your-bucket-name \
    embedcreds=no \
    encryption=none \
    host=s3.us-west-004.backblazeb2.com \
    protocol=https \
    signature=v4

Note: Git-annex will require your Backblaze B2 credentials. You can provide these through environment variables:

AWS_ACCESS_KEY_ID
AWS_SECRET_ACCESS_KEY

Alternatively, you can use a configuration file for managing multiple credentials.

File Size Configuration

To configure git-annex to only manage files larger than 10MB:

git config annex.largefiles "largerthan=10MB"

Copying Files to Remote

After committing files, copy them to the B2 remote:

git annex copy --to=b2-upload

Setting Up Public Access

To create a public download point:

git annex initremote b2-public type=httpalso --sameas=b2-upload \
    url=https://f004.backblazeb2.com/file/your-bucket-name/ encryption=none

There are two main ways to share your git-annex repository:

1. Direct Transfer Method

Drop the local copy of the files
Package and transfer the repository
At the destination, initialize git-annex

2. GitHub Method

When using GitHub:

Remember that files in the local repo are symbolic links to the actual files in .git/annex/objects/
Git-annex creates a separate git-annex branch for configuration
Make sure to push both main and git-annex branches to GitHub
On the new machine:
1. Clone the repository
2. Switch to git-annex branch
3. Run git annex init
4. Use git annex enableremote command
5. Get files using git annex get .

Removing Git-Annex

To remove git-annex from a repository:

Backup or remove files from .git/annex/objects/
Run git annex uninit

You can verify git configuration settings in .git/config.

Best Practices

Always ensure both main and git-annex branches are pushed when using GitHub
Backup important files before uninitializing git-annex
Keep track of your B2 credentials securely

This guide covers the basic setup and management of git-annex with Backblaze B2. Remember to replace placeholder values (like your-bucket-name) with your actual configuration details.

Note: All paths and bucket names in this guide have been anonymized for security purposes.