G2S3

G2S3 is a solution to regularly back up your Google data by copying it over to S3.

This is work in progress.

What's working:

• It can copy a single Google Drive folder with all files in it (non-recursively) to S3
• CDK deploys AWS Batch jobs, necessary compute environment, queues, SNS topics, subscriptions, etc.

What's not working yet:

• Recursive copy of a Google Drive folder, or entire Google Drive into S3.

There's also an accompanying blog post:

• Backing Up Google Photos to Amazon S3 published in Better Programming

"Quick" Start Guide

1. In Google Cloud Platform (GCP):

   1. create a new project (e.g. "g2s3")
   2. enable the Google Drive API
   3. create a OAuth 2.0 consent screen.
   4. Download credentials (TODO: add how)

2. Build and run the retrieve-google-tokens binary (TODO: add instructions)

3. In your AWS account

   1. "Store a new secret" in AWS Secrets Manager ("Other type of secret")
   2. Add CLIENT_ID, CLIENT_SECRET, and REFRESH_TOKEN into new secret as key/value pairs

4. Choose one of the following two options:

   • Option 1: Using a personal fork and your own Docker image:

     1. On GitHub, fork this repository, enable GitHub Actions in your forked repository, and wait for the Action to successfully complete.

     2. In your terminal, run:

        git clone [email protected]:<GITHUB_USER>/g2s3.git
        cd g2s3
        

   • Option 2: Using petergtz repo and the associated Docker image:

     1. In your terminal, run:

        git clone https://github.com/petergtz/g2s3.git
        cd g2s3
        

5. To configure your backup setup:

   1. In your terminal, run:

      cp cdk/bin/deployment-config.json.example cdk/bin/deployment-config.json
      

   2. Fill in proper values in cdk/bin/deployment-config.json (hopefully self-explanatory)

6. Assuming aws CLI is installed and configured correctly, run:

   ./scripts/cdk-deploy.sh
   

Architecture Diagram

Components

G2S3 consists of three main components:

• a binary built with Rust to copy files from Google Drive to S3, called back-up-drive-folder
• a Docker image built from a Dockerfile to package back-up-drive-folder, available as GitHub Package g2s3/g2s3.
• a CDK stack to deploy everything as AWS Batch job to regularly invoke back-up-drive-folder

CLI back-up-drive-folder

This CLI can be invoked locally for testing, or from the cloud when part of a regular backup. For instructions, on how to use it:

$ back-up-drive-folder --help

Docker Image

The Docker image (built via this GitHub Action) can be used directly using:

$ docker pull ghcr.io/petergtz/g2s3/g2s3:latest

Or it can be built using:

$ ./scripts/build-release.sh && ./scripts/build-container.sh

CDK Stack

The CDK stack can be found in ./cdk/lib.