Self-Hosted Serverless Cloud Storage

Host your own cloud storage at solely S3 rates, i.e., 50 TB /month at 1.6 INR/GB (~0.021 USD) per month.

Features

• Completely serverless (that's right, no Lambda either)
• Security: Firebase authenticated client, data access restricted by AWS S3 IAM Policies.
• View, Download, Organize a plethora of diverse files: from videos and images, to ipynbs and PDFs!

How to Set Up

• Create an AWS S3 bucket.
• Create your AWS credentials. (Make sure you write secure IAM policies in accordance with AWS S3 IAM Policies.)
• In the project directory, run: cat "REACT_APP_S3_BUCKET=<YOUR_BUCKET_NAME> REACT_APP_S3_ACCESS_KEY=<YOUR_ACCESS_KEY_ID> REACT_APP_S3_SECRET_ACCESS_KEY=<YOUR_AWS_SECRET_ACCESS_KEY> REACT_APP_S3_REGION=<YOUR_AWS_DEFAULT_REGION> " >> .env
• Create and Register a Firebase project.
• Create a Firebase config object.
• Create an allowed user for your config (https://console.firebase.google.com/project/<project-name>/authentication/users).
• From the config object, note the API_KEY, PROJECT_ID, SENDER_ID, APP_ID, and MEASUREMENT_ID.
• In the project directory, run: cat "REACT_APP_FIREBASE_API_KEY=<YOUR_API_KEY> REACT_APP_FIREBASE_PROJECT_ID=<YOUR_PROJECT_ID> REACT_APP_FIREBASE_SENDER_ID=<YOUR_SENDER_ID> REACT_APP_FIREBASE_APP_ID=<YOUR_APP_ID> REACT_APP_FIREBASE_MEASUREMENT_ID=<YOUR_MEASUREMENT_ID> " >> .env
• npm install
• npm start

Recommended IAM Policy:

Only allow public access to buckets and objects granted through new access control lists (ACLs)

{
    "Version": "2012-10-17",
    "Id": "http referer policy example",
    "Statement": [
        {
            "Sid": "Allow all kind of http requests originating from https://<YOUR_WEBSITE>",
            "Effect": "Allow",
            "Principal": "*",
            "Action": "s3:*",
            "Resource": [
                "arn:aws:s3:::<BUCKET_NAME>/*",
                "arn:aws:s3:::<BUCKET_NAME>"
            ],
            "Condition": {
                "StringLike": {
                    "aws:Referer": [
                        "https://<YOUR_WEBSITE>"
                    ]
                }
            }
        }
    ]
}

Sync with a Local Folder

• aws configure - enter all access details for your bucket
• vi ~/local-to-s3.sh and paste this:

  
  #!/bin/bash
  LOCAL_FOLDER="<abs-path-to-your-local-folder>"
  LTIME=$(stat -f%Z $LOCAL_FOLDER)
  while true
  do
  ATIME=$(stat -f%Z $LOCAL_FOLDER)
  if [[ "$ATIME" != "$LTIME" ]]; then
      echo "$(date): Syncing Changes" >> ~/logs/log.$(date +%B%Y)
      /opt/homebrew/bin/aws s3 sync $LOCAL_FOLDER s3://<BUCKET_NAME> --delete >> ~/logs/log.$(date +%B%Y)
      LTIME=$ATIME
  fi
  sleep 2
  done

* Replace the value for `LOCAL_FOLDER` with the folder you want to sync your S3 bucket with.
* To keep this script always running in the background, run `crontab -e` and add the line: `@reboot local-to-s3.sh`.

**NOTE**: This will only sync your local changes to the S3 bucket. 

* To sync S3 bucket changes to your local folder, run `crontab -e` and add the line: `*/30 * * * * aws s3 sync s3://<BUCKET_NAME> <abs-path-to-your-local-folder> --delete >> ~/logs/s3-to-local/log.$(date +%B%Y)`.
* This last sync isn't ideal and may lead to temporarily inconsistent data across the local folder and S3 bucket. Open to suggestions!