DynamoDB Backup

Building distributable package for upload to lambda

$ npm run build

Create a lambda function

1. Within lambda, press the 'Create a Lambda Function' button
2. Press the 'Skip' button to bypass the suggested blueprints
3. Enter the lambda function name dynamodb-backup
4. Select 'Node.js' as the Runtime
5. Upload the zip
6. Under 'Handler' add 'Index.handler'
7. Add two environment variables:
   1. TABLE_NAME which is the table to backup
   2. BUCKET_NAME which is the bucket to publish to
   3. (optional) CAPACITY_FACTOR a float less than 0.9 which controls how much of the provisioned read capacity to consume

More documentation on Lambda

Configure the access policy for your lambda role

Your lambda function will run as an IAM role. This is where we configure the permissions required.

Lambda function master policy

{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Effect": "Allow",
            "Action": [
                "logs:CreateLogGroup",
                "logs:CreateLogStream",
                "logs:PutLogEvents"
            ],
            "Resource": "arn:aws:logs:*:*:*"
        },
        {
            "Effect": "Allow",
            "Action": [
                "s3:GetObject",
                "s3:PutObject"
            ],
            "Resource": [
                "arn:aws:s3:::*"
            ]
        },
        {
            "Effect": "Allow",
            "Action": [
                "dynamodb:DescribeTable",
                "dynamodb:Query",
                "dynamodb:Scan"
            ],
            "Resource": [
                "${TABLE_NAME}"
            ]
        }
    ]
}

This contains permissions for:

1. Saving logs for your lambda execution.
2. Stroring zipped dynamo backups from dynamo to S3.
3. List dynamo tables and get dynamo data

Restore example

There is an example restore function in restore.js. Example usage

./restore.js -b table-backups -s instances -t stages

The usage is:

[aws-dynamodb-backup (master)]$ ./restore.js -h

  Usage: restore [options]

  Options:

    -h, --help                     output usage information
    -b, --bucketname <bucketname>  The name of the s3 bucket to restore from
    -t, --target <target>          The name of the table to create
    -s, --source <source>          The name of source file

All the options are required. The source is the directory and name of the backup file. The target is the name of the table to restore the data to.