meteor-backup | Backup script for meteor mongodb | Continuous Backup library

Create an S3 bucket for the backup, eg my-perfect-backup
For increased security we're using an instance IAM role, which lets us get away from storing actual AWS credentials on the host itself. AWS will take care of providing short lived one-time credentials to the s3cmd tool when it needs them. Create a role named MeteorMongoBackup. The Role Type should be Amazon EC2, next click on Custom Policy and insert the contents of the file MeteorMongoBackupRole.json (replacing my-perfect-backup with your own bucket name from step 0)
Boot up an EC2 micro instance (a micro should be enough). When you create this instance make sure you select MeteorMongoBackup as the instance role. This will allow the host to seamlessly access your S3 bucket without needing to hardcode your AWS credentials on it.
Once the host is running, ssh to this host and run the setup.sh script. Note that this script contains many sections (some of them are optional) so my suggestion is that you read it and run each line manually.
Edit the file backup.sh to update the SITE URL and the COLLECTIONS. The COLLECTIONS is only needed if you want to crate cleartext (json) backups using the meteor-backup tool. If you're only interested in binary backups (bson) using mongodump then no need to specify the collections, they will all be backed up automatically, in binary format. json backups are useful for diffing or looking at just one collection's data, or grepping through them, but they also consume some more space.
Next, test the backup.sh script by running /bin/bash backup.sh
Set up a cron job by running crontab -e and adding the contents of the crontab file from this repository