Simple node.js script to backup files and folders from a computer to an S3 bucket at regular intervals.
A list of files and folders are specified in the configuration. When backing up a specified file it will be compressed and uploaded as a gzip archive. When backing up a folder it will tar and then gzip the folder and upload the resulting archive.
The backup files are uploaded to the specified bucket in a folder hierarchy consisting of a root with the computer hostname and then a subfolder named as an ISO 8601 string representation of the current date/time.
Right now the script is set up to use Winser to install it as a windows service automatically during installation via npm. This should not affect its use on other platforms, but it is not currently set up to automatically create a daemon on any other platform.
All configuration settings are in the config.js file and will need to be set to your needs.
You may worry about the possibility of someone accessing sensitive S3 credentials in the configuration file. You can configure a bucket policy for a discreet backup user that will allow them to write backups to S3 but not list or download any previous backup data.
{
"Version": "2012-10-17",
"Statement": [
{
"Action": [
"s3:GetBucketAcl",
"s3:GetObjectAcl",
"s3:PutObject"
],
"Sid": "Stmt1373697306000",
"Resource": [
"arn:aws:s3:::bucket_name_here/*",
"arn:aws:s3:::bucket_name_here"
],
"Effect": "Allow"
}
]
}
- Handle paths with spaces
- Install automatic daemon on other OSs