Backup external server data to AWS over SSH

Question

Hello everybody,

I run an server outside of AWS which can be accessed over SSH.

I don't want to do backups from the server to AWS. It should work the opposite.

How can I run a serverless tool in aws, which is daily connecting to the server and copy all data to a cheap AWS storage.
Is there something build in or do I have to build something on my own with Lambda? and S3 Glacier? (Or how would you do this?

Thank you

Answer

While there is [AWS Transfer for SFTP](https://aws.amazon.com/blogs/aws/new-aws-transfer-for-sftp-fully-managed-sftp-service-for-amazon-s3/), it provides an SFTP server (and cannot login to existing SSH/SFTP servers).

So for your use case, the most serverless option is probably to create a (via EventBridge periodically executed) CodeBuild job which either
  - [mounts an EFS (with infrequent access) volume](https://docs.aws.amazon.com/codebuild/latest/userguide/sample-efs.html)
  - and then calls (in the buildspec) `rsync/scp myserver.com:/what/to/backup /efsmountpoint`

or - another possibility without EFS, but with S3 (and thus Glacier storage tier possible):
 - Create an AWS Transfer SFTP Server
 - calls (in the buildspec) `rsync/scp myserver.com:/what/to/backup awstransferendpoint:/`

You might also be able to remove the AWS Transfer SFTP step by directly mounting the S3 bucket in the CodeBuild job with [s3fs-fuse](https://github.com/s3fs-fuse/s3fs-fuse).

Cheers

Backup external server data to AWS over SSH

Contenido relevante