After years of running our own rsync server, I’m drinking the cloud cool-aid and outsourcing our backups. Here is what brought us here.
For years we’ve been using rsync to handle our daily backups to safeguard our data off-site, not only for our own servers and web sites, but also for clients servers. Historically we’ve had one linux server with some really big drives configured for RAID1. However now that I’m working on consilidating some servers, and have grown tired of dealing with hardware, I finally decided that it’s worth the money to give rsync.net a whirl.
In case you’re not familiar with rsync.net, it’s a cloud based backup service with the following features:
- ssh based access (ensures secure transport)
- zfs based filesystem (which allows easy snapshots on their end)
- unix native commands available on their end
The snapshot feature is what sold me on giving this a whirl. Recently a client brought a new developer on. That developer installed some updates on Friday, but the client didn’t discover problems until Monday. By that time our current rsync setup had overwritten the pre-Friday code. We did have another backup source that we could pull the files off of, but this had gotten me thinking I needed to impliment some sort of versioning snapshots, etc. The default rsync.net account features a week of daily snapshots, and weekly snapshots are available for an addition (nominal) fee. More information on their snapshots is available on their website.
To access your backups you can either use the same command line clients that your server uses, any sftp based transfer application (such as filezilla), sshfs, or mapping it as a drive in windows.
While it’s more expensive at $0.20/GB that other cloud storage services such as AmazonS3 or Rackspace Could Files, it’s meant as a backup service, not just a could storage service.