It’s good to keep backups, especially of servers in remote data centers using old hard drives.

rsync -vaEz --progress user@remote:/path /local/path

In my case, I’m doing it as root and just copying /, although, in hindsight, I think I should have used the –exclude=… option… It doesn’t make any sense to me to “back up” /proc or /dev, /tmp is iffy, and /mnt is usually not desired.

A few notes: I use –progress because otherwise it wants to just sit there, which is irritating.

-a is archive, which actually maps to a slew of options. -z enables compression. Note that this may or may not be desirable: on a fast link with a slower machine, this may do more harm than good. There’s also a –bwlimit argument that takes KB/sec as an argument. (–bwlimit=100 would be 100KB/sec, or 800kbps.)

Using rsync for backups is nothing new, but it’s still not used as widely as it could be. A seemingly-common option is to create a huge backup with tar, compress it, and then download the massive file. rsync saves you the overhead of making ludicrously-large backup files, and also lets you just download what’s changed, as opposed to downloading a complete image every time. It’s taking forever the first time, since I’m downloading about 40GB of content. But next time, it’ll be substantially quicker.

With backups this easy, everyone should be making backups frequently!

Related posts:

1. I/O Under Linux
2. The State of Linux
3. Torrent Hosting
4. Advice
5. Delete Old Files in Linux