It's also totally unmaintained, so it's time to move on.
What do you use, what should I use?
Huge preference to open-source tools, I'd like to host this on my own NAS (Linux-based).
[1] http://backuppc.sourceforge.net/
I've just got it to a point where I've been dogfooding it for a year or so, and feel comfortable promoting it more, but I'm not sure how to go about doing that (other than packaging it for various distros, and submitting it to them).
Bacula is actively developed. http://www.bacula.org/git/cgit.cgi/bacula/log/?h=Branch-7.0 No risk of being unmaintained at some point, since they build commercial stuff on top of community Bacula.
It is better at several large files than thousands of small files. In fact when you throw millions of small files at it it chokes pretty bad (CPU skyrockets). So if you need to backup an SQL server data store you'll be just fine, but thousands of XML files? Maybe not.
Regarding the small files performance - thanks for pointing it out. We are using Bacula to backup JIRA & Confluence data stored directly on the disk as files. Some customers have thousands of small attachments in their instances. I haven't seen any performance problems though. Maybe because not only there need to be thousands of small files, but also they need to change often, which is not a case for us. Will keep this in mind. Thanks.
Beside the maintenance regularity on it, what's wrong with it?
Let me know if you have any questions about how I got it to work (for the most part, the biggest issue I see you encountering is encrypting drives on OSX).
After going through the documentation, the only major turnoff that I see is that it fails to provide explicit documentation on the internals (ie when a backup occurs A->B->C happens to your data using formats very well documented in X, Y, & Z)