Been using Spideroak for backups for about six months now. They were doing a special offer with unlimited space, and they offer zero-knowledge encryption (so the key never leaves your computer). Overall it’s pretty good but I have had some annoying issues with the client software.
The first major issue is that the client needs to create temporary files that are as large as the largest file in the backup set. It creates them in the default temporary directory, which is a really really bad idea. On Windows this is the user’s profile directory, and it can easily cause your system drive to fill up. This makes the app hang and Windows starts to have problems. Other apps will crash too, and it’s easy to lose data. I back up some pretty big files, like 100GB disk images and virtual machines, and the system drive on my server is only 40GB total.
There is a work around, which is to run Spideroak with a CLI parameter that tells it to use a different directory for its profile and temporary data. No way to configure it in the client, and it doesn’t handle this and a few other error conditions at all… It just hangs.
The second issue is that the client seems to have major problems communicating with the server when it needs to re-download data after a severe crash. Often it just hangs again. Sometimes you get an error message, but often the UI just craps out and the background task keeps going so you can’t restart it without manually killing that off first.
I have about 3TB of data and over 800,000 “versions” (mostly individual files, some multiple versions of the same file) and it seems to be struggling. The app mostly copes okay but does start using a lot of memory and getting rather slow. It does de-duplication so it needs to keep a lot of metadata around, but I can’t help feeling that it could be more efficient.
If they could just sort the client out then I’d have no hesitation recommending Spideroak.