Backing up your research

[TheLostExploration]

Mod, please delete this post.

Basically, the poster wanted to know what people do to back up their data. Do you use Dropbox, Google Drive, etc.? And, how often do you back up your data? Are you worried about lost or corrupted data?

Edited December 1, 2015 by rising_star
to restore content

[rising_star]

Oh, I don't back up to Google Drive or Dropbox due to PRISM related concerns. I prefer things like SpiderOak or even setting up one's own server for backups. If I had to do it all over again, I would set up my own server. That said, I don't do anything other than have automatic backups which run constantly. I use SugarSync and it automatically syncs any changes to any file in the designated folders (which for me means all of the folders with documents in them). I also change the file name every day (I use the date in the file name) to ensure that I can keep track of what I worked on and when. What this means is I don't really worry about losing data or having corrupted data. 

[lewin]

This setup requires more startup cash and some tech savvy but storage is all local. I use a D-Link ReadyNAS with two 3TB WD Red harddrives (in raid) and Bittorrent Sync running on the NAS. A BTSync client runs on my work and home computers, immediately syncing all my work files between them (about 50 GB, I have large raw data files). It uses end-to-end AES-128 encryption. The NAS is encrypted so that it requires a USB to boot, which I keep separately. That is, if someone stole the box it wouldn't power up again. Soon I'm going to buy whole disk encryption for all the computers too.

Upsides: no cloud storage, as much storage as your local harddrive can hold, no ongoing costs

Downsides: no cloud backup, higher upfront costs, sync always requires an online peer [but the NAS is this if you leave it on], high memory use if you're syncing lots of files, occasional sync quirks (e.g., rarely, if I rename a large folder, it recreates the whole tree of empty subdirectories)

Bonus feature:  I have BTSync running on my lab computers so that when the research assistants collect data it's immediately backed up off site (to where the NAS sits at my house). I can also deploy experiment files from my work computer to the lab computers, e.g., if I need to update a program.

 

Edited December 5, 2015 by lewin

[lewin]

ETA: During grad school I used SpiderOak because they seemed to have the best privacy/security policies. I was lucky to get in on their $125/year unlimited backup plan. Everything is backed up instantly with client-side encryption. Upside: privacy/security. Downside: slow upload client, occasionally buggy, it's still on the cloud.