Backups

Backups are essential to anyone working on computers: hard drives fail, laptops get stolen, humans accidentally delete files, etc. The problems is, with the proliferation of operating systems, and the huge amount of data now being generated in astronomy and space physics, that it has become impossible for us to back everything up. What should matter to you, however, is how YOUR work is getting backed up. Here are a few pointers – my advice is to make sure your important files are being backed up at least a couple of different ways.

1) The best backup

My first note about backups is that in my opinion there is no better backup than one you completely control. No one cares about your work more than you do, and only you know when you’ve done enough work that it would be a crisis to lose it.

I personally have a small fleet of western digital portable drives. Here’s a link to one I currently use, a 1 TB drive currently costing about $144 (amazon.com). If/when the link becomes outdated just go to amazon.com (or whatever vendor you prefer) and search for western digital passport. Note that this version is a USB 3.0 drive, making it a bit faster but not quite as widely compatible (for example my xbox does not like it), so if that is an issue buy a USB 2.0 version. Keep these off-site – I backup my home stuff and keep those in my office, and vice versa.

You can store a lot on a 1 TB drive, and even if you have a huge amount of data you can buy a few of them and probably cover it. If you need to back up more than that you are probably dealing with research data which is an area we (the computer staff) should probably be handling for you.

One reason these can be preferable to the IS&T backups is those are stored off-site, and the recovery can take up to one week. It has been pointed out that I should mention that these drives are not designed to be run 24/7 – attach them, back up your files, and remove them. If you leave them running all the time you can pretty much expect them to fail – they are not designed for that purpose. (This is also the reasoning behind my fleet of drives – I do not place complete confidence in any one drive.)

2) IS&T backups

IS&T offers backup services. In the past we have used these almost exclusively on unix machines, but they do offer backups of windows and apple machines as well. However, the page describing these services indicates that they are “Available To Faculty, Researchers, Staff, Departments,” which I suspect means not to students. I have contacted them for clarification (no response thus far) – there might be a case where an RA’s machine is owned by a faculty member and thus eligible. Some links:

backup and restore page

IS&T notebook backup page

We endeavor to back up UNIX home partitions to disk. The problem with the BU backups is we generally exceed the network bandwidth, so we have to judiciously choose what to back up. IS&T also has limits on partitions sizes for the same reason – they do not want a backups, which are run overnight, to run on into the morning. For the same reason IS&T does not like to back up large static data sets, tmp partitions, etc.

The way I (as well as most of the staff) insure that my work files are being back up is to keep a copy of important files on a samba share that is a backed up partition shared from a UNIX machine. If your group has such a partition we can set up this scheme for you. The share I currently use is hosted by CAS, and students can also have limited space on this machine (currently 1 gb):

CAS fileserver

The list of what is and is not backed up by IS&T is restricted by our Kerberos passwords, but if you’d like to review the backups for your group at any time just send us an e-mail indicating the machine names (bugs at skynet.bu.edu) and we can send you the list.

3) Crashplan

BU supports Crashplan for backups, and you get a discount on the commercial version:

Crashplan homepage

IS&T discount

What BU will do for free is rather limited (10gb) and has some other restrictions. (I will point out that if this 10 gb were used to back up your current project it could be a lifesaver…) However, the free version of Crashplan has a rather interesting feature: for free you can back up between two computers. This means you and a friend could back up to each other’s computers, you could back up from work to your home computer, etc.

4) Cloud storage

Cloud storage is a wonderful development – Dropbox is my personal favorite (link). You can get some storage for free (I usually grab an account wherever offered), or you can pay a reasonable fee and get more. It’s also a useful way to keep a file you are working on available wherever you are. What you should remember about a service such as dropbox is that it is not really a backup unless you pay for their backup service, so this should not be your only copy of an important file.

If you are a bit more technically inclined, you might be interested in a solution I (David) use at home. I purchased a Synology diskstation (a DS211, which has now been superceded by the DS212), populated it with 2 3gb drives, and run a RAID configuration which simply mirrors one drive to the other. With a good UPS this all cost about $700 (in 2011), and gives me 3 TB of reliable storage at home which I use to back up our laptops, store our photographs, music files, etc., and access these files whereever I am. The box runs unix, so it can be made reasonably secure and I can pick and choose from many different services I can run. It happily talks to windows, mac, ipads, and even our Logitech Squeezebox. (Another highly amusing device if you enjoy music.)