NicJ.net

This sounds very cliche, but my life is very… digital. So many parts of my life revolve around bits of information stored on disks and drives in various places. Documents, pictures, music, emails, projects — the list goes on.

We live in an age where this is very common. But have you ever thought about what would happen if you lost it all? How would you feel? Sure, losing some of your MP3s isn’t the end of the world. But what about the other stuff? Maybe I’m an outlier on the bell curve for this, but there is so much data that I have created and I want to have forever.

How could you lose your data? Unfortunately, it’s pretty easy:

Problems:

1. Media failure: CDs get scratched, and will often only store your data for a few years. DVDs tell a similar story.
2. Hardware failure: Hard drives crash (the average lifetime is probably around the same as CDs).
3. User failure: “oops! I hit delete!”
4. Catastrophic events (flood, fire), but those are a bit harder to plan for.

I’ve always backed up bits and pieces to various places, but a few weeks ago I decided I needed to take a serious look at protecting all of this information. The problem will only get worse — I will continually produce data for the rest of my life, and most of it I will want to save. My “My Documents” folder is already 40GB and growing. Other data/media is nearly 500GB.

So what’s my backup plan? Let’s keep it simple — reduce the risks of Problems #1, #2 (media and hardware failure) and #3 (user error). Problem #4 is a bit harder, but there is an easy solution if you just want to plan for worst-case scenario.

But first, you have to decide what to backup. I can think of three categories of data:

Categories of Data:

1. My personal documents. This includes school work, source code, emails, financial information, etc. I want to be able to save this stuff forever. Additionally, data from the websites needs to be backed up. Losing it would be catastrophic. This is priority 1.
2. Media. Movies, music, pictures, TV shows. If I lost this stuff, not a big worry — I just lose time re-acquiring. I can imagine that in the future this won’t even be a problem (bandwidth will be irrelevent). So not mission-critical, but helpful if I had some sort of redundancy. Priority 2.
3. Operating systems. When an OS crashes, it usually just costs time, but that can be one of the most frustrating experiences and ruin a whole day. There are two classes of machines (for me): home and servers. I don’t mind losing my home machines for a day to do an OS rebuild, but losing my web server for a day costs revenue. Priority 3 for home machines, Priority 2 for servers.

So a backup plan needs to be multi-tiered. One solution won’t work for every thing. So how do you protect yourself from losing data? What options are available?

Solutions:

1. Media. CDs and DVDs.

   Advantages: Cheap (both media and burners). Portable.
   Negatives: Unreliable due to scratching and life expectancy. Easy to misplace. Low capacity (5GB for DVDs). Staleness of data (it’s a pain to do backups each week).

   This solution should work for most people if they don’t need a lot of stuff backed up.

2. Hard Drives (External hard drive). A USB or Firewire drive that is connected when you want to backup your data.

   Advantages: Portable. High-capacity (500GB and growing). Fast.
   Negatives: Costly. Same unreliability as media (3 years?). Potential to break (dropping). Same staleness problem as media unless you have it automated.

   Probably the best solution for most people who have more data than media will allow.

3. RAID. RAID basically trades storage on one of your drives for redundancy (a backup). The best part is, the redundancy is handled automatically — if one drive crashes, the other drives have enough information that they will continue to work with no service interruptions.

   Advantages: High-capacity (multiple drives can work together). Fast (depending on implementation). High availability.
   Negatives: Costs more than drives alone (you need RAID controller cards). Not portable (hard to move from computer to computer). Lose capacity (trading space for redundancy).

   Probably overkill for many people, but excellent when you have to protect a lot of data in an environment that you need high-availability.

4. Drives on other machines. Using free space on other computers (on a network) to backup data.

   Advantages: Cheap (utilize space that isn’t used elsewhere). Fast. Can be automated to provide daily (or better) backups.
   Negatives: You need extra computers to make this work.

5. Revision control (Keeping older versions of documents in case you need to revert to an earlier version).

   This solution is mainly geared toward Problem #3, User error.

   Advantages: Changes can be removed.
   Negatives: Extra space is needed for each version.

   I utilize this for source code where I may be required to back-out a change if I find problems.

My Backup Plan

So what do I do?

My personal documents: Since this is Priority 1 for me, I have four levels of backup.

1. My Documents resides on a 4 disk, 750GB RAID-5 array (4x 250GB Western Digital WD2500YD disks). These drives come with a 5-year warranty (most consumer drives carry 3 years). I can lose any one of 4 drives, and all of my data will still be safe.
2. Bi-weekly, I backup to an external hard drive. Afterward, this is placed in a fire-proof safe.
3. I’ve burned my most essential documents onto DVD and have sent them to my parents in another state. This protects against some catastrophic events (fire) — and if both my home and my parents go up in fire on the same day, I have more important things to worry about (the end of the world perhaps?).
4. Revision control for my source code and some other documents, keeping all old versions.

My web server data, which contains things such as the web pages, databases and server configuration files. Also Priority 1. This data has three levels of backup.

1. Important data from the web server is backed up nightly with a 7-day history.
2. This data is backed up onto a second hard drive on the same machine nightly as well. The second drive is essentially a mirror of the first drive.
3. Weekly, this data is backed up to my home server.

My media, which I would probably go “eh” if I lost it, but I’m utilizing free space from the above solutions to provide backup.

1. I’ve copied much of the data I had on CD and DVD onto the RAID-5 array (about 300GB total).
2. The rest is still stored on CDs and DVDs where I won’t touch them unless need-be.

My machine OSs. I have a home machine, a home server (which does backup jobs and TiVo stuff), and my web server (which powers this and other websites).

1. My home machine’s OS is not backed up. I don’t mind having to reinstall Windows if need-be. If I was worried, I could switch to RAID-1.
2. My home server backs up important configuration files to another drive nightly.
3. My web server has two drives. The 2nd drive does a nightly mirror of the first drive, so it can replace the first drive with minimal downtime.
4. My home server also downloads the web server’s important data weekly to its own backup drive.

The home machines are all backed up with UPS power supplies. UPS is very important for my home computer, which has 6 hard drives spinning constantly. When you lose power, some disks might have trouble “slowing down” and could break. The UPS connects to the home machine and provides ~5 minutes of backup power (complete with the LCD monitor and networking if need-be), then automatically shuts Windows down gracefully.

Oh, I also backup my parent’s documents over the internet weekly — I don’t think they care all too much, but I’d like to make sure they’re safe, regardless.

So how could I lose it all? I’m sure there are flaws in my plan, but I’m protected on many levels against multiple problems. Nothing is perfect, but I do feel a lot safer knowing my data is safe.

Phew.

What do you do?