Majnun Ben-David

Archive Backups: Why Time Machine Is Not Enough

The noted author Ayelet Waldman is more careful about backing up her computer files than most people. She uses both Dropbox and Time Capsule. Last night, she realized that she had lost all of her writing prior to 2013, even though Dropbox and Time Capsule worked exactly as designed. Her pain was palpable on Twitter, and her resolve to move forward and "write better" is impressive. I wish her the very best in those efforts.

I wish I could have helped Ms. Waldman. But it occurred to me that I could perhaps at least help others from suffering through the same thing. Because remember, Ms. Waldman did have backups. She was careful, but she got tripped up with an often-overlooked (even by techies) limitation to automatic backup solutions. And there's a simple and cheap solution to that limitation, which is the point of this blog post.

There are two kinds of backups, and discovering the difference can be a painful experience. Below I explain the difference between these two kinds of backups, and describe a cheap and easy solution for writers that dramatically reduces your odds of losing data.

TLDR: Periodically copy all your writing files onto a flash drive as an "archive backup" because automatic backup systems [including services like Carbonite] eventually replace old backups with new backups which can be very very bad.

One kind of backup system is what we'll call "automatic backups." Examples of this are if you work out of a Dropbox or iCloud or Windows Drive folder on your computer, or if you have Windows / Apple's Time Machine set to automatically backup your files to a second hard drive, or if you use a commercial service like Carbonite or Backblaze. The advantage of this method is that you don't have to do anything, but your files are still backed up.

Sort of. There's a vitally important twist here that is often overlooked. Automatic backup systems generally work by synchronizing the files on your computer with the files stored on the backup drive, whether that backup drive is in your house (Windows/Time Machine) or in the cloud (Dropbox / iCloud / Windows Drive).

Imagine you have a file called "My Best Short Story Ever" that is full of awesome text. One day, your cat deletes all the text in that file and replaces it with "feed me." No problem, because you have a backup, right?

Maybe. It depends on how soon you notice what happened. Because the automatic backup system has no way of knowing that you did not want to change your story to the minimalist and innovative "feed me." So it will happily back up the new version of "My Best Short Story Ever." Most automatic backup systems also keep older versions of files for a while, but generally not forever. Eventually, the old backups get erased by newer ones. If you don't notice the change in that file until after all the backups are of the new "feed me" version, then your original version is gone. The same process applies if your cat simply deletes the "My Best Short Story Ever" file.

How long the old backups stay around depends on the total size of all the files you are backing up relative to the size of the backup drive you are using. To take a simple example, imagine you have 1 gigabyte worth of files you are backing up, your backup drive (whether local or in the cloud) has a capacity of 25 gigabytes, and your computer is set to back up its files once a day. In theory, then, your "backups" only represent the last 25 days of activity. So if bad changes were made to a file, or a file was accidentally deleted, you'd only have 25 days to notice it and retrieve the "good" version from your backup drive. In practice, you have longer, because most backup systems only save copies of files that have changed, rather than all files, but you get the idea of why you only have a limited "window" of time to notice and fix a problem.

Commercial services generally have a much shorter time window for you to retrieve deleted files. For example, both Carbonite and Backblaze give you just 30 days before they will delete their backup of your file. So if you only realize on day 31 that you messed up or deleted a file, you are out of luck.

In Ms. Waldman's case, the window of time on her setup was about 8 months. Unfortunately, it took her 9 months to notice that she had accidentally deleted all her old writing. So all of her backups, to Dropbox and Time Machine, were of her computer as it looks today -- without the deleted files.

That's why you really need to have the other kind of backup, what I will call "archive backups." An archive backup is a backup that is NOT created automatically, and after it is created, it is left alone, unless you need to retrieve files from it. For writers, this is easy and cheap to do. I suggest doing it at least once a year, and perhaps as often as once a month.

Here's how you can create an archive backup in three easy steps. The overview of the steps is as follows, with details below. Step #1: Buy some flash drives. Step #2: Copy all of your writing files to a flash drive. Step #3: Put flash drive somewhere safe. Then, at least as long as the flash drive lasts (should be a very long time), you will always have an archive of your writing files as they existed when you copied them to that flash drive. Because it is not even connected to a computer, it is not subject to being overwritten. Repeat this periodically.

Step #1. Get several "flash drives" also known as a "thumb drives" or a "usb drives." Flash drives are on
Amazon here, and they are available at all kinds of stores (Target, Radio Shack, I even see them in some grocery stores). Don't worry about the USB 2.0/3.0 or any of that argle bargle in the description. All that should matter is the capacity, given in gigabytes (GB). Text files are small, so the smallest size flash drive should be plenty big enough for an archival backup of your writing. Right now, the smallest flash drive commonly sold seems to be about 8 gigabytes. That will hold at least 500,000 pages of text in Microsoft Word format. An 8 gigabyte flash drive can be had for about $5, or in bulk for even cheaper (at the time of this writing, Amazon was offering ten 8 gigabyte flash drives for about $25). For prolific writers, 32 gigabyte flash drives can be had for under $20 each.

If you want to include all your files, such as photos, you can use a regular hard drive in place of a flash drive to dramatically increase the storage capacity. Of course, regular hard drives also cost more. At the moment, you can get 2 or 3 terabyte (TB; one terabyte is a thousand gigabytes) external hard drives that also connect to the USB port on your computer
for around $100. One terabyte can hold about 200,000 photos at a typical resolution. In the following steps, I'm going to assume you got flash drives, but it works the same with an external hard drive.

Step #2: Plug one of the flash drives into the USB port of your computer. Copy all your writing files (or whatever files you want to archive) to that flash drive. If you're good with such things, you might want to change the name of the flash drive to something like "Archive Jan 2015." Also, it's also pretty easy on most systems to password protect a flash drive if you're worried about that (Google is your friend here). I would also suggest physically labeling the flash drive -- a quick and easy way is to toss the flash drive in a ziplock with a piece of paper that has the date of the archive and what it is, along with the key phrase "do not use!" (you don't want your kid using it to take their homework to school or something).

Step #3: Disconnect the flash drive from the computer and put the flash drive somewhere safe. Of course, there is no such thing as a perfectly safe place, but it's probably best to keep it in a different room than your computer. The big risk is probably not remembering where you put the backups, so I'd advise against hiding them. The advantage of flash drives being cheap is that you can make several archive backups for very little money. If you want, make several and keep them in different places, maybe one at a relative's house, one at work, etc. But be sure to label them and put "do not use" on them. And don't use them unless you need to recover a file -- that's the only time they should be connected to a computer. And don't re-use the same flash drive for multiple archive backups; it should be one and done (again, they are literally cheaper than a cup of [Starbucks] coffee).

At this point, you have an archive backup of all your writing files. Whatever happens with your automatic backup, you can always go back to your files as they stood on the day you made the archive. So if your cat replaces the text of "My Best Short Story Ever" with "feed me," and you don't notice it until all your automatic backups have the "feed me" version, you could still pull the original version out of your archive.

How often to make such an archive depends on how risk-tolerant you are. I would suggest once a year, at a minimum. Personally, I have a schedule of once every 3 months, because my automatic backups cover at least 9 months. Or you could customize this approach to fit how you work -- maybe every time you finish a submit a short story, you make a new flash drive archive. Again, with 8 gigabyte flash drives going for less than $5, there's no reason not to have multiple archive backups.

The downside of archive backups is that they rely on you to make them. That's why having an automatic backup system is crucial. But most all automatic backup systems will eventually replace old backups with new backups, which can lead to the loss of important data if mishaps are not recognized promptly. So get a flash drive and make an archive backup. Let's learn from the loss of others.

Parting Thought: Hemingway also lost most of his early work, despite having the best backup system then available (carbon copies). His wife at the time, Hadley, packed up all his writing (originals and carbon copies) to bring it to him in Switzerland in December 1922, but the bag with it gotten stolen off the train. More on that

Disclaimer: None of the above links are "referral" links (i.e., I don't make any money if you click on them). This is an entirely non-commercial blog.
blog comments powered by Disqus