My Backup Workflow for Large File Repositories (Images, Projects, Etc)
We shift files back and forth trying to find the best backup strategy.
But the activity of moving files around only increases our chance of loosing important data.
At one point or another, we've all lost a file forever. Maybe it wasn't a significant tragedy, or maybe it was 10 years of your child's photos. Or the project due tomorrow. Or your client's wedding photos. As we generate more and more data in our lives, it becomes increasingly more important to solidify our backup strategy -- and stick to it.
Workflow is personal.
Sometimes when folks ask me about workflow, I preface by saying "this might not work for you". We need workflows that we are comfortable with and simi-automatic -- otherwise we won't do it. Or we forget a step and then make matters worse when trying to play catch-up. But I'm happy to share with you my workflow, let me know in comments if it works for you, or if you have a better system.
3 copies, or it doesn't exist.
Some of us may be familiar with this concept. Basically it means, if you don't have 3 copies of a file, in 3 physically different locations, then it doesn't actually exist. The intent here is to focus your attention on backups before you begin other aspects of your workflow.
I'm always a little tickled when people post images from our group photography events, even before I drive home. But I also wonder if folks have finished the critical backup stage of their workflow before they start editing images.
My workflow, and my recoverable backups
This workflow may not work for everyone. It's a methodology that works for me and may give you some hints for your own process. For this example, I'm using a workflow from a photography shooting event.
STEP ONE copy files - do NOT import directly from your memory card!
The first thing I do after plugging in my card is "COPY" the images from my card to a staging directory on my computer, usually on my desktop. Many times I short-cut this step and just move the entire directory.
The most important part of this step is making sure ALL FILES are copied over. If you are using slower cards and an inexpensive card reader, the computer can sometimes "lose" connection as files are copied over. The copy process will halt and you might think the process is complete (and format your card not realizing you're losing LOTS of images).
This is significantly worse if you try to "import" into your image management software (iPhoto, Lightroom, etc). If your card suddenly ejects and instantly reconnects during import, the software will pop up a pretty dialog box saying all files were imported, when actually they were not.
STEP TWO import files into Lightroom.
I wrote a blog post about safe Lightroom Catalog Backups a few months ago, so check out that article for more details. Each time I import a new batch, I create a new directory on my Drobo, a multi-disk storage device. This is where my "working files" will live forever.
So in a file directory called "2013", I will create a new directory called "2013-10-05 Grand Prix of Houston". This way when I order by directory name, all of the directories fall into "date order".
Note that I say "working files" and not "backup files". Although the Drobo Mass Storage Device system is often used as a "backup" system, I get the best use out of it as my primary working space.
The Drobo is a black box that sits on your desktop. You plug in 4 or 5 empty (and cheap) hard drives, and your computer "sees" it as one drive. As time goes on and you start running out of space, you pop out whichever drive is the smallest, and pop in a new hard drive of any make, model or size. Presto, more hard drive space w/out having to copy files all over the place.
This also makes upgrading to a new computer MUCH easier, since my work lives on my Drobo rather than my computer's hard drive. So if my computer crashes (or needs to be repaired/replaced), I can plug the drobo into a new computer and I'm off-and-running like nothing happened.
The Drobo is also redundant. If one of the drobo's drives goes bad (and you know they do/will), the other drives contain enough data that anything lost can be instantly recovered. All I have to do is run down to Best Buy, find a cheap replacement drive and plug it in. Presto. (Ok, not presto, it takes a while for the device to be ready to use again -- but at least I didn't lose any projects or data).
Finally, if the Drobo itself goes bad (mine is going on about 5 years old, so it won't be long), I can order another from the manufacturer and pop in the old disks. This is very similar to RAID devices. If your RAID enclosure burns out, your data is still safe on the drives. Unfortunately because each manufacturer uses their own modified version of the linux operating system, you are usually forced to buy another RAID enclosure from the same manufacturer, otherwise the drives have to be reformatted. Similarly, if your Drobo device fails, you have to buy another Drobo to continue using the data on those drives.
One of the criticisms for using external storage (like USB drives, etc) is that it slows your workflow. Drobo products you can buy today have Thunderbolt and USB 3.0. I don't experience a significant issue using the drobo for editing images. However, as mentioned in the other article, I do keep my Lightroom catalog data on my local hard drive.
Ok, enough about Drobo.
At this point, I have a single copy of the project files, stored on a redundant drive system. This is good and safe, but not quite safe enough. This is where Time Machine comes in. If you are using Windows, there are several Time Machine-esk utilities on the market. Time Machine makes a backup of my computer a couple of times a day. In my configuration settings, I added my Drobo to the list of things being backed up.
This means I have one huge Time Machine, and eventually I'll have to revisit this step. But for now, its a completely hands-off backup system. I now have two copies of my project: a copy on redundant file protection system, and a backup on Time Machine.
It's great that I have two copies. Unfortunately they are sitting right next to each other. If my house burns down, or hurricane or tornado, etc, then both will be lost. The third location needs to be "off site", somewhere completely different. Some folks use two Time Machines, and leave one at the office. Then they just swap them out once a week or month.
I've also heard of people using a safety deposit box. But that means having to retrieve it every time you want to update the backup. Kinda messy and painful to make a run to the bank over and over. Another option is to trade backups with family. You hold a USB drive of their backup and they hold one for you. That works, but still the same problem as the bank, just not as bad.
A much more realistic, and hands-off, approach is "cloud" storage. There are several companies offering secure cloud backup services, the one I chose with is CrashPlan. I went with their "Unlimited" plan, which is about $60 per year. However they are currently offering a 20% discount for new subscribers.
I installed the product on my computer, and configured which directories I wanted backed up. This took a long, long time. I was limited by my bandwidth, and I have LOTS of files (about 3 TB). When I say a long time ... it was measured in months, not hours. But once it was complete, a simple backup refresh only takes a few minutes.
CrashPlan will send you a hard drive that you can fill and send back to get your first backup started off quick. Since I was already using another system for off-site backups, I didn't pay the additional money for that service. However, if my house burns down, I will likely pay for their rapid return service, where they send you a copy of your data on a hard drive if you need it in a hurry.
So you may ask, what if CrashPlan goes out of business? That's about the same risk exposure as a failed hard drive. Since I have three copies, losing one copy means I have work to do to replace it, but not a tragedy. If CrashPlan goes out of business, or prices themselves out of my budget, then there are plenty of competitors I can switch to.
STEP FIVE - Productions
So now I have my three backups of all my files. However there are auxiliary and less obvious backups. For example, several of the places I upload photos to have them printed will keep files available for some period of time. Also, I have almost every image I've produced on my Zenfolio web site, MEDIA.24Moves.com. Likewise, I have fairly large versions of my images stored on Flickr as well.
Now this isn't all of my work, just the "good stuff" that was worth displaying. Just like you I have LOTS of other stuff from shoots that should probably be deleted rather than backed up. If I lost those source files, it would be sad, but having the productions available on Zenfolio and Flickr is a nice bit of extra security.
I do *NOT* get any warm feelings about the security or safety of using Facebook as a backup system. When you upload images to facebook, their image hosting software takes great liberty with any image that seems "too big". This includes down-sampling file quality and even reducing the file dimensions in some cases.
End of braindump. What's your system?
See something here that you might be able to improve? Or do you have a system that works better for you? Anything here helpful enough that it changes your workflow for backups? Please leave a comment and encourage my good behavior.