Choose a backup strategy

The main driving factor in deciding how you’re going to be backing up your data is really the data itself: what the data is, how urgently you need the data back if it was gone, where you’re backing up to, how much you’re backing up, how much you’re willing to pay.

Let’s look at each of these, then look at some strategies you can use to backup your files.

Things to consider

What kind of data are you backing up?

Not just what kind of data are you backing up, but:

Here are some examples of the kinds of data you could consider backing up, and some answers to the questions above:

Where are you backing up to?

Once you’ve thought about the data that you want or need to back up, you should start thinking about where you’re going to back up that data.

How important or replaceable your data is and the urgency that you have to get it back affect the choices you’re making here.

If you’re backing up multiple different things, you might be backing up those things to different places.

When you’re thinking about where you’re backing up to, you should consider a few things:

• How easy it is to use (“I already do stuff like this all the time” to “I need to learn something completely new”).
• How expensive it is (“I already have all required equipment or software” to “I need to buy new expensive hardware or software”).
• How reliable it is (“I still lose everything if my hardware fails” to “my house could burn down and I’m still OK”).

Here are some examples of where you could back up to:

How are you actually going to back up your data?

When you’ve decided where you’re backing up to, then you can think about how you’re going to make the back ups. There are two approaches to backing up data from a computer:

1. Image-based (drives or partitions).

   This is making a byte-for-byte copy of your entire disk.

   This is good because you’re copying literally everything that’s on your drive. Ideally this means that you can take a complete image, remove the drive from your computer (or move to a different computer), then restore the image to the new drive or computer and immediately have an exact copy of the whole thing.

   This is bad because it takes a lot of space (you’re backing up a lot of stuff you don’t care about) and because you’re effectively treating all the data on your drive as the same kind of data (your operating system is not as important as your pictures). You also can’t trivially extract a single file from a drive image (you can, but not trivially).

2. File-based.

   This is copying just files and folders (and maybe their metadata), and usually this means that you can open the folder in your file browser and actually see the files.

   This is good because you’re only taking as much space as you need to back up what you’re trying to back up, and it means that you can be more granular about your approach to backing up your files (you can put pictures on tape in a safe deposit box and just not back up your downloaded music at all).

   This is bad because you have to intentionally back up what you want to back up. You also can’t (or shouldn’t) back up an entire operating system this way.

Software

Let’s look at a few examples of software for backing up your files and evaluate them.

Copy and paste

This is… copy and paste.

cp -r source destination

You can be more complicated than just simple copy-and-paste, though: something like rsync copies files, optionally preserves metadata, and can delete files in the destination that have been deleted in the source (synchronizing the contents of directories).

Evaluation:

• ✔️ Super easy to use, you don’t need any special software.
• ✔️ Free!
• ✔️ Granular, you can choose exactly what you want to copy.
• ✔️ You’re in control of your data at all times.
• ❌ Doesn’t preserve metadata (unless you’re using something like rsync).
• ❌ Isn’t versioned (you don’t have the current version and old versions in your back up, just the current version).

Live-synchronizing folders

Some examples of software enabling live-synchronizing of folders include:

• Dropbox
• OneDrive
• iCloud
• Seafile
• Backblaze

Evaluation:

• ✔️ Generally very easy to use, even if you need special software.
• ✔️ Usually free to use, but costs are also minimal (usually single digits per month).
• ✔️ Granular, you can usually choose which folders to synchronize.
• ✔️ Versioned, you can usually go back to older versions of files.
• ❌ You have to trust someone else to host your data (Seafile can be “self-hosted”, see the Seafile manual for more information).

Disk imaging

Some examples of software for disk imaging are:

• CloneZilla
• Acronis Cyber Protect Home Office
• Redo Rescue
• dd for Linux/macOS can copy entire disks (see man dd).

Evaluation:

• ✔️ ❌ Some are easy to use (Redo Rescue, Acronis), some are not (CloneZilla, dd).
• ✔️ ❌ some a free to use (dd, Redo Rescue, CloneZilla), some are not (Acronis).
• ❌ Not at all granular, you copy an entire disk or partition.
• ❌ Not versioned.
• ✔️ You’re generally in control of your data.

3-2-1

Software enables you to actually copy or back up software, but there’s still strategy for how you manage your back ups.

One strategy is called “3-2-1”, and it’s summed up as:

• Have at least 3 copies of your data.
• Have 2 local or on-site copies of your data on different devices (two different computers, one computer and one USB drive, one computer and CDs or DVDs).
• Have at least 1 copy be remote or off-site.

You can read a lot more about “3-2-1” and its variants like “3-2-1-1-0 or 4-3-2 backups” (e.g., Seagate, Carbonite, Veeam), but this is a pretty straightforward strategy.

This strategy tries to mitigate or prevent several different things:

• Having 1 copy of your data off-site means that your data is resilient to a physical disaster where the other copies of your data are stored (e.g., fire, flood).
• Having 2 local copies means that your data is resilient to hardware failures (e.g., your disk dies, the CDs or DVDs you burned aren’t readable later).
• Having 3 copies in total means… you’ve got 3 copies in total. The likelihood of one failure happening is moderate, the likelihood of two failures happening is lower.

Restoring and verifying backups

Having your data copied to one or more other places is great, but you can’t be confident that your back ups are actually back ups until you’ve demonstrated that you can restore the back up (you can actually get the data back you thought you were backing up).

How you restore and verify your backups depends on whether you’re doing file-based backup or image-based backup.

File-based restore and verify

Verifying that your files are backed up can be as straightforward as opening your file explorer or web browser, navigating to the location where your files are backed up to, and then just manually inspecting if you have actual photos, files, and folders in that location (i.e., literally open the files and look at them).

That’s effective and easy, but not very scalable — checking hundreds or thousands of photos is tedious and painful.

We’ve seen the idea of diff before in the context of comparing plain-text files, but we can also use diff-like-tools to compare entire folders for differences.

Here are some tools that will allow you to compare entire folder structures for differences:

• Meld (Meld works on all of Linux, macOS, and Windows)
• WinMerge (WinMerge only works on Windows)
• KDiff3 (KDiff3 works on all of Linux, macOS, and Windows)
• FileMerge (FileMerge is only for macOS with Xcode)

The general strategy here is this:

1. Make the entire backed-up folder structure available on your local machine. This might be downloading your entire backed-up folder structure from the cloud provider you’re using (Dropbox, OneDrive), inserting a CD or DVD, plugging in an external USB drive, or using rsync or scp to copy your files back to your computer.
2. Open your visual diff tool of choice and compare the two folders.

Here’s what Meld looks like when you compare two folders to one another where the two folders have no differences:

Here’s what Meld looks like when you compare two folders to one another and there’s one or more differences:

Ideally, you should see no differences between what you got from your backup and what you have locally.

Image-based restore and verify

When you back up your data using an image-based approach you wind up with a single, very large file (many GBs).

This file contains all the bytes that are either on your entire drive, or all the bytes that are on individual partitions that are on your drive.

There are two ways to verify the contents of these files:

1. Mount the images and use the same approach as File-based restore and verify.
2. Write your image(s) back to the same machine or a different machine.

Mount the images

This approach is primarily limited to Linux or macOS. On Linux, you can treat an individual file as a complete drive by using a “loop” device. You can take an individual file and “mount” it a though it were a drive. This is a feature of the idea that “everything is a file”.

You can do this on Windows, provided that you have something like the Windows Subsystem for Linux installed (which you should now!).

Here’s an example of how you might mount a partition image from Windows (using the NTFS file system), this might not work depending on the distribution that you’re using:

mount -t ntfs disk1.img /mnt/windows

Once your disk is mounted, you can compare the contents using the approach described in File-based restore and verify.

Write the images

The other option is to use the same software that you used to create the images in the first place (CloneZilla, Redo Rescue, Acronis, etc) to write them back to either the same device (dangerous!) or a different device that has similar characteristics.

If you can write the images back and boot your machine again, and then use the same approach as in File-based restore and verify, you’ve got a working backup 🎉!