A Cautionary Tale about Cloud Storage

I was an early adopter of cloud-based document storage. Who doesn't want immediate access to all of their documents in one place? Direct access to your data without the painstaking process of logging into a server? Today, Cloud-based document storage is seamless and reliable, and we, as people and small businesses, may put more trust in them more than they deserve.

Robust backups are more critical than ever to maintain data integrity in case of rare but disastrous failures.

The state of cloud-based document storage in 2021

I've tried many of the typical cloud-storage applications, and they all have pros and cons. Dropbox integrates with your devices almost seamlessly but has a heavy presence on your device. It is a resource-hungry application and is very pervasive in your local system. Despite a tendency to convert any document into a Google cloud format, Google has excellent collaborative storage options. Google's native files are slower and less full-featured than most of the desktop applications I use. And I feel they produce ugly documents. Also, Google will often slurp up any data it's allowed for advertising targeting. I always worry about my documents' privacy when using their services. Microsoft has come on strong with OneDrive if you are a Windows user. Box is another Dropbox-like service I've tried. Other companies come and go. But if you are storing your files online, you need to find a company you can trust. You can even spin up an AWS or Linode server inexpensively for a simple FTP-style server. Just don't forget to set the AWS bucket password to avoid having all of your data leaked on the internet.

As an all Apple device user, I settled on Apple's iCloud storage. It works seamlessly across all of my devices and part of the operating system. Additionally, the pricing structure works for me, as I pay for a family plan which allows shared usage of the 2TB bucket for all family members. As I've been using Apple's iCloud for many years, I feel comfortable with the data integrity.

A Cautionary Tale

When a program works flawlessly for some time, it is easy for one to become complacent and trust it implicitly. Apple's iCloud has become one of those trusted applications that I knew /just worked/. And I believe many other people feel the same way in choosing their cloud storage systems. Many people or businesses trust these services with all of their critical personal photos and documents.

Recently, after upgrading to a new M1-chipped MacBook, I noticed a sync problem on a settings file between my two Macs. Data I wrote into the file on one machine didn't show up in the other machine's file. Further investigation made it look like one of the devices wasn't feeding new files to the cloud. I signed out of my Apple ID and logged back in again to reauthenticate the computer and upload the data I wanted to store in the cloud, per a support document.

However, the behaviour of the log back in did not do what I expected. It blew away all of the files I had created since the disconnect happened. All gone! My desktop, where I keep my working files, was empty. My documents folder, missing months' worth of essential data. As the sync problem hadn't uploaded these files to the cloud, logging back in removed the files as I created since the last cloud sync point. Unexpected at best, and I'm sure I had a few other words in the meantime.

I feel this is a massive mistake in programming. There is no undo button. And no Command-Z would fix this one.

The case for backups

Luckily I grew up in an age where wearing both a belt and suspenders was a term for being overly cautious. Early personal computers were a lot less reliable than the quality of today's systems. I often experience data loss from a computer crashing or me accidentally opening a document and forgetting to save it as a new file. All mistakes from which I learned.

Previous data loss created my good habits for document storage and archive. I am so glad that these habits have persisted.

I use a couple of systems of backup. The first is that Apple's iCloud system replicates my local documents. The second is that I periodically hook up an external drive to my laptop to create an external Time-Machine backup using Apple's built-in tools. And the third is I use a trusted cloud-based third-party backup system.

A failure of iCloud caused my data loss. My backup disk was a few days old, as I only backup weekly. But my constantly running cloud-based backup had a full backup of everything I'd been working on, up to the current documents. Phew.

Backblaze to the rescue

Online third-party backup systems are available. They do come with an additional yearly expense (about $100 Canadian annually). I've used a service called Backblaze for many years, which backs up my hard drive data and directly connected local drives. The standard plan comes with 30-day retention and unlimited storage. I've never had to use the service for document recovery before, except when I first started using it to understand how the service worked and test my setup.

But am I ever glad I subscribe to this service! I logged into my account on a web-browser (including the 2FA token) and had access to all of my data. I even had options of requesting a drive of all of my lost data (in case of total loss) or downloading just the affected areas. That was all I needed, and I was back up and running in about 15 minutes. But what a lesson and a heart-stopping moment.

Recommendations

There are several tools for online backup services available. Get one. Use a hybrid approach of backups - as the old saying goes, "Two is one, and one is none." This expression is especially true for data redundancy and back-ups.

If using a cloud-based archive system, make sure that one computer always downloads your documents onto its hard drive. Then, back up that hard drive both physically (this is easier if you have a desktop) and remotely. It'll seem like overkill. And for many years, my system was. Until that one day, when I needed it, and it was there.