There aren’t a ton of overlaps between my day job and my secret life as an author, but the one that seems to tie them inexorably to each other is the work that goes into maintaining my website. I’ve probably selected a path that is atypical for most authors in that the solution I chose a number of years ago is a balance between being hosted in the cloud and having full control over the actual content there; my reasoning at the time was that websites are kind of my thing (hence wanting full control) but I had no desire to actually maintain the hardware to run them (hence the hosting). For the most part that has worked out quite nicely, though every now and again I run up against some sort of oddity that requires me to don my Tech Nerd hat.

Case in point: somewhere along the line last spring, I lost access to the open source image library my blog often uses. It wasn’t a big deal as I have plenty of my own photos to use, but the convenience of finding an image that “fit” whatever article I was writing instead of hunting through my own files was pretty nice. Then, maybe in mid-summer, I began to see an error when crafting these posts; it was incredibly generic, though, and after a number of fruitless web searches, I had absolutely no idea what was causing it to appear. Posting content didn’t seem to be affected, so like a terrible end user, I kept on grooving.

Okay: in fairness the techie inside my brain was paying attention and worried that the message was some sort of canary in a coal mine. And, as it turns out, said canary was right on the money, for in August I suddenly lost the ability to share my posts across all of my social media platforms. That impacted me far more than anything else, so I dove back into everything again and pulled back far more layers of the onion than ever before. The culprit appeared to be a base module in my website, one responsible for both preventing spam but also was keeper of the statistics; that the reports on web traffic suddenly all slipped to zero around the same time felt like a clue even my fictional detectives might not have missed.

And yet, armed with such knowledge I still had no idea how to fix the problem. Endless searches across the web resulted in answers ranging from delete your entire website and start over to jettison that tool and buy ours for the low, low monthly fee of $x per year. None of them seemed right, especially to someone used to looking through error logs and not predisposed to assume the initial error is actually what’s going on. I’ve written plenty of code in my years as a software developer, enough to have suffered a multitude of self-inflicted bugs that have long since made me paranoid and suspicious.

So, back to the logs I went, hunting for what I presumed would be hiding there: much like Vasily scouring the financials from a suspect, I scrolled endlessly looking for some sort of anomaly that, on first blush, might seem innocuous. This past weekend, that review finally paid off when I discovered my troubles started about when this module updated last spring; each successive increase in problems coincided with another update, almost as though the features in the package were somehow slowly becoming incompatible with my website.

Now, you might be asking yourself, how can that happen?

That’s a great question, and the answer lies in how a developer might handle what we think of as legacy applications. In my case, the software running my website is so-called backwards compatible to several older versions of server operating software. Unfortunately — or, at least as far as I can divine without decompiling the code in question — this module that is integrated into my website is making use of newer features in newer versions of that server operating system, one that, as it turns out, I was not running.

At least, not until late Sunday afternoon. Once I tracked down how to request the upgrade from my hosting company, it took another few minutes to reset a few authentication tokens and just like that… no more errors. Not only that, both the open source image library and my ability to post to social media also magically returned. From what I can tell, it seems my willingness to allow the website software to automatically keep itself up-to-date ran up against the hosting company’s policy of deferring to the client on such things. Is that a bad thing? Probably not, though it also tells me I need to be more vigilant when the website software proudly informs me it’s now running the latest and greatest whatever.

Oddly, that is so not how I do things in my day job; we try extremely hard not to upset the applecart with any of our updates, nor do we embrace enhanced features until we are sure all of users can, well, use them. Some of that is just policy I’ve developed from years of being in this business, but the rest is a healthy understanding of what a mistake can do to a client. I feel fortunate that I was able to troubleshoot my personal website, but I am keenly aware not all authors are comfortable in that space. I’m sure that’s why there are hosting solutions out there that require far less technical expertise, but the tradeoff is a severe lack of flexibility. Maybe it’s worth it to some, but for me, for now, I’ll stick it out with what I have currently.

At least, until the next thing breaks.