As I mentioned in my last post, one of the things I did recently was move Sett to a new server. This is a task that I had every reason to do five years ago, but had been dreading and putting off. It was never that urgent, wasn’t moving me closer to any major goal, but most importantly it just sounded like a miserable project.
The most daunting part of it all was that all of the software that Sett relied on was horribly out of date. I was two major versions of PHP behind, and each of its 5-10 dependencies was certainly either obsolete or out of date. Of course, as time went on this disparity became even greater, making me even less likely to want to do it.
At first this cost me about $170 a month, then I finally downgraded our server (but used the same image) to save about $40 a month. Overall, it probably cost me about $10,000 to not move servers! Even greater than this cost is the constant burden of knowing in the back of my head that I should move it and having to make the decision of whether or not to do the work.
Finally, in quarantine, I decided to take a stab at it. I resolved to spend half a day working on it and reassessing from there. If it was going to require too much of a rewrite I would try something else. It was hard to know exactly how long it would take, but it felt like a 5-7 day project to me.
My half a day of effort got it about 95% there. Our search backend, solr, took a few frustrating hours to upgrade, but other than that it was shockingly easy. In fact, it worked so well that I had to check multiple times to make sure it had actually moved to the new server and I wasn’t just loading the old one. I was even a bit surprised when it kept working when I turned off the old server. Over the next day I got it to 100% and made some additional bug fixes and improvements.
Clearly, I should have just done this earlier. Had I done so, it would have taken even less time since I could have done smaller incremental upgrades. What I thought would take a miserable week could have been done in a reasonably easy day.
So what caused me to make this $10,000 mistake?
First, I massively overestimated how hard it was going to be. I’m not exactly sure why I did this, but I think that it was just distance from the Sett code. When I was working on it daily I knew every nook and cranny. Six months after touching it I forgot how it all worked and wasn’t particularly excited to dive back in.
The right remedy for this problem is to do what I did five years too late: allocate some small amount of time to working on the problem and give myself an out if it looked too nasty. Had I dedicated even a couple hours to it in the first year of shutting down Sett (when I knew it would no longer grow), I could have saved all of this money and cognitive load.
I’ve written about this before, but I should have also applied the principle of “I’m going to do this anyway, so I might as well do it now”. When we shut down Sett I agreed to take sole custody of it because my blog was on it so I actually cared. There was no way anyone other than me was ever going to move the server, so I should have just bit the bullet and done it. I’m sure there were weeks or months that I was too busy to add it to my plate, but was there a week somewhere in the past five years where I could have just scheduled this? Of course.
Last, I underestimated how great I’d feel once I completed the project. It’s great to save money, but the biggest benefit is that I just don’t have to think or worry about it anymore. I no longer have to deal with an expensive and out of date server and all of my sites are now in one easy to administer server. I think that we all consistently underappreciate how much nagging todo list items affect us.
Though none are are close in magnitude to moving Sett, many of the projects on my coronavirus list fell into this broad category of “non urgent tasks I probably should have already done”. Even something as small as making a bigger bed has been something I thought about at least weekly. Not having to decide not to do these things all the time frees up my mental space considerably.
I suspect that all of us have tasks like this, and most of us have an unprecedented opportunity to tackle them right now during coronavirus. As I’ve mentioned before, I’m constantly thinking, “how can I make this forced quarantine the best period of my life”, and knocking out all of these sorts of tasks is a major component of that.
What tasks do you consider doing but keep putting off? What overdue tasks weigh on you? Pick one and spend an hour or two making inroads on it today. If you’re like me, that may lead to you realizing it’s not so bad and maybe that will snowball into you clearing your todo list as I have.
Photo is another drone photo from Lake Mead. I promise there will be pictures of other places as soon as I can travel again!
I started a Patreon. Thanks so much to the people who have supported it so far!
I like to answer questions on Twitter. I have started working on defeating spam on Sett again so I may test opening up comments again, but until then I’m available on Twitter.