WordPress Permalinks suddenly not working, here's one possible fix

Yesterday, I suddenly spotted some 404 errors at TinkerTry. Folks out there were searching for particular stuff. Their search engine was kind enough to spit out a direct link to one of this site's 276 articles. Often about stuff you won't find anywhere else. But after clicking on a link, visitors were sometimes greeted with a 404 error, instead of the actual article. Especially for my most popular articles. Not good. Must fix. Quickly.

Even worse, internally, some links to some articles in categories views, such as the Hyper-V category:

https://tinkertry.com/category:Hyper-V

were broken too. Uh oh. While I worked to fix this issue, at least the main TinkerTry home page kept cruising right along.

I may never really know exactly what brought on my broken permalink problem. Perhaps it was recent dabbling in Bluehost+CloudFlare integration, since that required changing tinkertry.com to www.tinkertry.com, explained nicely here, and changed over on Settings, General. Then I got to thinking that seemed to be a very unlikely culprit, since that change happened back at TinkerTry re-design launch, well over a week ago. This permalink breakage seemed to be much more recent. So I started the troubleshooting by focusing my efforts, using tests on one broken link to try to reveal what was going on.

At some point months ago, after publishing VMware ESXi 5.1 can run Microsoft Hyper-V Server 2012 VMs, nice!, I made a change to the permalink. The original URL was:

https://tinkertry.com/esxi-5-1-running-hyper-v-server-2012

but at some later point in time, I had changed it to:

https://tinkertry.com/esxi-5-1-can-run-hyper-v-server-2012

Thing was, only the original URL was working:

https://tinkertry.com/esxi-5-1-running-hyper-v-server-2012

Normally, this sort of renaming is not a big deal at all, even if not advised. Anybody or site using the old URL would seamlessly and automatically get redirected to the new URL. If I rename again, the original, the second, and the third URL would magically just keep on working. Nice!

Well, here's where things got whacky.

Just to see if I could workaround the issue, I tried renaming the URL to:

https://tinkertry.com/esxi-5-1-can-run-hyper-v-server-2012-in-a-vm

This new URL didn't work either. This worried me. How to recover? In the midst of working on some CloudFlare related DNS clean-up with Bluehost technical support, I casually asked if the representative if he had any ideas for recovering a messed up permalink structure in WordPress, just in case my recent changes had somehow contributed to confusing one of my WordPress Plugins, for example. I realize Bluehost isn't my Wordpress support. But amazingly enough, the fix Bluehost suggested was that I give a super-simple fix a try. And it seems to have worked!

I simply went to my WordPress admin menu, under Settings, Permalinks:

https://tinkertry.com/wp-admin/options-permalink.php

and was advised to change nothing, merely clicking on the Save Changes button, then waited 3 seconds, and tada, it says "Permalink structure updated." pictured below.

Skeptical it'd do anything, I tested the URLs above again. Not fixed. Now more skeptical. Flushed all caches through the usual W3 Total Cache Performance, Empty All Caches menu. Still not fixed. Turns out waiting about 15 minutes, fixed. Everthing just started working again. Could rename the permalinks. Excellent. Old permalinks worked, new ones worked too, like it should.

Phew, this was a relief. Worst-case could have been to have to restore my 1.7GB site on Bluehost from an earlier time, back when permalinks worked. This would have brought the site down for an hour or two, and been a bit time-consuming for me. So yeah, happy the fix wound up being so simple. Perhaps it'll work for you too, and doubt it'd hurt anything to try.

With that bullet dodged! Now I can go back to focus on returning to virtualization and home server based articles, and work on the back-end to reduce my page load times, considering W3 Total Cache + Amazon CloudFront.

Here's the screenshot of the fix: