In response to a posting on Chris Scott's blog, Boss Ross points out what happened with Blogrolling yesterday:

Acutally, it was a Tucows mistake that caused the problem. A few weeks ago, we had to fail over to a non-production box because one of the servers that Jason had colocated us on died without much warning. It was the only way to keep the service alive on short notice.

Problem was, it was running slightly outdated code - the production server, at the time, had already been patched by Jason. By failing over to the older development box and turning it live without patching it properly, we reintroduced the problem that Jason had kindly fixed for us.

The box with the patch was never brough back online because we are in the process of moving off of the colo that we're currently at and into the regular Tucows environment at IBM - no sense in going through more trouble than it was worth we thought.

So the long and the short of it is that I really messed this up. Jason has been an amazing source of support through all of this (post-purchase) and I'd really hate for my mess-ups to reflect on him somehow. Tucows, and more to the point, me is the root cause of the problem that we saw today.

And now we need to pick ourselves up, dust everything off and move on - without repeating the same mistakes that we've just fixed.