Oh no! Not now!


Right in the middle of NaNoWriMo, the server got the hiccups.

First, the blogs stopped responding. Then the main site went down. I did what any stressed dev might do and restarted the server.

Bad mistake!

The MySQL server refused to come back up. Just totally refused.

After some time I found an error log file that had gotten a bit big. I removed it but the server still reported a file space problem.


dev in painOkay, stay calm.

I will not panic, I will not panic…


It turns out that sometimes you also need to restart services that were using the file so they give the space back. I’m sure there is a good reason for that but… Right now I am no fan of that feature. Actually, that’snot true. I know exactly why things work like that and it is a really good design. It comes up so infrequently that I had forgotten about it.

Anyway panic over I restarted the database and everything was fine, right?

No of course not.

One of the blogs – my blog – had a corrupted table and so was still acting like it was down.

I tried not to headbutt my nice new desk and ran the table repair. MySQL has this “magical” ability to fix itself if you “ask really nicely”.

Panic over…

Note to self, get with migrating your older stuff that makes big log files to something a bit less aggressive.

The image used for this post was first used by one of our bloggers. It seemed apt.

Not sure what your hosting platform is, but don’t the log files get recycled? (not sure what the technical term is). Normally log files would be gzipped after a period of time, or when they reach a certain size, only a limited number would be kept.

You are right Danny. Log Rotate would have eventually taken care of the problem. Not logging a lot of nonsense was just more efficient. It was pretty much just an edge case mistake on my part.


Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

no spam
A few potential bug fixes and spam defences

I’ve been looking into two seemingly unrelated bugs that I now suspect share a root cause. This evening, I added some new directives that control cache expiry. In both cases, this is for JSON content which is what the cache shows you when you want to see a post and …

CSS error on the front page

Something has gone screwy with the CSS for the front page. I’m already working on fixing it.

Rare JSON/HTML bug

I have been made aware of a rare bug that may only impact Safari users. The exact cause is unknown but most likely cache-related. It can cause JSON to be served up instead of HTML for new posts. I believe that 99% of users will most likely never see the …

Skip to toolbar