If you actually follow this blog, you might have noticed it was down…
for a week… I know, I know…
It was a combination of “boy cries wolf” fallacy, weird MYSQL behaviour
and possibly the image on Digital Ocean being too small. I’ll go over
each one for future reference.
The blog runs on WordPress and also has the JetPack adding installed.
One of the useful features of JetPack, provided for free by
Wordpress.com, is that it sends a heartbeat to the mothership, and I get
an email if the blog skips a beat. However because I’ve been running
this server with 1GB RAM, doing way to many things with it, mysqld
sometimes chokes and holds it’s breath for 2-3mins. This means that
every day I get 5-10 emails, in twofold one for holding it’s breath and
one for taking one, that the blog was down but is now back up.
Which brings us to the little shepherd. A week ago I got the email that
the blog was not responding, but didn’t notice that I never got a
confirmation that it was back up. Mysql decided to asphyxiate, if I may
stretch this metaphor by it’s sheep tail
I finally managed to look into it today, thinking it would be a simple
matter of restarting the mysqld. Nope! Didn’t work; apparently with no
useful message either. Up until I did this:
sudo tail -n 20 -f /var/log/mysql/error.log
and the problem was revealed. For some reason having the bind-address as
localhost, while working just fine in the past, didn’t seem to
work anymore. Changed it to:
bind-address = 127.0.0.1
and “Hey presto!”: mysql was back online.