I’ve been meaning to update the blog for a while as some important stuff has happened.
The MySQL High Availability project went live in early January and is a success 🙂 We made some changes though. Sadly, we decided against read query scale-out because we found that extremely large queries took a noticeable time to propagate. So now read and write queries both loopback. A shame as the application felt faster with read scale out and I deliberately upped the query cache on the read server. Of course, being reliant on a single database for reads also introduces a single-point-of-failure (SPOF). Thinking about SPOFs in other arenas has lead me to reliable ssh tunnel projects like autossh, which is something I can make use of elsewhere.
Debian 5.0, Lenny, was released! As a sysadmin it is both exciting… and a bit of a pain in the bum. Probably in equal measure. In non-production tests, Etch to Lenny dist-upgrades seem to be pretty painless apart for some small oddity in MySQL not coming back up (just start manually). Also we are faced with an upgrade of Apache1.3 to Apache2.2. Lets see what happens there.
I’m getting into Catalyst these days. Essentially, to Perl what Rails is to Ruby. I have plans to implement an existing Perl/mod_perl web app in Catalyst as a way of learning it. It’s quite exciting stuff. Tomorrow I attend an Advanced Perl class which also touches on Catalyst. I have been hacking it on and off for the last 6 months but I just feel I need to kick it into a higher gear and maybe this class will enthuse me. Another idea I had was to create my own blog in Catalyst (a bit like Jono Bacon). Part of the logic in doing that is that in theory it will be more resistant to automated attacks.