So about a week ago I switched hosting providers for this and 17 other personal low-traffic sites. In the process I took the chance to get familiar with Nginx, MariaDB, AWS EC2 Instances & EBS Volumes, and Capistrano. It was a great exercise, and surprisingly it worked on the first go! (at least for the majority of the week — read on)
For running and configuring WordPress on Nginx, I found the codex article a little confusing, but found this well-written article (with pictures!) by a WPMU contributor. Though I’m generally not a fan of WPMU’s plugins, this article put everything nicely in order for my mind and helped a great deal!
With all this newness came a growing pain yesterday. I’m running a micro instance on AWS. With the default MariaDB config and using InnoDB tables, it was trying to allocate 256M of memory when there wasn’t that much available with the other services running on the server. The way it manifest itself though was the worst…silently failing and then not allowing me to start the MariaDB server again. After several hours of Googling and downgrading from MariaDB 10.0 to MariaDB 5.5, I discovered the memory issue and got it working again! I needed to remove (rename) the existing InnoDB log files and modify the
/etc/mysql/my.cnf file to use only 50M (like I said, pretty small sites):
-innodb_buffer_pool_size = 256M +innodb_buffer_pool_size = 50M
#protip initialize a local git repository before you change system config files!
The adage is true — the best way to learn something is to break it! Let me know if you find anything amiss or just downright broken ;)