Just thought I’d do a quick post on the stability of this server, given the hours of downtime today.
Long story short, one of the cronjob tasks that goes through and closes expired orders for EVE Metrics did so, then crashed and burned quite rapidly, chewing up all the RAM on the server and expertly smashing some queries into MySQL which promptly made MySQL cough up blood and fall over. Several uncacheable SQL-heavy API calls to EVE Metrics later and Apache became inundated with slow-completing requests and started spawning more threads, so on and so forth, and there goes the server. What _should_ have happened is that the process resource monitor script should have caught the rake task eating a gig of RAM and killed that, but for some reason I’m still trying to work out it didn’t. In either case, the server is now back up and stable again.
I run this server entirely out of my own pocket for services I provide to the community, and with the exception of one person who bought a GTC through the link on this site, I don’t get any financial aid on that front. So for now, a single point of failure is the way it’ll have to be unless I can convince my host to provide me with a second server and load balancer for free 🙂
If you want to help me manage costs and improve my stability, then buying GTCs via the link on this blog is a good way to go about that; alternatively, feel free to donate via the link on the sidebar, where I’ve set up a Pledgie campaign for the donations. Every little helps (Unless it’s a really little chunk in which case PayPal will steal all of it and give me pennies, which aren’t very useful), so please do donate. Anything left over at the end of each month will be donated to Cancer Research UK rather than being spent on mroe bree.