For the second time this month, my “managed” web server provider has, with no explanation, allowed the web service on my server to fail in such a way that I couldn’t restart it even by rebooting. As a result, the blog was unavailable for extended periods of time. Because the provider manages the server for me, in order to fix something, I have to submit a support ticket and wait for them to respond. The latest outage was for an entire day.
Even though web server administration is part of my day job, I chose managed hosting because I wanted to spend my free time blogging, not administering. With these recent issues, I’ve come around to the old notion that, if you want a job done right, you have to do it yourself.
The good news is that virtual server technology has changed a lot since I put my blog back online in 2005. I was able to follow an excellent how-to guide and get an Amazon EC2 instance provisioned, with database and web services running and tested, in about 90 minutes this morning. I hope to finish the configuration over the next few days and migrate the blog this weekend. With a bit of luck, the next blog interruption will be a momentary hiccup as I update the DNS entries from the old server to the new one.
I’ve decided to hold off on bad-mouthing my current provider by name, at least until the migration is complete. Of course you could probably figure it out by doing a traceroute.