After a switching away from Apache some time ago, our primary web server had been running Cherokee for quite a while - since September 2011, in fact, looking back at the configuration history. More recently, however, I’ve switched us again. This time to Nginx - with impressive improvements in performance and configurability (10x for some static files) — and reliability.
See my technical guide for technical info about switching.
Originally, the selection of Cherokee as web server software looked promising - the Cherokee project had been around for a while, was gathering support, had a reasonable level of uptake, and consistent, active development. All promising signs. It also had a variety of novel features at the time - flexible conditional rules, an app marketplace (something it no longer has, but I’ll come to this shortly), a GUI for configuration, and was very shiny for want of a better word. In terms of performance, it had outdone the other options, so I went through the rather tedious process of rebuilding the configuration from Apache into Cherokee. Something that took a rather long time. That set up served us well, despite a few issues - ongoing or otherwise - for the last 18 months.
However, as with many software projects, times do change. I began to be somewhat concerned with the project when its lead developer (and copyright holder) decided to pursue other interests, leaving the project in somewhat of a state of flux. Community support has wavered through this time, bug reports left to be ignored, and development slowed to a near stand-still. Hoping to not give up Cherokee just yet, I even contributed a number of significant fixes (well, I’d define them such) for things like correcting HTTP method handling, correcting timeout handling, fixing tests, and more. I suppose needing to do this probably should have been a concern in itself, but I enjoy open source development — and someone needs to fix those issues. Other things like lacking an official release process, complete lack of interest from package maintainers in Debian/Ubuntu/EPEL, and more, fuelled my concerns further.
More recently, there’s been a bit of renewed interest from several developers, but with the number of issues present in the 100s, a disjointed experience from website to release process to issue tracking, and little uptake across the web these days (see http://w3techs.com/technologies/overview/web_server/all), I foresee a lot of work needed by a lot of people. I’m sure some out there that are still using Cherokee may disagree with what I write - yes, the project still has some interest about it - but I believe my concerns are valid. I’m also not trying to disparage the Cherokee project itself either - it’s had a tough life and I can appreciate that - but after my experiences, I don’t feel that it is a good fit any longer, and I can’t devote any more of my time to it.
Technical issues aside for now, looking at a page like http://nginx.com/company.html, it makes me a little more comfortable in the fact those companies are surely contributing back to the development. Even if not from the perspective of money, but by simply running Nginx in production, they’ll be identifying basic issues well, well before I even know about them. According to those statistics above, almost 16% of the web uses Nginx. That’s a bigger user base, support base, better security, money and given overall usage - this is possibly most important - a guarantee of needing to keep up with new web technologies. I just don’t see Cherokee being able to keep up in any of these aspects, so hence the switch to Nginx was a logical one.
In addition, from my dev-ops perspective, Cherokee and Nginx are very similar, so it’s not like I’m stepping back to Apache. They work - to the best of my knowledge - on a very similar event-driven set of principles and during my configuration conversion, found they’re configured almost identically in some cases. Nginx does what I’ve come to like of Cherokee (speed, certain modules, functionality, and so forth), but with benefit of a wider community behind it. It doesn’t have a shiny GUI, but the simplicity of its configuration means it doesn’t need one in my opinion. Separately, Nginx isn’t nearly as long lived as Apache, but I’m sure it won’t have trouble sticking around for that long.
In short, in running a production-grade system, you want to be focused on actually being able to deliver a service, and not having to watch over your shoulder every moment out of fear that your service will fall over. An issue like this: https://github.com/cherokee/webserver/issues/952 is not something I want to have to think about again, if I can at all help it.
So farewell Cherokee, and hello Nginx.
Note: My next post will be the technical side of the configuration changeover.