Comparison of Rails Deployment Stacks | Hacker News Reader

Comparison of Rails Deployment Stacks

webroar.in

31 pointsluckystrike9 comments

I clicked through to this expecting it to be about Capistrano vs. Chef vs. [...], but the post is really about nginx vs. Apache vs. [...].

A more appropriate title might be, "Comparison of Rails Production Environments"

It's not about nginx vs Apache. It's about webroar vs passenger vs thin vs mongrel, all of which are rails production-grade servers.

Yeah, I think the point of his comment was that "deployment" in the title is misleading.

When it comes down to it for rails deployment your going to maximize the number of workers you have based on the amount of available memory. I'd like to see a test where they set the number of workers equal to the amount that fit within 512MB's of ram or something similar. I'd say using passenger + ree you'd be able to double or triple the number of workers for the same amount of memory and blow the rest of the stacks out of the water.

You are assuming that the processes are memory bound. While this is a frequent situation for Rails apps deployed on VPS machines, they are just as frequently CPU bound when under load.

Forgive me if I'm just out of date here, but last time I checked, REE offered something like a 30% memory savings per forked worker. By my math (Workers_ree = Workers_mri / 0.7) that gives you something like a 1.4X increase in number of available workers.

Now, a 40% boost in running threads within the same memory is nothing to sneeze at. It's just a far cry from a 200% increase in the same metric.

Furthermore, for many deployments running on bare hardware (or even virtualized in a private datacenter), RAM isn't really the most precious resource. An modern app server can easily be packing 16-32GB of RAM, at which point CPU and I/O throughput can easily challenge memory for "most precious resource".

I've never heard ofr webroar before, but I'm not surprised that the people who wrote it can set it up better than they can set up passenger.

At least the bias is transparent, but I can't take benchmarks directly from the creator of any piece of software too seriously.

I'd love to see some numbers for nginx + passenger.

I'd be interested in that too, since people tend to overestimate the inefficiency of apache.

Reply on news.ycombinator.com