Could someone please clarify for me what is mongrel cluster vs. proxy vs. all that other stuff you use with mongrel? I'm basically confused why I can't simply run one mongrel server and be happy. Why do I need a cluster? Why can't I just run: mongrel_rails start -n 3 wouldn't that mean I have 3 mongrel servers running and they can handle more load? Or do I still need something in front of these 3 servers to distribute requests to them. Why do people use Apache infront of Mongrel? So what about Pound? Basically I'm very confused with all these options out there to run in production environment using Mongrel. Please help me and explain to me why all this is needed. Thank you in advance.
on 15.04.2008 19:33
on 15.04.2008 20:09
On Tue, Apr 15, 2008 at 11:33 AM, Greg Lazarev <lists@ruby-forum.com> wrote: > Why do I need a cluster? Why can't I just run: > > mongrel_rails start -n 3 That starts 3 mongrel instances, each listening on a different port. > wouldn't that mean I have 3 mongrel servers running and they can handle > more load? Or do I still need something in front of these 3 servers to > distribute requests to them. Why do people use Apache infront of > Mongrel? So what about Pound? People put something in front of the mongrels because they need a way to balance the load, which is all coming into one ip/port, across multiple processes, each of which are listening on a separate port. This mechanism can be a web server with proxying capabilities, like Apache or nginx, a proxy with web serving capabilities, like Swiftiply, or a straight proxy like Pound or HAProxy. Kirk Haines
on 15.04.2008 21:14
Then what is the point behind mongrel_cluster?
on 15.04.2008 21:46
On Tue, Apr 15, 2008 at 1:14 PM, Greg Lazarev <lists@ruby-forum.com>
wrote:
> Then what is the point behind mongrel_cluster?
It is an attempt to make it simpler to manage the mongrel processes.
Kirk Haines