On Fri, Feb 25, 2011 at 7:23 PM, Russ Erickson <li...@ruby-forum.com> wrote:
> Hello!

Hello Russ!

Sorry for the long delay in the reply.

> I'm just wondering if anyone has any recommendations for running
> mongrel/mongrel_service under Windows (Client requires Windows Server
> 2008, SQL Server 2008, IIS7, no budget for a *nix box of any kind).  I
> have a working setup, but wondering if anyone has any further insight.
> Current working setup:
> Ruby 1.87 (via rubyinstaller.org)
> Rail 2.3.11
> mongrel 1.1.15 (mingw32)
> mongrel_service 0.4.0

With the requirement of using Rails 2.3.11 or even going to Rails 3, I
will recommend avoid mongrel in this case.

There are some reports of multiple cookies issues that haven't been
addressed due the inability to reproduce them or provide concrete test

I've suggested someone port mongrel_service to provide the same
support for Thin, which could be pretty trivial, but noone stepped up.

With that you should be able to setup multiple services in different
ports and then make the service dependent on IIS or TCPIP service so
the automatic startup do not affect.

Another point, completely unrelated to mongrel and mongrel_service is
ensure your server have Ruby executable in the DEP exclude list.

More details at this entry of RubyInstaller troubleshooting:


> I'll be looking at using the ARR features of IIS7 to do some load
> balancing between a few instances of mongrel running on different ports
> on the same box.

Sounds good, I've done the same approach in the past.

> I have read a lot of posts from people discussing troubles in getting
> Ruby 1.92 for windows working with mongrel 1.1.20pre and
> mongrel_service.  Does anyone have a working setup of this and is it
> worth fighting for?

1.9.2 has some startup performance issues on Windows. starting a Rails
application can take way longer than 1.8.7, but then request
performance on production mode is good and pretty much up-par Linux in
this area.

But, mongrel 1.2.0.pre is another story. I've mentioned the cookie
issue above, which could be problematic.

Again, sorry for the long delay in my reply and hope these answers help you out.
Luis Lavena
Perfection in design is achieved not when there is nothing more to add,
but rather when there is nothing more to take away.
Antoine de Saint-Exupéry
Mongrel-users mailing list

Reply via email to