Hello!

On Fri, Feb 22, 2013 at 04:16:03PM -0500, jims wrote:

> nginx 1.2.7 on CentOS 6.3. Reverse-proxy with two upstream servers. One
> server should be default, the other should only be used when called from a
> link on a specific web server.
>
> I have the reverse proxy working fine for the default upstream, but not
> quite right on the second, site-specific upstream. Ignoring that I am also
> rewriting for http to force https (yeah, I know, it's discouraged to go
> encrypted between the nginx proxy and its upstreams, but it's a corporate
> security thing) I am using a two server block configuration, one server name
> for the default and another for the site-specific server. The site-specific
> server will link to the default server URL but should rewrite to the
> site-specific upstream's proxy's URL.
>
> It seems to work - not sure if it gets stuck in the proxy or on the upstream
> for the site-specific setup -, but it was suggested that a single server,
> selecting between upstream servers, would be preferable. I would like to
> know how that would be configured - just a quick overview or link to a
> documentation page would help a lot.
>
> If I can throw together a quick test of a single-server-block reverse-proxy
> config with multiple upstreams which are selected based on calling server, I
> could test to see if it's my multi-server-block setup or a problem on the
> upstream.

Normal nginx aproach is to use distinct server blocks for servers
with different configuration.

--
Maxim Dounin
http://nginx.com/support.html

_______________________________________________
nginx mailing list
nginx@nginx.org
http://mailman.nginx.org/mailman/listinfo/nginx