# Tunneling in capistrano twice **URL:** \ **Category:** Rails\ **Created:** [June 7, 2006, 4:10pm UTC](https://www.ruby-forum.com/t/tunneling-in-capistrano-twice/62720 "2006-06-07T16:10:56Z")\ **Posts on this page:** 8\ **Page:** 1
**Author:** ![henster](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/h/a183cd/32.png) [@henster](https://www.ruby-forum.com/u/henster)\ **Post date:** [June 7, 2006, 4:10pm UTC](https://www.ruby-forum.com/t/tunneling-in-capistrano-twice/62720/1 "2006-06-07T16:10:56Z")
Hi there, The default recipe expects you to be able to access your web server directly from your local machine. On our web server, for security purposes, it is only possible to ssh to it through one particular ip (which is also a remote machine.) The capistrano source is so compact and tidy it looks very hackable. But rather than messing around I thought I’d ask first if anyone else has any ideas? Cheers Henry ---
**Author:** ![henster](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/h/a183cd/32.png) [@henster](https://www.ruby-forum.com/u/henster)\ **Post date:** [June 7, 2006, 6:50pm UTC](https://www.ruby-forum.com/t/tunneling-in-capistrano-twice/62720/2 "2006-06-07T18:50:59Z")
Henry T. wrote: > has any ideas? > Seems doable to me. I’d probably leave the deploy task alone and create before\_deploy and after\_deploy tasks that did whatever you need to do. If you can tunnel through the intermediate host, this is pretty easy to do. If you have more restrictive networking rules that block port forwarding, and you can run code on the intermediate server, you could just do ssh user@particularip ‘cd /path/to && rake deploy’ My memory is that this will send the sudo prompt back to you, but I could be wrong about that. In some production environments, this might be the easiest way to go. – Ray ---
**Author:** ![henster](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/h/a183cd/32.png) [@henster](https://www.ruby-forum.com/u/henster)\ **Post date:** [June 7, 2006, 7:13pm UTC](https://www.ruby-forum.com/t/tunneling-in-capistrano-twice/62720/3 "2006-06-07T19:13:12Z")
I remeber having had the same problem, but I don’t remember the solution. But why don’t you proxy to mongrel and serve the static files with apache ? That is much easier to setup and works very well for me. ---
**Author:** ![henster](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/h/a183cd/32.png) [@henster](https://www.ruby-forum.com/u/henster)\ **Post date:** [June 7, 2006, 6:57pm UTC](https://www.ruby-forum.com/t/tunneling-in-capistrano-twice/62720/4 "2006-06-07T18:57:17Z")
Greetings, I’ve set up apache as a proxy for my lightTPD server as describe in JDD essay: [http://duncandavidson.com/essay/2006/01/railsonapache](http://duncandavidson.com/essay/2006/01/railsonapache) Everything works great except if I try to access my site like this: [http://softdist6.berkeley.edu](http://softdist6.berkeley.edu) Then my webrowser tries to download the dispatch.fcgi from the webdir of my app. I have no idea why this is happening. If I do softdist6.berkeley.edu/all or any sub directory of softdist6.berkeley.edu then everything works fine. My lighthttpd.conf and vhost.conf files are below. Anyone have any ideas? [lightTPD] server.document-root = “/srv/www/ror-apps/software\_central/public” server.modules = (“mod\_rewrite”, “mod\_accesslog”, “mod\_fastcgi”) server.port = 81 server.username = “wwwrun” server.groupname = “www” server.pid-file = CWD + “/tmp/lighttpd.pid” accesslog.filename = CWD + “/log/access\_log” server.errorlog = CWD + “/log/error\_log” server.indexfiles = ( “index.html” ) mimetype.assign = ( “.css” =\> “text/css”, “.gif” =\> “image/gif”, “.html” =\> “text/html”, “.jpeg” =\> “image/jpeg”, “.jpg” =\> “image/jpeg”, “.js” =\> “text/javascript”, “.pdf” =\> “application/pdf”, “.png” =\> “image/png”, “.txt” =\> “text/plain”, ) var.appname = “/srv/www/ror-apps/software\_central” $HTTP[“host”] == “softdist6.berkeley.edu” { server.document-root = CWD + “/public” url.rewrite = ( “^/$” =\> “index.html”, “^([^.]+)$” =\> “$1.html” ) server.error-handler-404 = “/dispatch.fcgi” fastcgi.server = ( “.fcgi” =\> ( “localhost” =\> ( “min-procs” =\> 5, “max-procs” =\> 5, “socket” =\> CWD + “/tmp/sockets/fcgi.socket”, “bin-path” =\> CWD + “/public/dispatch.fcgi”, “bin-environment” =\> ( “RAILS\_ENV” =\> “development” ) ) ) ) } [/lightTPD] [apache] \ ServerName softdist6.berkeley.edu DocumentRoot /srv/www/htdocs/softdist6 ProxyPass /wsspurch ! ProxyPass /i [http://softdist6.berkeley.edu:81/](http://softdist6.berkeley.edu:81/) ProxyPassReverse / [http://softdist6.berkeley.edu:81](http://softdist6.berkeley.edu:81) ProxyPreserveHost on [/apache] ---
**Author:** ![henster](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/h/a183cd/32.png) [@henster](https://www.ruby-forum.com/u/henster)\ **Post date:** [June 7, 2006, 7:35pm UTC](https://www.ruby-forum.com/t/tunneling-in-capistrano-twice/62720/5 "2006-06-07T19:35:35Z")
Well, if I can’t fix this problem then I may consider doing that, but for the time being I’d like find a working solution for lightTPD. ---
**Author:** ![henster](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/h/a183cd/32.png) [@henster](https://www.ruby-forum.com/u/henster)\ **Post date:** [June 7, 2006, 9:32pm UTC](https://www.ruby-forum.com/t/tunneling-in-capistrano-twice/62720/6 "2006-06-07T21:32:32Z")
“Why would you prefer to use three things when two will do?” ??? Who is using 3 things? There are two options that were discussed: 1. Apache being the proxy for lightTPD 2. Apache being the proxy for mongrel At least that’s how I understood the suggestion. -Steven ---
**Author:** ![henster](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/h/a183cd/32.png) [@henster](https://www.ruby-forum.com/u/henster)\ **Post date:** [June 7, 2006, 9:19pm UTC](https://www.ruby-forum.com/t/tunneling-in-capistrano-twice/62720/7 "2006-06-07T21:19:35Z")
Why would you prefer to use three things when two will do? – – Tom M. ---
**Author:** ![henster](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/h/a183cd/32.png) [@henster](https://www.ruby-forum.com/u/henster)\ **Post date:** [June 7, 2006, 10:38pm UTC](https://www.ruby-forum.com/t/tunneling-in-capistrano-twice/62720/8 "2006-06-07T22:38:17Z")
On Jun 7, 2006, at 11:31 AM, Steven H. wrote: > “Why would you prefer to use three things when two will do?” > > ??? > > Who is using 3 things? There are two options that were discussed: > > 1. Apache being the proxy for lightTPD Lighty then passes Rails requests to FastCGI, which makes it three. > 1. Apache being the proxy for mongrel Where mongrel handles the Rails requests, making it two. – – Tom M.