Typo on mongrel + apache + ubuntu

Ruby Apps Typo
1

I have typo running on mongrel proxied thru apache2. But I have two
problems:

  • If a page request comes in while mongrel/typo is down I get an error
    but
    when typo comes back up I see it on the port mongrel uses but
    apachecontinues to give the error. I have to restart apache to get
    things
    back in sync. Why?
  • What form should my sysv init.d scripts take? I created the one below
    to
    get around my apache restart problem but I see other talk about starting
    mongrel instead of typo on reboot. Which is right? Can someone point
    me to
    a good init.d script that can be used for typo?

#!/bin/bash

Starts the ruby mongrel server that runs the occamsmachete blog

case “$1” in
start) cd /home/pat/blog
typo start .
/etc/init.d/apache2 restart
;;
stop) cd /home/pat/blog
typo stop .
;;
restart) cd /home/pat/blog
typo restart .
/etc/init.d/apache2 restart
;;
sweep_cache) cd /home/pat/blog
typo sweep_cache .
;;
*) log_action_msg “Usage: /etc/init.d/typo
{start|stop|restart|sweep_cache}”
exit 2
;;
esac
exit 0

I guess this is a mongrel question but thought I¹d try first here in
case
another typo user has run into this.

2

Pat,

Unfortunately that is an Apache thing, what it does is determine that
the service it is connecting to as a proxy is down, and it remembers
that it is down.

You would see in the apache error_log (for example)

[Sat Feb 16 11:55:38 2008] [error] (111)Connection refused: proxy:
HTTP: attempt to connect to 127.0.0.1:4485 (xxx.xxxxx.xxx) fad
[Sat Feb 16 11:55:38 2008] [error] ap_proxy_connect_backend disabling
worker for (xxx.xxxxx.xxx)
[Sat Feb 16 11:55:38 2008] [error] proxy: HTTP: disabled connection
for (xxx.xxxxx.xxx)

Sometimes, a hard refresh after starting/stopping typo will resolve it
without restarting apache. However that is really a limitation of
apache and has nothing to do with mongrel or typo. :frowning:

As far as your init script, it looks fine… you might have a case
where when you stop apache it doesn’t kill all of it’s children and
the restart fails… not very likely since it’s a restart but I’ve
seen it happen.

BTW, I just did that, it disabled it, re-started typo and it started
working again without restarting apache… running Apache 2.2.8

Hope that helps.

Scott

3

On Feb 16, 2008, at 15:16, Scott L. wrote:

[Sat Feb 16 11:55:38 2008] [error] proxy: HTTP: disabled connection
for (xxx.xxxxx.xxx)

Hmm, I’ve had typo not start on me on more than one occasion and my
apache hasn’t done this. I wonder what’s different about my config.

I’m using:

     RewriteEngine On
     RewriteRule /(.*) http://localhost:8080/$1 [P,L]
     ProxyPassReverse / http://localhost:8080/

-Bill

4

I¹m using the virtual host def that came with the most recent typo shown
below. Do you have sysv init.d scripts to start typo?

<VirtualHost *>
ServerName blog.occamsmachete.com
ServerAlias www.blog.occamsmachete.com

Change this to your email address

ServerAdmin [email protected]

Change these to be valid paths for your host. The DocumentRoot path

isn’t very important because we don’t actually use it for anything.

For security’s sake, it’s best that it points to an empty directory,

but that’s not critical.

DocumentRoot /var/www/blog
ErrorLog /var/log/apache2/blog_error.log
CustomLog /var/log/apache2/blog_access.log combined

ServerSignature On

This is the important part–it sets up proxying.

ProxyRequests Off
<Proxy *>
Order deny,allow
Allow from all

ProxyPass / http://localhost:4155/
ProxyPassReverse / http://localhost:4155/
ProxyPreserveHost On

5

Oops, sorry, missed your reply.

On Feb 22, 2008, at 11:08, Pat Ferrel wrote:

Do you have sysv init.d scripts to start typo?

No, I have to admit to having some rc.local lines. :stuck_out_tongue:

-Bill