Connection refused: proxy: HTTP: attempt to connect to failed [error] ap_proxy_connect_backend disabling worker

Question

Connection refused: proxy: HTTP: attempt to connect to failed [error] ap_proxy_connect_backend disabling worker

Hi all,

I'm investigating cause of a wedged wiki this morning. The MacMini (10.7.5) was working well for the last year but we're not sure when it gave up the ghost.

I've already applied the fixes detailed at Mac OS X v10.5, Mac OS X Server v10.5: Web, Mail, Wiki, Personal Web Sharing, Parental Controls services may not start after restoring a Time Machine backup but it has't helped.

The error we see when we visit the top level of the server is:

Service Temporarily Unavailable

The server is temporarily unable to service your request due to maintenance downtime or capacity problems. Please try again later.

Apache/2.2.22 (Unix) PHP/5.3.15 with Suhosin-Patch mod_ssl/2.2.22 OpenSSL/0.9.8r DAV/2 Server at sacmini.local Port 80

The message in the httpd error logs is:

Connection refused: proxy: HTTP: attempt to connect to failed [error] ap_proxy_connect_backend disabling worker

Let me know if you have any ideas. Thank you.

--Athonia

Mac mini, OS X Server

Posted on May 7, 2013 10:21 AM

Reply

Answer 1

May 7, 2013 2:19 PM in response to fusionstormAC

It would appear that one of the worker processes may have tipped over. If that's actually the case, then why, of course, is the next question. Check with Console.log or check the logs directly, and see if you can tease out a few more details from the tsunami of log data lurking there; some more details on the failed process(es).

Two very general tests follow, to see if there's something wonky with the server. Launch Terminal.app and issue the following diagnostic commands, and see if there are DNS errors, and also if the Apache configuration files have correct syntax:

sudo changeip -checkhostname

sudo apachectl configtest

Two general suggestions:

I would generally not recommend applying OS X Server 10.5 fixes to OS X Server 10.7. FWIW.

Reply

Answer 2

May 7, 2013 2:27 PM in response to MrHoffman

Thanks for the reply.

So, there were some Ruby related log file entries. Do these have any meaning to you? I'm checking to see what collabcored2 does exactly.

May 7 14:22:33 sacmini com.apple.collabcored2[22106]:	from /System/Library/Frameworks/Ruby.framework/Versions/1.8/usr/lib/ruby/1.8/rubygem s/custom_require.rb:31:in `require'
May 7 14:22:33 sacmini com.apple.collabcored2[22106]:	from /usr/share/collabd/coreclient/config/environment.rb:11
May 7 14:22:33 sacmini com.apple.collabcored2[22106]:	from /System/Library/Frameworks/Ruby.framework/Versions/1.8/usr/lib/ruby/1.8/rubygem s/custom_require.rb:31:in `gem_original_require'
May 7 14:22:33 sacmini com.apple.collabcored2[22106]:	from /System/Library/Frameworks/Ruby.framework/Versions/1.8/usr/lib/ruby/1.8/rubygem s/custom_require.rb:31:in `require'
May 7 14:22:33 sacmini com.apple.collabcored2[22106]:	from /usr/share/collabd/coreclient/config.ru:13
May 7 14:22:33 sacmini com.apple.collabcored2[22106]:	from /Library/Ruby/Gems/1.8/gems/rack-1.2.1/lib/rack/builder.rb:46:in `instance_eval'
May 7 14:22:33 sacmini com.apple.collabcored2[22106]:	from /Library/Ruby/Gems/1.8/gems/rack-1.2.1/lib/rack/builder.rb:46:in `initialize'
May 7 14:22:33 sacmini com.apple.collabcored2[22106]:	from /usr/share/collabd/coreclient/config.ru:1:in `new'
May 7 14:22:33 sacmini com.apple.collabcored2[22106]:	from /usr/share/collabd/coreclient/config.ru:1

May 7 14:22:33 sacmini com.apple.launchd[1] (com.apple.collabcored2[22106]): Exited with code: 1

May 7 14:22:33 sacmini com.apple.launchd[1] (com.apple.collabcored2): Throttling respawn: Will start in 9 seconds

Reply

Answer 3

May 7, 2013 3:41 PM in response to fusionstormAC

That's the wiki server.

No, I don't know what has happened to cause it to tip over.

If you've not already done so, try stopping and restarting it, or rebooting and allowing it to restart.

Reply

Answer 4

May 7, 2013 5:43 PM in response to fusionstormAC

My colleague here found what seems to be the problem. Below is what we're looking through in an attempt to recover the server. although these links are helpful we still don't have a solution. We'd still appreciate input from the community:

Wiki server gives 503 error after...: Apple Support Communities

Lion Server: Do not update RubyGems when running the Profile Manager or Wiki services

I'm attempted to upgrade the 'rails' component which upgraded the gems but this didn't fix anything. Is the only option do roll back Ruby?

--Athonia

Reply

Answer 5

May 8, 2013 5:05 AM in response to fusionstormAC

Apple suggests using bundler and not altering the installed Ruby configuration when the Profile Manager or the Wiki Server is in use. (TS4042)

Reply