Here is my pound.cfg file

Daemon 1
User "www"
Group "www"
LogLevel 3
Alive 30

ListenHTTP
        Address 192.168.0.2
        Port 80
End
Service
        HeadRequire "Host: .*www.mydomain.com/squirrelmail.*"
        BackEnd
                Address 127.0.0.1
                Port 8080
        End
        Session
                Type BASIC
                TTL 300
        End
End

Service
        HeadRequire "Host: .*mail.mydomain.com.*"
        BackEnd
                Address 127.0.0.1
                Port    9100
        End
        Session
                Type BASIC
                TTL 3000
        End
End

The first service is a php app that I run using fastcgi on lighttpd.  I
previously ran this as my webserver only.  With pound, it is noticable
slower, and now am getting Connection Time Outs in the broswer.  The second
is a ruby on rails app that I was also running with lighttpd with fastcgi,
but have changed the configuration over to pound -> mongrel clusters.  This
also is slower and am receiving connection time out errors.


When I do a fgrep pound /var/log/messages, I get the following output
Apr 17 13:16:08 titus pound: bad header from
216.253.246.194(JCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQiMKAwgA+A8BAPgPASOCCDkDggQRAENDQ0Mg8P1/U1ZXZoHsgACJ5ujtAAAA/zZoCRLWY+j3AAAAiUYI6KIAAAD/dgRoa9AryujiAAAAiUYM6D8AAAD/dgRo+pcCTOjNAAAAMdtoEAQAAFP/0InDVot2EInHuRAEAADzpF4xwFBQUFNQUP9WDItGCG
Apr 17 13:16:08 titus pound: bad header from
216.253.246.194(0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDI4IEIAMJAOsGkJCQkJCQA4IEEQBERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERER
Apr 17 13:29:34 titus pound: error read from 68.236.176.200: Operation timed
out
Apr 17 14:27:37 titus pound: bad header from
216.253.246.194(JCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQiMKAwgA+A8BAPgPASOCCDkDggQRAENDQ0Mg8P1/U1ZXZoHsgACJ5ujtAAAA/zZoCRLWY+j3AAAAiUYI6KIAAAD/dgRoa9AryujiAAAAiUYM6D8AAAD/dgRo+pcCTOjNAAAAMdtoEAQAAFP/0InDVot2EInHuRAEAADzpF4xwFBQUFNQUP9WDItGCG
Apr 17 14:27:37 titus pound: bad header from
216.253.246.194(0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDI4IEIAMJAOsGkJCQkJCQA4IEEQBERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERER
Apr 17 14:42:18 titus pound: error read from 72.232.207.42: Operation timed
out
Apr 17 20:10:10 titus pound: error copy chunk cont: Broken pipe
Apr 17 23:50:43 titus pound: error read from 72.232.207.42: Operation timed
out
Apr 18 03:17:31 titus pound: bad header from
216.253.246.194(JCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQkJCQiMKAwgA+A8BAPgPASOCCDkDggQRAENDQ0Mg8P1/U1ZXZoHsgACJ5ujtAAAA/zZoCRLWY+j3AAAAiUYI6KIAAAD/dgRoa9AryujiAAAAiUYM6D8AAAD/dgRo+pcCTOjNAAAAMdtoEAQAAFP/0InDVot2EInHuRAEAADzpF4xwFBQUFNQUP9WDItGCG
Apr 18 03:17:31 titus pound: bad header from
216.253.246.194(0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDQ0NDI4IEIAMJAOsGkJCQkJCQA4IEEQBERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERER
Apr 18 07:26:47 titus pound: error read from 72.232.207.42: Operation timed
out


When I'm hitting the box, I'm not getting any errors? What am I doing wrong?

Thanks,

MB

On Wed, 2007-04-18 at 10:00 -0500, Matt Bjornson wrote:[...]

That is meaningless and won't match anything - all requests will go to
the second service. You probably want:

        HeadRequire "Host: .*www.mydomain.com.*"
        URL "^/squirrelmail.*"
[...]

Fix this and we'll look at the results.[...]

On Apr 18, 2007, at 8:00 AM, Matt Bjornson wrote:
<snip>
[...]
I've been having similar issues, I thought increasing the TimeOut  
value for the BackEnds would help but I don't see much improvement.  
I've also increased the Client timeout with  the same result. Here's  
a rundown from today's logs (since logrotate ran):

This error has occurred 26 times, spaced out throughout the day so far:
Apr 18 10:15:57 lb001 pound: error copy server cont: Connection timed  
out

This error has occurred 21 times
Apr 18 06:28:53 lb001 pound: error copy chunk cont: Connection reset  
by peer

There are a few others as well but they appear to be happening far  
less frequently:
Apr 18 09:43:38 lb001 pound: error flush headers to ###.###.###.###:  
Connection reset by peer
Apr 18 10:03:45 lb001 pound: error copy client cont: Success
Apr 18 07:08:09 lb001 pound: error read from ###.###.###.###:  
Connection reset by peer

These seem to all be happening most often when a POST occurs.  Such  
as when users upload images/audio/video to our service.  From the  
first-hand accounts this happens very quickly, not after waiting a  
for content to upload, etc.

-Blake

Robert,

Thanks for the modification.  I can't say that's made any difference though.
I"m still getting the Connection Time out errors.  I should have included
this before.....

FreeBSD 6.1-RELEASE #0
Pound 2.2
Lighttpd 1.4.15
Mongrel 1.0.1
Mongrel cluster 0.2.1

Dell 2400 Dual CPU 1GB of RAM
Plenty of freespace.....

Here's the latest from /var/log/messages

Apr 18 13:16:19 titus pound: received signal 15 - exiting...
Apr 18 13:16:19 titus pound: received signal 15 - exiting...
Apr 18 13:16:25 titus pound: starting...



On 4/18/07, Robert Segall <roseg(at)apsis.ch> wrote:[...]

On Apr 18, 2007, at 1:31 PM, Matt Bjornson wrote:
[...]


If you're seeing the request hitting your mongrels, it's most likely  
not pound.

What is lighty doing in there btw? static content host? Are you sure  
your uploads aren't going through pound->light->mongrel (they  
shouldn't).


-mml

lightty is serving up the php app, squirrelmail.  I'd not think I'd be going
pound -> lightty -> mongrel..

I should have posted this...

Lightty running on 8080

Mongrel_cluster on 9000, 9001

I think I've pound configured to send the squirrelmail  to lightty, and the
mailr to 9000 (mongrel_cluster).  I'm not seeing any errors in
/var/log/lighttpd.error.log....  Though I did find an error in the mailr
code, so ultimately, it could be app errors.  Thanks,


On 4/19/07, McClain Looney <m(at)loonsoft.com> wrote:[...]