From tony at games-master.co.uk  Tue Oct 11 19:30:32 2005
From: tony at games-master.co.uk (Tony Spencer)
Date: Tue, 11 Oct 2005 20:30:32 +0100
Subject: Piranha and radius
Message-ID: <200510111929.j9BJTifo016941@linux.tonyspencer.co.uk>

I'm having problems with running piranha to load balance radius on 2
servers.

Piranha is installed and working as far as port 80, 21 and 25 are concerned.

However whenever I try to get it running for radius it doesn't work.

I'm guessing the problem is that the nanny process can't verify that the
service is up.

 

Oct 11 18:47:33 lvs2 nanny[17185]: READ to 192.168.0.4:1645 timed out

Oct 11 18:47:33 lvs2 nanny[17186]: READ to 192.168.0.5:1645 timed out

Oct 11 18:47:38 lvs2 nanny[17145]: READ to 192.168.0.4:1645 timed out

Oct 11 18:47:38 lvs2 nanny[17146]: READ to 192.168.0.5:1645 timed out

 

So I created a script that was run from the "send" line for the virtual
service:

 

#########

#!/usr/bin/perl

 

$host = $ARGV[0];

 

open (RAD,"radtest test test $host 1645 testing123 \| grep rad_recv |");

while (<RAD>){

if (/^rad_recv: (Access-Accept).*/){           

print "OK";

}

}

 

#########

 

And included these lines in the lvs.cf file for the radius virtual service:

 

     send = "/root/rad-test.pl %h"

     expect = "OK"

 

But still get the same errors from the nanny process.

 

The script works as I can run it from the command line and give the IP
address of the servers running radius and it outputs "OK";

Am I missing something or configured something wrong?

 

Thanks in advance

 

Tony

 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/piranha-list/attachments/20051011/b1ca75cb/attachment.htm>

From sebastien.bonnet at experian.fr  Wed Oct 12 08:29:41 2005
From: sebastien.bonnet at experian.fr (=?windows-1252?Q?S=E9bastien_BONNET?=)
Date: Wed, 12 Oct 2005 10:29:41 +0200
Subject: Piranha and radius
In-Reply-To: <200510111929.j9BJTifo016941@linux.tonyspencer.co.uk>
References: <200510111929.j9BJTifo016941@linux.tonyspencer.co.uk>
Message-ID: <434CC975.9090509@experian.fr>

Hi Tony,

> So I created a script that was run from the ?send? line for the virtual 
> service:

You have to use send_program instead of send. Your piranha might be 
buggy regarding this feature and you might need the attached patch.

-- 
S?bastien BONNET     --    Ing?nieur syst?me
Tel: 04.42.25.15.40      GSM: 06.64.44.58.98
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: piranha-0.7.0.diff
URL: <http://listman.redhat.com/archives/piranha-list/attachments/20051012/4ebdfd98/attachment.ksh>

From tony at games-master.co.uk  Wed Oct 12 16:15:16 2005
From: tony at games-master.co.uk (Tony Spencer)
Date: Wed, 12 Oct 2005 17:15:16 +0100
Subject: Running squid and squidguard over lvs
In-Reply-To: <434CC975.9090509@experian.fr>
Message-ID: <200510121614.j9CGELtJ024361@linux.tonyspencer.co.uk>

We wanted to start running squid over 2 servers and get the lvs to load
balance them both.

After implementing this it worked fine, until we decided to implement
squidGuard as well.
Although squid still serves page fine squidGuard just doesn't block
anything.

I know that squid and squidGuard is working because if we disconnect one of
the squid servers from the LVS network and plug it directly into an internet
facing switch squid and squidGuard works as it should.

The LVS is connected into a backend switch that is on a 192.168.0 subnet as
are the 2 squid boxes. 

Has anyone had this type of setup working?

Thanks
Tony 

-- 
No virus found in this outgoing message.
Checked by AVG Anti-Virus.
Version: 7.0.344 / Virus Database: 267.11.13/123 - Release Date: 06/10/2005