[squid-users] Squid as LB: how to achieve 503 answer causes to try next peer when using cache-peers?

Hi,

We use squid as LB for HTTP Traffic (by defining cache-peers),
It works quite well when we have a connectivity failure between squid
and one of the application servers so then it will try next possible
cache-peer

Our cache-peer are defined with round-robin:

cache_peer <AS_IP> parent <AS_PORT> 0 no-query originserver name=asAccel1
round-robin
cache_peer <AS_IP2> parent <AS_PORT2> 0 no-query originserver name=asAccel2
round-robin

 However  we would like to handle following scenario:

If our applications is not deployed on application server (glassfish)
but the application server is started then we receive 503 Service
unavailable answer from the application server.
 On that case we would like our squid to try next possible node too
however it keeps using the cache-peer answering 503.

 Are we using some bad setting on squid.conf? So, how can we achieve
to switch to next possible node when we receive a 503 answer?

Thanks,
Received on Mon Feb 13 2012 - 10:54:52 MST