On Tue, 14 Jun 2005, Adam Clark wrote:
Squid is setup to forward any request to the IWSS then IWSS makes squid
do the final Request, so squid - IWSS - squid. We require to be like
this so accounting and error Pages are handled correctly.
This all works very nicely but has a side effect of forwarding loop
errors in cache.log.
So the cache log becomes my largest log. I have a feeling that if I
want to remove this Side effect I will need to run another instance of
squid.
Correct.
Is there another way?
You could edit the source to not emit the warning if the request was seen
from the IWSS server and "our" name is only seen once in the header.
This is however a very odd setup. You will most likely be better off
running an additional Squid instance.
Squid 1 -> IWSS -> Squid 2
and only one of the two in caching mode. Where depends on your scanning
policy, but most people select to cache on the Squid closest to the
clients even if this risks having viruses in the cache and not scanned
again after a virus definitions update.
Access controls on the second Squid instance should normally be set to
limit access to coming from IWSS only.
Regards
Henrik
