Re: persistent forwarding loop

From: Henrik Nordstrom <[email protected]>
Date: Thu, 13 Jan 2000 01:03:35 +0100

I see three alternatives:

a) The parent hasn't generated it's digest yet

b) There is a naming error, either in the cache_peer line on the
child/sibling, or the parents visible_hostname.

c) The parents primary http port isn't 3128. The primary http port is
the first http_port entry in squid.conf.

From looking at the request, (b) seems most likely. It looks like you
have configured the child/sibling cache using the IP address of the
parent in the cache_peer line.

Configuration options:

a) Configure the sibling/child cache using the parents visible_hostname
as the peer name in cache_peer.

b) Add the IP address as an known alias in the parents squid.conf
(hostname_aliases).

c) It may work if you add a always_direct entry forcing the
sibling/child cache to go direct to the parent as a web server when
fetching internal objects from it.

/Henrik

Matija Grabnar wrote:

> Anyway, here is the same message for digest.... ( I didn't notice that I was
> getting two different messages until you pointed it out.) And before you ask,
> the parent IS compiled with --enable-digest
>
> GET /squid-internal-periodic/store_digest HTTP/1.0
> Accept: application/cache-digest
> Accept: text/html
> Host: 193.2.1.88:3128
> Via: 0.0 rzenik.arnes.si:3128 (Squid/2.3.STABLE1), 1.0 proxy.arnes.si:3128
> (Squi
> d/2.2.STABLE5)
> X-Forwarded-For: unknown, 193.2.1.232
> Cache-Control: max-age=259200
> Connection: keep-alive
Received on Wed Jan 12 2000 - 18:19:20 MST

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:50:21 MST