On Tue, 11 Nov 2003, Emilio Casbas wrote: > 2003/11/11 10:11:18| WARNING: Forwarding loop detected for: > GET / HTTP/1.0 > Host: host_ip > User-Agent: check_http/1.24 (nagios-plugins 1.3.0) > Via: 1.0 www.mysite.com:80 (squid/2.5.STABLE4) > X-Forwarded-For: x.x.x.x > Cache-Control: max-age=259200 > Connection: keep-alive
This indicates unique_hostname was not set correctly, and is entirely different from the entry below. > 2003/11/11 10:22:14| WARNING: Forwarding loop detected for: > GET / HTTP/1.0 > Accept: image/gif, image/x-xbitmap, image/jpeg, image/pjpeg, */* > Accept-Language: en-us > User-Agent: Mozilla/4.0 (compatible; EMonitor 6.1 Windows NT) > Host: x.x.x.x > Via: 1.0 server2:80 (squid/2.5.STABLE4), 1.0 server1:80 (squid/2.5.STABLE4) > X-Forwarded-For: x.x.x.x, x.x.x.x > Cache-Control: max-age=259200 > Connection: keep-alive This request is a forwarding loop. The request path was client -> server2 -> server1 -> [the server giving this error] Such loops can occatioanlly happen in sibling relations, but Squid should recover automatically. If you want to prevent it from happening ever then use cache_peer_access to deny the use of the sibling if the request was received from a sibling. Alternatively you can use always_direct to do the same (but this approach is not compatible with Squid-3. I recommend using the cache_peer_access approach) Regards Henrik
