> Amos, > I can't regenerate the problem but, I also noticed that the cache.log > file is filled with: Okay, these are several other problems. Lets work on those even if the segfault can't be traced yet. > > 2009/01/12 10:01:40| WARNING: Forwarding loop detected for: > GET / HTTP/1.0 > Accept: image/gif, image/x-xbitmap, image/jpeg, image/pjpeg, > application/x-shockwave-flash, application/vnd.ms-excel, > application/vnd.ms-powerpoint, applicat > ion/msword, */* > Accept-Language: es-do,en-us;q=0.5 > Accept-Encoding: gzip, deflate > User-Agent: Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1) > Host: 192.168.2.1:3128 > Via: 1.1 trahersaWiFi (squid/3.0.STABLE11), 1.0 trahersaWiFi > (squid/3.0.STABLE11), 1.0 trahersaWiFi (squid/3.0.STABLE11), 1.0 This is a configuration error in the reverse-proxy setup. I suspect an old 2.5-style setup is being used. Please check your config matches the current how-to: http://wiki.squid-cache.org/SquidFaq/ReverseProxy NOTE: http_port options. The cache_peer entry has a direct IP address of the 'real' server and uses dstdomain for access routing. Redirectors and 'transparent' are NOT involved. This eliminates any DNS looping, firewall looping, or name re-writing looping that may crop up. > > Also I have this about the same time the squid server restarted it self: > > > 2009/01/12 10:01:40| clientProcessRequest: Invalid Request > 2009/01/12 10:01:41| ctx: enter level 0: 'http://192.168.2.1:3128/' > 2009/01/12 10:01:41| HttpMsg::parse: Too large reply header (20496 > > 20480 > 2009/01/12 19:40:02| ctx: exit level 0 > 2009/01/12 19:40:02| WARNING: All dnsserver processes are busy. > 2009/01/12 19:40:02| WARNING: 5 pending requests queued > 2009/01/12 19:40:02| Consider increasing the number of dnsserver > processes in your config file. This is due to the --disable-internal-dns option. Drop it from your build options and rebuild Squid. > i98.photobucket.com > 2009/01/12 19:40:25| dnsSubmit: queue overload, rejecting > lads.myspace.com > (squid)[0x8105fab] > [0xffffe400] > (squid)[0x8137417] > (squid)[0x80fa044] > (squid)[0x80f6279] > (squid)[0x80f62be] > (squid)[0x807f76f] > (squid)[0x8080b3a] > (squid)[0x8081d34] > (squid)[0x8051c76] > (squid)[0x8052064] > (squid)[0x8080d06] > (squid)[0x808183f] > (squid)[0x80765c8] > (squid)[0x80777df] > (squid)[0x8078cd7] > (squid)[0x8113f22] > (squid)[0x8113f4a] > (squid)[0x8113f7d] > (squid)[0x8098bd4] > (squid)[0x8098cc8] > (squid)[0x80d0dd8] > /lib/tls/i686/cmov/libc.so.6(__libc_start_main+0xc8)[0xb7d2cea8] > (squid)(__gxx_personality_v0+0x109)[0x804c081] > FATAL: Received Segment Violation...dying. Sigh. Not even any useful info in the stack trace. That might have been useful for finding the segfault even without a core. When re-building squid for the dns issue, see if you can get it to produce debug symbols in that trace. Amos > > > Sorry to send too much stuff but, I'm trying to get you the most > information I can provide. > > Thanks for your help. > No worries. Amos
