RE: [squid-users] segfault after signal 6

> 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
Received on Wed Jan 14 2009 - 21:46:54 MST