On Mon, 20 Jun 2011 11:32:27 -0500, Brent Norris wrote:
I am running squid-3.1.11-1 and it keeps locking up on me. When I
then tell squid to restart using the init.d scripts it floods the
logs
with messages like this:
1308587241.668 24741 10.76.16.15 TCP_MISS/502 74187 GET
http://10.76.16.15:8089/array.dll? - DIRECT/10.76.16.15 text/html
which the squid machines ip address is 10.76.16.15
I was experiencing this behavior in any release prior to 3.1 and I am
still using my config from those prior versions, but I have looked at
the default config that comes with the package and I didn't see
anything that stood out as the reason it would be doing this.
Can anyone give me a clue to what I need to look at? I would like to
stay up on these newer versions but I can't get around this error.
Brent
Malware causing forwarding loops. access.log entries only occur *after*
the completion of a request.
This vulnerability is a side effect of removing the HTTP protocol Via:
header. You can catch such requests early by erasing the "via" or
"x_forwarded_for" from your 3.1 config. The defaults are to use loop
protection.
If this is an interception proxy make sure you also have firewall
protection preventing visitors from directly connecting to the squid
listening NAT/"intercept" port.
Or, you can use the "deny to_localhost" security access control next to
your "deny !Safe_ports" one. You may need to add the Squid box public IP
to the list of prohibited localhost IPs.
Amos