Re: how does this work exactly networking wise?

From: Matt Ashfield <[email protected]>
Date: Thu, 9 Mar 2000 09:49:15 -0400

Hi,

In my squid.conf I do have the following uncommented line:
pinger_program /usr/local/squid/bin/pinger

I have been doing some sniffing of our network, and for some of the sites
that are requested, there is definetely an ICMP echo packet sent to the
destination host. Followed by some TCP handshaking, then the HTTP Request.
The ICMP echo is of lenght 5E0, whis is 1504 bytes long. Looks like all
binary zeroes. This isen't a major problem, but I'm just confirming that it
is in fact a feature of squid. It appears to only send these ICMP packets to
hosts that are requested from the cachebox. So that rules out a possible
hack. (I hope!)

Any comments are greatly appreciated!

Matt Ashfield
mda@unb.ca

-----Original Message-----
From: Duane Wessels <wessels@ircache.net>
To: Matt Ashfield <mda@unb.ca>
Cc: squid-users@ircache.net <squid-users@ircache.net>
Date: Thursday, March 09, 2000 12:57 AM
Subject: Re: how does this work exactly networking wise?

>
>
>On Wed, 8 Mar 2000, Matt Ashfield wrote:
>
>> HI All,
>>
>> I am running squid on AIX. I got a call the other day from a person on my
>> campus complaning of receiving 1500byte pings from my webcache box. I did
>> some monitoring and noticed that it seems when a request hits my
webcache,
>> my webcache then does a ping of the server being requested. Is this
standard
>> practice? Or am I reading this wrong? If this is standard practice, why
does
>> the ping packet have to be so large?
>
>squid does send ICMP, only if you compile with --enable-icmp.
>
>It should not be sending 1500 byte ICMP packets.
>
>I suggest you run tcpdump to confirm that they are indeed 1500 bytes.
>If so, we have a problem and we can track it down.
>
>Duane W.
>
Received on Thu Mar 09 2000 - 06:55:18 MST

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:51:59 MST