Re: [nsp] 12.1.2 and WCCP

From: Lincoln Dale (
Date: Wed Jul 05 2000 - 21:44:38 EDT

[TAC engineer also cc'd on email discussion from cisco-nsp list]

At 21:28 05/07/00, Roy wrote:
>Nothing that simple.
>It has hit at least one other site and I think its related to CSCdk10038 since
>I am using 802.1Q and the "no ip route-cache" does change the behavior. I
>have an open problem with TAC.


CSCdk10038 has long since been fixed (as-in June 1998).

looking at the case-notes for TAC case you've opened, i see that you're
running WCCPv1 with squid.

the actual "problem description" doesn't say much other than " Symptons are
slow web pages or errors that say web
  page does not exist."

since you're running both (a) wccp, and (b) 802.1Q with (c) squid, it may
well be that your squid host is NOT handling PMTU correctly.
can you confirm what your tcp MSS gets negotiated to on your
browser-to-cache tcp connections?

i wouldn't be surprised if that is the root of your problems. (the MSS
_has_ to be negotiated less than the typical default of 1460 due to the
fact that you have both 802.1q and WCCP and GRE headers to compete with the
payload size too).
cisco's own caching products do this automatically. i know that squid
doesn't, and the host running squid (linux?/freebsd) probably is unaware of
such issues.



   Lincoln Dale           Content Services Business Unit          cisco Systems, Inc.       |         |
                                                    ||        ||
   +1 (408) 525-1274      bldg G, 170 West Tasman  ||||      ||||
   +61 (3) 9659-4294 <<   San Jose CA 95134    ..:||||||:..:||||||:.. 

This archive was generated by hypermail 2b29 : Sun Aug 04 2002 - 04:12:14 EDT