Hi Hank,
At 10:18 AM 20/12/2000 +0200, Hank Nussbacher wrote:
>Running on 12.0(11) with wccp v1 I get the following (the cache is Cacheflow):
>
>Dec 18 13:05:32: %WCCP-5-CACHEFOUND: Web Cache n.n.n.n acquired
>5w1d: WCCP-PKT: Sending I_See_You packet to n.n.n.n w/ rcvd_id 000023AB
>5w1d: WCCP-PKT: Sending I_See_You packet to n.n.n.n w/ rcvd_id 000023AC
>5w1d: WCCP-PKT: Sending I_See_You packet to n.n.n.n w/ rcvd_id 000023AD
>5w1d: WCCP-PKT: Sending I_See_You packet to n.n.n.n w/ rcvd_id 000023AE
>Dec 18 13:06:02: %WCCP-1-CACHELOST: Web Cache n.n.n.n lost
>Dec 18 13:06:12: %WCCP-5-CACHEFOUND: Web Cache n.n.n.n acquired
>Dec 18 13:06:42: %WCCP-1-CACHELOST: Web Cache n.n.n.n lost
>Dec 18 13:06:42: %WCCP-5-CACHEFOUND: Web Cache n.n.n.n acquired
><snip>
>
>Problem does not exist when running with wccp v2 (on 12.1(2)). Anyone
>know of this as a known wccp v1 bug?
can you provide a full debug trace (including "debug ip wccp events" and
"debug ip wccp packets") ?
the problem is either one of:
(a) bugs in the CacheFlow WCCP implementation.
enabling the above debugging will show whether this is the case or not.
(b) you have significant packet-loss between the cache and the router,
causing
sufficient WCCP messages to be dropped causing the protocol to fail.
note that this would have to be _significant_ packet-loss, 3 x HERE_I_AM
messages (at 10-second intervals) from the cache being lost in a row as
well as 3 REMOVAL_QUERY messages (at 1 second intervals) not being
answered.
cheers,
lincoln.
This archive was generated by hypermail 2b29 : Sun Aug 04 2002 - 04:12:24 EDT