[nsp] dCEF repair pressure group (was: Re: performance impact on rate-limit)

From: Christian Panigl, ACOnet/VIX/UniVie (panigl@cc.univie.ac.at)
Date: Fri Aug 18 2000 - 05:37:39 EDT


    George et al.,
    
    I'm singing this blues since two years now, upgraded to RSP4 in the
    meantime, changed all VIPs to VIP2-50 with 128MB DRAM, changed from
    11.1CC to 12.0S but still, enabling dCEF cumulatively disconnects my
    customers from the Internet. In our environment, running
    (non-distributed) CEF on our 7507 is fine, just dCEF is the problem.
    I got the impression that specifically configurations with many
    sub-interfaces (ATM, FrameRelay, VLANs) do have problems with dCEF.
    
    Well, is there any interest in forming a pressure group to increase the
    motivation @ Ci$co to finally throw some manpower onto solving those more
    than annoying (d)CEF problems ?
    
    Or is everybody just waiting for her Junipers to be delivered ?
    
    Cheers
    CP

===========================================
>Date: Thu, 17 Aug 2000 14:40:04 -0400 (EDT)
>From: George Robbins <grr@shandakor.tharsis.com>
>
>CEF works fine it some situations, but is unstable in the more
>complex 7500 multi-PA/multi-VIP/multi-IP environment, typically
>it works fine for a while, then you find 50% of the packets to
>some destination are being dropped (or routed to the wrong place)
>and the damage is cumulative until you reboot or turn CEF off and
>back on again. Customers don't appreciate this, and there's not
>a big handle to make it easy to detect and/or debug, so the usual
>response is to turn off CEF, try to manage the CPU loading and
>then try the next release that comes out in the vain hope that just
>maybe Cisco fixed something.
>
>[...]
===========================================



This archive was generated by hypermail 2b29 : Sun Aug 04 2002 - 04:12:15 EDT