[c-nsp] 7513 - RSP4 - 122-25.S15 - MLPPP / Dcef Weirdness

Pete Templin petelists at templin.org
Tue Nov 11 10:33:15 EST 2008


Gregory Boehnlein wrote:
>> Isn't 12.2(25)S really really not recommended on 7500?  I seem to
>> remember several exchanges where this was mentioned by cisco people here.
> 
> I'm going to look through the list archives and see if I can find those
> references. Everything that I've seen revolves around earlier iterations of
> the code, not the S15 release that has been out for a year.
> 
> I'm happy to consider upgrading to a different IOS version.. just looking
> for recommendations on what I should be looking at for a 7515 w/ Dual RSP
> 4+, 5 VIP cards and the need for LLQ, OSPF, BGP, VLANs, MLPPP etc..

I've been very happy with 12.0(27)S5 for MLPPP, LLQ, OSPF, BGP, MPLS. 
VLANs could be an issue - we had problems with subinterfaces not being 
fully CEF-switched in earlier 12.0(27)S releases and abandoned that 
configuration.  SSO is quite good.  It'd be 100% stable if it weren't 
for VIP2-50s having memory issues and bombing out occasionally, but 
that's not a code issue.  Lucky guess, the first two routers I checked 
have uptimes of 2y13w.

I've been somewhat happy with 12.0(32)S[7,8,10] for "simple" core 
routing.  MPLS Traffic Engineering is garbage, at least when talking to 
GSRs, and we've now officially abandoned MPLS TE on 7507s entirely.

That said, I like Jon Lewis' suggestion to switch to enough 7206s to 
carry the PAs you're using.  Single forwarding engine on a clean, very 
well baked architecture means simple and reliable.  We're moving to 
7206s for CT3 aggregation, GSRs for DS3 and OCx, and 
6500/7600/Sup720-3BXL for Ethernet.

pt



More information about the cisco-nsp mailing list