[c-nsp] slow convergence for full bgp table on a Cisco 7613/SUP720-3BXL

[Emanuel Popa]

Hi Ryan,

We are running 12.2(18)SXF4 IOS version which is not modular. Although
according to documentation smaller hold-queue won't drop routing
protocol updates, we tried with 1000 and 2000 and there was no
difference.

Regards,
Emanuel


On 3/13/07, Ryan Harden <hardenrm at uiuc.edu> wrote:
> I have two suggestions as we saw a similar issue when moving from
> Foundry BI-4000s to 6500s. Your results may vary.
>
> 1. Try issuing 'hold-queue 2000 in' on each peer interface. I believe
> the default is 70. We did this at the suggestion of TAC. I'm not sure
> how much it helped but it seemed to reduce the number of discards we
> were seeing on the peering interfaces and might have sped up convergence
> a bit.
>
> 2. Make sure you're not running the Modular version of IOS. Our 6506s
> came with -vz onboard and we figured we'd try it out. After a week of
> CPU >50% and 30-45 minutes to converge each BGP peer, we went back to
> -mz. CPU dropped to ~15% and we converge in 2-3 minutes now.
>
> /Ryan
>
> Emanuel Popa wrote:
> > hi,
> >
> > we are using as border router a cisco 7613 with sup720-3bxl,
> > WS-X6724-SFP+DFC linecards and WS-X6704-10GE+DFC linecards.
> >
> > we had 2 x GE links with our upstream and one BGP session with full
> > routing table on the loopback interfaces. everything was fine until we
> > needed an upgrade. so we upgraded from 2 x GE to 1 x 10GE. our biggest
> > issue is that our router needs around two hours to receive the full
> > bgp table over the new session. the old session takes a maximum of
> > five minutes to converge. the only difference is that one session ends
> > on a 6724-SFP linecard and the other session ends on a 6704-10GE
> > linecard. both of them have DFCs.
> >
> > this border router also receives the full routing table from four
> > other upstream providers. the RP CPU is pretty overloaded and we are
> > planning an upgrade for this machine but this is not the main issue
> > right now.
> >
> > we thought that the problem comes from the BGP configuraton on the
> > provider side, but we tried the same upgrade with another provider and
> > we have the same kind of problem.
> >
> > i was wondering if any of you had the same problem and if you can
> > share the workarounds.
> >
> > have a nice day,
> > emanuel
> > _______________________________________________
> > cisco-nsp mailing list  cisco-nsp at puck.nether.net
> > https://puck.nether.net/mailman/listinfo/cisco-nsp
> > archive at http://puck.nether.net/pipermail/cisco-nsp/
>
> --
> Ryan M. Harden, BS, KC9IHX              Office: 217-265-5192
> CITES - Network Engineering             Cell:   630-363-0365
> 2130 Digital Computer Lab               Fax:    217-244-7089
> 1304 W. Springfield                     email:  hardenrm at uiuc.edu
> Urbana, IL  61801
>
>          University of Illinois - Urbana/Champaign
>                      -  All your Base -
>
>
>