Re: [OSL | CCIE_RS] BGP design question

On Sun, Mar 28, 2010 at 17:38, Narbik Kocharians <narbikk_at_gmail.com> wrote:
> Marko,
>
> When you tested this feature, did you test the latency, like whenB the
> primary comesB up after a failover and the back up letting go of the routes?
> OR when the B/up actually takes over?
>
> I tested this feature when it first came out in 12.3 T something code (I
> don't remember) and i had to play around with timers to speed up
> convergence, and even then, itB was kind of slow to converge between PRI to
> B/up and or from B/up to Prim failover, and having theB routers inject the
> routes with the correct next-hop.
> I will try to find that lab (Don't hold your breath, it was a long time
> ago), But when i tested it, the second BGP detected an adjacency change, it
> terminated the session, but this termination was NOT immediate (Once again,
> PRI---B/up and/or B/up to PRIM and advertising the correct next-hop to the
> other routers), this was between Scan intervals, and when i tried it few
> times to see the range, i noticed that once awhile it took 15 seconds to 30
> seconds.

Let us not mix and match things here going into fine-tuning
reconvergence times. When neighbor goes down, router will send updates
to its peers with the new best-path, if it has one. That was not the
question in this thread.

--
Marko Milivojevic - CCIE #18427
Senior Technical Instructor - IPexpert
YES! We include 400 hours of REAL rack
time with our Blended Learning Solution!
Mailto: markom_at_ipexpert.com
Telephone: +1.810.326.1444
Fax: +1.810.454.0130
Web: http://www.ipexpert.com/
Blogs and organic groups at http://www.ccie.net