[c-nsp] Reasons for "random" ISIS flapping?

[Mark Tinka]

On Wednesday, August 28, 2013 10:11:02 PM Peter Rathlev 
wrote:

> We actually started using minimal ISIS timers instead of
> BFD generally when BFD for SVIs became unavailable after
> SXF. We have since started using BFD again but haven't
> thought about raising the ISIS timers again. I'll
> definitely make sure to take a look at correcting this.

Traditionally, "aggressive" IGP timers in conjunction with 
BFD have been such that convergence happens as soon as BFD 
signals its client (the IGP, in this case) of an issue on 
the link.

If BFD is fast but the IGP is still waiting around to run 
its SPF process, the benefits of BFD become marginalized.

Then again, with (r)LFA, perhaps the case for BFD is slowly 
going away. It's something I'm seriously thinking about. 
(r)LFA has its constraints, but on face value, it seems 
redundant to have both (r)LFA and BFD enabled.

> We saw some amount of BFD false positives with int
> 100/mul 3 though, possible (probably even) because of
> the same CPU overload. Which is why it's int 200/mul 4
> now, which works well for us.

I've generally run BFD at 250ms/3s within the PoP, and 
250ms/5s between PoP's. It's a real trade-off between speed 
and stability.

Mark.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: This is a digitally signed message part.
URL: <https://puck.nether.net/pipermail/cisco-nsp/attachments/20130829/31c15226/attachment.sig>