On Wed, 11 Nov 2015, Daniel Walton wrote:
That is a much much bigger can of worms that would have to go through
IETF which can take a really long time.
We can do it for Quagga first, and it will work for Quagga with that
capability and not bring up sessions to peers without it IF the order has
been changed - which is exactly what is needed.
No up-front IETF work needed, though it could always be brought to IDR
later.
Donald and I were chatting about this....I think we are overcomplicating
this knob:
- We've had this behavior for 2 years and have not had any customers hit
any issues with it
You mean, you've not had customers report they've had this issue - see
below. ;)
- On the off chance that the stars aligned and a prefix oscillated back
and forth between A--B one could always set local-pref, origin, etc (any
attribute that is compared earlier than cluster-list length) to stabilize
that prefix. I think the odds of that happening would be slim though, just
because the IGP cost to the nexthop would likely cause A to always choose
the E path.
The thing is, customers might not even be able to diagnose they have this
problem. They might just have flaky routing every now and then and not
really know why.
It took a long time to diagnose MED and IGP oscillations too (AFAIK),
never mind get a grip on the root cause of it.
So we can avoid adding a new knob and we already have a way to work-around
this corner case scenario.
We can handle this completely transparently to the end-user too, and it's
not difficult at all. Making routing simpler for operators by making BGP
just Do The Right Thing surely is the best option? :)
I'll whip up a patch even...
regards,
--
Paul Jakma [email protected] @pjakma Key ID: 64A2FF6A
Fortune:
A political man can have as his aim the realization of freedom,
but he has no means to realize it other than through violence.
-- Jean Paul Sartre
_______________________________________________
Quagga-dev mailing list
[email protected]
https://lists.quagga.net/mailman/listinfo/quagga-dev
