[6bone] [NOTIFY] XS26 service/peering outage

[Pim van Pelt]

| Lets see - They want free v6 service (we're talking about 6bone
| here)...  They'll have to deal with the fact that routing policy dictates
| that each city in the ASN has a block of addresses from the aggregate from
| which space is assigned.
| 
| If they don't like it, they can get their own _portable_ address space.
| 
| Sorry.  I don't, nor will I ever, overcomplicate network topology when the
| only "good" thing it yields is easy portability for non-paying
| sunscribers.
Here's an important difference between you (and me) on the one hand, and
XS26 people on the other hand. They have set out to have a /48 per
person, and let them migrate these prefixes between cities, indeed. This
way, nobody has to renumber.

Mind you, a tunnel up/down event (gif or sit), did not trigger SPF
recalculations when I tried it back in 1999. This might have changed in
the meanwhile.

Coming back to the original post, I too have seen Zebra forget to write
routing updates to the kernel, so that the bgpd's view on the FIB was 
different to the kernel's view on the FIB. That sucked, and only an
entire restart of bgpd solved this. I consider it a bug in zebra, but
a rare one which only surfaced like, once in the last two years.

Anyway, good luck with your new routing suite, perhaps you can
opensource it so other people can learn from it!

groet,
Pim
--
---------- - -    - - -+- - -    - - ----------
Pim van Pelt                 Email: pim@ipng.nl
http://www.ipng.nl/             IPv6 Deployment
-----------------------------------------------