[j-nsp] /32 host routes on down interfaces

[Pavel Lunin]

Hi experts,

Can anyone actually explain what for those /32 routes are intended at all?
Even if the iface is up. Ain't they to attract traffic addressed to control
plane?

My way of thinking is that they are alive at link down time for the same
reason as they exist when the link is up.

I haven't been working with JUNOS as long as with ScreenOS, so don't know
its story so well. But I also was a bit surprised when first saw /32
interface routes in somewhat ScreenOS 5.x, though didn't got much confused
about this. AFAIR, It happened definitely just after NetScreen had became a
part of Juniper.

I've just checked if the same behavior is present in ScreenOS 6.x—yes, it
is. When a link goes done, corresponding iface host route stays up. So, you
know, I would not make a conclusion about JUNOS architecture or a track of
old code or any other ‘clumsy done’ conclusion so easily. Of course, one can
say, this was because Juniper reused JUNOS code or architecture ideas in
ScreenOS, but I don't much believe in such a hypothesis.

My version is this is done intensionally for something.

BTW, is it OK to make an unnumbered interface to not respond to ARP if the
interface, which it borrows an address from, goes down?

--
Regards,
Pavel


2010/4/22 Richard A Steenbergen <ras at e-gerbil.net>

> So I just noticed an interesting behavior which I think is a bad thing,
> but I want to see what other people think.
>
>