Peter Memishian wrote:
> Since this is analogous to the dhcp "renew-lease" for static addresses,
> we can have a renew-* set of commands for the addreses to reverify
> the address.
Calling that "renew" seems confusing. It's also worth considering what
such a DAD mechanism would look like if you had the ability to bring
addresses down/up with ipadm. (I haven't looked at the latest design
document, so apologies if it covers this).
I don't think we should preserve the notion that an address can be down
or up, since AFAIK that was an accident in 1991 when all the SIOC*
ioctls were made to apply to ipifs and not ills. (And we've already
fixed some of those accidents as part of IPMP over they years.)
The notion that an *interface* can be operationally enabled/disabled
matches the ifAdminStatus in the MIB, but there is nothing similar for
an address.
For addresses we have the IpAddressStatusTC in the MIB, but that relates
to the operational status and not administrative status. See RFC 4293
for details. (I was positively surprised to see that it has the
inaccessible status which covers the case when the interface has
disappeared.)
Erik
_______________________________________________
networking-discuss mailing list
[email protected]
