> I fear there's still a significant chance for confusion. In particular,
> having a namespace that appears cosmetically similar to hostnames, and in
> simple cases acts like a hostname, but is not actually a hostname and will
> not be affected if the hostname->address mapping changes seems perilous.
> Stepping back a little bit, have we really created something that's
> noticeably better than logical interfaces? For all their problems,
> logical interfaces do make it clear what interface is associated with a
> given address slot, and also do not clash with hostnames.
>
Why? I can just as easily create a /etc/nodename that has "net0:1"
and boot my machine to respond with
# hostname
net0:1
which then lets me do (after suitably editing /etc/hosts, and plumbing net0)
# ifconfig net0:1 plumb net0:1
# ifconfig net0:1
net0:1: flags=1000842<BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 3
inet 1.2.3.4 netmask ff000000 broadcast 1.255.255.255
At least the "label" namespace doesn't pretend to be an "interface",
like logical interfaces do.
the alternative is
- not to support hostname mapping at all- several people felt this would
be a big handicap (see PSARC inception notes)
- to always do the hostname lookup- this is also perilous in that the
returned value for a delete-addr or set-addrprop could be different
than the one used at create-addr. Moreover, the "label" is used
to look up DNS for static addresses, but means something else for
dhcp/ipv6-*conf- that asymmetry is also confusing
The advantage in having a "label" for all static addresses is that
we can now collapse all the common commands like the delete-* commands
to just one command that takes the label.
--Sowmini
