On 6/9/2015 3:38 PM, Paul Jakma wrote:
On Tue, 9 Jun 2015, Jafar Al-Gharaibeh wrote:
The current "table" command seems to be more of a global state
controlling what is the "current" kernel table Quagga is
communicating with by default.
Yeah, that's what I thought. It's never really been terribly useful
though, AIUI.
I can see it useful in cases where the user doesn't want Quagga to
change/mess-up with the default table (for policy/security/etc. reasons)
or run multiple Quagga instances and feed their routing information into
different routing tables. That is without/before the VRF patch.
This still makes sense in a VRF-aware world and the same semantics
can still be honored. However, since this never worked well, killing
it off might not be a bad idea especially that VRF should be able to
provide a way to control a specific table.
Hmm, so could also be left. Maybe deprecate and hide the command for now.
Or make it forward compatible with VRFs somehow(did we decide to name
them logical tables?!). You'd think that VRF support should make it
easier to get even the old table command semantics to work, because it
adds the ability to manage multiple routing tables and provides the
correct plumbing to hookup with different kernel routing tables - which
seems to be the reason why the table command never worked before anyway.
Regards,
Jafar
_______________________________________________
Quagga-dev mailing list
[email protected]
https://lists.quagga.net/mailman/listinfo/quagga-dev
