On Mon, 1 Jun 2015, Paul Jakma wrote:
How, if you are required to be able to support *ONE* Zserv instance for
*ALL* VRFs (as we would be if we take this patch train as it stands),
would you implement that using a zebra process per VRF in a non-horrific
way?
I.e., in order to assauge my review comments, either:
Explain how the above will work to me.
OR
Make the single-daemon approach use a separate ZServ instance per VRF
(e.g. a different socket for each VRF on the zebra side,
/var/run/quagga/zserv.$VRF.api for VRF>0 say, or whatever). This API it
seems to me might be easier to implement by /either/ way.
OR
Be up-front about the fact that the approach of this patch set may cause
significant problems for anyone who later tries to implement
multiple-daemons - potentially a blocker. (Because it was presented with
claims that it doesn't stop the other way at all - though it seems to me
it does, at least until someone answers the quoted question).
regards,
--
Paul Jakma [email protected] @pjakma Key ID: 64A2FF6A
Fortune:
More are taken in by hope than by cunning.
-- Vauvenargues
_______________________________________________
Quagga-dev mailing list
[email protected]
https://lists.quagga.net/mailman/listinfo/quagga-dev
