On Fri, Apr 04, 2014 at 09:28:09AM +0900, YAMAMOTO Takashi wrote: > > We have had inaccuracies in OpenFlow flow stats for similar reasons, for > > years, and it doesn't seem to bother people. I would like perfectly > > accurate stats, but no one yet has come up with an inexpensive way to do > > that. Given that, and because the existing stats are basically broken, > > I am inclined to accept Simon's approach. What do you think? > > i'm not necessarily against his approach. just a question. > i agree that it's an improvement over the current situation.
OK. I think the answer to your question is "yes", for what it's worth. It sounds like we're all on the same page. Now I need to go back and properly review the patch. (I'd also like to get better per-table stats.) _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
