> On Jun 10, 2016, at 7:03 PM, Ben Pfaff <b...@ovn.org> wrote: > > On Fri, Jun 10, 2016 at 04:09:31PM -0700, Justin Pettit wrote: >> >>> On Jun 9, 2016, at 3:17 PM, Ben Pfaff <b...@ovn.org> wrote: >>> >>> On Thu, Jun 09, 2016 at 12:12:39AM -0700, Justin Pettit wrote: >>> Most of our *-add and *-del commands fail, by default, if there is >>> already a duplicate or if there is nothing matching to delete, >>> respectively, unless given --may-exist or --if-exists, respectively, but >>> these new commands do not behave that way. (In particular adding a >>> duplicate static route seems like annoying behavior.) >> >> That's fair. While adding this functionality, I decided to normalize >> the IP addresses, too, which makes the output more consistent. >> >>> It would be nice for lr-route-list to warn about an invalid route >>> prefix. Just ignoring them is likely to confuse users. >> >> Good suggestion. >> >> This was a pretty invasive change, so I've appended an incremental for >> a sanity check. Let me know if you'd prefer me to just send it as a >> v2. > > Thanks. > > It's a little hard to tell just from looking at the incremental, but I > think that lr-route-add does not call "verify" on the static_routes > column. It should do that now, because its behavior depends on this > column's content. To be really and truly consistent and isolated, it > should also verify all the columns it looks at in the > Logical_Router_Static_Routes table (or we could make them immutable; > that is one reason that many "name" columns in Open_vSwitch database > tables are immutable).
You're right. It handled it for the addition case, but not the modify case. I've updated it to also verify the columns with the static router table. > Acked-by: Ben Pfaff <b...@ovn.org> Thanks! --Justin _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
