============================================================ #opendaylight-openflowplugin: OpenFlow Plugin Meeting Jun 16 ============================================================ Meeting started by abhijitkumbhare at 15:05:26 UTC. The full logs are available at http://meetings.opendaylight.org/opendaylight-openflowplugin/2016/openflow_plugin_meeting_jun_16/opendaylight-openflowplugin-openflow_plugin_meeting_jun_16.2016-06-16-15.05.log.html . Meeting summary --------------- * Agenda (abhijitkumbhare, 15:05:56) * Li migration changing default (abhijitkumbhare, 15:06:47) * Flow removed notification (abhijitkumbhare, 15:07:20) * Code reviews (abhijitkumbhare, 15:07:54) * Flow removed notification (abhijitkumbhare, 15:09:15) * Manohar says some use cases depend on flow removed notification (abhijitkumbhare, 15:10:51) * Previously yang notifications were present in the He design (abhijitkumbhare, 15:11:30) * Can be worked around in Li design - but it depends currently on the stats collection being present (abhijitkumbhare, 15:12:21) * In general Manohar agrees with having data tree change - but for the short term need to have the yang notification on a config parameter (default no yang notification) (abhijitkumbhare, 15:14:30) * In general need we need to give more thought on this - on how to solve this (abhijitkumbhare, 15:15:28) * vishnoianil asks why yang notification for flow removed is a bad idea - currently we already have pkt in & some other yang notifications (abhijitkumbhare, 15:16:55) * vishnoianil says some apps using RPCs for flow programming; it would be odd for them to listen for data change notification (abhijitkumbhare, 15:26:04) * vishnoianil says that stats manager based approach can be performance intensive (abhijitkumbhare, 15:26:54) * vishnoianil says muthu, shuva and he were discussing we should add a verifier mode - which does not need stats mgr for operational data store update (abhijitkumbhare, 15:28:48) * jbacigal concerned about the performance of yang notification - and thinks it needs to be characterized (abhijitkumbhare, 15:30:49) * vishnoianil thinks even if we are not going with yang notification - to update to the data store we will need to process the flow removed notification to update the data store (abhijitkumbhare, 15:34:14) * vinayak & vishnoianil think that yang based notification would be higher performance (abhijitkumbhare, 15:36:59) * ACTION: Manohar to summarize why we need the flow removed notifications; and the cons and send the email to the list on June 17 (abhijitkumbhare, 15:44:11) * Li migration changing default (abhijitkumbhare, 15:44:38) * abhijitkumbhare to send the weather report (abhijitkumbhare, 15:48:00) * Code reviews (abhijitkumbhare, 15:48:38) * LINK: https://git.opendaylight.org/gerrit/39102 (hideyuki, 15:51:24) * LINK: https://git.opendaylight.org/gerrit/39524 (hideyuki, 15:51:30) * LINK: https://git.opendaylight.org/gerrit/40429 (hideyuki, 15:51:35) * some patches needing reviews as per hideyuki (abhijitkumbhare, 15:52:07) * LINK: https://git.opendaylight.org/gerrit/#/c/36452/ (hideyuki, 15:52:07) * LINK: https://git.opendaylight.org/gerrit/#/c/35359/ (hideyuki, 15:52:11) * LINK: https://git.opendaylight.org/gerrit/39102 (abhijitkumbhare, 15:52:20) * LINK: https://git.opendaylight.org/gerrit/39524 (abhijitkumbhare, 15:52:30) * LINK: https://git.opendaylight.org/gerrit/40429 (abhijitkumbhare, 15:52:42) * LINK: https://git.opendaylight.org/gerrit/#/c/36452/ (abhijitkumbhare, 15:52:58) * LINK: https://git.opendaylight.org/gerrit/#/c/35359/ (abhijitkumbhare, 15:53:10) * LINK: https://git.opendaylight.org/gerrit/#/c/40346/ (hideyuki, 15:54:13) * LINK: https://git.opendaylight.org/gerrit/#/c/40181/ (hideyuki, 15:54:25) Meeting ended at 16:00:20 UTC. People present (lines said) --------------------------- * abhijitkumbhare (32) * hideyuki (7) * odl_meetbot (4) * ManoharSL (2) Generated by `MeetBot`_ 0.1.4