Agree Luis on all the points. On Friday, February 19, 2016, Luis Gomez <[email protected]> wrote:
> There are at least 3 areas Li plugin performs better than He plugin: > > - Flow bulk operations > - Statistics collection > - OF messages ordering > > For these reasons I think it is good to move to Li plugin but not before > we fix some of the stability and cluster issues it has. > > BR/Luis > > On Feb 19, 2016, at 5:16 PM, Abhijit Kumbhare <[email protected] > <javascript:_e(%7B%7D,'cvml','[email protected]');>> wrote: > > Ashutosh, > > About your question: > > >> Is it possible to validate Li-plugin (performs equal-to-or-better than > He-plugin) before asking the projects to migrate? > > There were comparisons sent out earlier showing the performance was better > or equal than the Helium design in most areas. > > Thanks, > Abhijit > > > On Fri, Feb 19, 2016 at 3:51 AM, Ashutosh Bisht < > [email protected] > <javascript:_e(%7B%7D,'cvml','[email protected]');>> wrote: > >> Hi Abhijit >> >> We had tested Li-plugin earlier.. We also ran into stability issues >> >> *Bug 4925* <https://bugs.opendaylight.org/show_bug.cgi?id=4925>* -* >> [Clustering]: >> Openflow connections unstable with Lithium plugin. >> >> >> >> We limited our testing on Li-plugin since it was decided to use He-plugin >> for Be release. >> >> Is it possible to validate Li-plugin (performs equal-to-or-better than >> He-plugin) before asking the projects to migrate? >> >> >> >> With regards >> >> Ashutosh >> >> >> >> *From:* [email protected] >> <javascript:_e(%7B%7D,'cvml','[email protected]');> >> [mailto:[email protected] >> <javascript:_e(%7B%7D,'cvml','[email protected]');>] >> *On Behalf Of *Luis Gomez >> *Sent:* Friday, February 19, 2016 1:15 PM >> *To:* Abhijit Kumbhare >> *Cc:* [email protected] >> <javascript:_e(%7B%7D,'cvml','[email protected]');> >> *Subject:* Re: [openflowplugin-dev] OFP-Li and OFP-He in Beryllium SRs >> >> >> >> Personally I think it is less risky changing master and if that works, >> cherry pick change to stable/beryllium. >> >> >> >> BR/Luis >> >> >> >> >> >> On Feb 18, 2016, at 6:48 PM, Abhijit Kumbhare <[email protected] >> <javascript:_e(%7B%7D,'cvml','[email protected]');>> wrote: >> >> >> >> Yes - agree. But the work to add the support for the Lithium design can >> start from the projects side in parallel. We should also explore creative >> ways to see how we can make this easier on the projects: >> >> >> >> Since Beryllium has just been released & now the development of most >> projects will shift to the master branch may be we should change the >> default features to be Li version in stable/beryllium and ask projects to >> see what breaks? Does that make sense? >> >> >> >> On Thu, Feb 18, 2016 at 2:00 PM, Luis Gomez <[email protected] >> <javascript:_e(%7B%7D,'cvml','[email protected]');>> wrote: >> >> Hi ofp devs, >> >> >> >> I agree with the plan to move projects to Li design, however I think >> before asking projects to effectively move we have to address few issues >> here: >> >> >> >> 1) OF1.0 issue: >> >> >> >> >> https://jenkins.opendaylight.org/releng/view/openflowplugin/job/openflowplugin-csit-1node-flow-services-lithium-redesign-only-beryllium/ >> >> https://bugs.opendaylight.org/show_bug.cgi?id=5328 >> >> >> >> 2) Cluster issues: >> >> >> >> >> https://jenkins.opendaylight.org/releng/view/openflowplugin/job/openflowplugin-csit-3node-clustering-only-beryllium/ >> >> https://bugs.opendaylight.org/show_bug.cgi?id=5388 >> >> >> >> 3) Stability issues: >> >> >> >> >> https://jenkins.opendaylight.org/releng/view/openflowplugin/job/openflowplugin-csit-1node-periodic-longevity-lithium-redesign-only-beryllium/ >> >> https://bugs.opendaylight.org/show_bug.cgi?id=5271 >> >> >> >> BR/Luis >> >> >> >> >> >> On Feb 18, 2016, at 12:42 PM, Abhijit Kumbhare <[email protected] >> <javascript:_e(%7B%7D,'cvml','[email protected]');>> wrote: >> >> >> >> Hideyuki, >> >> >> >> Thanks for bringing this topic up! I was planning to initiate a >> conversation on this later today - since the Beryllium release was >> officially approved today. First of all - I think we should dedicate a TWS >> session for this topic regarding the following discussion and feedback: >> >> >> >> 1. OpenFlow plugin project voted to do the following: >> >> OF Plugin recommends dependent ODL projects to all move to Li plugin >> design in the Be SR 1 timeframe ( >> https://lists.opendaylight.org/pipermail/openflowplugin-dev/2016-February/004607.html >> ) >> >> 2. What do projects see as issues & impediments to move to the Lithium >> design in the SRs (target SR 1) and how the OpenFlow plugin project can >> address them? Example of the issues - your request regarding the stats >> RPCs. >> >> 3. Develop a timeframe for completing this migration based on the >> issues/impediments. >> >> I would like to note that the key requirement from OpenFlow plugin >> perspective is that the Boron development should not be done on two >> different designs. The timeframe should be based on this requirement. Hence >> OFP thinks it will be important to do this change in the SR1 (or SR2 if >> cannot be managed). This way OFP developers can concentrate on the Lithium >> design during the Boron release timeframe (other than occasional high >> priority bug fixes on Helium design in the subsequent SRs). However, we >> (OFP & dependent projects) will need to work together for this timeframe. >> >> >> >> I will try to get this topic on the TWS call on Monday (assuming we can >> get the following folks from OFP on Monday (alphabetical order) - Anil >> Vishnoi, Jozef Bazigal, Kamal Rameshan, Michal Rehak & hopefully >> Muthu/Shuva & Luis. >> >> >> >> I think the above answers all of your questions except one: >> >> Do you plan to continuously support the OFP-He of Beryllium until >> OpenDaylight project stops to support Beryllium? >> >> The answer to that is - yes. OpenFlow Plugin has to support OFP-He in >> Beryllium - as per ODL release / deprecation requirements. >> >> >> >> Thanks, >> >> Abhijit >> >> >> >> On Thu, Feb 18, 2016 at 11:43 AM, Tai, Hideyuki <[email protected] >> <javascript:_e(%7B%7D,'cvml','[email protected]');>> wrote: >> >> Hi OpenFlow Plugin project, >> >> Could you share your plan about the migration to the OFP-Li in Beryllium >> SRs? >> Should all projects migrate to the OFP-Li by Beryllium SR1? >> Do you plan to continuously support the OFP-He of Beryllium until >> OpenDaylight project stops to support Beryllium? >> (In my understanding, OpenDaylight project officially stops to support >> Beryllium (4th) after it releases Carbon (6th).) >> >> If we need to migrate to the OFP-Li by Beryllium SR1, I would like to >> start to test VTN project's features with the OFP-Li as soon as possible. >> Though, to migrate the VTN features from OFP-He to OFP-Li, we need the >> "get-stats-.." RPC by default in the OFP-Li as I explained in the following >> mail. >> https://lists.opendaylight.org/pipermail/vtn-dev/2016-February/001278.html >> >> Regards, >> Hideyuki Tai >> >> >> >> >> _______________________________________________ >> openflowplugin-dev mailing list >> [email protected] >> <javascript:_e(%7B%7D,'cvml','[email protected]');> >> https://lists.opendaylight.org/mailman/listinfo/openflowplugin-dev >> >> >> >> _______________________________________________ >> openflowplugin-dev mailing list >> [email protected] >> <javascript:_e(%7B%7D,'cvml','[email protected]');> >> https://lists.opendaylight.org/mailman/listinfo/openflowplugin-dev >> >> >> >> >> >> >> > > >
_______________________________________________ openflowplugin-dev mailing list [email protected] https://lists.opendaylight.org/mailman/listinfo/openflowplugin-dev
