Hi Abhijit, We brought this discussion during last integration call, and we concluded that from user perspective as well as from consumer applications, it would have been better to create a switch in the config system rather than a new feature (-li plugin). Can we allocate some time next Monday call to talk about this? I already asked Anil about this possibility and he thinks it is difficult because of how the config system works but lets talk about this next call.
BR/Luis > On Dec 21, 2015, at 7:52 PM, Abhijit Kumbhare <[email protected]> wrote: > > As Luis and I just discussed - it would be good if we can have the following > possibility: > > 1. Have feature names odl-openflowplugin-A & odl-openflowplugin-B (the A & B > names to be named as per the characteristics of the designs - not > odl-openflowplugin* & odl-openflowplugin-li*). > 2. Have top level features odl-openflowplugin* - pointing to either > odl-openflowplugin-A or odl-openflowplugin-B. Hopefully we can do this on > some config setting > > I am not sure if the above is doable or not - but can help to migrate toward > a single design as we can set a default and move it to another. > > > On Mon, Dec 21, 2015 at 6:18 PM, Luis Gomez <[email protected] > <mailto:[email protected]>> wrote: > This time to the right openflowplugin list :) > > > On Dec 21, 2015, at 6:16 PM, Luis Gomez <[email protected] > > <mailto:[email protected]>> wrote: > > > > Hi all, > > > > Thanks to Vratko we have Beryllium distribution verify job back to business: > > > > https://git.opendaylight.org/gerrit/#/c/31720/ > > <https://git.opendaylight.org/gerrit/#/c/31720/> > > > > As you see in this self-explanatory patch we had to remove few features > > from the integration test. Note projects features repos are still in the > > distribution as they only impact the controller run-time (distribution > > build is fine): > > > > 1) Features of Centinel and Sxp > > > > Yang model errors are seen in karaf.log for those projects. For this we > > will just need the projects to address their yang issues before we can test > > the distribution with their features > > > > 2) Features of DIDM and NIC > > > > Those features are bringing in odl-openflowplugin-*-li, which lead to > > BindExceptions, which lead to red dot. This is a more tricky issue as today > > we cannot have He and Li openflowplugin implementations in the same > > controller run-time. The possible solutions to this: > > > > - If He and Li implementations are going to exist for a while and we expect > > a gradual project migration He->Li, we will have to break down our csit > > "compatible with all" (*-all-*) tests in 2 groups: "compatible with He > > plugin" and "compatible with Li plugin". > > - If there is a plan for migrating He-> Li and this transition is expected > > to happen quick, we could survive with the current -all- test pulling He > > plugin and change this to Li plugin when migration is completed. > > > > I am cc-ing openflowplugin-dev people to see if we can get more inside info > > on the above. > > > > BR/Luis > > > > _______________________________________________ > release mailing list > [email protected] <mailto:[email protected]> > https://lists.opendaylight.org/mailman/listinfo/release > <https://lists.opendaylight.org/mailman/listinfo/release> >
_______________________________________________ openflowplugin-dev mailing list [email protected] https://lists.opendaylight.org/mailman/listinfo/openflowplugin-dev
