Hi Numan,
By keeping both schemes I mean we can develop the re-factor and address bugs found in scale environment and broader testing(which aren't run as part of CI) while keeping the earlier/existing scheme around for bug fixes + backports. Finally I would prefer one way of doing this (with the refactor in context of contrail-neutron-plugin) and we switch over to it after we are satisfied with testing. thanks ajay ________________________________ From: Dev <[email protected]> on behalf of Numan Siddique <[email protected]> Sent: Wednesday, May 27, 2015 10:28 PM To: [email protected] Subject: Re: [opencontrail-dev] New Neutron OpenContrail plugin Hi Ajay, What do you mean by keeping both the schemes around ? Do you mean to have both the schemes and make it configurable to choose one of them until a final decision is made which one to keep ? Shall we abandon the review https://review.opencontrail.org/#/c/10123/ on api_server and work on adding this refactor code to contrail-neutron-plugin or you want this to be merged (ofcourse after the code reviews and CI passes :) ) first ? Thanks Numan On 05/27/2015 01:12 AM, Hampapur Ajay wrote: Hi Edouard, Placing the adaptation from neutron to contrail representation outside of api-server seems logical(and beneficial since 2 cores can be used by the 2 processes) but we have to consider the fact that this layer of code is performance sensitive (especially for list) and being restricted to neutron-server's environment (eventlet based + changing at a pace that we don't influence) may be limiting.? We may have to keep both schemes around (vnc_openstack with neutron_plugin_db.py running in context of api-server and the refactor to neutron_plugin_db.py running in context of neutron-server) till they achieve parity for testing + scale performance. Thanks ajay ________________________________ From: Édouard Thuleau <[email protected]><mailto:[email protected]> Sent: Sunday, May 24, 2015 2:03 AM To: [email protected]<mailto:[email protected]> Cc: Nachi Ueno; Hampapur Ajay; Numan Siddique; Babu Shanmugam Subject: Re: [opencontrail-dev] New Neutron OpenContrail plugin Oops the mail was sent to quick... I re-write it from scratch: Hi all, As you already know Numan and Babu refactored an re-organize the vnc_openstack code [1]. And you also probably know, the OpenStack Neutron project decided to move all plugins and service plugins outside the Neutron tree [2]. Nachi already proposed a review to split the OpenContrail plugin code in the neutron tree [3]. With that patch, Neutron tree will contain only a plugin folder that adds python dependency to contrail-neutron-plugin [4]. That means all the Contrail plugin code is external to Neutron repository. So what I propose it's to merge new vnc_openstack code to the contrail-neutron-plugin repository and deprecated the old vnc_openstack code into the Contrail API to remove in the future. Any thoughts? [1] https://www.mail-archive.com/[email protected]/msg00065.html [2] https://github.com/openstack/neutron-specs/blob/master/specs/kilo/core-vendor-decomposition.rst [3] https://review.openstack.org/#/c/163218/ [4] https://github.com/Juniper/contrail-neutron-plugin Regards, Édouard. On Sun, May 24, 2015 at 1:31 AM, Édouard Thuleau <[email protected]<mailto:[email protected]>> wrote: Hi all, As you already know Numan and Babu refactored an re-organize the vnc_openstack code [1]. And you also probably know, the OpenStack Neutron project decided to move all plugins and service plugins outside the Neutron tree [2]. Nachi already proposed a review to split the OpenContr It's why I propose to create a [1] https://www.mail-archive.com/[email protected]/msg00065.html [2] https://github.com/openstack/neutron-specs/blob/master/specs/kilo/core-vendor-decomposition.rst https://review.openstack.org/#/c/163218/ _______________________________________________ Dev mailing list [email protected]<mailto:[email protected]> http://lists.opencontrail.org/mailman/listinfo/dev_lists.opencontrail.org
_______________________________________________ Dev mailing list [email protected] http://lists.opencontrail.org/mailman/listinfo/dev_lists.opencontrail.org
