17:04:06 #startmeeting tacker 17:04:07 Meeting started Tue Mar 8 17:04:06 2016 UTC and is due to finish in 60 minutes. The chair is sridhar_ram. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:04:09 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:04:11 The meeting name has been set to 'tacker' 17:04:27 sorry, few mins later... 17:04:31 *late 17:04:38 hello 17:04:42 I too was 4 minutes late 17:04:42 #topic Roll Call 17:04:45 who is here ? 17:04:45 o/ 17:04:50 o/ 17:04:52 o/ 17:04:54 o/ 17:04:54 o/ 17:04:56 o/ 17:05:00 o/ 17:05:01 o/ 17:05:02 o/ 17:05:27 o/ 17:05:39 howdy everyone! 17:05:46 lots of attendees, impressive! 17:05:46 lets roll.. 17:05:54 #topic Agenda 17:06:07 #link https://wiki.openstack.org/wiki/Meetings/Tacker#Meeting_Mar_8.2C_2016 17:06:19 #topic Annoucement 17:06:42 Tacker is now an official OpenStack big-tent project .. woot !! 17:06:52 #link https://review.openstack.org/#/c/276417 17:07:00 congratulations to one and all in the tacker team 17:07:08 great! 17:07:13 nice! 17:07:23 Woot congrats sridhar_ram. 17:07:52 yep, what a journey so far.. we now have the guidance of TC to take this forward...! 17:07:55 awesome 17:08:46 great 17:09:02 one of the key take aways to evolve tacker into something that could be even applicable beyond Telcos ... ! 17:09:21 we will keep that in mind as we progress in this journey! 17:09:39 next.. 17:09:43 We are entering Mitaka Release R-4 17:09:50 #link http://releases.openstack.org/mitaka/schedule.html 17:10:22 it is going to be a tight fit.. into Mitaka ! 17:10:42 we should try everything possible to squeeze in ... 17:11:07 .. and more importantly line up properly for Newton ! 17:13:17 #topic VNFFG-SFC update 17:14:18 trozet: can you provide an update on the plugin side of things ? 17:14:38 sridhar_ram: sure, do you guys use #info? 17:14:49 s3wong: can you provide an update on the networking-sfc driver part ? 17:15:00 trozet: not much ! 17:15:19 ok 17:15:28 trozet: you can use #info --- I know they do that a lot for ODL IRC meetings :-) 17:15:51 (though ODL IRC is mostly for logging info from the actual meeting mostly on webex) 17:15:52 trozet: yeah, feel free to use it .. 17:16:09 trozet: for key bits of info 17:16:14 #info i pushed https://review.openstack.org/#/c/286802/ which is just the vnffg extension piece. It isn't fully working yet, and was setting up an environment to debug it, but wanted feedback from tacker team on it 17:16:54 #info the plugin I can also post in parallel, or wait for some feedback on the extension - I don't think that piece will be difficult 17:17:36 #info as for the driver, I would like to talk with s3wong and figure out if I should focus on the networking-sfc driver for Tacker, or focus on an ODL driver for networking-sfc 17:18:13 trozet: my understanding is between s3wong and louisef tacker networking-sfc driver has coverage... 17:18:17 im a bit fuzzy on how we will translate networking-sfc architecture -> ODL arch (translating port pairs, etc) 17:18:36 trozet: that's definitely a longer discussion.. 17:19:07 sridhar_ram: ok is there any links to the code from s3wong or louisef on the networking-sfc driver? 17:19:11 I'm hoping to see where we stand w.r.t to tacker-vnffg -> networking-sfc as it stands today (without ODL support) 17:19:21 s3wong: can you please fill us in ? 17:19:26 for me, I had the spec ready 17:19:47 but was (pleasantly) surprise that LouisF also had written a spec 17:20:08 so we have merged a spec, and a preliminary draft should be posted either today or tomorrow 17:20:36 s3wong: this is a spec for the ODL driver for networking-sfc? 17:20:42 trozet: no code yet, unfortunately 17:21:00 trozet: no --- a networking-sfc driver spec for Tacker VNFFG 17:21:12 s3wong: ok got a link? 17:21:19 trozet: we would still need to write one for networking-sfc (and goes into networking-sfc repo) 17:21:30 trozet: haven't posted in on gerrit, will do that soon 17:21:40 trozet: and will put you in as a reviewer 17:21:49 s3wong: I'd like to contribute to that ODL driver as well, if I can 17:22:03 trozet: sure --- no one is working on that at the moment :-) 17:22:15 trozet: we can draft a spec together 17:22:33 trozet: I can bug LouisF and Cathy_ to review it :-) 17:22:38 so sridhar_ram: would you like me to just get the vnffg extension, plugin, networking-sfc driver working with just plain networking-sfc? then move onto ODL later? 17:22:53 trozet: actually I think that would be better 17:23:06 trozet: exactly .. that is the current initial scope 17:23:23 based on the ML email sent after the tacker big-tent discussion.. 17:23:26 trozet: looking at your code, for example, I think Neutron client instance should be part of the driver (specifically the networking-sfc driver) instead of inside the plugin 17:23:30 s3wong: networking-sfc is pretty stable and wokring? 17:23:44 s3wong: yeah agree 17:24:01 trozet: it is working :-) 17:24:12 ODL support in networking-sfc is a different discussion .. it needs to happen in networking-sfc 17:24:21 sridhar_ram: got it 17:24:42 I'd couldn't recommend some better than trozet to take that work! 17:24:47 trozet: networking-sfc meeting are on the same time, same channel on Thursdays 17:24:57 okay... decision time ... 17:24:57 sridhar_ram, s3wong: then I'll also work on the plugin next and something pu for review, and debug both on my setup 17:25:02 sridhar_ram: agreed 17:25:24 sridhar_ram, s3wong: and keep me in the loop on any networking-sfc driver for tacker info, I might hack one up myself just to test the extension and plugin 17:25:43 trozet: sure --- currently networking-sfc only has one driver: OVS 17:25:53 ok 17:25:58 trozet: absolutely...! again, we should plan for a spec in networking-sfc project.. 17:26:28 yeah 17:26:30 trozet: you contribution at networking-sfc would be very welcome 17:26:34 now we talk timeline :) ? 17:26:52 s3wong: yeah will start looking through networking-sfc code this weekend 17:27:04 alright.. given these facts and where we are in these work items.. it is quite clear we need to move to out of Mitaka .. into early Newton. 17:27:25 trozet, given the Mitaka cycle schedule that sridhar_ram posted earlier, I think it is highly unlikely that this can land in Mitaka 17:27:29 trozet: what would be the impact in OPNFV side ? 17:27:30 (just being realistic) 17:28:23 sridhar_ram: you mean OPNFV with SFC? 17:28:37 trozet: yes... 17:28:52 sridhar_ram: the SFC project in OPNFV focuses on ODL, so either way I will need to support that 17:29:18 trozet: for VNFM I thought tacker is pulled as a post-install option ? 17:29:32 * sridhar_ram we need to move to next agenda item in couple of mins 17:29:37 sridhar_ram: it does, in the next OPNFV release (August) it will be part of the installer 17:29:49 sridhar_ram: OPNFV will allow us to put MANO into our installers for that release 17:30:01 trozet: that will be based on Mitaka correct ? 17:30:06 sridhar_ram: yeah I think so 17:30:40 sridhar_ram: since we are release independent, can we pull a Mitaka release for Tacker several weeks after Mitaka official? 17:30:44 sridhar_ram: I can use a github fork of any progress we make upstream tacker, and include the ODL driver 17:30:56 trozet: okay.. it will miss Tacker VNFFG option for Aug release as well.. we gottu plan any demos based on off repo code 17:31:15 trozet: exactly.. make sense 17:31:24 sridhar_ram: and i think that is perfectly OK for OPNFV 17:31:32 sridhar_ram: like i said we have more flexibility there 17:31:41 trozet: there is a bigger release cadence mis-match between OS and OPNFV .. that we can't solve here 17:31:44 trozet: great..! 17:32:28 sridhar_ram: but given the time frame for Mitaka, i agree I don't think we can get the SFC complete in that time frame 17:32:35 s3wong: we are planning to go w/ release-with-intermediary... we want to bit more structured for downstream projects to take us seriously 17:32:51 trozet: agreed.. 17:33:07 sridhar_ram: OK 17:33:10 trozet: s3wong: please propose you specs to tacker/newton 17:33:19 sridhar_ram: OK 17:33:23 I'll create a dir in tacker-specs repo for newton.. 17:33:34 will do 17:33:53 general shout out to everyone.. time to "soft-land" all features in flight ASAP in Mitaka... 17:34:10 and get started on Newton features 17:34:20 anything else quick on SFC ? 17:34:45 alright, lets move on.. 17:34:57 #topic NFV Catalog evolution 17:35:22 the context of this .. as discussed in mitaka midcycle.. VNFD catalog is buried in VNFM. 17:35:46 we wanted to split that into someplace common across VNFM and (the emerging) NFVO.. 17:35:57 Sonata team has ideas and requirements in this area 17:36:05 Sonata folks - please take over.. 17:36:26 sridhar_ram: what is Sonata? (sorry for my ignorance) 17:36:34 Hi everyone, we've prepared a Google Doc upon the advice of Sridhar 17:36:48 s3wong: they got introduced in the last weekly meeting .. eavesdrop is your friend! 17:37:02 sridhar_ram: OK :-) 17:37:02 https://docs.google.com/document/d/1SDq_ujOIPba7O7jgcecUVvUBq3D8rCgNP00PzPBMN50/edit 17:37:27 We have a short intro there too for more context. 17:38:26 JimA__: that's a quite well written doc... 17:38:41 So the workplan context, as Sridhar knows, is that our first implementation of this catalog is for our April milestone, going in the integrated open source release this summer 17:38:52 thanks to my colleagues here in the chat 17:39:28 JimA__: do you mind introducing some of the catalogue features envisioned for Sonata ? 17:40:43 on p3 there is an overview of those features, would you like us to mention them directly here in the chat? 17:41:23 (bare with us, we're new to the chat-only meetings :) ) 17:41:26 JimA__: no.. that's okay.. I was trying to extract some key aspects for the crowd here.. 17:41:51 JimA__: totally understand.. we all were there once :) 17:42:08 let me paraphrase then.. 17:42:39 catalog features in no particular order... 17:42:55 In the SONATA project, we have two main catalogues .. SDK-Catalogue (local to the developer) and SP-Catalogue (for service deployment/also available to other developers) 17:43:39 Shuaib: lets zoom in SP-Catalogue for a min... 17:44:22 the document talks about an ability to activate / deactivate NSD / VNFD .. what you mean by this ? 17:46:06 This activate/deactivate feature is to unpublish a NSD/VNFD (not delete) so it is not available to other developers .. 17:46:57 the running instances will keep running but new instances couldnt be launched 17:47:19 for the deactivated NSD/VNFD 17:47:35 Shuaib: cool.. 17:48:05 I can clearly see its use..! 17:49:03 Shuaib: you mention "developers".. I want to get the "personas" depicted in the architecture correctly... are developers the one who author VNFD and NSD templates ? 17:49:16 Shuaib: how about the role of "operators" in this architecture ? 17:49:54 yes .. the developers are who author NSD/VNFD 17:50:28 sridhar_ram: can you please help us explain how this would be useful? 17:50:39 Operators will be the one running the SONATA service platform 17:51:20 Shuaib: sure.. I was trying to get the nomenclature correct to read the doc better 17:52:23 any questions from other in tacker team on sonata / catalog ? 17:52:49 sridhar_ram: i posted a question above 17:53:08 sripriya: oops.. missed it 8( 17:53:14 In this project, we are addressing from NS development to NS deployment and execution. 17:53:17 Another quick link for context: https://drive.google.com/file/d/0B_eHzl365Ks0Z2Y0bjhjYzMzb0U/view 17:53:18 sridhar_ram: np 17:53:30 (overview and arch.) 17:54:02 I think it is fair to take some time to digest the architecture and come up w/ opinions based on that... 17:54:16 yes, very much agree 17:54:34 the interesting exercise would be to make some of the ideas into tacker asks, requirements and eventually to a spec that we can implement 17:55:01 Ok 17:55:03 for e.g. the ask of activate/deactive is quite useful... 17:55:29 How best can we follow-up between the weekly meetings? 17:55:41 another thing I find useful in the "ask" is the ability to "version" the catalog entry 17:56:06 JimA__: email to mailing list ... 17:56:28 JimA__: .. to openstack-dev with [tacker] in the subject line.. 17:56:52 SONATA is considering versioning as well in the catalogues 17:57:22 Shuaib: I find that I interesting... 17:57:43 JimA__: we can also discuss this in following weekly meetings as an agenda item... 17:57:59 we anyway are entering Newton feature planning,... so this is good time 17:58:12 lets wrap NFV catalog for today.. 17:58:33 are there some initial progress on the catalog side from Tacker up until now that we could see to sync? 17:58:40 e.g. requirements list? 17:59:15 JimA__: yes, I can point you to that... 17:59:23 we are out of time for today... 17:59:34 JimA__: Shuaib: thanks a lot for introducing us to Sonata.. I'm excited about this collaboration! 17:59:43 #topic Open Discussion 17:59:46 thanks to you all 17:59:58 Thank you .. we are glad to collaborate! 18:00:08 markus_z: Error: Can't start another meeting, one is in progress. Use #endmeeting first. 18:00:15 bye 18:00:17 byt all 18:00:19 bye all 18:00:23 #endmeeting