16:00:53 <sridhar_ram> #startmeeting tacker 16:00:54 <openstack> Meeting started Tue Jul 5 16:00:53 2016 UTC and is due to finish in 60 minutes. The chair is sridhar_ram. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:55 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:58 <openstack> The meeting name has been set to 'tacker' 16:01:06 <sridhar_ram> #topic Roll Call 16:01:13 <vishwanathj> o/ 16:01:16 <dkushwaha__> o/ 16:01:17 <tbh> o/ 16:01:20 <janki> o/ 16:01:26 <sripriya> o/ 16:01:59 <sridhar_ram> howdy all ! 16:02:15 <manikanta_tadi> o/ 16:02:20 <sridhar_ram> KanagarajM_: are you here ? 16:02:33 <KanagarajM_> sridhar_ram, hi 16:02:50 <sridhar_ram> alright, lets start.. 16:02:58 <sridhar_ram> #topic Agenda 16:03:08 <sridhar_ram> #link https://wiki.openstack.org/wiki/Meetings/Tacker 16:03:17 <s3wong> o/ 16:03:31 <sridhar_ram> #topic Annoucements 16:03:39 <sridhar_ram> master based python-tackerclient is released 16:03:47 <sridhar_ram> #link http://lists.openstack.org/pipermail/openstack-announce/2016-June/001259.html 16:04:04 <sridhar_ram> going forward we will make regular client releases based off master 16:04:39 <KanagarajM_> Nice :) 16:04:39 <sridhar_ram> Doodle poll to narrow down the date / time for Midcycle meetup is at.. 16:04:43 <sridhar_ram> #link http://doodle.com/poll/qzsagbhnhqbuurth 16:05:01 <sridhar_ram> Please respond to the poll 16:05:14 * sridhar_ram notes to send this poll link to ML 16:05:39 <sridhar_ram> OpenStack Barcelona Summit talk submission is coming up - July 13th! 16:05:53 <sridhar_ram> moving on... 16:06:04 <sridhar_ram> #topic Audit Event log 16:06:15 <sridhar_ram> vishwanathj: KanagarajM_: please take it away 16:06:36 <vishwanathj> sridhar_ram thanks .... 16:07:34 <vishwanathj> I think at this time, KanagarajM_ and I believe that the spec is in a reasonable shape for other reviewers including core reviewers to start reviewing and provide feedback ...... 16:08:11 <sripriya> vishwanathj: can you please share the link? 16:08:32 <vishwanathj> https://review.openstack.org/#/c/321370/ 16:08:56 <sripriya> vishwanathj: thanks 16:09:17 <sridhar_ram> vishwanathj: KanagarajM_: are Horizon UI changes just going to be per resource ? 16:10:06 <sridhar_ram> are you envisioning any general "Events" tab that spans across all tacker resources ? 16:10:34 <vishwanathj> sridhar_ram, yes we are envisoning a general "Event" tab..... 16:10:57 <sridhar_ram> vishwanathj: okay 16:11:00 <vishwanathj> however, would want to first focus on getting the per resources implementation working..... 16:11:22 <sridhar_ram> vishwanathj: sure, that can be a follow on.. 16:11:44 <vishwanathj> the Events tab display in Horizon would be easier given the way data is structured in the database.... 16:12:00 <sridhar_ram> agree 16:12:04 <sridhar_ram> on my side, i'm good to sign off on this one .. i don't see any major blocker at this time 16:12:20 <vishwanathj> also we are making an attempt to keep the paradigm for accessing events in Horizon and tacker client similar.... 16:13:31 <vishwanathj> the idea in Horizon for e.g; when you click on a VNF, it would display VNF details in one tab and would have another tab for list of asscociated events.... 16:13:49 <sripriya> vishwanathj: will we be retaining events data for ever or is there some kind of an expiry on the data? 16:14:21 <michael_bredel> o/ 16:14:25 <vishwanathj> similarly in tacker client when the tacker vnf-show <id> command is executed, we are thinking of displaying VNF details info along with VNF event details info 16:14:58 <neel> o/ 16:15:14 <vishwanathj> KanragarajM_ feel free to chime in..... 16:15:18 <sridhar_ram> vishwanathj: VNF visibility is a huge deal for NFV ... 16:15:45 <anshukch> o/ 16:15:51 <sridhar_ram> vishwanathj: we shd make this a prominent 'first-class' feature of any Tacker resource 16:16:15 <vishwanathj> sripriya, there will be a command provided to purge entries from database.... 16:16:17 <sridhar_ram> lets not relegate this into a show command ooption 16:16:50 <vishwanathj> sridhar_ram I see your point for the VNFs 16:16:55 <sripriya> vishwanathj: i see that, thanks 16:17:35 <sridhar_ram> sounds good... 16:17:40 <vishwanathj> sridhar_ram, there will be a separate events level support for all resources.... 16:18:09 <vishwanathj> additionally for VNF resources, we plan to have vnf-show-events style command as per your feedback 16:18:54 <sridhar_ram> cool, I'd encourage other reviewers to provide their comments / +1s / +2s.. 16:19:22 <vishwanathj> KanagaragM_ anything that you want to add I might have missed .... 16:19:40 <manikanta_tadi> +1 for vnf-show-events style command 16:19:42 <KanagarajM_> vishwanathj, i think all covered 16:20:01 <sridhar_ram> okay, let's move on.. 16:20:23 <sridhar_ram> #topic Network Services Descriptor (NSD) 16:20:33 <sridhar_ram> #link https://review.openstack.org/304667 16:20:47 <sridhar_ram> dkushwaha__: tbh: please take over.. 16:21:22 <dkushwaha__> sridhar_ram, thanks. 16:22:19 <dkushwaha__> sridhar_ram, I have committed WIP prototype patches for nsd 16:22:55 <dkushwaha__> in which tbh also working with me 16:23:18 <sridhar_ram> the main thing IMO to be resolved for this effort is (a) scope of this initial NSD work and (b) tosca-parser dependency 16:24:26 <sridhar_ram> on the scope, the spec should describe what is the plan for .. 16:24:41 <sridhar_ram> (a) VNFFG embedded in NSD .. will this be supported ? 16:24:41 <tbh> sridhar_ram, regarding the tosca-parser dependency, do you think we have any dependencies on tosca-parser apart from substitution_mappings? 16:24:59 <sridhar_ram> (b) can each VNF be placed in different VIMs ? 16:25:39 <sridhar_ram> tbh: substition_mapping is probably the main thing.. but there might be others that you would get to know only when you starting trying :) 16:27:07 <sridhar_ram> back to scope, my suggestion is to introduce this with limited scope - to support just VNFs, VLs and not support VNFFG in the initial NSD 16:27:16 <tbh> sridhar_ram, sure, for VNF placement I think we are taking --vim-id input, dkushwaha__ am I correct? 16:27:24 <sridhar_ram> also restrict it, all VNFs in the same VIM 16:27:35 <KanagarajM_> tdh: this would need nested template support generation from heat-translator. 16:27:55 <dkushwaha__> sridhar_ram, tbh yes. 16:28:20 <KanagarajM_> sridhar_ram, as VNFD already supports to place across VIM, will it be an issue for supporting in NSD ? 16:28:28 <dkushwaha__> sridhar_ram, currently I planed scope with to support just VNFs, VLs 16:28:36 <sridhar_ram> KanagarajM_: no, we shouldn't take that approach 16:28:58 <sridhar_ram> Tacker-NSD feature should call Tacker-VNFM to stand-up each VNF specified in the NSD 16:29:18 <tbh> KanagarajM_, can you elaborate on nested template support? 16:29:43 <sridhar_ram> I've placed a midcycle meetup topic to discussion Tacker NFVO and Tacker VNFM separation 16:30:18 <KanagarajM_> tbh, As NSD is containing VNFD and VNFD is already an HOT tempalte, you may endup in placing multiple VNFD hot template inside NSD HOT tempalate 16:30:41 <sridhar_ram> we should avoid Tacker NFVO features to be tightly coupled with Tacker VNFM 16:30:41 <KanagarajM_> tbh, but i see the sridhar_ram point too 16:31:22 <sridhar_ram> i envision we might end up having two git repos in the near future .. one for tacker-nfvo and one for tacker-vnfm 16:31:25 <KanagarajM_> sridhar_ram, yes. that is good to address 16:32:14 <tbh> KanagarajM_, got it thanks 16:32:23 <sridhar_ram> this is one of the learnings from the recent OPNFV summit :) 16:33:11 <sridhar_ram> tbh: dkushwaha__ : please continue to shape / iterate on the spec... describe the scope clearly.. what it will do and *more importantly* what it will NOT do :) 16:34:09 <sridhar_ram> we also could make use of a workflow engine like mistral to implement.. 16:34:09 <dkushwaha__> sridhar_ram, yup, I will update the spec this week 16:34:33 <sridhar_ram> .. instead of coding all the workflow steps in python.. 16:34:49 <sridhar_ram> there are pros and cons to such an approach.. 16:34:58 * sridhar_ram wonders if bobh is here .. 16:34:59 <tbh> sridhar_ram, sure will update the spec based on recent learnings 16:35:33 * bobh is on a conf call and not really paying attention.... 16:35:43 <sridhar_ram> tbh: dkushwaha__ : sounds good.. 16:36:06 <michael_bredel> Sonata includes Mistral as workflow engine. If you want, we can share our experiences at some point 16:36:19 <sridhar_ram> we can make this a recurring agenda item for next few weekly meeting to track its progress 16:36:31 <sridhar_ram> michael_bredel: that would be great! 16:36:50 <dkushwaha__> sridhar_ram, yes sure 16:38:01 <sridhar_ram> IMO, if we can embrace a full-featured worflow engine like Mistral, we can achieve lot more feature velocity by generating worflow templates for features instead of coding lot of python code to achieve the same 16:38:18 <sridhar_ram> perhaps this could be another Midcycle topic.. 16:38:45 <sridhar_ram> michael_bredel: would you be able to join our midcycle meetup to share you Sonata<->Mistral experience ? 16:38:52 <KanagarajM_> sridhar_ram, I completely agree on it ! 16:39:40 <sridhar_ram> anything else on NSD ? 16:40:26 <dkushwaha__> sridhar_ram, currently no from my side. 16:40:30 <sridhar_ram> okay, lets move on... 16:40:51 <sridhar_ram> tbh: dkushwaha__: nice to see you both picking this piece up! 16:41:05 <sridhar_ram> #topic Midcycle Meetup Topics 16:41:26 <sridhar_ram> #link https://etherpad.openstack.org/p/tacker-newton-midcycle 16:41:48 <sridhar_ram> Please RSVP on this etherpad if you plan to join 16:43:00 <sridhar_ram> Need less to say, if you would like to discuss a particular topic .. please add them 16:43:18 <sridhar_ram> based on the content, we can make it 1-day or 2-day event 16:43:50 <KanagarajM_> sridhar_ram, shall we keep the things to discuss on newton cycle priority 16:44:44 <KanagarajM_> sridhar_ram, you have listed many topics for meetup ! 16:44:54 <sridhar_ram> KanagarajM_: you mean the things currently in flight for Newton ? 16:45:18 <KanagarajM_> sridhar_ram, currently in and planned for newton 16:45:33 <sridhar_ram> KanagarajM_: sure, we can discuss them.. 16:45:58 <sridhar_ram> We have always tried to balance with things in flight + the upcoming roadmap 16:46:45 <sridhar_ram> Next, the biggest challenge I've is to find a timeslot that would ensure max participation 16:47:04 <sridhar_ram> I see we are spread across vast diff timezones 16:47:14 <KanagarajM_> sridhar_ram, yeah. timezone really seems difficult part 16:48:15 <sridhar_ram> I've picked early morning PDT that overlaps with late afternoon / evening in India.. but is late evening for Korea 16:48:45 <KanagarajM_> it comes between 7:30 PM to 1:30 AM for india 16:49:02 <sridhar_ram> US East coast and CET (Central European) gets reasonable 16:49:14 <sridhar_ram> *has reasonable timing 16:49:48 <sridhar_ram> perhaps we can go one hour earlier ? 16:50:06 <sridhar_ram> or make it two-days and reduce the hours 16:50:14 <tung_doan> sridhar_ram: np.. still keep time if it is fine for almost guys :) 16:50:16 <sridhar_ram> may be both :) 16:50:43 <dkushwaha__> sridhar_ram, I missed few weekly meetings due to timezone only, as it is 1AM JST :( 16:50:49 <sridhar_ram> tung_doan: I'm trying my best to see if you can join..! 16:50:59 <sridhar_ram> dkushwaha__: ouch! 16:52:01 <tung_doan> sridhar_ram: thanks :) 16:52:17 <sridhar_ram> okay, should we do 2-day / 5-hour slots starting at 6AM PDT, 6:30PM IST ? 16:52:41 <KanagarajM_> sridhar_ram, its better choice 16:52:54 <janki> sridhar_ram: better option 16:52:59 <sridhar_ram> any violet objections ? 16:53:05 <sridhar_ram> *violent 16:53:26 * sridhar_ram looks at sripriya 16:53:47 * sridhar_ram .. and s3won 16:54:07 <sripriya> sridhar_ram: 2-day works fine IMO 16:54:19 <sridhar_ram> sripriya: okay 16:54:29 <sridhar_ram> let's spread the suffering :) 16:54:43 <sridhar_ram> lets pencil this in.. 16:54:49 <sridhar_ram> will send a note to ML 16:54:50 <sripriya> sridhar_ram: coffeee to the rescue 16:54:59 <sridhar_ram> sripriya: absolutely! 16:55:23 <sridhar_ram> anything on this subject ? 16:55:41 <sridhar_ram> #topic Open Discussion 16:56:21 <sridhar_ram> tung_doan: we should also circle back to your monitoring spec.. lets discuss that in the next week's meeting to see if we can wrap that up 16:57:03 <KanagarajM_> please start to review https://review.openstack.org/#/q/status:open+branch:master+topic:tacker-scaling 16:57:03 <tung_doan> sridhar_ram: yes, right.. i really want. thanks, sridhar 16:57:13 <sridhar_ram> looks we are done for today.. 16:57:19 <sridhar_ram> thanks everyone for joining 16:57:27 <sridhar_ram> #endmeeting