16:00:44 #startmeeting tacker 16:00:44 Meeting started Tue Jul 19 16:00:44 2016 UTC and is due to finish in 60 minutes. The chair is sridhar_ram. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:45 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:48 The meeting name has been set to 'tacker' 16:00:53 #topic Roll Call 16:00:59 hi 16:01:08 o/ 16:01:10 o/ 16:01:11 o/ 16:01:14 o/ 16:01:14 o/ 16:01:25 hi all 16:01:32 o/ 16:01:34 o/ 16:01:50 howdy folks! let's start... 16:02:14 #topic Agenda 16:02:19 #link https://wiki.openstack.org/wiki/Meetings/Tacker 16:02:42 packed agenda.. 16:02:51 #topic Annoucements 16:02:57 sridhar_ram, shall we discuss the current items first, I saw today's meeting agenda, its big :) 16:03:40 KanagarajM: yes, planning to take the items in flight first.. 16:04:08 back to annoucements.. 16:04:26 Since we are meeting next week for midcycle meetup.. 16:04:54 #info No weekly meeting next week, on July 29th 16:05:41 Newton Feature Freeze is coming close - Aug 29th, just 7-weeks away 16:05:50 #link http://releases.openstack.org/newton/schedule.html 16:06:22 all features and RFEs need to merge by that date 16:06:41 only bug fixes and integrations issues can be addressed after that.. 16:07:18 Please get your patchsets for specs and RFEs to post as soon as possible... 16:07:37 we have a busy review cycle upon us.. 16:07:50 #topic Project Mascot 16:08:04 #link http://lists.openstack.org/pipermail/openstack-dev/2016-July/099507.html 16:08:23 We need to pick an animal mascot for Tacker project.. 16:08:38 here is the etherpad to collect inputs.. 16:08:47 #link https://etherpad.openstack.org/p/tacker-mascot 16:09:06 an animal is a must? 16:09:52 janki, +1 16:09:54 janki: it can be anything from the natural world 16:09:59 i like this idea :) 16:10:03 animal is the popular choice 16:10:21 there is an interesting option in the etherpad to use a Squid.. 16:10:26 i think only heat had some sort of logo earlier. good to have for all 16:10:28 I had an idea (shared on ehterpad) but it doesnot have animal/natural thing 16:10:50 Ya, squid is nice one 16:11:06 few of us created this earlier for Tacker (with some professional help), 16:11:08 #link https://wiki.openstack.org/w/images/2/21/Squid_rev7.jpg 16:11:14 how does this look ? 16:11:32 sridhar_ram: +1, its really cool 16:12:00 sridhar_ram: nice! 16:12:01 again, openstack logo team will rework this to make it something that is identifies it as openstack project.. 16:12:33 any significance for containers in one of the tentecle 16:12:33 janki: s3wong: thanks! 16:12:36 manikanta_tadi: Tacker eventually orchestrating VNFC in Containers 16:13:07 hopefully we get there sooner with integration with projects like Magnum & Zun/Higgins 16:13:24 please leave ur comments in etherpad.. 16:13:33 moving on.. 16:13:53 #topic VNF Scaling 16:14:05 #link https://review.openstack.org/#/q/status:open+branch:master+topic:tacker-scaling 16:14:18 KanagarajM: over to you 16:14:32 this current patch set have the end-end functioanlity 16:15:15 sridhar_ram, would be great if we could get cores time on these patches 16:15:32 KanagarajM: sure 16:15:43 hopefully this week or early next week we could target to merge them 16:16:23 cores: please help to land this by early next week 16:16:39 sridhar_ram, thanks. 16:16:51 sridhar_ram, i will land the doc patches sooner 16:17:00 for the scaling 16:17:12 KanagarajM: I'm pulling the patchset to try.. now, how do you typically trigger the scaling to test ? 16:17:36 please pull both cli and server patches 16:17:51 and it should work simply by using the tacke vnf-scale cli 16:18:05 KanagarajM: okay... 16:18:21 there are two samples, and one with unit testing would be simple to test 16:18:41 sounds good 16:18:42 any questions on vnf-scaling ? 16:18:49 *any other 16:19:03 i will and the patch for functional test once client patch get merged 16:19:20 land* 16:19:31 KanagarajM: we might need to make a client release w/ scaling changes ? 16:19:54 sridhar_ram, yes. i feel so 16:20:27 VNFFG will bring in slew of client changes.. not sure if we can club both into one release 16:21:01 i'll take up the client release part 16:21:05 VNFFG is targeted for Newton release? 16:21:16 janki: yes! 16:21:27 sridhar_ram, sure. but i feel it might delay the functional test. but i feel its ok 16:21:41 sridhar_ram: cool 16:21:55 let's briefly discuss other things in the pipeline.. 16:22:10 apologies for impromptu change.. 16:22:17 sridhar_ram, i will try to keep the patch ready, so that once the client rlease is done, we could take immediate action on functioanl test patch 16:22:34 KanagarajM: perfect! 16:22:57 KanagarajM: thanks a lot ! couldn't be happier how vnf-scaling has come together 16:23:18 #topic VNFFG-AlarmMon-Status 16:23:59 tung_doan: do you've an idea when you might be able to post the alarm-monitoring patchsets for review ? 16:24:02 sridhar_ram, I think we missed NSD 16:24:21 dkushwaha: will got to it in few mins 16:24:48 sridhar_ram, ok 16:25:20 sridhar: I already devided some parts for impemtation 16:25:38 sridhar: hope to finish soon 16:26:26 tung_doan: okay, again please keep in mind all the code-review and devref need to finish by Aug 29th.. otherwise this feature need to move to O-release.. 16:26:49 s3wong: along the same lines, how is your networking-sfc driver coming along ? 16:26:57 sridhar: i know this time is very important.. thanks for reminding 16:27:09 tung_doan: sure 16:28:05 s3wong: u there ? 16:28:20 s3wong: did you catch the question above ? 16:29:00 * sridhar_ram will sync up w/ s3wong offline 16:29:04 #topic NSD 16:29:26 #link https://review.openstack.org/304667 16:29:34 dkushwaha: take over, please 16:29:42 sridhar_ram, Last week, I have updated the spec. 16:30:03 sridhar_ram, i have some questions 16:30:20 dkushwaha: shoot 16:30:33 sridhar_ram, and want o know team's opinion 16:30:52 sridhar_ram, As in spec, I am mentioning VirtualLink and ConnectionPoint tables separately, is it make sense? or we need to keep it attributes part of NetworkServiceTemplate ? 16:31:33 dkushwaha: IMO, db tables for VLs and CPs doesn't make sense 16:31:51 sridhar_ram, +1 I too felt the same 16:32:02 sridhar_ram: I agree 16:32:06 dkushwaha: handling it as attributes similar to how we are doing it for VNFD should be sufficient 16:32:39 dkushwaha: now, what is the advantage of having this in DB ? 16:33:15 there is a reason you are proposing this and I'd like to understand that 16:33:36 sridhar_ram, Actually I am thinking to invoke create_vnfd inside nsd onboard operation. and handle vld and cp separately 16:34:28 sridhar_ram, thats why i thought to keep it in different tables 16:34:42 dkushwaha: these VLD and CPd are part of NSD yaml template itself, correct ? 16:34:57 sridhar_ram, yes 16:35:24 dkushwaha: do we allow any API / CLI operations directly on VLD or CP ? 16:35:42 * sridhar_ram thinks most likely not 16:36:20 sridhar_ram, we can allow for some operations 16:36:32 sridhar_ram, like add vld 16:36:56 dkushwaha: yes, but that is probably for a future work, and it shd be out of scope for NSD work 16:37:39 sridhar_ram, agree. make sense to me 16:38:07 sridhar_ram: given that VNFFD and VNFD is in place and now NSD is coming in, are we thinking about a separate catalog framework for these templates? 16:38:26 btw, on a related but a different idea.. i'd like you to explore using Mistral workflow for your NSD work 16:39:09 Mistral support for Tacker got recently merged.. 16:39:19 #link https://review.openstack.org/#/c/331827/ 16:39:36 and there are some same workflow to create VNFD / VNF is posted recently.. 16:39:38 sripriya, now glare is ready ! so it would be an choice to consider :) 16:39:45 #link https://review.openstack.org/#/c/341164/ 16:40:10 sripriya, http://lists.openstack.org/pipermail/openstack-dev/2016-July/099553.html 16:40:50 sripriya: good point in separating call the catalog out of VNFM 16:40:59 s/call/all/ 16:41:23 KanagarajM: thanks for the pointer, they mention tacker as well in the mail 16:41:30 sripriya, yes 16:42:05 we shd engage more w/ glare team in Barcelona, we missed sync w/ them in Austin... 16:42:55 Folks: can nfv-catalog split out be taken up as part of NSD work ? 16:43:18 there shouldn't be any end user facing VNFD impact ? 16:43:22 sripriya: ? 16:43:42 sridhar_ram: i guess no 16:44:15 dkushwaha: please assess the impact of nfv-catalog split out as part of your NSD work 16:44:32 sridhar_ram, yes sure 16:44:47 dkushwaha: all: please think about using mistral for NSD 16:45:04 we can discuss that in the midcycle 16:45:14 sridhar_ram, I will llik into that 16:45:23 sridhar_ram, is invoking create_vnfd inside onboard nsd is correct way? 16:46:24 dkushwaha: in the initial phase, NSD can just reference VNFDs already onborded into the catalog 16:47:32 dkushwaha: in a follow-on, we can consider supporting in-ling VNFDs within NSD. Then we need create_vnfd() with some properties to make it special (not visible in VNFD list, only available for that specific NSD) 16:47:44 *in-line / embedded 16:48:38 dkushwaha: let's continue the discussion in the midcycle.. you shd lead this topic, obviously 16:48:52 sridhar_ram, yes sure 16:48:55 moving on.. 16:48:59 #topic VNFC 16:49:14 Hi All, Myself and bharath tried to understand how heat software component can be fit for VNFC support. 16:49:36 #link https://review.openstack.org/#/c/339798 16:49:46 One question on our mind is , How vnf upgrade can be supported on fly if we opt for heat software component 16:50:38 KanagarajM: have some thought here ? 16:51:04 sridhar_ram : we are exploring various drivers apart from ssh driver what we had in the spec 16:51:33 basically, we need to know if heat can do "uninstall" of s/w in VMs 16:52:24 manikanta_tadi: can you & tbh plan to present some slides on this topic for next week's midcycle... 16:52:30 KanagarajM, pls provide your inputs from heat perspective 16:52:39 please go over all the options in that prezo 16:52:49 sridhar_ram : Sure! 16:53:20 I'll also lookup a bit more on Heat SoftwareComponent 16:53:20 sridhar_ram : We will come up with that... 16:53:48 openstack-service --> VMs spawned is a hot topic these days in the ML 16:54:04 sridhar_ram: sorry, was out of desk a while ago 16:54:11 there is a concern different projects are resorting to different methods 16:54:18 something to keep in mind 16:54:36 s3wong: np 16:54:37 sridhar_ram: currently aiming to have the patches out before the mid cycle next week; sorry for the delay 16:55:05 s3wong: sure, release dates are approaching faster than i thought it would :) 16:55:15 moving on 16:55:29 #topic Midcycle Meetup Topics 16:55:56 #link https://etherpad.openstack.org/p/tacker-newton-midcycle 16:56:11 last call for topics to disucss in the midcycle.. 16:56:30 i'd like to invite the team to actively participate and be.. vocal ! 16:56:46 I have a question 16:56:46 I'd rather talk less and .. have the community speak :) 16:56:50 janki: sure 16:57:07 Are we planning to have an ansible playbook for tacker as well? 16:57:11 * sridhar_ram 3mins left 16:57:32 janki: yes, someone from Brocade is working on it.. 16:57:41 ohk, cool 16:57:47 janki: need to check how far it has progressed.. 16:58:05 sridhar_ram: sure, let me know if help needed 16:58:13 janki: sure 16:58:17 sridhar_ram: I also have an idea/feature development 16:58:25 not sure how valuable it is 16:58:37 anything else on midcycle ? 16:59:15 janki: write a spec / blueprint 16:59:26 sridhar_ram: sure, will do 16:59:28 janki: don't worry about the usefulness.. 16:59:34 janki: you never know :) 16:59:34 :) 16:59:38 alright times up 16:59:46 thanks for joining everyone 16:59:49 bye 16:59:51 talk to all next week 16:59:59 #endmeeting