16:00:31 #startmeeting tacker 16:00:32 Meeting started Tue Sep 13 16:00:31 2016 UTC and is due to finish in 60 minutes. The chair is sridhar_ram. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:34 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:36 The meeting name has been set to 'tacker' 16:00:44 #topic Roll Call 16:00:54 Hi folks! 16:00:55 o/ 16:00:56 o/ 16:01:01 o/ 16:01:03 o/ 16:01:31 o/ 16:01:36 o/ 16:01:38 o/ 16:02:13 Greeting everyone.. ! 16:02:20 checking few folks... 16:02:30 s3wong: tung_doan: are you here ? 16:02:44 #chair sripriya tbh bobh 16:02:45 Current chairs: bobh sridhar_ram sripriya tbh 16:03:55 #topic Agenda 16:04:12 #link https://wiki.openstack.org/wiki/Meetings/Tacker#Meeting_Sep_13th.2C_2016 16:04:35 we can take up "other" items in open discussion 16:04:47 #topic Announcements 16:05:09 We are in Newton RC1 week - 16:05:18 #link https://releases.openstack.org/newton/schedule.html 16:05:28 o/ 16:05:35 we will discuss more on this.. 16:05:54 nothing else to announce from my side 16:06:40 except.. i know bunch of Tacker devs got travel support from OpenStack Foundation and going to go to Barcelona ! 16:06:55 moving on.. 16:07:06 #topic Newton Release 16:07:38 sridhar_ram: I didn't get travel support unfortunately :( 16:07:38 lets wait to see if s3wong and tung_doan joins to get a status on those two last-minute feature 16:07:56 diga: ah, too bad... hopefully next time 16:08:18 sridhar_ram: hope so 16:08:32 sridhar_ram, me too :( 16:09:15 dkushwaha: unfortunate.. the decision making is still a bit of black box. Again, keep trying 16:09:28 back to newton release... 16:09:39 Here is the plan.. 16:09:56 most patchsets identified for Newton needs to land by this week.. 16:10:44 cores: please use this etherpad to review and merge - https://etherpad.openstack.org/p/tacker-newton-release-priority 16:10:47 hello; sorry, late 16:10:56 s3wong: hi, np 16:11:17 s3wong: taking up .. newton release priority, will circle back to vnffg soon 16:11:28 sridhar_ram: sure 16:11:53 Folks - anyone missing any *MUST HAVE* in newton in the above etherpad ? 16:12:21 s3wong: my setup is ready, need your help for e2e VNFFG testing 16:12:47 sridhar_ram: https://review.openstack.org/#/c/334256/ ? 16:12:47 diga: hang on one min, we will get to vnffg shortly 16:12:50 diga: great. Let's work on that offline 16:13:29 sridhar_ram: and also https://review.openstack.org/#/c/329759/ ? i will respin it 16:13:32 sripriya: it is already in the list 16:13:42 sridhar_ram: ah ok, sorry i missed it 16:13:57 sripriya: np.. 16:14:22 sridhar_ram, https://review.openstack.org/#/c/346665/ 16:14:35 sripriya: for https://review.openstack.org/#/c/329759/ .. please respin and add it to the etherpad please 16:14:56 sridhar_ram: sure. will do 16:15:08 dkushwaha: add it to the list.. 16:15:16 One general word of caution.. 16:15:45 .. your patchsets in this list need to in the healthy state.. CI shd pass fine at the minimum.. 16:16:21 we can't afford to take any collateral damage at this state.. 16:16:41 anything else folks ? 16:17:14 Now, starting next week... we need to test the heck out of the master branch.. 16:17:52 I need all your help to hammer different features, horizon ui and even install steps (particularly manual install) and report any issues 16:18:17 I'll request a newton release by 2nd half of next week 16:18:57 any general questions on newton release ? 16:19:28 sridhar_ram: the one thing i would request is if we cherry pick any patch to stable release and it changes the manual installation, other tham chaning the main doc, good to add that change to manual doc’s stable version 16:19:58 sripriya: that's a good point.. 16:20:32 I remember there is a patchset (from venkatamahesh) on doing various cleans in the manual install doc.. 16:20:53 I'll find that and add it to the priority list.. 16:21:14 Having the good install story is very, very important... ! 16:21:17 sridhar_ram: yes, for sure our doc needs to improve. thanks 16:21:27 moving on 16:21:40 #topic Newton-VNFFG 16:21:48 s3wong: can you give an update ? 16:21:54 diga: ^^ 16:22:03 sridhar_ram: I think we are ready for integrated / end to end test 16:22:07 hi 16:22:12 yes 16:22:22 sridhar_ram: diga has volunteered to get things going (Thanks!) 16:22:46 sridhar_ram: the one thing missing in the patchset now is to make it a mixin with openstack infra driver 16:23:13 (previously trozet has volunteered to make that change... but he didn't get to it before his vacation) 16:23:32 sripriya: you've handled the infra_driver refactoring.. what do think of the above ^^ ? 16:23:58 s3wong: which timezone you live in ? so that I can work accordingly 16:24:16 diga: pacific daylight time 16:24:20 sridhar_ram: sure i can take a stab at it 16:24:27 ok 16:24:33 diga: but I usually work on non-day-job related stuff after 10pm PDT 16:24:49 s3wong: :) 16:24:51 diga: what timezone are you at? 16:25:10 s3wong: IST(India STandard Time) 16:25:31 s3wong: will plan for this after the meeting 16:25:36 diga: so 10pm here should work well with you? 16:25:52 s3wong: yep 16:26:01 diga: cool 16:26:09 diga: let's get that going :-) 16:26:15 diga: as you are doing this, can you also "write up" a devref user guide doc.. 16:26:18 yes 16:26:32 sridhar_ram: will do 16:26:40 sridhar_ram: so hopefully we will have good news on integration test / end2end by next meeting 16:27:09 s3wong: okay 16:27:23 sridhar_ram: that's all I have 16:27:51 From blueprint point of view (assuming diga will take care of the docs).. 16:28:00 (other than that trozet has made sure everything works from VNFFG TOSCA to no-op driver .... GOOD JOB trozet!!!) 16:28:00 .. the pending item will be functional test 16:28:15 yes 16:28:56 s3wong: Agree, kudos to trozet .. starting from last fall where he demo'd Tacker + FFG using ODL in the OPNFV Summit.. he followed through! 16:28:58 sridhar_ram: will take care of it whichever remaining on the VNFFG side 16:29:12 diga: great, thanks.. 16:29:21 sridhar_ram: welcome! 16:29:43 diga: given you are testing .. both doc and func test will be appropriate for you to take up.. 16:29:57 sridhar_ram: yes +1 16:29:57 i know we might come short in time.. but please try your best 16:30:07 sridhar_ram: sure 16:30:49 Beyond that, given this (and alarm monitor) are coming in late.. I think it is appropriate to mark these two features as "experimental" .. 16:31:08 this is to set appropriate expectation to our user community 16:31:15 sridhar_ram: +1 16:31:54 we will use the next cycle to fix any outstanding / corner case issues and remove the "experimental" tag 16:32:07 moving on.. 16:32:16 #topic Newton-Alarm-Monitor 16:32:35 looks tung_doan is not here... 16:32:39 sridhar_ram, I pinged him, but no response 16:32:55 but tung_doan and dkushwaha are putting some amazing effort here.. 16:33:04 dkushwaha: can you give an update ? 16:33:26 sridhar_ram, Mainly I were involved in testing the patch and notified to Tung, and he keep updating accordingly. 16:33:42 dkushwaha: good collaboration..! 16:34:11 sridhar_ram, its running successfully 16:34:39 dkushwaha: can you describe the scenario you tested ? 16:35:16 sridhar_ram, I tried by puting high cpu load on create vm 16:35:34 dkushwaha: what action did it trigger ? 16:35:45 and respawn is success now 16:36:07 dkushwaha: good.. did you had a chance to try "scale" action ? 16:36:09 i tested it for respawn only for now 16:36:35 sridhar_ram, not yet, but i have to test it 16:36:35 dkushwaha: okay.. 16:36:43 sridhar_ram: dkushwaha: the scale url and alarm action url are different 16:37:19 sridhar_ram: i did not see a way to handle the scale action in create_alarm_url logic 16:38:44 sripriya: process_alarm / receiver should bubble up to the plugin level to invoke whatever action specified in the TOSCA template 16:39:02 sripriya: if that is not happening, we have a serious "workflow" problem 16:39:22 bobh: did you had a chance to review this ? 16:39:46 sridhar_ram: not yet - on my list for today/tonight 16:40:34 bobh: dkushwaha: sripriya: i'm trying to assess if this feature ... 16:40:36 (a) has progressed enough to become an "experimental" feature of Tacker 16:40:48 (b) doesn't introduce any collateral damage 16:41:35 if, as reviewers, can provide some feedback in next couple of days on these two aspects.. i'd really appreciate that 16:41:43 sridhar_ram: will do 16:41:59 sridhar_ram: i have not yet looked into code in depth, but scale api expects a certain url and body which needs to be handled. i may be missing something. will get back. 16:42:24 bobh: sripriya: thanks.. 16:42:46 anything else on alarm mon ? 16:43:10 #topic Ocata Grooming 16:43:24 #link https://etherpad.openstack.org/p/tacker-ocata-grooming 16:43:50 mike_m: can you share you ideas around EM BP ? 16:43:57 sure 16:43:59 #link https://blueprints.launchpad.net/tacker/+spec/em-module 16:44:27 the premise of what I think this blue print is about is related to a multi-vm strategy, and coordination between them. 16:44:49 think of a network service where there are different VM's, and there is a relationship between them. 16:44:59 (this is just one example though of FCAPS, the Config part) 16:45:29 if you look at juju canonical, it has a "relation" concept to associate different VM's, its very similar to that. 16:45:34 sridhar_ram, sripriya : I have pushed API framework spec - https://review.openstack.org/#/c/368511/ 16:46:08 the ETSI NVF framework also discusses this API, the "em" interface. 16:46:59 mike_m: if we take up the "C" part of FCAPS in this "relation" .. is it about supporting a sequence of "configuration" for a set of related VMs ? 16:48:02 yes it could be. But then the VNFM starts to become more specific, complicating the VNFM. Imagine instead there was a EM interface, where the VNFM told some entity (the defined EM), about all the VNFs. 16:48:45 so maybe the VNFM tells the EM, VNF "x" is instatiated, this is the mgmt ip 16:48:55 and so forth 16:48:59 mike_m: I see, you are proposing an Tacker VNFM interfacing with an "external 16:49:09 ... "external" EM 16:49:16 yes exactly. 16:49:27 scaling decisions should also be allowed by this external EM 16:49:55 just wanted to make sure that EM is Element Manager, right? 16:50:02 I see this interface becoming an East/West API 16:50:16 mike_m: mgmt_driver framework that we have served this purpose to some extent.. announce a VNF is instantiated to the ODL controller to do "something" 16:50:16 thats right, Element Manager as per ETSI NFV 16:51:17 yes, have been looking at the mgmt_driver, so that it could do something like that, but i think it needs additional APIs. 16:51:18 mike_m: here is my suggestion.. i think we could use a small google doc writeup (with some simple diagrams) to illusrtate this further.. is that something you can take up ? 16:51:45 sridhar_ram: sure I can do that 16:52:16 mike_m: thanks.. as we wind down newton release, we can use more mtg time to dig this further.. 16:52:43 sridhar_ram: thanks for the time to explain. 16:53:01 diga: noted, good to see API framework spec starting.. 16:53:41 I'd like to step back a bit and ask the team to think the bigger picture for Ocata... 16:54:09 so far we have kind of went with the flow of a roadmap and picked the top things on the list and executed.. 16:54:58 Going forward it will good to have a specific "theme" for an upcoming release and have the whole team converge and "make it happen" instead of the community spreading thin on 10 to 15 different things 16:55:16 Thoughts ? 16:55:43 +1 16:55:47 +1 16:56:04 * sridhar_ram notes 4mins left 16:56:19 Here is my suggestion along those lines .. 16:57:01 bucket-1) over-arching theme for Ocata - enable an healthy eco-system of infra drivers - vmware, aws, kubenetes 16:57:20 bucket-2) wrap up left overs - VNFC, NSD, TOSCA CSAR 16:57:41 bucket-3) project maintenance - API framework, OSC, deprecation 16:58:08 Please think in these term / buckets 16:58:20 sridhar_ram: +1 to the bucket slots 16:58:27 there could be other buckets too.. but we can't have too many as we will lose focus 16:58:55 I'll put Ocata as the top agenda for next meeting... 16:59:23 please come w/ your thoughts so that we can start making the final "goals" for Tacker Ocata release 16:59:24 sridhar_ram: +1 - nice plan 16:59:39 sripriya: diga: thanks! 16:59:46 we are out of time today.. 17:00:00 talk to you all next week .. specially on Ocata! 17:00:06 *specifically 17:00:15 bye everyone 17:00:18 #endmeeting