04:31:32 <gongysh> #startmeeting tacker
04:31:33 <openstack> Meeting started Wed Jun  7 04:31:32 2017 UTC and is due to finish in 60 minutes.  The chair is gongysh. Information about MeetBot at http://wiki.debian.org/MeetBot.
04:31:34 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
04:31:36 <openstack> The meeting name has been set to 'tacker'
04:31:40 <gongysh> hi
04:31:57 <janki> Hi gongysh
04:32:31 <gongysh> #topic roll call
04:32:44 <gongysh> janki, welcome
04:32:53 <YanXing_an> hello
04:34:20 <tung_doan> hi all
04:34:24 <janki> gongysh, thanks :)
04:34:50 <YanXing_an> ping tung_doan
04:34:51 <trinaths> o/
04:35:07 <tung_doan> YanXing_an: Hi there
04:36:20 <gongysh> we are toward deadline of pike-2
04:36:44 <gongysh> hope some core will join us later
04:37:00 <gongysh> #topic bp
04:37:13 <gongysh> tung_doan, let talk about your bp first
04:37:20 <gongysh> https://review.openstack.org/#/c/434974/10/specs/pike/vnffg-autohealing.rst
04:37:38 <tung_doan> gongysh: I am trying to wrap this bo
04:37:47 <gongysh> today, I want to do meeting reviewing
04:37:49 <tung_doan> gongysh: only one concern is abt API
04:37:59 <YanXing_an> when is deadline? not 2017-6-5?
04:38:03 <gongysh> so please go to this bp review page.
04:38:28 <tung_doan> I need Tacker guys about API in my bp
04:38:30 <gongysh> first, what is vnffg-ha meaning?
04:38:30 <gongysh> tung_doan, hi
04:39:03 <tung_doan> gongysh: vnffg-ha is meaning high availability for VNFFG
04:39:20 <tung_doan> it includes both scaling and autohealing support
04:39:37 <gongysh> ha in general means there are two or more for this kind of stuff.
04:39:48 <gongysh> do we have two vnffg to do ha?
04:40:06 <tung_doan> gongysh: exactly, i just did not have other names for it
04:40:39 <gongysh> so we will have two vnffg if we run 'tacker vnffg-create xxx'
04:40:41 <gongysh> ?
04:41:15 <tung_doan> gongysh: in case of scaling, we have 2 paths with the same vnffg
04:41:58 <YanXing_an> gongysh, IMO, we do not have two vnffg to do ha
04:42:07 <gongysh> scaling just scaling member vnf's vdu, right?
04:42:08 <tung_doan> YanXing_an: +1
04:42:13 <diga> Hi
04:42:20 <gongysh> diga, welcome
04:42:30 <diga> gongysh: Hi
04:42:31 <tung_doan> gongysh: it looks like a different use case
04:42:35 <gongysh> diga, we are talking about https://review.openstack.org/#/c/434974/10/specs/pike/vnffg-autohealing.rst
04:42:47 <diga> gongysh: okay, I will go through it
04:43:14 <tung_doan> gongysh: pls look at this diagram: https://review.openstack.org/#/c/469975/2/specs/pike/vnffg-scaling.rst@73
04:43:15 <gongysh> tung_doan, vnffg-ha is the first question.
04:44:04 <tung_doan> gongysh: we still have the same vnffg but we have a group of VNF for scaling case
04:44:20 <YanXing_an> maybe vnffg-supports-vnf-ha is a better name
04:44:20 <gongysh> tung_doan, so it is not vnffg ha itself.
04:44:33 <gongysh> we need a different name
04:44:59 <tung_doan> gongysh: ok. pls review and give me your suggestion. thanks
04:45:30 <gongysh> second question is: why do we need a url?
04:45:44 <gongysh> **POST on  /v1.0/vnffghas/<vnf-id>/<monitoring-policy>/<action-name>/<params>**
04:45:57 <tung_doan> gongysh: when vnf is respawned, it is just in VNFM level
04:46:34 <tung_doan> gongysh: vnffg is in NFVO level, we need to specify which vnffg VNF is belong to
04:46:53 <tung_doan> gongysh: that's why we need to have API in NFVO
04:48:50 <tung_doan> gongysh: Is it clear to you?
04:49:00 <gongysh> tung_doan, it seems reasonable.
04:49:17 <YanXing_an> tung_doan, in policy action, we call the rest api to notify NFVO?
04:49:24 <tung_doan> gongysh: thanks.
04:49:35 <gongysh> it is based on nfvo and vnfm separation.
04:49:36 <tung_doan> YanXing_an: correct!
04:50:33 <tung_doan> YanXing_an: one vnf can belong to several vnffgs or NSs. We need to have api to notify them
04:51:30 <gongysh> tung_doan, ok make sense.
04:51:39 <YanXing_an> tung_doan, it's nice method, i think.
04:52:06 <tung_doan> gongysh: YanXing_an: happy to know that :)
04:52:30 <gongysh> trinaths, please have a look at this spec too
04:52:30 <gongysh> diga, you too.
04:52:39 <trinaths> gongysh: sure.
04:52:44 <trinaths> gongysh:  looking now
04:52:47 <diga> gongysh: yes, going through it
04:53:07 <gongysh> I will try to merge the spec with your review, even without core's score.
04:53:10 <trinaths> I believe vnfs and vsfs must be sharable
04:53:42 <gongysh> tung_doan, https://review.openstack.org/#/c/469975/
04:54:03 <gongysh> please work on it too.
04:54:18 <gongysh> let's get them merged as soon as possbile.
04:54:20 <tung_doan> gongysh: work in parallel with vnffg-healing
04:54:32 <tung_doan> gongysh: got it
04:54:53 <gongysh> janki, hi
04:56:20 <gongysh> it seems janki  is not here.
04:56:48 <gongysh> next bp is from me:
04:56:48 <gongysh> https://review.openstack.org/#/c/467479/
04:57:11 <gongysh> also please have a look at it.
04:57:38 <trinaths> sure
04:57:42 <gongysh> this is for bp
04:58:00 <gongysh> do you guys have other bp hanging?
04:58:32 <tung_doan> gongysh: your bp should be prior. will look at this
04:58:50 <gongysh> YanXing_an, yes the pike2 is 6-5.
04:59:07 <trinaths> when is pike-3 ?
04:59:28 <janki> gongysh, hey
04:59:34 <gongysh> trinaths, I will move all unmerged stuff into pike-3
04:59:45 <trinaths> ok
04:59:58 <gongysh> janki, do you have anything to talk about tacker bp?
05:00:19 <YanXing_an> p3 is 7-24
05:00:29 <janki> gongysh, yes. I did some research and looks like magnum won't be of use to us
05:00:37 <janki> we can use zun to create containers
05:00:45 <janki> but zun doesnot create docker image
05:01:03 <trinaths> janki: the work is related to cVNF management ?
05:01:08 <janki> so I need to figure out a way to create a docker image from tosca template
05:01:16 <janki> trinaths, what do you mean by cvnf?
05:01:18 <gongysh> janki, nice to hear that. you can have a demo implementation.
05:01:19 <tung_doan> janki: why magnum cannot be a candidate?
05:01:26 <trinaths> containerized vnf
05:01:51 <janki> tung_doan, magnum proivdes COEs. and not a way to create container
05:01:56 <janki> trinaths, yes. it is for cVNF
05:02:11 <trinaths> janki: okay.
05:02:22 <gongysh> next topic
05:02:26 <janki> if anybody has any idea about converting tosca template to docker image, please reach out to me
05:02:59 <tung_doan> janki: i think any update from your bp will be good to see :))
05:03:01 <gongysh> janki, what do you mean by 'convert tosca temmplate into docker image'?
05:03:18 <diga> janki: make sure, we should provision containers on COE
05:03:40 <janki> gongysh, from a VNFD we need a docker image
05:03:51 <diga> in zun we provide, magnum support as COE & on top of that, we can create containers
05:03:52 <janki> tung_doan, I will update once I have some understanding on this part
05:04:02 <tung_doan> janki: +1
05:04:16 <gongysh> janki, go on the research.
05:04:18 <gongysh> thanks
05:04:29 <trinaths> janki: can you elaborate on VNFD to DI ?
05:04:30 <gongysh> #topic codes and bug
05:04:59 <janki> trinaths, lets take this discussion furhter on #tacker after this meeting
05:05:07 <gongysh> YanXing_an, https://bugs.launchpad.net/tacker/+bug/1667905
05:05:08 <openstack> Launchpad bug 1667905 in tacker "can't delete vnffg if it get stuck at PENDING_DELETE state" [Medium,In progress] - Assigned to sajuptpm (sajuptpm)
05:05:09 <YanXing_an> janki, does it mean, launch a container from vnfd, to replace kvm VM?
05:05:09 <trinaths> janki: okay
05:05:20 <janki> YanXing_an, yes
05:05:31 <trinaths> YanXing_an: yes
05:06:09 <gongysh> https://review.openstack.org/#/c/469304/
05:06:18 <YanXing_an> gongysh, have no much time on it this week
05:06:48 <gongysh> YanXing_an, but I have seen you had a state transition diagram draw
05:06:59 <trinaths> YanXing_an: fixing that bug stated above ?
05:07:46 <YanXing_an> yes, it can help to fixing this bug
05:08:15 <trinaths> YanXing_an: i suggest to own the bug on your name. its will be good for tracking.
05:09:17 <gongysh> next code patch:
05:09:18 <gongysh> https://review.openstack.org/#/c/465080/
05:09:19 <trinaths> ther are x`https://bugs.launchpad.net/tacker/+bug/1667905 https://bugs.launchpad.net/tacker/+bug/1689426
05:09:20 <openstack> Launchpad bug 1667905 in tacker "can't delete vnffg if it get stuck at PENDING_DELETE state" [Medium,In progress] - Assigned to sajuptpm (sajuptpm)
05:09:21 <openstack> Launchpad bug 1689426 in tacker "VNF in PENDING_DELETE status can not be deleted" [Undecided,In progress] - Assigned to Bao Fangyan (baofangyan)
05:09:26 <trinaths> duplocate ?
05:10:15 <YanXing_an> vnffg and vnf both have this bug
05:10:35 <trinaths> YanXing_an: okay.
05:10:38 <trinaths> i was confused
05:11:02 <gongysh> https://review.openstack.org/#/c/465080/ is almost done
05:11:03 <YanXing_an> we can fix it together
05:11:15 <trinaths> YanXing_an: good
05:11:22 <gongysh> I have tested it. and just one nit left over.
05:11:38 <gongysh> once it is done and get CI passed, I will +2
05:11:52 <gongysh> hope you guys have a review and test too.
05:12:12 <gongysh> YanXing_an,  well done
05:12:23 <YanXing_an> gongysh, thanks
05:12:41 <gongysh> do you guys have other code patch or bugs to talk about?
05:13:09 <janki> gongysh, I want to talk about integration with ooo
05:13:22 <gongysh> ooo?
05:13:27 <tung_doan> triple )
05:13:31 <janki> sorry, Tripleo
05:14:00 <gongysh> so what will we need to do for it?
05:14:08 <tung_doan> janki: I heard that tacker already integrated in OOO, right?
05:14:25 <janki> gongysh, tacker is integrated in tripleo.
05:14:43 <janki> now all services are getting containerised in TripleO
05:14:46 <janki> tung_doan, yes
05:15:01 <janki> gongysh, we need to containerise Tacker in TripleO
05:15:20 <janki> tung_doan, it is integration but disabled by default
05:15:21 <gongysh> janki, yes, that is awesome.
05:15:30 <gongysh> janki, do you want to do it?
05:15:32 <janki> gongysh, I can take that up
05:15:45 <tung_doan> janki: nice
05:15:47 <gongysh> janki, fine.
05:16:01 <YanXing_an> janki, it's good to deploy tacker in container for me
05:16:13 <gongysh> if you think it needs a spec bp, you can do one.
05:16:23 <janki> gongysh, i will open a bug about this on tripleo
05:16:30 <gongysh> or just start a launchpad to do it.
05:16:35 <tung_doan> janki: +1
05:16:42 <janki> gongysh, no spec needed. just a bug. the code will go in tripleo and kolla
05:17:00 <janki> I odnt think there would be any Tacker code change
05:17:03 <trinaths> in opnfv-models setup, tacker is containerized.
05:17:14 <gongysh> janki, kolla already has it.
05:17:19 <janki> trinaths, can you point the url?
05:17:32 <janki> gongysh, ohh, then it is just TRipleo changes we need
05:17:49 <janki> gongysh, let me reconfirm it on tripleo front too
05:18:08 <gongysh> janki, put a bug on tacker and write a install doc in tacker
05:18:23 <YanXing_an> I have tried to launch tacker in contrainer several months ago, but failed, no time to fix it
05:18:25 <diga> First verify, what support we have in tripleo then go ahead and file the bug
05:18:30 <gongysh> we need guys to know how to install it using kolla.
05:18:58 <janki> gongysh, the bug will go in Tripleo launch pad. there won't be any doc needed as this will be part of Tripleo installation
05:19:19 <janki> and the doc should already be in tripleo repo
05:19:24 <diga> gongysh: I know kolla, will take a look at it
05:19:27 <trinaths> janki: https://wiki.opnfv.org/display/models/Testing
05:19:33 <janki> trinaths, thanks
05:20:10 <gongysh> janki, diga you can change the tacker's installation doc about kolla or tripleo installation.
05:20:15 <gongysh> diga, thanks
05:20:38 <diga> gongysh: sure
05:21:02 <gongysh> guys,  taker installation should not need nova, cinder components.
05:21:19 <gongysh> in devstack installation, we name it as standalone installation.
05:22:00 <gongysh> so the kolla or tripleo way to install tacker should be the 'standalone' way.
05:22:08 <diga> okay
05:22:22 <gongysh> #topic open discussion
05:23:00 <gongysh> any other stuff to talk?
05:23:39 <gongysh> if not we can end the meeting earlier
05:23:55 <tung_doan> gongysh: I am the way to harmonize scaling code between Tacker and heat-translator. pls have a look
05:24:12 <tung_doan> https://review.openstack.org/#/c/471049/
05:24:30 <tung_doan> https://review.openstack.org/#/c/462690/
05:24:52 <gongysh> tung_doan, I was trying, but it is not easy to do the review since I don't know what is the point.
05:25:13 <tung_doan> ok.. i really need Kanagaraj and tbh
05:25:25 <tung_doan> but seems they are not here. will ping them latter
05:25:42 <gongysh> trinaths, heard opnfv meeting is on near.
05:25:54 <gongysh> do you want to share?
05:26:00 <gongysh> trinaths, hi
05:26:29 <trinaths> gongysh: I have 3 presentations planned 14 and 15th  jue
05:26:31 <trinaths> june
05:26:39 <gongysh> in beijing?
05:26:43 <trinaths> presentation stuff got approved yesterday.
05:26:46 <trinaths> gongysh: yes.
05:27:00 <janki> gongysh, what is our plan for PTG?
05:27:05 <gongysh> ok, but I have no ticket there.
05:27:06 <trinaths> gongysh: JW marriot and Ritz Carlton  hotel
05:27:30 <gongysh> trinaths, I have to get a ticker first.
05:27:46 <gongysh> janki, do you want to organize the PTG this time?
05:28:06 <janki> gongysh, I am not sure whether I would be travelling or not
05:28:12 <janki> so can't commit
05:28:13 <trinaths> i have got 20 minutes of talk
05:28:15 <trinaths> gongysh: https://events.linuxfoundation.org/events/opnfv-summit/attend/registration
05:28:30 <gongysh> janki, I think we will doing the virtual PTG again.
05:28:43 <janki> gongysh, sounds good :)
05:29:08 <trinaths> gongysh: 'idea nest' is a informal talk for 20 minutes.
05:29:23 <gongysh> time is up. lets go to tacker channel
05:29:26 <trinaths> gongysh: I want to propose tacker - ongoing and future work.
05:29:44 <tung_doan> trinaths: nice
05:29:44 <gongysh> trinaths, sure, that's nice.
05:29:52 <gongysh> #endmeeting