04:31:14 #startmeeting tacker 04:31:15 Meeting started Wed Dec 6 04:31:14 2017 UTC and is due to finish in 60 minutes. The chair is gongysh. Information about MeetBot at http://wiki.debian.org/MeetBot. 04:31:17 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 04:31:19 The meeting name has been set to 'tacker' 04:31:43 Hi all 04:31:59 phuoc__, hi 04:32:12 hi gongysh 04:32:35 Hello :) 04:32:50 MinWookKim, hi 04:32:57 yamahata, helo 04:33:00 gongysh, Hello 04:33:23 hello 04:33:34 o/ 04:33:52 o/ 04:34:42 o/ 04:34:50 helo, all 04:35:01 o/ gongysh 04:35:27 #topic commit stats 04:35:40 http://stackalytics.com/?metric=commits 04:35:58 our tacker ranks at 79th in commit metric 04:36:38 we are lagging too mcuh behind. 04:37:08 I hope can put some commits in Queens version 04:37:38 56 tacker 198 in reviwe 04:38:00 its 91 rank 04:38:44 diga, your are right. 04:38:59 so we need put more efforts. 04:39:23 gongysh: there are many reasons for this 04:40:08 lets look at the review data in our tacker projects. 04:40:09 http://stackalytics.com/?metric=marks&module=tacker-group 04:40:30 end of 2017? 04:40:59 all the data is bad for tacker projects 04:41:36 so what do you guys have ideas to improve these data? 04:42:18 I think we can fix some errors, or add new features 04:42:34 phuoc__, you are nice to have many contribution. 04:43:26 main problem is, nobody is actively working on tacker except phuoc__ 04:43:27 diga, any idea? 04:43:56 I am owning some stuff and have planned weekly target for tacker for my org 04:44:01 https://bugs.launchpad.net/tacker/+bugs?field.tag=low-hanging-fruit 04:44:06 diga, you are right 04:44:14 1 bug is resolved and 3 bugs are in progress 04:44:26 will finish those by end of this week 04:44:52 API-framework base patch should get merged, because its a base template 04:45:20 gongysh: need your view on that 04:45:35 diga , any way, queens focus is k8s vim and vnfs 04:46:01 gongysh: I think every organisation should commit some time for tacker 04:46:03 diga, fill my name on the patch. 04:46:09 gongysh: ok 04:46:26 gongysh, diga, YanXingAn can you guys look at https://review.openstack.org/#/c/516698/ 04:46:27 gongysh: https://review.openstack.org/#/c/520495/ 04:46:48 I think it looks good to be merged this spec 04:47:27 #topic annoucement 04:47:38 queens-2 milestone is close to deadline 04:47:57 I will move all the tasks queued to queens-3 04:48:12 ok 04:48:41 we have finished k8s vim now. in queens-3, lets focus on c-vnf impelmentation. 04:49:04 so that we have something to be announced. 04:49:37 of course the vnffg bug and enhancement is also welcom 04:49:38 https://review.openstack.org/#/c/525089 04:49:49 #topic bp 04:50:33 I found a bug in VNFFG operation while I was trying to deploy VNFFG 04:50:35 https://review.openstack.org/#/c/525089 This link is about how to define VNFC in TOSCA parser 04:50:57 phuoc__, could you please take your time to discuss the c-vnf bp? 04:51:57 first, I bring to tosca-parser project to define VNFC 04:52:10 VNFC will be a part to define c-vnf in Tacker 04:52:32 I will update c-vnf spec to describe this in detail 04:52:53 phuoc__: so many things are pending on tosca side, my k8s setup is not working 04:53:18 phuoc__: will some patches this week if everything goes well 04:54:43 phuoc__, https://review.openstack.org/#/c/511991/10/specs/queens/kubernetes-type-for-containerized-VNF.rst 04:54:57 there is no vnfc at all there? 04:55:31 I will update this later 04:55:58 phuoc__, thanks 04:55:59 phuoc__, we have no much time 04:56:03 another about https://review.openstack.org/#/c/516698/ 04:56:18 mardim will start coding when this patch is merged 04:56:24 we had better have a helo vnf implemented in c-vnf format. 04:56:49 gongysh, sure 04:57:11 I will describe it in template 04:57:35 next bp is zabbix 04:57:43 MinWookKim, hi 04:57:53 gongysh, hello 04:58:01 I have a comment on it about the token 04:58:18 it will be a design fault 04:58:38 gongysh, I'm thinking of a solution to Token comments. 04:58:49 gongysh, But it seems to be a difficult problem. 04:59:19 gongysh, Can I ask you a question? 04:59:28 MinWookKim, I have question: what is trigger in zabbix? 05:00:15 is it constructed by tacker into zabbix when defining the monitor? 05:00:16 gongysh, trigger is a notification that occurs according to a specified condition. 05:00:58 gongysh, Yes, that's right. 05:00:59 and it will be triggered in zabbix server after a condition matched, right? 05:01:08 so the token will be there a long time. 05:01:40 gongysh, yes right. 05:03:19 MinWookKim, you can start with implementation like ceilometer alarm driver. 05:03:43 gongysh, do you consider this issue for security? 05:04:01 although it is also a security risk. 05:04:09 however, Zabbix is not a openstack project, so it has a limitation. 05:05:03 Hyunchic, I am considering how zabbix called into tacker to trigger a action, such as respawn and scall 05:05:25 gongysh, In fact what do you think putting aodh between zabbix and tacker? 05:06:03 guys, I am leaving now, have some urgent work to do 05:06:06 see you guys later 05:06:16 diga, thanks 05:06:23 diga, thanks 05:06:23 I will put API base patch in. 05:06:53 MinWookKim, what is your idea? 05:07:02 our ceilometer driver is using aodh 05:07:23 aodh has its own monitoring and trigger mechanism 05:08:10 it overlapping with zabbix server. 05:09:04 gongysh,you're right. I think we'll need a little more investigation. 05:09:09 MinWookKim, you can first merge with app action, no need to call into tacker 05:09:31 and then we can consider how to support scale, respawn action. 05:09:43 yes it's better. 05:09:55 gongysh, Okay. I'll make that BP edit. 05:10:21 MinWookKim, ok, thanks. break it into pieces and finish one by one. thanks 05:10:32 #topic open discussion 05:10:39 gongysh, thanks a lot 05:10:53 anything do you guys want to talk? 05:11:09 long time ago, 05:11:14 gongysh, are we going to PTG? 05:11:28 i updated my code for port monitoring. 05:11:40 janki, I will not 05:11:54 so virtual PTG then 05:11:59 if you guys want, you can have a meeting there. 05:12:24 For update vnffg:tacker vnffg-update --vnffgd-file vnffgd-template.yml 05:12:41 In Sydney, we got a Use cases from NTT for HA feature over multi-compute nodes 05:12:44 gongysh, do you have any comment for that command 05:12:57 Gongysh: Do you remember about it? 05:13:11 longkb, yes 05:13:57 https://review.openstack.org/#/c/516698/ phuoc__ 05:14:01 this bp? 05:14:19 gongysh, yes 05:14:50 ok, I will review after this meeting 05:15:27 As I can see from Tacker features, Tacker only support to deploy VNF over VIMs from CLI. In the case of vendors want to deploy VNF over compute-node, they have to define it in VNFD, right? 05:16:23 I think we shoud enable VNF creation over availability zone from Tacker client without az inside VNFD 05:16:57 longkb, you are talking about openstack nova command 05:17:25 No, I mentioned about Tacker command 05:18:00 longkb, but you can implement it via tacker cli too. define argument in vnfd and generate vnf with parameter. 05:18:41 Ok, I will make a path about it :) 05:19:01 by the way, I am voting for individual director, if you guys have fundation member ship, please support me. 05:19:03 https://www.openstack.org/community/members/profile/28514/yongsheng-gong 05:19:49 gongysh, sure :) 05:20:01 phuoc__, thanks 05:20:13 ok if nothing else, we can end this meeting. 05:20:20 #endmeeting