15:59:58 <inc0> #startmeeting kolla 15:59:59 <openstack> Meeting started Wed Jun 14 15:59:58 2017 UTC and is due to finish in 60 minutes. The chair is inc0. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:00 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:04 <openstack> The meeting name has been set to 'kolla' 16:00:12 <inc0> #topic rollcall, w00t! 16:00:22 <Jeffrey4l> w00t 16:00:22 <inc0> a bit WOOT today please 16:00:29 <pbourke> WOOT 16:00:34 <duonghq> w00t 16:00:50 <egonzalez> w00t! 16:01:34 <spsurya_> w00t! 16:01:48 <rwellum> WOOT 16:02:48 <jascott1> woot 16:02:57 <krtaylor> 0/ aka woot 16:03:46 <mandre> w00t! 16:04:05 <zhubingbing__> w00t 16:04:21 <zhubingbing__> WOOT 16:04:40 <inc0> that's the spirit:) let's get with meeting 16:04:46 <inc0> #topic Announcements 16:05:06 <inc0> I started voting for spsurya_ core team addition, cores, please cast your votes 16:05:51 <inc0> I will start separate vote for Surya to add him to kolla-kubernetes core team too 16:06:10 <inc0> but since these core teams are separate, consider this a separate vote 16:06:25 <inc0> another announcement 16:06:38 <inc0> We released 5.0.0b2, so images are now tagged 5.0 in master 16:06:48 <inc0> Jeffrey4l: how are we looking in terms of 4.0.2? 16:07:13 <Jeffrey4l> need collect info today. if no other issue, should be release today 16:07:43 <inc0> cool, if there are any outstanding bugs left, let us know. And everyone, please prioritize these 16:07:54 <inc0> ok, that's it from me, any community announcements? 16:08:57 <inc0> guess not 16:09:00 <inc0> let's get to agenda 16:09:12 <inc0> #topic Bump min Ansible version to 2.2.0 (duonghq) https://review.openstack.org/#/c/452067/ 16:09:17 <inc0> duonghq: you have the floor 16:09:23 <duonghq> roger 16:09:40 <duonghq> As you can see, the ps is merged (for a few hours) 16:09:53 <duonghq> so we should start working on remove all warning of Ansible 16:10:02 <inc0> good point 16:10:10 <inc0> lots of thigns will break on 2.4 16:10:29 <duonghq> the original intend of the topic is I want to know if we want to bump the Ansible to 2.2.0, but the ps is merged, so the point is as above 16:10:41 <inc0> duonghq: can you create bp for that? and let's start gathering warnings that needs fixing 16:10:49 <duonghq> inc0, sure 16:10:52 <inc0> thanks 16:11:06 <inc0> we need to distribute this work 16:11:22 <inc0> but first we need to figure out what types of changes need to be made 16:11:30 <inc0> as in "remove all when: {{..." 16:11:41 <inc0> then we can distribute work across roles 16:11:41 <duonghq> #link https://launchpad.net/bugs/1695142 16:11:43 <openstack> Launchpad bug 1695142 in kolla-ansible "In Ansible 2.3, when statement should not contain delimiters" [Medium,In progress] - Assigned to Duong Ha-Quang (duonghq) 16:11:54 <Jeffrey4l> gate is a good place to collect all warnings. 16:12:38 <duonghq> Jeffrey4l, yes, I'll collect warning in tomorrow 16:12:43 <inc0> thanks 16:12:45 <Jeffrey4l> cool. thanks. 16:13:05 <Jeffrey4l> btw, this patch need to be merge in ocata. 16:13:08 <Jeffrey4l> https://review.openstack.org/473206 16:13:21 <Jeffrey4l> which is break by ansible 2.3. 16:13:37 <inc0> it's merged 16:13:39 <Jeffrey4l> now it is block by its parent patchq https://review.openstack.org/473205 :) 16:13:42 <duonghq> hmm, egonzalez did w+1, but why is not merged? 16:13:51 <inc0> ah ok 16:13:55 <Jeffrey4l> parent patch ;) 16:14:08 <inc0> I'll review this after meeting, thanks Jeffrey4l 16:14:16 <Jeffrey4l> thanks. 16:14:41 <inc0> anything else on the topic duonghq ? 16:14:56 <duonghq> inc0, no, thanks 16:15:01 <inc0> cool, moving on 16:15:04 <inc0> #topic remove kolla-k8s related code (jeffrey4l) 16:15:08 <inc0> Jeffrey4l: go ahead 16:15:19 <Jeffrey4l> hrm, actually this is mainly for you inc0 ;) 16:15:29 <inc0> lol 16:15:36 <inc0> yeah, I guess 16:15:36 <Jeffrey4l> before kolla-k8s is depends on kolla-ansible genconfig tools 16:15:45 <inc0> now it doesn't I know;) 16:15:58 <Jeffrey4l> but now, seem you have implement configuration generate in kolla-k8s' side. 16:16:04 <inc0> I was thinking of doing this but it slipped my mind :( 16:16:19 <Jeffrey4l> and i have confirmed with kolla-k8s guys. 16:16:33 <inc0> anyway, yes, we can get rid of k8s code from kolla-ansible conf 16:17:37 <inc0> so, if anyone feels like helping with that, it would be appreciated:) 16:18:13 <spsurya_> inc0: sure 16:18:13 <Jeffrey4l> seems we can move on? 16:18:38 <inc0> yeah 16:18:49 <inc0> #topic bug triaged this week (spsurya) 16:18:55 <inc0> go ahead Surya 16:19:02 <spsurya_> yes 16:19:25 <inc0> #link https://etherpad.openstack.org/p/kolla-bug-triage-week 16:19:28 <spsurya_> so this week i was triaging these bugs 16:19:29 <inc0> let's gather all bugs here 16:19:36 <spsurya_> https://www.irccloud.com/pastebin/l9sp8WM3/ 16:20:06 <spsurya_> inc0: you can find here too 16:20:19 <spsurya_> some of them are in progress 16:20:23 <inc0> cool, I copied it to ehterpad so we can have one list 16:20:27 <spsurya_> for reporter response 16:20:41 <Jeffrey4l> cool. it will be more clearly for kolla-ansible. 16:20:44 <Jeffrey4l> btw, here is the kolla-k8s' implementation link https://github.com/openstack/kolla-kubernetes/tree/master/ansible 16:20:44 <spsurya_> one of them i found invalid 16:21:12 <spsurya_> I will be trying more to triage from upcoming weeks 16:21:15 <Jeffrey4l> ( network issue, ignore mine ;( ) 16:21:17 <inc0> spsurya_: can you please use etherpad and list out bugs which are ready to fix? 16:21:33 <inc0> that would be super useful for anyone who looks for contribution opportunity 16:21:38 <spsurya_> that is all from my side 16:21:46 <spsurya_> inc0: sure 16:22:15 <inc0> and if someone ever asks for "how to contribute" or we ourselves want to do something with our free time, it would be good list to take from 16:22:56 <spsurya_> inc0: you may continue 16:23:01 <inc0> that's it 16:23:05 <inc0> #topic open discussion 16:23:11 <spsurya_> I will keep adding in the same link 16:23:13 <inc0> anyone wants to share anything?:) 16:23:18 <inc0> great Surya, thank you 16:23:20 <Jeffrey4l> i have one. 16:23:21 <spsurya_> which are ready to fix 16:23:22 <Jeffrey4l> about https://review.openstack.org/473168 16:23:34 <Jeffrey4l> about 16:23:35 <Jeffrey4l> Flatten neutron_* variables 16:23:57 <vhosakot> o/ (sorry for the delay.. I'm one person attending 3 meetings ;)) 16:23:58 <Jeffrey4l> we have already move all container variable into default/main.yml file. 16:24:28 <Jeffrey4l> but it is a big dict now. and it is hard for end-user to do some customization. 16:24:46 <Jeffrey4l> for example. add a environment variable into neutron_server container. 16:24:52 <inc0> yeah, maybe we should do that across the board 16:24:55 <inc0> it's good idea. 16:25:10 <Jeffrey4l> flatten the neutron_servers will be helpful for this kinda case. 16:25:13 <zhubingbing__> +1 16:25:34 <inc0> but if we do it here, we need to do it everywhere 16:25:40 <Jeffrey4l> yep. 16:25:46 <inc0> so customization logic will be consistent 16:26:10 <duonghq> +1 16:26:19 <Jeffrey4l> this is a example modificaiton. need change other roles if this change is necessary. 16:26:46 <Jeffrey4l> another related thought is: move config.json into default/main.yml, then end-user could add whatever config file into container by simple change a dict. 16:27:40 <inc0> that's gonna be harder 16:27:47 <inc0> well 16:27:54 <Jeffrey4l> i.e. config.json is render base on default/main.yml file. 16:27:55 <inc0> ansible still needs to render .json 16:28:16 <inc0> we need to keep ABI consistent, but yes, that's doable 16:28:24 <Jeffrey4l> yes. config.json is still required. but it is render base on default/main.yml configration. 16:28:29 <inc0> we can have one template for config.json and render it all from vars 16:28:42 <inc0> yeah, +2 from me 16:28:43 <Jeffrey4l> yes. 16:28:47 <Jeffrey4l> thanks. 16:29:05 <duonghq> should we have one bp about this? 16:29:19 <inc0> if we're doing this, yes 16:29:23 <Jeffrey4l> i think so. 16:29:24 <inc0> to organize work 16:29:30 <inc0> and add work items for all the roles 16:29:36 <inc0> so we can distribute 16:29:38 <Jeffrey4l> we have many roles now ;( 16:29:47 <inc0> yeah 16:29:54 <duonghq> I think we can categorize by alphabetic order 16:30:05 <inc0> that's why distribution of work is crucial, and for that we need to keep track what's done what is not 16:30:12 <duonghq> and distribute task by range 16:30:22 <Jeffrey4l> yes. use work item to track each role. 16:30:31 <inc0> or etherpad 16:30:33 <duonghq> as daidv did in his bp 16:30:52 <Jeffrey4l> work for me. 16:30:56 <inc0> cool 16:31:13 <inc0> anything else Jeffrey4l ? 16:31:25 <Jeffrey4l> i have another new about hub.docker.com. 16:31:34 <inc0> one in China?:) 16:31:52 <Jeffrey4l> seem docker company are trying work with aliyun ( one china company ) to provide docker iamges service. 16:31:53 <Jeffrey4l> yep. 16:32:06 <inc0> cool 16:32:06 <Jeffrey4l> but i searched news, seem it is not ready now. 16:32:14 <inc0> oh I see 16:32:27 <inc0> well, we will have publisher code hopefully soon 16:32:28 <spsurya_> inc0: update the triaged bugs link https://etherpad.openstack.org/p/kolla-bug-triage-week one is ready to fix, two of them are under review and one of the bug is invalid 16:32:34 <Jeffrey4l> anyway, once it is ready. we can push image into it . 16:32:35 <inc0> it would be good to have Chinese mirror somewhere 16:32:40 <Jeffrey4l> yes. 16:33:04 <inc0> I have one topic too 16:33:06 <inc0> https://review.openstack.org/#/c/466815/ 16:33:15 <inc0> so last few weeks I was working with gates 16:33:30 <inc0> just so you see, adding new gate scenerio will be super easy after this merges 16:33:40 <inc0> let say we want to add designate to scenerio 16:33:53 <Jeffrey4l> i checked the logs. seems the nova-compute in subnodes report errors. 16:33:57 <berendt> o/ 16:34:20 <inc0> you create gate-kolla-ansible-dsvm-deploy-designate-centos-binary-centos-7-2-node-nv job in infra 16:34:30 <inc0> (notice deploy-designate) 16:34:51 <inc0> then $ACTION in gates will have value "deploy-designate" 16:35:25 <inc0> and you can add change to config template https://review.openstack.org/#/c/466815/13/tests/templates/globals-default.j2 like this 16:35:53 <inc0> {% if 'designate' in action %}enable_designate: "yes"{% endif %} 16:35:56 <inc0> that's it 16:36:10 <Jeffrey4l> is it possible to use "deploy-heat-magnum-senlin" in jobs name? 16:36:26 <inc0> yeah 16:36:36 <Jeffrey4l> cool. 16:37:22 <duonghq> nice 16:37:23 <inc0> it's gonna need infra patch and that's kinda tricky to find all required places to add this 16:37:26 <Jeffrey4l> aha, you even didn't change project-config. nice idea. 16:37:26 <inc0> but I can help with that 16:37:36 <inc0> no, you need, project-config 16:37:51 <Jeffrey4l> i mean the magic ACTION. 16:39:11 <inc0> https://github.com/openstack-infra/project-config/blob/master/jenkins/jobs/projects.yaml#L6370 you'll need to add another option for this action 16:39:36 <Jeffrey4l> yes 16:39:48 <inc0> but that's arguably hardest part of adding new scenerio 16:40:10 <inc0> and if you do it this way, as in heat-magnum-senlin 16:40:30 <inc0> you can modify globals in a way that if heat in action: enable heatr 16:40:48 <inc0> so if we ever add heat-ceilometer-aodh, that will just work 16:41:23 <inc0> but let's be prudent with this, every new gate consumes resources and makes contribution harder 16:41:32 <Jeffrey4l> should we use some test scenario like puppet does? like scenario 1 for heat related , and scenario for ceph related. 16:41:50 <inc0> so let's focus on covering common use cases, we can have slew of experimental gates for more...exotic ones 16:42:01 <inc0> yeah, that was my thinking 16:42:10 <Jeffrey4l> then we may use 2-3 scenrio to cover all roles. 16:42:12 <Jeffrey4l> coo. 16:42:14 <duonghq> It's better if we can gate only related-change 16:42:36 <duonghq> but it's required dependency tree between service 16:42:37 <inc0> well change to kolla-ansible will be super simple now 16:42:51 <Jeffrey4l> duonghq, yes. 16:43:08 <inc0> also I'd like to not add new aio gates 16:43:15 <inc0> let's do everything multinode from now on 16:43:25 <Jeffrey4l> inc0, +1 16:43:36 <duonghq> +1 16:43:51 <inc0> cool, that's it from me 16:44:01 <inc0> keep an eye on this change 16:44:47 <inc0> anyone else? 16:45:04 <egonzalez> yep, i have one topic regarding gnocchi 16:45:08 <inc0> go ahead 16:45:28 <egonzalez> as most may know, gnocchi has been retired from openstack and now is an independent project 16:45:46 <egonzalez> this means, there is no more tarballs to download/install in source images 16:46:04 <inc0> uhh that's a good point 16:46:09 <inc0> git clone from github? 16:46:59 <inc0> we can just change default in build (just for master because I assume ocata tarballs will stay on) 16:47:29 <egonzalez> not sure about how backports will work 16:47:39 <inc0> probably wont... 16:48:02 <Jeffrey4l> seem the gnocchi in releases project is still updated https://github.com/openstack/releases/blob/master/deliverables/_independent/gnocchi.yaml 16:48:41 <Jeffrey4l> when gnocchi is retired? 16:49:02 <inc0> we should ask release team how gnocci is handled 16:49:09 <Jeffrey4l> hrm, ok. ignore me. gnocchi is release 4.0 already. 16:49:16 <inc0> ah 16:49:21 <inc0> I guess git clone then? 16:49:34 <egonzalez> whole repo was removed, just missing removal from gobernance https://review.openstack.org/#/c/466320/ 16:49:55 <Jeffrey4l> here is the new repo: https://github.com/gnocchixyz/gnocchi/ 16:50:18 <Jeffrey4l> i think it will now follow global-requirements.txt too. 16:50:43 <Jeffrey4l> git clone should be best solution now. 16:51:02 <inc0> let's just do git clone->pip install 16:51:18 <Jeffrey4l> how about master? 16:51:18 <inc0> we already have plumbing for tha 16:51:20 <inc0> t 16:51:30 <inc0> gnocci master github 16:51:32 <inc0> ? 16:52:29 <Jeffrey4l> another question is which version should we use when pike is released ;( 16:52:50 <inc0> we can ask gnocci community:) 16:53:12 <Jeffrey4l> yep 16:54:07 <inc0> ok, are we done?:) 16:54:25 <egonzalez> thats all from me 16:54:33 <inc0> anyone else? 16:55:10 <inc0> guess not:) 16:55:17 <inc0> thank you all for coming! 16:55:20 <inc0> #endmeeting kolla