14:00:42 <fdegir> #startmeeting Cross Community CI 14:00:42 <collabot> Meeting started Wed Jan 10 14:00:42 2018 UTC. The chair is fdegir. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:42 <collabot> Useful Commands: #action #agreed #help #info #idea #link #topic. 14:00:42 <collabot> The meeting name has been set to 'cross_community_ci' 14:00:50 <fdegir> hi everyone 14:01:02 <fdegir> let's have the first meeting of 2018 14:01:07 <fdegir> #topic Rollcall 14:01:48 <electrocucaracha> #info Victor Morales 14:01:53 <joekidder_> #info Joe Kidder 14:02:17 <hwoarang> #info Markos Chandras 14:02:30 <david_Orange> #info David Blaisonneau 14:03:05 <mbuil> #info Manuel Buil 14:03:07 <ttallgren> #info Tapio Tallgren 14:03:30 <fdegir> #topic Scenario/Feature Status: os-odl-sfc 14:03:39 <fdegir> mbuil: anything new here? 14:05:26 <mardim> #info Dimitrios Mrakou 14:05:35 <mardim> Markou* 14:05:38 <fdegir> I suppose not 14:05:43 <mbuil> fdegir: Nothing new. master is working, stable/Pike is working. We are waiting for the new SHAs to be bumped and for the jjobs to be ready 14:05:56 <fdegir> #info Change for post-merge job is waiting to be reviewed 14:05:58 <fdegir> #linkm https://gerrit.opnfv.org/gerrit/#/c/49291/ 14:06:02 <fdegir> #link https://gerrit.opnfv.org/gerrit/#/c/49291/ 14:06:05 <mardim> fdegir, same here nothing new 14:06:10 <epalper> epalper* 14:06:19 <fdegir> thx mbuil mardim 14:06:34 <jmorgan1> #info Jack Morgan 14:06:43 <fdegir> I'll ping others for +2 for 49291 so we can move on with the jobs 14:07:12 <fdegir> #topic Scenario/Feature Status: os-odl-bgpvpn 14:07:21 <fdegir> epalper: any progress? 14:07:49 <epalper> not much update. waiting for upstream blueprint spec to get merged 14:08:09 <fdegir> #info The blueprint is still under review 14:08:20 <hw_wutianwei_> #info Tianwei wu 14:08:25 <fdegir> #link https://review.openstack.org/#/c/523171/ 14:08:32 <fdegir> epalper: thx 14:08:51 <fdegir> #topic Scenario/Feature Status: Kubernetes in XCI 14:09:10 <fdegir> hw_wutianwei_: how is it going? 14:09:35 <hw_wutianwei_> fdegir: I submit a new patch to use kubespray to deploy kubernetes. 14:09:49 <hw_wutianwei_> I am working in progress according the suggestion 14:10:11 <hwoarang> hw_wutianwei_: fdegir. i have some concerns about two things 14:10:12 <hw_wutianwei_> #link https://gerrit.opnfv.org/gerrit/#/c/50213/ 14:10:28 <hwoarang> 1) kubespary does not deploy kubernetes master. is that OK from the XCI point of view? 14:11:01 <hwoarang> 2) kubespray does not support openSUSE so we miss the 'support 3 distro' goal (at least for now) 14:11:27 <fdegir> regarding 1) unfortunately cncf/k8s people are not responsive 14:11:33 <hw_wutianwei_> hwoarang: the kubernetes version is beta release v1.9.0 currently when using kubespray to deploy. 14:11:34 <fdegir> they haven't answered any of the mails I sent 14:12:02 <hwoarang> fdegir: what was the question to them? 14:12:03 <fdegir> so we have to keep looking for what we can use and until we find it, we might need to stick to kubespray 14:12:16 <fdegir> hwoarang: I asked them what they use for installing k8s in their ci 14:12:16 <hwoarang> for a tool that deploys master? 14:12:18 <fdegir> yes 14:12:19 <hwoarang> ok 14:12:47 <fdegir> I'll let everyone know if I get a(ny) response 14:12:53 <hw_wutianwei_> I have a question 14:13:01 <hw_wutianwei_> what name we can use about the kubernetes scenario, k8-nosdn-nofeature or other? 14:13:15 <hwoarang> that should be fine 14:13:40 <fdegir> +1 14:13:52 <hw_wutianwei_> if it is ok, I will use k8-nosdn-nofeature later 14:13:59 <fdegir> I also saw this blog post: https://aws.amazon.com/blogs/opensource/running-bleeding-edge-kubernetes-on-aws-with-kops/ 14:14:13 <fdegir> which is for aws but I at least talks about deploying from master 14:14:31 <fdegir> in case if anyone wants to read 14:14:35 <fdegir> so moving on 14:14:39 <hwoarang> interesting. i will have a look 14:14:39 <fdegir> thx hw_wutianwei_ 14:14:58 <hw_wutianwei_> fdegir: that's all 14:14:59 <fdegir> #topic Senario/Feature Status: Congress in XCI 14:15:11 <fdegir> I'm not sure if Taseer is around 14:15:34 <fdegir> #info The final piece of congress integration is still in progress 14:15:43 <fdegir> #link https://review.openstack.org/#/c/522491/ 14:16:06 <fdegir> #topic Senario/Feature Status: Promise/Blazar in XCI 14:16:18 <fdegir> #info Blueprint for OSA integration is under review 14:16:20 <fdegir> #link https://review.openstack.org/#/c/528567/ 14:16:32 <fdegir> #topic Senario/Feature Status: HA/Masakari in XCI 14:16:36 <OPNFV-Gerrit-Bot> Merged pharos: .yamllint: Add initial config based on releng https://gerrit.opnfv.org/gerrit/50317 14:16:54 <fdegir> #info Masakari team will create the blueprint themselves and then start working on creating the role for OSA 14:17:36 <fdegir> #info Once that is done, then we need to see how we can support them and HA team in OPNFV 14:17:55 <fdegir> that was all about scenarios/features 14:18:01 <fdegir> #topic General Framework Updates: CentOS Support 14:18:08 <fdegir> ttallgren: your turn 14:18:28 <ttallgren> I had all the patches lined up and it passed the tests 14:19:02 <ttallgren> Then I made some changes, base on code review comments, and it no longer passes 14:19:13 <fdegir> #info The patches introducing CentOS support are open for review and there are few issues to fix 14:19:50 <ttallgren> The problem is that OSA installs epel packages from the Internet to all containers and if any fails, then the whole thing fails 14:19:50 <hwoarang> ttallgren: would it be possible to group them all in the same topic/branch so CI can test them together? I mean individual commits like you have them now but all belonging to the same branch 14:20:08 <hwoarang> some fixes went in for epel mirror issues 14:20:11 <ttallgren> I am not sure how to do that 14:20:27 <hwoarang> ttallgren: just create a branch, 'git cherry-pick' them there and then git review 14:20:32 <ttallgren> Ok 14:20:52 <hwoarang> git review can send multiple commits for review no need to send one at a time 14:20:52 <ttallgren> Will there be a local epel cache? 14:20:55 <hwoarang> yes 14:21:11 <ttallgren> Ok. Also, the Centos cloud vm could be somewhere locally 14:21:22 <hwoarang> hmm it may be an 'epel' cache or a cache for the lxc repo 14:21:24 <hwoarang> i will have to check 14:21:29 <fdegir> hwoarang: isn't dependency between changes needed for ci to test them in right way? 14:21:39 <fdegir> hwoarang: not just same topic branch 14:21:52 <hwoarang> yes but the last commit in the branch effectively test them all together 14:22:17 <hwoarang> because its parents are all the previous commits in that branch 14:22:29 <fdegir> now I got it 14:22:52 <ttallgren> What about using submodules for bifrost and OSA? 14:23:25 <fdegir> what do you mean? 14:23:33 <fdegir> like having them as git submodules to xci or? 14:23:39 <ttallgren> Yes 14:23:42 <hwoarang> why? 14:23:49 <ttallgren> Faster, more reliable 14:23:56 <hwoarang> that means they should be moved to a separate repo right? 14:24:13 <hwoarang> why would it be more reliable? 14:24:18 <hwoarang> or faster? 14:24:28 <ttallgren> No need for external network connection 14:24:39 <hwoarang> ah you mean the upstream repos 14:24:54 <fdegir> but those are just to among many 14:25:08 <fdegir> two* 14:25:10 <hwoarang> you still need external connectivity for everything else :) 14:25:18 <hwoarang> containers, openstack services blah blah 14:25:24 <ttallgren> Yes 14:26:09 <fdegir> I think this is kind of part of a bigger topic 14:26:15 <fdegir> upstream supporting offline deployments 14:26:23 <hwoarang> yeah i agree 14:26:23 <fdegir> which is in their roadmap 14:26:36 <fdegir> so if they fix this, we might need to worry about reliability/speed 14:27:08 <fdegir> ok, moving to the next topic then 14:27:26 <fdegir> #topic General Framework Updates: Improving Stability 14:27:37 <fdegir> david_Orange: did you get any reviews to your changes? 14:28:11 <david_Orange> fdegir, yes hwoarang , Alexandru and Dimitrios 14:28:18 <david_Orange> thanks to them 14:28:28 <fdegir> good 14:28:35 <fdegir> #link https://gerrit.opnfv.org/gerrit/#/q/status:open+project:releng-xci+branch:master+topic:stabilization 14:28:44 <fdegir> #info The changes are under review 14:28:53 <fdegir> david_Orange: anything else to add? 14:28:54 <david_Orange> i made a few changes to take in account some comments, and split 2 patches as requested by Markos 14:29:35 <fdegir> ok 14:29:36 <david_Orange> i understand that this is a big set of patches, but i hope it will be a based for next step 14:29:54 <david_Orange> thoses patches will need adding later 14:30:01 <david_Orange> for centos and suse support 14:30:23 <fdegir> david_Orange: I haven't had time to look at them unfortunately so this might be a silly question 14:30:32 <david_Orange> now i just switch to this new code for ubuntu 14:30:40 <fdegir> david_Orange: if they all get merged today, what might stop working? 14:30:55 <hwoarang> the patchset can still be tested in the CI right? 14:31:03 <hwoarang> even if it's just ubuntu 14:31:05 <david_Orange> fdegir, ubuntu 14:31:30 <david_Orange> hwoarang, yes, this is the case 14:31:45 <hwoarang> ok so if it passes ubuntu then we can see how to make it pass in the rest of the distros 14:31:52 <hwoarang> without merging it and breaking the rest of them 14:31:55 <david_Orange> fdegir, builds today are failing with jenkins 14:32:29 <david_Orange> it was before the patch before christmas, and i add a bad line today, i need to check what fails now 14:32:40 <fdegir> ok 14:32:41 <david_Orange> hwoarang, yes 14:32:44 <hwoarang> ok 14:33:09 <fdegir> if that's all, thx david_Orange 14:33:12 <david_Orange> who will do suse and centos integration ? 14:33:30 <hwoarang> i will provide comments for both so you can amend the patches 14:33:36 <hwoarang> once ubuntu is green 14:33:37 <david_Orange> that i just prepare the mail to point config that need to be done 14:33:46 <david_Orange> hwoarang, sure 14:34:25 <fdegir> moving to the last few/quick topics 14:34:27 <fdegir> #topic General Framework Updates: Multiple NFVi deployer support 14:34:53 <fdegir> #info We switched calling the tools we use installer and the rest remains same 14:35:00 <fdegir> it was previously called NFVi 14:35:09 <hwoarang> i think the ground work for this bit is done 14:35:33 <fdegir> yes and thank you for that 14:35:55 <fdegir> it would be useful for k8s and others such as kolla if they appear 14:36:04 <hw_wutianwei> yes 14:36:07 <hwoarang> hw_wutianwei_'s patch is based on that so it's good 14:36:20 <hw_wutianwei> it is more convenient 14:36:25 <electrocucaracha> I can start adapting the kolla patch for that 14:36:25 <hw_wutianwei> hwoarang: thanks 14:37:01 <fdegir> electrocucaracha: can we wait until Kolla wars ends? 14:37:17 <electrocucaracha> fdegir: heheheh, okay 14:37:30 <david_Orange> Kolla wars ? 14:37:31 <fdegir> I still haven't heard anything from Orange for the Kolla they use 14:37:40 <fdegir> david_Orange: there are 3 Kollas... 14:37:43 <fdegir> https://etherpad.opnfv.org/p/xci-kolla 14:37:56 <hw_wutianwei> what's the wars about kolla? 14:38:14 <fdegir> it's a joke from opnfv plugfest 14:38:22 <david_Orange> fdegir, it is not a war, just a familly of kolla :) 14:38:32 <fdegir> yes but which one we will adopt :) 14:38:56 <fdegir> you together with Intel and CableLabs need to pick one of them 14:39:11 <fdegir> based on what the purpose/use case for bringing kolla in to xci 14:39:21 <david_Orange> fdegir, i propose to work on it with electrocucaracha 14:39:30 <fdegir> david_Orange: please talk to Randy as well 14:39:31 <jmorgan1> consolidation is a better term i think, in stead of competing kolla implementations 14:39:46 <david_Orange> fdegir, i can be a good thing to integrate PDF/IDF in the same time 14:39:52 <david_Orange> without impacting osa 14:40:02 <david_Orange> fdegir, sure 14:40:02 <fdegir> david_Orange: randyl was asking about that part so it might be good to reach him 14:40:25 <fdegir> moving to the next topic 14:40:27 <david_Orange> fdegir, my main issue for now is that i have no pod for testing 14:40:29 <jmorgan1> fdegir: what we need is an owner to lead consolidation of kolla implementations 14:41:12 <fdegir> jmorgan1: my hope was and is that people who want Kolla to come in can have the first agreement/understanding within themselves 14:41:22 <fdegir> and then one of them steps up taking the lead 14:41:39 <fdegir> it is not about having multiple versions either 14:41:47 <david_Orange> we will :) 14:42:11 <fdegir> there might be differences when it comes to which version of openstack or what scenarios they already have 14:42:21 <fdegir> so all these depend on the reasons of wanting kolla in 14:42:42 <jmorgan1> understood 14:42:43 <fdegir> we will be waiting until things are clarified 14:42:56 <david_Orange> electrocucaracha, i suppose the source from intel is your patch ? not the link on the pad that point to the kolla doc 14:43:42 <fdegir> the orange text was put there by Kolla PTL so you can involve him as well 14:43:46 <electrocucaracha> david_Orange: yes, that's right 14:44:33 <fdegir> #topic Additional CI Loops 14:44:46 <david_Orange> electrocucaracha, i will look at sources and see what is common, maybe you and randyl can do the same 14:44:59 <fdegir> #info On top of the scenario related jobs, we will have periodic jobs for osa 14:45:11 <fdegir> #link https://gerrit.opnfv.org/gerrit/#/c/49271/ 14:45:23 <fdegir> #topic Intel Lab Maintenance 14:45:34 <fdegir> jmorgan1: are the dates correct on the etherpad? 14:45:53 <jmorgan1> yes 14:46:11 <jmorgan1> all power in the building will be shutdown 14:46:12 <fdegir> #info XCI resources from Intel Lab will be unavailable between Friday Jan. 12th - Tuesday Jan. 16th due to maintenance 14:46:41 <fdegir> #topic AoB 14:46:47 <fdegir> anyone wants to bring anything up? 14:46:55 <hw_wutianwei> fdegir: could we integrate opencontrail in XCI? opencontrail is one of the SDN. 14:47:15 <fdegir> hw_wutianwei: we already had a meeting with them 14:47:19 <fdegir> and presented XCI 14:47:39 <fdegir> their infra/ci team/people were interested but asked time for them to finish their work with their infra 14:47:53 <fdegir> I mean they wanted more time before coming to xci 14:48:04 <hw_wutianwei> ok, thank you 14:48:21 <hw_wutianwei> fdegir: if we integrate kolla, it seem to be easy to integrate opencontrail 14:48:22 <fdegir> hw_wutianwei: but if you believe you can restart this conversation, I can get you in touch with them and you take the lead 14:48:23 <jmorgan1> you mean upstream or opnfv folks? 14:48:30 <fdegir> jmorgan1: upstream 14:48:35 <jmorgan1> ok 14:48:36 <hw_wutianwei> fdegir: https://github.com/Juniper/contrail-ansible 14:48:40 <fdegir> i joined to their infra meeting about 2 months ago 14:48:52 <fdegir> I'll ping them again to see where they are 14:49:23 <jmorgan1> the opencontrail/opnfv person was working on a demo for plugest with joid 14:49:26 <fdegir> hw_wutianwei: that would be one of the ways 14:49:54 <fdegir> i specifically asked for deploying the latest version rather than an earlier/old one 14:50:05 <fdegir> so I am not sure what opencontrail joid or kolla uses 14:50:28 <jmorgan1> maas for deployments, i think 14:50:41 <fdegir> maas is for provisioning the nodes 14:50:53 <fdegir> but we can check with narinder 14:51:06 <fdegir> and with Kolla people to see what they do 14:51:08 <jmorgan1> yes, might check with him 14:51:24 <fdegir> if nothing more, thanks a lot for joining 14:51:29 <fdegir> and happy new year to everyone again 14:51:47 <fdegir> I believe this year will be even more exciting than 2017 14:51:54 <fdegir> you will hear more about this coming weeks 14:51:59 <fdegir> #endmeeting