14:02:01 <hwoarang> #startmeeting Cross Community CI 14:02:01 <collabot> Meeting started Wed Dec 13 14:02:01 2017 UTC. The chair is hwoarang. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:02:01 <collabot> Useful Commands: #action #agreed #help #info #idea #link #topic. 14:02:01 <collabot> The meeting name has been set to 'cross_community_ci' 14:02:09 <David_Orange> hwoarang: hello 14:02:15 <jmorgan1> hello 14:02:16 <hwoarang> #topic Rollcall 14:02:23 <David_Orange> #info David_Orange 14:02:24 <jmorgan1> #info Jack Morgan 14:02:26 <hwoarang> #info Markos Chandras 14:02:29 <mardim> #info Dimitrios Markou 14:02:30 <joekidder_> #info Joe Kidder 14:02:42 <hwoarang> #link Agenda is available on https://etherpad.opnfv.org/p/xci-meetings 14:02:42 <hw_wutianwei> #info Tianwei Wu 14:02:43 <mbuil> #info Manuel Buil 14:02:56 <hwoarang> lets wait 1 minute for people to add additional items 14:03:52 <hwoarang> ok lets do it 14:03:53 <mbuil> is fdegir travelling again? 14:04:01 <hwoarang> he has some family business 14:04:10 <mbuil> ah ok 14:04:12 <hwoarang> #topic Scenarios/Feature Status: os-odl-sfc 14:04:16 <hwoarang> mbuil: mardim stage is yours 14:04:27 <mardim> ok I will start 14:04:31 <mbuil> ladies first ;) 14:04:36 <mardim> hahhah :P 14:04:49 <mardim> so the only thing I want to add 14:04:58 <mardim> is that the OVS-NSH patch 14:05:00 <mardim> is merged 14:05:10 <mardim> so it will be available in queens 14:05:12 <hwoarang> in OSA you mean? 14:05:13 <hwoarang> ok 14:05:23 <hwoarang> #info ovs-nsh patch is merged in OSA and will be available in Queens 14:05:26 <mardim> https://review.openstack.org/#/c/517259/ 14:05:27 <hwoarang> great 14:05:29 <David_Orange> mardim: great 14:05:34 <hwoarang> #link https://review.openstack.org/#/c/517259/ 14:05:43 <mardim> thanks that's all from me 14:05:49 <hwoarang> thank you mardim 14:05:54 <mardim> but mbuil has alot of stuff 14:05:58 <mbuil> #info on the upstream side, we are waiting for two related patches to be merged. First patch I think is ready and just needs somebody to merge it 14:06:13 <mbuil> #link https://review.openstack.org/#/c/525264/ 14:06:13 <hwoarang> link so i can look at it after meeting? 14:06:13 <mbuil> #info second patch is waiting for a tempest fix to get a +1 from Zuul 14:06:16 <hwoarang> great 14:06:21 <mbuil> #link https://review.openstack.org/#/c/510909/ 14:06:21 <mbuil> #info this is the tempest fix 14:06:22 <mbuil> #link https://review.openstack.org/#/c/527686/ 14:06:32 <mbuil> #info SFC scenario was successfully merged into releng-xc 14:06:32 <mbuil> #link https://gerrit.opnfv.org/gerrit/#/c/43469/ 14:06:32 <mbuil> #info it works with opensuse and ubuntu with master and pike. To launch it, here are the instructions: 14:06:32 <mbuil> #link https://wiki.opnfv.org/display/sfc/Deploy+OPNFV+SFC+scenarios 14:07:07 <mbuil> this time I wrote it before the meeting :P 14:07:20 <mardim> mbuil, we realized that :P 14:07:31 <hwoarang> thank you mbuil 14:07:33 <mbuil> I guess the first step is to get the deployment automated in a daily jjob 14:07:43 <mbuil> The second step would be to also trigger the functest testcases but for that we need to use a bigger VM because we would need ~70G of memory 14:08:00 <durschatz> #info Dave Urschatz 14:08:13 <hwoarang> #info need some daily CI jobs for deployment 14:08:20 <ttallgren> #info tapio tallgren 14:08:27 <hwoarang> #info need a way to get a bigger VM so functest can run 14:08:33 <epalper> #info Periyasamy Palanisamy 14:08:45 <jmorgan1> hwoarang: how big if bigger? 14:08:52 <jmorgan1> s/if/si 14:08:54 <jmorgan1> is 14:09:01 <hwoarang> we use 48G, mbuil says they need 70 14:09:16 <hwoarang> so we need at least 140G on the host (because we run max 2 VMs) per slave 14:09:22 <hwoarang> so we should be ok to raise that 14:09:23 <mbuil> hwoarang: I ran successfully on a 65G host but it was swapping 1G 14:09:35 <jmorgan1> this is space not ram, correct? 14:09:42 <hwoarang> no it's RAM ;p 14:10:05 <jmorgan1> do we know why sp much ram? 14:10:30 <jmorgan1> just curious 14:10:50 <hwoarang> so we have this big VM right? 14:10:51 <David_Orange> hwoarang: this is for which test case ? you try VIMS ? 14:10:59 <hwoarang> in it we create 3 VMs 14:11:15 <hwoarang> all 3 VMs have the same memory because we can't set memory per VM 14:11:28 <hwoarang> it's only a temp problem until we can set memory per VM 14:11:31 <hwoarang> David_Orange: for SFC 14:11:38 <jmorgan1> hwoarang: ah, ok 14:11:40 <David_Orange> hwoarang: ah ok 14:11:48 <mbuil> jmorgan: xci VMs need around 12G, ODL requires around 4G, that means we need 16G VMs. Besides, our testcases use 4 non-cirros VMs which require 2G of RAM each 14:11:50 <jmorgan1> one idea is run vm in container instead of vm 14:12:27 <hwoarang> yep 14:12:32 <hwoarang> for not right now 14:12:39 <jmorgan1> this is how google does it learned this week 14:13:15 <hwoarang> we will see how to improve it. it's not too far away 14:13:22 <hwoarang> ok so anything else about SFC? 14:13:24 <epalper> jmorgan1: vm in container ? 14:13:34 <mbuil> hwoarang: nothing else 14:13:36 <jmorgan1> epalper: yes, kubernetes based 14:13:47 <epalper> jmorgan1: can you please send the link ? 14:14:10 <epalper> i.e about running vm inside a container that saw it 14:14:13 <jmorgan1> epalper: i don't have one, you might need to investigate. If I find it, I will send it out 14:14:20 <epalper> ok, sure 14:14:21 <hwoarang> #topic Scenarios/Feature Status: os-odl-bgpvpn 14:14:24 <hwoarang> epalper: ^ 14:14:43 <epalper> I have raised blueprint spec in osa 14:14:44 <epalper> https://review.openstack.org/#/c/523171/ 14:14:55 <hwoarang> #info blueprint for OSA has been created 14:14:59 <hwoarang> #link https://review.openstack.org/#/c/523171/ 14:15:06 <epalper> its under review and raised few patches around osa, os neutron and xci 14:15:25 <hwoarang> #link blueprint under review, WIP patches for osa, os_neutron and xci 14:15:28 <hwoarang> great :) 14:15:36 <epalper> I request you to review this patch: https://review.openstack.org/#/c/523907/ 14:15:38 <hwoarang> will review the latest revision of that spec this afternoon 14:15:46 <epalper> Testing for the above patch is pending 14:16:03 <epalper> thanks hwoarang 14:16:10 <hwoarang> #link request for reviews on WIP https://review.openstack.org/#/c/523907/ 14:16:14 <hwoarang> thank you epalper 14:16:32 <hwoarang> #topic Scenarios/Feature Status: os-nosdn-nofeature 14:16:52 <hwoarang> #info fdegir is working on bringing ovs and ceph pieces under this scenario 14:17:17 <hwoarang> #topic Scenarios/Feature Status: Kubernetes in XCI 14:17:25 <hwoarang> hw_wutianwei: ^ 14:17:27 <hw_wutianwei> hwoarang: hi 14:18:28 <hw_wutianwei> I find there are two patch https://gerrit.opnfv.org/gerrit/#/c/48711/ and https://gerrit.opnfv.org/gerrit/#/c/48739/, after they are merged, I will update the patch. 14:18:43 <hw_wutianwei> #link https://gerrit.opnfv.org/gerrit/#/c/46153/ 14:18:56 <hw_wutianwei> and request you to review this 14:19:29 <hw_wutianwei> I am trying to deploy the kubernetes using master branch code. but I have met some trouble. I am trying to solve. 14:19:54 <hwoarang> #info need to disassociate XCI and OSA to make room for more NFVIs 14:19:59 <hwoarang> #link https://gerrit.opnfv.org/gerrit/#/c/48711/ 14:20:12 <hwoarang> #info WIP to deploy k8s from master 14:20:19 <hwoarang> hw_wutianwei: that would be helpful indeed thanks! 14:20:21 <hw_wutianwei> the patch 46153 use kubespray to deploy the v1.8.4 of kubernetes 14:20:35 <jmorgan1> hw_wutianwei: do you have this documented anywhere? I'd like to look at kubernetes in xci 14:21:04 <hw_wutianwei> do you mean deploy master branch? 14:21:37 <David_Orange> hw_wutianwei: we will need to deploy both (for ONAP use case) 14:22:17 <hw_wutianwei> David_Orange: I am trying 14:23:00 <hwoarang> so ok to move on? 14:23:20 <hw_wutianwei> David_Orange: I will ask you fo help, if need 14:23:31 <hwoarang> ok moving on. more k8s talk after the meeting :) 14:23:32 <hwoarang> #topic Scenarios/Feature Status: Congress in XCI 14:23:36 <hwoarang> Taseer: you around? 14:23:37 <David_Orange> hw_wutianwei: you will succed, sure for the help 14:24:16 <hwoarang> guess not 14:24:30 <hwoarang> ok so for the OSA bit we still need os_congress merged 14:24:38 <hw_wutianwei> jmorgan1: this is the plan for k8s in XCI: https://etherpad.opnfv.org/p/xci-k8s 14:24:39 <hwoarang> #info os_congress role WIP 14:24:41 <hwoarang> #link https://review.openstack.org/#/c/522491/ 14:24:50 <jmorgan1> hw_wutianwei: thanks, just found it myself 14:25:19 <hw_wutianwei> jmorgan1: jsut found it :) 14:25:26 <hw_wutianwei> *just 14:25:34 <hwoarang> #topic Scenarios/Feature Status: Upcoming work 14:25:46 <hwoarang> #info Promise/Blazar: Taseer will start working on blueprint for blazar for OSA. 14:25:53 <hwoarang> #info HA/Masakari: we met and will meet again with HA/Masakari team to decide the way forward 14:26:10 <hwoarang> #topic General Framework Updates: CentOS support 14:26:13 <hwoarang> ttallgren: ^ 14:26:42 <ttallgren> I was planning to submit my patches this wek, but then I got some problems after rebasinh 14:27:16 <hwoarang> #info CentOS support is ready for review but need to resolve rebase problems 14:27:17 <ttallgren> I have time to work on it again later this week 14:27:26 <hwoarang> ok then thank you! 14:27:53 <hwoarang> moving on 14:27:55 <hwoarang> #General Framework Updates: Improving stability 14:28:01 <hwoarang> #topic General Framework Updates: Improving stability 14:28:06 <hwoarang> David_Orange: ^ 14:28:20 <David_Orange> #info i worked on it during plugfest, testing it on our pods and on cengen 14:28:54 <David_Orange> #info i am doing the last changes to make it runs with actual os deploy (example root access on opnvf_host :() 14:29:37 <David_Orange> #info it runs with ubuntu, prepared for centos and suse, but some stuff need to be checked for those 2 OS 14:30:10 <David_Orange> opnvf_host is based on ubuntu cloud image, i suppose you have the same for opensuse and centos 14:30:30 <hwoarang> yes 14:30:44 <hwoarang> but we can build and host images in artifacts.opnfv.org 14:30:51 <David_Orange> #info i split the code to post several patch, one for the VM creation on PDF/DF 14:30:52 <hwoarang> we already host stuff there 14:31:28 <David_Orange> #info then one for bifrost 14:32:00 <David_Orange> i will follow the model of OSA split, if you are ok 14:32:45 <hwoarang> sure 14:32:55 <hwoarang> ok sounds promising. Looking forward to it 14:32:58 <hwoarang> anything else? 14:33:08 <David_Orange> hwoarang: has you want, we can set an option to dl them or build them. for now i DL the official ubuntu cloud image for opnfv_hosts, the build the image for bifrost 14:33:21 <hwoarang> yeah we will see 14:33:38 <David_Orange> nothing more, i hope it will not be too big patchs for the review 14:34:05 <David_Orange> one last thing 14:34:39 <hwoarang> yes? 14:34:40 <David_Orange> i made this code to be fully separated from actual xci structure, to avoid perturbations 14:35:12 <David_Orange> i will post the patches, the work on merging folders if this works is ok for you 14:35:26 <hwoarang> well lets see first 14:35:31 <David_Orange> s/the work/then work/ 14:35:39 <hwoarang> i know many things are moving around these days :( 14:35:39 <David_Orange> hwoarang: sure 14:36:37 <David_Orange> that all 14:36:45 <hwoarang> great thanks! 14:36:52 <hwoarang> moving on 14:36:53 <hwoarang> #topic Enabling additional CI loops 14:37:01 <hwoarang> #info this work will proceed soon. 14:37:15 <hwoarang> #topic AoB 14:37:30 <jmorgan1> what is additional CI loops? 14:38:08 <hwoarang> so i guess it means more CI jobs but fdegir put it there so i may not interpret it correctly :) 14:38:22 <hwoarang> like more scenarios tested, ha, no-ha etc 14:38:23 <jmorgan1> ok, i'll ask him later. thanks 14:38:28 <hwoarang> #info Intel POD16-node{4,5} machines will be offline between 9am-12pm PST for RAM adjustments 14:38:49 <jmorgan1> hwoarang: it looks like just node4, node5 need 256Gb ram in pod16 14:39:11 <jmorgan1> hwoarang: just those two nodes will be down (if that is correct) 14:39:11 <hwoarang> aha 14:39:23 <hwoarang> yep 14:39:34 <jmorgan1> ok, so minimal impact i htink 14:39:45 <hwoarang> yes i believe so 14:39:47 <jmorgan1> as those are not being used too much yet 14:40:03 <hwoarang> not sure if they are hooked up to jenkins yet 14:40:11 <hwoarang> i only know about node1 and 3 14:40:13 <jmorgan1> no, not yet 14:40:16 <hwoarang> ok 14:40:20 <jmorgan1> i saw same 14:40:32 <hwoarang> great 14:40:49 <jmorgan1> i concur 14:40:57 <hwoarang> #info Anyone needs resource to do their work on? talk to Fatih :) 14:41:18 <hwoarang> so my understanding is that pod16 is going to be used by CI, and pod21 for devs but master fatih can confirm later on 14:41:32 <hwoarang> sooooooooo 14:41:37 <hwoarang> anything else? :) 14:41:37 <jmorgan1> hwoarang: yes, i beleive so 14:41:47 <hwoarang> btw jmorgan1 thank you for the shiny new hardware :) 14:42:16 <jmorgan1> hwoarang: no problem, enjoy 14:42:18 <hwoarang> you are our santa claus :) 14:42:20 <hwoarang> hehe 14:42:24 <David_Orange> in blue 14:42:28 <hwoarang> lol 14:42:45 <hwoarang> okkk 14:42:49 <hwoarang> i guess nothing else so... 14:42:59 <hwoarang> #endmeeting