13:00:12 <fdegir> #startmeeting Cross Community CI 13:00:12 <collabot> Meeting started Wed Jul 11 13:00:12 2018 UTC. The chair is fdegir. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:00:12 <collabot> Useful Commands: #action #agreed #help #info #idea #link #topic. 13:00:12 <collabot> The meeting name has been set to 'cross_community_ci' 13:00:35 <fdegir> I know hwoarang and mbuil_ are probably busy this week and others might be on holiday already 13:00:45 <jmorgan1> hey 13:00:52 <fdegir> hello jmorgan1 13:01:20 <hw_wutianwei_> #info Tianwei Wu 13:01:33 <epalper> #info Periyasamy Palanisamy 13:01:43 <joekidder> #info Joe Kidder 13:02:04 <fdegir> no agenda today so we can start with the 2 scenarios and if anyone wants to bring another topic, we can do that 13:02:10 <fdegir> hi joekidder epalper 13:02:15 <fdegir> #topic k8-nosdn-istio 13:02:15 <jmorgan1> #info Jack Morgan 13:02:26 <fdegir> hw_wutianwei_: so the patches are up for review 13:02:39 <fdegir> #link https://gerrit.opnfv.org/gerrit/#/c/58357/ 13:02:44 <fdegir> #link https://gerrit.opnfv.org/gerrit/#/c/58471/ 13:02:44 <hw_wutianwei_> fdegir: yep 13:03:24 <fdegir> hw_wutianwei_: I see the 2nd patch failed on jenkins 13:03:25 <fdegir> https://build.opnfv.org/ci/job/xci-verify-ubuntu-deploy-virtual-master/1814/console 13:03:45 <fdegir> is it due to code or should we do a reverify? 13:03:48 <hw_wutianwei_> you need merged first patch 13:04:04 <fdegir> hw_wutianwei_: i think the way you did should work 13:04:28 <fdegir> hw_wutianwei_: sorry 13:04:42 <fdegir> hw_wutianwei_: have you tried verifying using the refspec without merging the actual scenario? 13:04:58 <joekidder> hi fdegir epalper jmorgan1 hw_wutianwei_ 13:05:25 <hw_wutianwei_> fdegir: I have tested on my local enviroment 13:05:45 <epalper> yes fdegir, refspec works for scenarios 13:06:07 <fdegir> hw_wutianwei_: can i amend the change and get the scenario tested? 13:06:18 <hw_wutianwei_> fdegir: you can check the previous patchset 13:06:19 <fdegir> hw_wutianwei_: so we merge that one and then amend the 2nd change to switch tomaster 13:06:55 <hw_wutianwei_> fdegir: I think so 13:06:59 <fdegir> hw_wutianwei_: yep, i've seen that but this might be a good opportunity to demostrate how an open change can be verified 13:08:01 <OPNFV-Gerrit-Bot> Fatih Degirmenci proposed releng-xci: [xci] integrate istio in XCI https://gerrit.opnfv.org/gerrit/58471 13:08:25 <fdegir> hw_wutianwei_: if you look at the new patch and to diff 13:08:26 <fdegir> https://gerrit.opnfv.org/gerrit/#/c/58471/5..6/xci/opnfv-scenario-requirements.yml 13:08:45 <fdegir> ci will now verify your patch https://gerrit.opnfv.org/gerrit/#/c/58357/ 13:09:22 <fdegir> doing stuff on github is fine 13:09:49 <fdegir> this is just to show how things work if the change is on opnfv gerrit and not merged 13:10:18 <fdegir> hw_wutianwei_: anything else? 13:10:43 <hw_wutianwei_> no 13:10:56 <fdegir> hw_wutianwei_: thanks for fixing this scenario as well! 13:11:03 <fdegir> #topic k8-calico-onap 13:11:10 <fdegir> jmorgan1: how is it going? 13:11:14 <jmorgan1> good 13:11:31 <jmorgan1> electrocucaracha and I are working on the scenario 13:11:40 <fdegir> great 13:11:52 <jmorgan1> we have a patch that needs to be sent to gerrit after clean up 13:12:05 <fdegir> i learnt that the onap deployment is a bit shaky 13:12:12 <fdegir> i'm not sure if you also see similar issue 13:12:30 <jmorgan1> we got some errors with bifrost 13:12:48 <fdegir> when was this? 13:12:54 <jmorgan1> last Friday 13:13:01 <joekidder> jmorgan1: how many nodes in your k8 cluster? 13:13:02 <fdegir> hwoarang was working on integrating pdf and idf 13:13:03 <jmorgan1> not sure how to report defects 13:13:09 <fdegir> and things should be fine 13:13:14 <fdegir> i deployed this week - monday 13:13:21 <fdegir> could be temporary problem 13:13:42 <jmorgan1> fdegir: ok, i will try again today to see if we still have the problem 13:13:52 * electrocucaracha hopefully it is not a bifrost problem 13:14:01 <jmorgan1> fdegir: otherwse, how to report issues like this? 13:14:19 <fdegir> better to chat with hwoarang and yolanda_ here first 13:14:24 <jmorgan1> joekidder: its the virtual deployment 13:14:35 <fdegir> and if it is a genuine issue upstream then you can either send a fix there 13:14:41 <fdegir> or open a bug report on launchpad 13:15:02 <jmorgan1> ok 13:15:17 <fdegir> it seems they migrated to storyboard 13:15:17 <fdegir> https://storyboard.openstack.org/#!/project/941 13:15:28 <joekidder> jmorgan1: thanks - even so, how many nodes do you create in your virtual deployment? 13:15:41 <electrocucaracha> finally? woo 13:16:15 <jmorgan1> joekidder: just 3 i beleive while testing 13:17:24 <joekidder> jmorgan1: cool. thanks. one level of virtualization? I tried deploying k8/onap on a beefed up virtual OPNFV pod in LaaS, and it seemed cpu utilization was crippling 13:19:58 <fdegir> joekidder: regarding your first question - as jmorgan1 says we started with 3 node cluster; 1 master and 2 workers 13:20:08 <fdegir> joekidder: this is based on the Orange implementation 13:20:32 <fdegir> joekidder: https://gitlab.com/Orange-OpenSource/onap_oom_automatic_installation/ 13:20:33 <joekidder> fdegir: thanks 13:20:54 <fdegir> but they changed the no of nodes/specs recently 13:22:02 <fdegir> https://gitlab.com/Orange-OpenSource/kubespray_automatic_installation/commit/8331566683e902bb3a62c2c7e65c93908a620f36 13:22:07 <joekidder> fdegir: yes, ONAP Beijing seems to have gotten bigger... more k8s pods 13:22:45 <fdegir> jmorgan1: electrocucaracha: you might want to look at the 2nd link as well 13:22:58 <fdegir> electrocucaracha probably knows this so just to be sure 13:23:05 <fdegir> moving on 13:23:07 <jmorgan1> fdegir: ok, will do. Hopefully we will make better progress this week 13:23:17 <fdegir> #topic os-nosdn-osm 13:23:32 <fdegir> after switching to pdf and idf and using opnfv vm for driving stuff 13:23:43 <fdegir> osm installation started failing with lxd configuration 13:24:08 <fdegir> i'll look at it after the holidays 13:24:32 <fdegir> #topic PDF and IDF Implementation 13:24:52 <fdegir> XCI now uses PDF and IDF for virtual deployments so we are one step closer to full dynamic CI 13:25:37 <fdegir> baremetal work is also going on so once this is fixed, I believe we will be done with the Infra WG dynamic CI, PDF, IDF, doing things in VM etc. 13:26:11 <fdegir> that was all I had 13:26:18 <fdegir> anyone wants to bring anything else? 13:27:11 <joekidder> fdegir: minor comment on arm64 xci deploy... 13:27:21 <fdegir> joekidder: yes 13:27:29 <joekidder> hit a problem with trying to load syslinux by bifrost. 13:27:45 <joekidder> apparently syslinux doesn't make sense on arm64... 13:28:14 <joekidder> I noticed that https://storyboard.openstack.org/#!/story/2002934 indicates they might remove syslinux from bifrost role 13:28:49 <fdegir> ok, once that goes in then things should be fine 13:28:52 <joekidder> so hopefully that would unblock until the next thing;) 13:29:03 <joekidder> yes, there's usually only one problem!:) 13:29:06 <fdegir> joekidder: on the other hand, don't you use x86 jumphost for baremetal deployments? 13:29:24 <fdegir> i know this was for virtual btw 13:29:32 <joekidder> fdegir: not anymore. for armband, it's all arm64 13:29:33 <jmorgan1> joekidder: you mean no arm64 support for syslinux? 13:29:42 <fdegir> joekidder: i didn't know that 13:29:42 <joekidder> jmorgan1: yes 13:30:11 <fdegir> then we can wait for a bit to see if there is any progress with that storyboard item 13:30:17 <fdegir> if not, we can start pinging 13:30:19 <joekidder> jmorgan1: AlexAvadanii said that syslinux doesn't make sense to him for arm64, as it's pretty x86-specific low level 13:30:22 <jmorgan1> joekidder: is it a lack of a arm64 build or development is needed? 13:30:49 <jmorgan1> ok, makes sense 13:31:19 <fdegir> thanks for the update joekidder 13:31:45 <fdegir> so, I'll be on holiday until August 2nd and the next meeting will be on July 25th, chaired by mbuil_ 13:31:58 <joekidder> jmorgan1: not sure. The first mention I saw was the person from bifrost (Julia) saying they'd look at syslinux for arm64, and recently in the storyboard she seemed to think it made more sense to remove the dependence on syslinux. 13:32:06 <fdegir> I hope you also get some time off and talk to you in few weeks 13:32:12 <joekidder> fdegir: no problem. Have a great holiday!!! 13:32:18 <fdegir> thanks 13:32:20 <fdegir> #endmeeting