13:00:16 <fdegir> #startmeeting Cross Community CI 13:00:16 <collabot`> Meeting started Wed May 16 13:00:16 2018 UTC. The chair is fdegir. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:00:16 <collabot`> Useful Commands: #action #agreed #help #info #idea #link #topic. 13:00:16 <collabot`> The meeting name has been set to 'cross_community_ci' 13:00:22 <fdegir> anyone is up for a quick meeting? 13:00:33 <jmorgan1> #info Jack Morgan 13:00:34 <OPNFV-Gerrit-Bot> Manuel Buil proposed releng-xci: [WIP] PDF/IDF https://gerrit.opnfv.org/gerrit/57079 13:00:45 <fdegir> good morning jmorgan1 13:00:54 <hw_wutianwei_> `#info Tianwei Wu 13:01:01 <jmorgan1> fdegir: good morning 13:01:21 <fdegir> we have a light agenda this week as well 13:01:34 <jmorgan1> yeah, i didn't see anything on etherpad 13:01:43 <fdegir> jmorgan1: just updated :) 13:01:47 <fdegir> copy/paste from last week 13:01:58 <jmorgan1> right ;) 13:01:58 <epalper> #info Periyasamy Palanisamy 13:02:01 <joekidder> #info Joe Kidder 13:02:09 <fdegir> let's start and we ping people about the things they are working on 13:02:16 <fdegir> first, there is no update for openci 13:02:23 <fdegir> so moving to the next topic 13:02:39 <jmorgan1> question 13:02:48 <fdegir> #topic PDF, IDF, and Baremetal 13:02:50 <fdegir> jmorgan1: yes 13:03:12 <jmorgan1> for openci, i don't see any activity? are they using same mailing list / tools or moved to something else or just slow 13:03:15 <jmorgan1> ? 13:03:24 <fdegir> slow 13:03:35 <fdegir> i think openstack guys are preparing for the summit and opendev 13:03:37 <jmorgan1> ok, just making sure I'm not left out ;) 13:03:41 <fdegir> mainly zuul stuff 13:03:45 <mbuil> #info Manuel Buil 13:03:48 <jmorgan1> right right 13:03:51 <fdegir> you know, they will announce zuul as a project 13:04:02 <fdegir> lfn guys such as us are working on the prototype 13:04:14 <fdegir> cncf guys are waiting for us to go back to them 13:04:24 <fdegir> and that's all the communities 13:04:25 <hwoarang> #info Markos Chandras 13:04:28 <jmorgan1> cool 13:04:30 <jmorgan1> thanks 13:04:51 <fdegir> you're welcome 13:05:23 <fdegir> mbuil: hwoarang: how are things with pdf, idf, and baremetal? 13:05:44 <hwoarang> i am trying to make k8s understand dynamic inventory and idf. almost there 13:05:49 <hwoarang> *idf/pdf 13:06:05 <fdegir> good 13:06:08 <fdegir> hwoarang: question 13:06:19 <mbuil> fdegir: slowly progressing on booting the VMs using idf/pdf. I hope I can be done before the end of the week 13:06:19 <fdegir> hwoarang: when will you start consuming pdf 13:06:30 <fdegir> hwoarang: i mean you have the dynamic inventory stuff for os already 13:06:34 <fdegir> and k8s will be done soon 13:06:40 <fdegir> the next step is pdf I suppose? 13:07:00 <hwoarang> fdegir: i already am consuming pdf 13:07:07 <hwoarang> *i am already 13:07:25 <fdegir> good to hear 13:07:28 <hwoarang> i consume both 13:07:30 <fdegir> i need to look at things again 13:07:43 <fdegir> the reason i asked this is that i remember seeing some hardcoded ips etc in one of the scripts 13:08:33 <mbuil> fdegir, hwoarang: to boot the VMs (opnfv, controller00 and compute00), I am only consuming pdf so far 13:08:57 <fdegir> mbuil: you create vms yourself using pdf, aren't you? 13:09:03 <mbuil> fdegir: yes 13:09:05 <fdegir> mbuil: meaning that we stopped using the bifrost testvm stuff 13:09:13 <fdegir> good 13:09:15 <fdegir> we have progress 13:09:24 <fdegir> thanks for that 13:09:34 <fdegir> anything else you want to say about this? 13:09:37 <mbuil> fdegir: I am based on that testvm stuff. I changed a few things but most of the stuff is coming from them 13:09:47 <hwoarang> fdegir: some things are still hardcoded but but that's because i didn't get around to convert everything yet 13:09:52 <hwoarang> will do after the k8s is fixdd 13:09:54 <hwoarang> *fixed 13:09:55 <fdegir> mbuil: do you mean vms are created by bifrost? 13:10:29 <mbuil> fdegir: vms are created by libvirt ansible plays 13:10:40 <fdegir> mbuil: this one: https://github.com/openstack/bifrost/tree/master/playbooks/roles/bifrost-create-vm-nodes 13:11:00 <OPNFV-Gerrit-Bot> Markos Chandras proposed releng-xci: xci: xci-deploy.sh: Only check for xci_overrides files in scenarios cache https://gerrit.opnfv.org/gerrit/57469 13:11:01 <electrocucaracha> # info Victor Morales 13:11:15 <mbuil> fdegir: yes, I based my patch on that code 13:11:23 <fdegir> mbuil: if that's the case, we need to ensure we create our vms based on our pdfs 13:11:34 <mbuil> fdegir: exactly 13:11:36 <fdegir> mbuil: because the test vms created by bifrost get same specs etc 13:11:53 <fdegir> so we have ability to assign different cpu/ram/disk based on the role of the node 13:12:17 <fdegir> mbuil: just wanted to ensure we are thinking in same way 13:12:23 <mbuil> fdegir: My patch is based on that code but the template that describes the libvirt xml for the vms take the variables from the pdf 13:12:30 <fdegir> ++ 13:12:56 <fdegir> i suppose that's all for this topic 13:12:57 <mbuil> fdegir: as soon as I have it working, I'll add you guys for review 13:13:05 <mbuil> hopefully tomorrow 13:13:13 <fdegir> that would be great 13:13:19 <fdegir> thanks mbuil hwoarang 13:13:38 <fdegir> so moving to jmorgan1 and electrocucaracha 13:13:45 <fdegir> #topic k8-nosdn-nofeature 13:13:53 <fdegir> jmorgan1: electrocucaracha: how is it going? 13:14:27 <electrocucaracha> k8s-nosdn-onap 13:14:35 <fdegir> #undo 13:14:35 <collabot`> Removing item from minutes: <MeetBot.ircmeeting.items.Topic object at 0x2a5e0d0> 13:14:40 <fdegir> #topic k8-nosdn-onap 13:14:42 <fdegir> thx electrocucaracha :) 13:15:12 <electrocucaracha> Well, jmorgan1 is going to address your comments on the spec 13:15:37 <jmorgan1> for per senario pdfs 13:15:44 <fdegir> i think we are pretty close to getting it in 13:15:50 <fdegir> just some more clarification 13:16:02 <jmorgan1> https://gerrit.opnfv.org/gerrit/#/c/56815/ 13:16:21 <jmorgan1> we can discuss in more detail next week if needed 13:16:49 <fdegir> jmorgan1: per scenario pdf you mean? 13:17:40 <jmorgan1> right 13:17:54 <fdegir> ok 13:18:01 <fdegir> just something to mention that 13:18:11 <fdegir> we need something for k8-nosdn-onap scenario 13:18:18 <jmorgan1> as well as the flavors 13:18:19 <fdegir> whether it is per sccenario pdf or something else 13:18:44 <fdegir> what about flavors? 13:19:34 <jmorgan1> flavors and per scenario pdfs are the two items that you provided feedback for in that patch 13:19:55 <fdegir> yes 13:20:05 <jmorgan1> for flavors, i think it was about reducing the footprint of resources needed for onap if possible 13:20:13 <fdegir> the comment about the flavor was that, the aio is not really the one we use much 13:20:26 <fdegir> it is mainly for developers who want to get a feeling with it 13:20:36 <fdegir> for ci, the focus is on mini/noha, and ha 13:20:41 <mbuil> I was assuming we will have different pdf per flavor, is that wrong? 13:21:26 <fdegir> we can have that handled in single pdf, can't we? 13:21:41 <fdegir> if if it is mini, first 2 nodes in pdf get used 13:21:42 <hwoarang> we can 13:21:49 <fdegir> if noha, first 3 13:21:51 <fdegir> and so on 13:21:51 <hwoarang> all flavors are a cut-down version for the HA one in the end 13:22:02 <hwoarang> so a single pdf can accomodate all of them 13:22:06 <jmorgan1> mbuil: you mean vpdf - idea right now is to have a per flavor vpdf and possible per scenario vpdf that are different later 13:22:21 <jmorgan1> is this not correct? 13:22:50 <mbuil> hwoarang, fdegir: ok, I need to add that logic to my patch then 13:23:19 <mbuil> jmorgan1: not sure 13:23:22 <hwoarang> even for vpdf, we can still have a single one 13:23:23 <epalper> mbuil: cpu, ram parameters are different for each vms in case of onap 13:23:38 <hwoarang> we need a vpdf per scenario not per flavor 13:23:43 <hwoarang> probably... 13:24:23 <jmorgan1> anyway, having at least one now to get going is needed and we can work out details of the rest later 13:24:24 <fdegir> we can use baremetal pdf as example for flavor case 13:24:40 <fdegir> we can deploy noha or ha on baremetal PODs and the nodes that are not used stay as unused 13:24:50 <fdegir> i see it similar to vPDF for flavor case 13:25:23 <fdegir> but the scenario stuff needs thinking since we have ability to adjust the specs based on the scenario as oppose to how it is done if the same scenario goes to a baremetal POD 13:26:12 <fdegir> let's continue talking about these on patches 13:26:16 <jmorgan1> this type of discussion we can have next week ;) 13:26:29 <fdegir> i think it is easier to see what are the things we need to think and our options 13:26:51 <fdegir> jmorgan1: yes 13:26:58 <fdegir> moving to the next scenario 13:27:05 <fdegir> #topic os-nosdn-onap 13:27:11 <fdegir> epalper: how is this one? 13:27:26 <jmorgan1> besides commenting on the spec, i've not started working on k8s-nosdn-onap, not sure about electrocucaracha 13:28:03 <epalper> Have addressed the review comments in spec. did you guys get a chance to look at it ? 13:28:12 * electrocucaracha me neither 13:28:37 <epalper> Needed few enhancements in os_neutron role. raised in https://review.openstack.org/#/c/567809/ 13:28:48 <epalper> now addressing mbuil review comments 13:28:56 <mbuil> ;) 13:29:25 <fdegir> epalper: this is for something else i suppose 13:29:31 <fdegir> not os-nosdn-onap 13:29:44 <fdegir> or is it? 13:30:13 <epalper> no, to run onap heat templates we need to have public network connectivity from openstack's tenant vm's 13:30:24 <fdegir> got it 13:30:32 <epalper> right now os_neutron doesn't do it automatically 13:31:01 <fdegir> ok 13:31:07 <fdegir> i think the spec is nearly done 13:31:08 <fdegir> https://gerrit.opnfv.org/gerrit/#/c/56895/ 13:31:20 <fdegir> if anyone wants to provide last minute comments, please do that 13:31:42 <fdegir> and we will probably submit it in a day or two 13:31:59 <epalper> ok, thank you :) 13:32:04 <fdegir> thx epalper jmorgan1 electrocucaracha for looking into onap scenarios 13:32:18 <fdegir> these are important for us to start bugging them with real stuff 13:32:48 <fdegir> anyone else wants to talk about other scenarios like os-odl-sfc and/or os-odl-bgpvpn? 13:32:52 <electrocucaracha> epalper: I haven't checked but didn't onap requires more openstack services? 13:33:29 <joekidder> there was a need for a designate feature in OS 13:33:33 <joekidder> for dcae 13:33:35 * electrocucaracha I'm trying to remember the dns as a service name... maybe designate 13:33:59 <fdegir> joekidder: designate is supported in osa 13:34:01 <fdegir> joekidder: http://git.openstack.org/cgit/openstack/openstack-ansible/tree/playbooks/setup-openstack.yml 13:34:02 <joekidder> but that might be going away...as they redo dcae a bit for containerization...but I'm not really on top of that. 13:34:11 <electrocucaracha> thanks joekidder 13:34:16 <fdegir> we don't enable it to keep the installation to core services 13:34:34 <fdegir> but we can do it for os-nosdn-onap scenario if it is still needed 13:34:57 <fdegir> moving on 13:35:01 <fdegir> #topic SHA Bump 13:35:11 <fdegir> so we did the first attempt to move to rocky this morning 13:35:22 <fdegir> https://gerrit.opnfv.org/gerrit/#/c/57441/ 13:35:32 <fdegir> deployment on ubuntu failed 13:35:36 <fdegir> on opensuse it passed 13:35:43 <OPNFV-Gerrit-Bot> wutianwei proposed releng-xci: xci: Bump Kubespray SHAs and install ansible modules https://gerrit.opnfv.org/gerrit/57471 13:35:45 <fdegir> but the functest healthckeck failed for that one 13:36:00 <fdegir> it needs some troubleshooting 13:36:24 <fdegir> and as you've perhaps seen, hw_wutianwei_ just attempted bumping shas for k8s/kubespray too 13:36:52 <fdegir> once we get the basic scenarios, we need to verify other scenarios to see if they still work 13:37:00 <fdegir> if not, they will require fixes as well as usual 13:37:24 <mbuil> fdegir: what scenario you tried? 13:37:32 <fdegir> mbuil: os-nosdn-nofeature 13:37:33 <epalper> for SHA Bump, we need to add systemd_service role a-r-r file 13:37:41 <epalper> now neutron role uses that 13:38:02 <epalper> in a-r-r file* 13:38:14 <mbuil> epalper: neutron role uses systemd master? 13:38:34 <fdegir> epalper: it is there 13:38:35 <fdegir> epalper: https://gerrit.opnfv.org/gerrit/#/c/57441/1/xci/installer/osa/files/ansible-role-requirements.yml 13:39:09 <epalper> fdegir: ok, cool 13:39:18 <fdegir> #topic AoB 13:39:29 <fdegir> anyone wants to bring an additional topic before we end the meeting? 13:39:47 <electrocucaracha> fdegir: are you going to cancel the next week meeting? 13:40:00 <electrocucaracha> for the openstack summit 13:40:11 <fdegir> electrocucaracha: that might happen if i can't find someone else to chair the meeting 13:40:25 <fdegir> i generally look for chair a day or two before the meeting 13:40:40 <fdegir> and depending on the topics and cancel if can't find chair or topics 13:40:48 <electrocucaracha> got it 13:41:02 <fdegir> ending the meeting then 13:41:11 <fdegir> thanks a lot for joining and talk to you later 13:41:13 <fdegir> #endmeeting