13:00:16 #startmeeting Cross Community CI 13:00:16 Meeting started Wed Sep 20 13:00:16 2017 UTC. The chair is fdegir. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:00:16 Useful Commands: #action #agreed #help #info #idea #link #topic. 13:00:16 The meeting name has been set to 'cross_community_ci' 13:00:22 hi everyone 13:00:31 hi 13:00:34 #info Al Morton 13:00:45 please type in your name if you are joining the xci meeting 13:00:56 #info qiliang 13:00:58 #info David Blaisonneau 13:01:01 #info Tapio Tallgren 13:01:03 Hi everyone 13:01:04 #info Tianwei Wu 13:01:04 #info Jack Morgan 13:01:05 while we do that, I can put the link to the agenda as a reminder 13:01:08 hey David_Orange 13:01:11 #link https://etherpad.opnfv.org/p/xci-meetings 13:01:48 lets start and others can join on the way 13:02:03 the first topic for today is an issue we met today 13:02:12 jmorgan1: hi 13:02:18 #topic Issues with diskimage-builder 13:02:25 hello 13:02:36 both me and hwoarang faced issues with bifrost provisioning today 13:02:42 David_Orange: sorry for not getting to your patches yet, I will today 13:02:55 jmorgan1: np :) 13:02:56 some of you might have issues if you attempt using xci on ubuntu today 13:03:13 hwoarang: yolanda: can you give a quick summary about dib issue we have? 13:03:31 #info Markos Chandras 13:03:33 fdegir: Students are happy learning on their XCI deployed OpenStack! 13:03:43 durschatz: happy to hear! 13:03:54 fdegir: yeah basically root=LABEL=foobar doesn't work anymore on ubuntu for unknown reason 13:03:56 info# Dave Urschatz 13:04:06 actually i think the reason is that the dib doesn't inject an initrd on the ubuntu image 13:04:12 #info Stephen Wong 13:04:12 #info root=LABEL=foobar doesn't work anymore on ubuntu for unknown reason 13:04:14 i don't know why maybe something changed in the upstream packages 13:04:29 #info the reason might be that the dib doesn't inject an initrd on the ubuntu image 13:04:31 so, not able to make the meeting today... 13:04:40 yolanda: no worries 13:05:00 so the CI is broken due to this issue as well since we clean the images for each run 13:05:15 which results in jenkins giving -1 13:05:31 i want to use dracut to rebuild an initrd but the element doesn't play well if you don't already have an initrd present 13:05:32 the beauty of living on the edge 13:05:57 #link https://bugs.launchpad.net/diskimage-builder/+bug/1718415 13:05:59 upstream report^ 13:06:22 Then let me reconnect 13:06:46 hwoarang: can we assume that you will not give up on ubuntu yet and will continue trying to fix it? 13:06:55 it's a risky assumption 13:07:00 ;p 13:07:08 #action hwoarang to fix dib issues on ubuntu 13:07:10 thx hwoarang 13:07:14 moving on 13:07:22 #topic Multi-distro support 13:07:31 hwoarang: is it missing some driver support? 13:07:32 hwoarang: I think you have some news regarding SUSE 13:07:53 jmorgan1: i don't know what initrd is missing still investigating. the fact that initrd is missing is worrying enough 13:08:05 hwoarang: agreed 13:08:11 fdegir: yes suse support merged, makes it all the way to OSA but had to kill it so i can fix ubuntu 13:08:31 i mean, networking plugging etc works so that's good enough. if something fails then that's an OSA thing 13:08:35 #info SUSE support has been merged but the work is put on hold while troubleshooting dib issues on ubuntu 13:08:50 i still have a few patches to send but i am waiting for jenkins to be green again 13:09:02 #info If something goes wrong with SUSE, it is probably due to OSA which needs to be fixed as well 13:09:24 #info Few more patches are pending to be sent but we need CI before doing that 13:09:37 that's it 13:09:40 I think we are getting there 13:09:41 hwoarang: thx! 13:09:45 ttallgren: how about centos? 13:10:20 I can run the virtual deployment with noha on Centos 13:10:25 ttallgren: are you building ontop of the suse stuff which added multi distro support to roles? i guess the thing that needs fixing is just the networking bits with vlans etc? 13:10:39 #info The virtual deployment with noha flavor runs on Centos 13:10:50 Yes, that is already in master 13:10:55 ok that's good 13:11:13 ttallgren: have you had chance to try using the osa version I passed? 13:11:19 so you can progress further perhaps 13:11:48 Nope. For some reason the role haproxy_endpoints is not loaded 13:11:56 ok 13:12:19 as I mentioned, it might be safer to use the version we pinned since that version didn't have the issue you had 13:12:30 at least you can see if the problem you had is specific to centos or not 13:12:48 so centos is progressing as well 13:12:50 thx ttallgren! 13:12:57 we already talked about ubuntu 13:13:04 it's been fine until today and then dib happened 13:13:13 next topic is 13:13:16 #topic Testing 13:13:36 #info The work to enable functest for xci-verify and osa-periodic is progressing slowly 13:14:03 I took over the patch from mbuil 13:14:04 #link https://gerrit.opnfv.org/gerrit/#/c/42069/ 13:14:14 which I'll try to finish asap 13:14:39 #info Once the healthcheck is enabled and confirmed working, we can enable testing for sfc scenario once the scenario gets merged 13:14:49 that's all for the testing 13:15:02 #topic Release Readiness 13:15:12 some of you might know that we will release XCI as part of Euphrates 13:15:28 and in order for us to do that, we need to ensure CI exists/up & running 13:15:35 we need documentation 13:15:36 and so on 13:16:00 #info the CI parts is progressing and the documentation will soon be moved to docs.opnfv.org 13:16:18 I'll send few patches for documentation and add everyone as reviewers 13:16:31 please review them as they will be used by our users, like durschatz 13:16:36 and we need to treat our users well :) 13:16:48 :-) 13:17:02 that's all for the release for the timebeing 13:17:15 moving to the scenarios 13:17:29 #topic Scenario/Feature Status: os-odl-nofeature 13:17:34 epalper: are you with us? 13:17:45 fdegir: link for docs please 13:18:14 jmorgan1: http://docs.opnfv.org/en/latest/#infrastructure 13:18:22 jmorgan1: it's empty as the patches are not merged 13:18:48 https://gerrit.opnfv.org/gerrit/#/c/40329/ 13:18:51 fdegir: this is your patch 40329 13:18:53 https://gerrit.opnfv.org/gerrit/#/c/40533/ 13:18:53 right 13:19:01 thanks 13:19:10 I'll submit 40329 to see how it looks on docs.opnfv.org 13:19:29 epalper doesn't seem to be around so I give an update 13:19:29 i will provide review today 13:19:44 but don't have to wait for me 13:20:01 #info epalper's patch for os-odl-nofeature is under review 13:20:03 #link https://gerrit.opnfv.org/gerrit/#/c/39239/ 13:20:20 #info But we need to decide how to move forward with handling scenarios first before merging that patch 13:20:27 #info Which will happen soon 13:21:01 #topic Scenario/Feature Status: os-odl-sfc 13:21:06 mbuil: mardim: any updates? 13:21:47 fdegir: This week not many. You took over the xci-functest patch and we agreed on a way forward to connect scenarios to xci 13:21:56 fdegir: I need to move my pending patch to the sfc repo 13:22:16 #info The agreed way forward with scenarios need to be applied to sfc 13:22:26 fdegir: but I am currently very busy fixing ODL, so I am not sure if I will manage to do that before october 13:22:27 #info The sfc scenario needs to be moved to sfc repo 13:22:44 #link https://gerrit.opnfv.org/gerrit/#/c/39517/ 13:22:55 mbuil: that's fine 13:23:04 fdegir: and also the tacker patch 13:23:18 #link https://gerrit.opnfv.org/gerrit/#/c/39511/ 13:23:18 https://gerrit.opnfv.org/gerrit/#/c/39511/ 13:23:35 fyi for everyone 13:23:53 what we are trying to do with odl and sfc scenarios is actually finding a sustainable way forward with scenarios 13:24:00 by pushing scenarios to project repos 13:24:26 you can read the details about this on https://wiki.opnfv.org/pages/viewpage.action?pageId=12390152 13:24:33 #link https://wiki.opnfv.org/pages/viewpage.action?pageId=12390152 13:24:49 the first attempt will be a bit messy and we will fix it properly once we get the first scenarios running 13:24:54 thx mbuil 13:25:11 yeah, and we will create a proper documentation once that is done 13:25:19 fdegir: you are welcome 13:25:24 #topic Kubernetes in XCI 13:25:34 before I pass the word to s3wong, let me introduce him 13:25:45 Stephen joined us to work with k8s in XCI 13:26:01 and we will start a prototype to bring k8s in 13:26:09 s3wong: stage is yours 13:26:19 fdegir: Thanks 13:26:36 welcome s3wong! 13:26:38 any updates? 13:26:55 welcome s3wong 13:27:07 I am currently looking at some playbooks from kubespray 13:27:22 Personally I normally set up k8s cluster using kubeadm 13:27:46 hwoarang: Thanks 13:27:52 #info s3wong is looking into kubespray for the existing playbooks 13:28:06 CNI for networking? 13:28:34 durschatz: Yes, with either Calico and Canal for CNI plugin 13:28:43 :-) 13:29:23 As Calico (especially) and Canal are well supported and basically are defaults now for k8s networking 13:29:50 #info For networking, either Calico and Canal for CNI plugin will be looked into 13:30:25 Will first try to do AIO with ubuntu as OS 13:30:42 #info First try will be on ubuntu with the flavor aio 13:31:18 I would like to try if you have anything working down the road. 13:31:28 durschatz: certainly 13:32:02 s3wong: anything else? 13:32:09 s3wong: i have to look at k8s for ONAP needs, so i will test it too 13:32:24 s3wong: (and Welcome :)) 13:32:48 David_Orange: yes, I noticed from a link you provided last week 13:32:53 let me put the link to the etherpad again 13:32:53 David_Orange: Thanks 13:32:59 #link https://etherpad.opnfv.org/p/xci-k8s 13:33:09 and highlight some key points/our ambition level 13:33:17 I'm also interested in what is going on with XCI and ONAP 13:33:19 fdegir: that's all for now, will update progress (and the etherpad content) as I move forward 13:33:22 fdegir: sorry, got into some other task. thanks for giving an update ! 13:33:43 #info With k8s, we have similar approach as openstack 13:33:52 #info multi-distro support; ubuntu, suse, centos 13:34:08 #info consume k8s from master - probably using the weekly releases done by k8s community 13:34:26 #info bring up a prototype which is what s3wong has been working on 13:34:48 #info reach out to k8s community and present what we are trying to achieve based on real stuff 13:35:13 like how we did with openstack community (bifrost & osa) 13:35:21 thanks for the update s3wong 13:35:32 fdegir: Thanks! 13:35:34 #topic Congress in XCI 13:35:42 Taseer: are you around? 13:35:52 fdegir: Yes 13:36:03 Taseer: please give us an update regarding congress 13:36:17 #link https://review.openstack.org/#/c/499066/ 13:36:31 Blue print has been accepted by the OSA community 13:36:53 #info Blueprint for congress has been accepted & merged by the OSA community 13:36:59 #link https://github.com/Taseer94/os_congress/ 13:37:22 creating a role from scratch since there is none upstream 13:37:46 #info Taseer is creating the role for congress from scratch since it doesn't exist in the upstream 13:38:07 #link https://review.openstack.org/#/c/503971/ 13:38:15 integration patch in progress as well 13:38:31 #info Integrating the role to osa is in progress as well 13:38:43 Taseer: I suppose you had a look at bryan_att's ansible playbook for congress 13:38:44 That's from me 13:38:57 fdegir: Yes 13:39:05 thx Taseer 13:39:15 fdegir: you are welcome 13:39:17 #topic Extending CI 13:39:39 as we discussed at the beginning of the meeting, CI is currently blocked due to dib issues 13:39:56 apart from that we are working on extending it 13:40:06 jmorgan1: pm now? 13:40:17 hwoarang: yup 13:40:28 #info osa-periodic jobs will be enabled soon and they will have the usual mini deployment together with functest healthcheck to chase the osa master 13:40:48 #info Apart from having the basic jobs, we will enable CI for all the distros we support; ubuntu, centos, and suse 13:41:07 jmorgan1 might have a present for us soon... 13:41:49 #info Once the basic jobs are in place, jobs for the scenarios will be created following what is explained in CI Evolution 13:42:11 #info No more (re)running things if things don't change and instead take them to more extensive testing 13:42:38 that's all for CI parts 13:42:45 #topic XCI on Baremetal 13:42:51 David_Orange: any updates? 13:42:57 fdegir: yes 13:43:12 #info i just pushed my baremetal proto 13:43:22 link or it didn't happen... :) 13:43:53 #link https://gerrit.opnfv.org/gerrit/#/c/42267/ 13:44:29 #info as previous version, the idea is only to have the upstream bifrost+osa installation 13:44:45 #info this version is based on PDF and its actual evolutions 13:45:18 now everyone is welcome to review, ans test it 13:45:28 David_Orange: a question 13:45:58 fdegir: yes 13:46:02 David_Orange: if we have pdf for one of the xci pods, I assume and hope we don't need to change too much and this should just work automagically? 13:46:09 David_Orange: since I see you templated stuff and so on 13:46:46 fdegir: yes it should... i test it on another pod tomorrow 13:47:02 David_Orange: let me know if you want to give a try on one of the xci pods 13:47:05 as intel pod18 is not available 13:47:28 i will make pdf for all intel pod 13:47:29 fdegir: sure, we can plan a test next week 13:47:35 before end of month 13:47:38 David_Orange: ok, I'll contact you offline 13:47:45 jmorgan1: thanks 13:47:51 fdegir: thanks 13:47:57 thx David_Orange 13:48:01 thx jmorgan1 13:48:02 thanks David_Orange 13:48:08 so the last topic is 13:48:11 #topic AoB 13:48:12 fdegir: your most welcome! 13:48:30 #info XCI has been presented to ETSI last week and I think it was good 13:48:32 AoB? 13:48:40 durschatz: any other business 13:48:46 :-) 13:49:19 #info We also had conversations with OpenStack QA thanks to yolanda and talked about what we aim to achieve with XCI 13:49:48 #info I think we have lots of (new) opportunities to collaborate further with OpenStack community around XCI 13:50:05 that was all from me for AoB 13:50:11 anyone else wants to add anything? 13:50:32 When will ONAP be part of XCI? 13:50:32 hw_wutianwei: I see you have some patches to xci 13:50:39 fdegir: integrating ceph are in progress 13:50:55 #info hw_wutianwei is working on integrating ceph 13:51:01 and some bug need to be fixed 13:51:10 hw_wutianwei: bug in osa or in xci? 13:51:26 #link https://gerrit.opnfv.org/gerrit/#/c/42503/ 13:51:32 in xci configuration 13:51:42 is this the one: https://gerrit.opnfv.org/gerrit/#/c/41517/ 13:51:46 hw_wutianwei: ^ 13:52:03 fdegir: https://gerrit.opnfv.org/gerrit/#/c/41517/ 13:52:32 #info A fix has been issued for the bug in xci configuration 13:52:37 #link https://gerrit.opnfv.org/gerrit/#/c/41517/ 13:52:44 thx hw_wutianwei 13:53:01 durschatz: making ONAP part of XCI is in our roadmap 13:53:11 fdegir: you are welcome 13:53:23 but want to get it working fist, no 13:53:24 durschatz: we have been discussing this and things will hopefully start happening soon 13:53:48 durschatz: at least on a more detailed level 13:54:01 durschatz: when things start happening for real depends on our progress 13:54:01 I have a student who will be interested in ONAP. So may want him to understand what is being done so far in XCI 13:54:15 thx 13:54:50 durschatz: then I'll bug you further 13:54:57 :-) 13:55:13 any other updates from anyone? 13:55:16 comments/questions? 13:55:38 jmorgan1: did you mean we should get ONAP working first? 13:55:43 cause xci works perfectly... 13:55:52 fdegir: no xci working first 13:55:53 No. Just want to know status. 13:56:18 jmorgan1: we will make it work 13:56:21 The student may get ONAP working here and could leverage anything already being done. 13:56:43 durschatz: we had some conversations around LaaS/XCI and ONAP 13:56:44 Also provide feedback. But TBD also 13:57:00 these are TBD as well 13:57:30 :-) 13:57:35 sounds good 13:57:43 just a final note; as you noticed, we have the XCI meetings on a weekly basis moving forward 13:57:56 #info XCI meetings will be held on a weekly basis from now on 13:58:06 thank you all for joining! 13:58:13 with pleasure 13:58:15 thanks 13:58:19 and thanks for the great work 13:58:22 thanks 13:58:25 Thanks all! 13:58:34 thanks all 13:58:38 you should know that it is appreciated by many and we are on our way to change things 13:58:40 just saying... 13:58:45 #endmeeting