14:00:12 <fdegir> #startmeeting Cross Community CI 14:00:12 <collabot`> Meeting started Wed Mar 7 14:00:12 2018 UTC. The chair is fdegir. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:12 <collabot`> Useful Commands: #action #agreed #help #info #idea #link #topic. 14:00:12 <collabot`> The meeting name has been set to 'cross_community_ci' 14:00:19 <fdegir> hi everyone 14:00:33 <fdegir> #topic Rollcall 14:00:36 <hwoarang> #info Markos Chandras 14:00:52 <hw_wutianwei> #info Tianwei Wu 14:01:36 <electrocucaracha> #info Victor Morales 14:01:39 <mardim> #info Dimitrios Markou 14:02:10 <fdegir> as you know, we have Kolla topic today so the usual scenario/feature part of the agenda will be quick to leave enough time for kolla 14:02:30 <fdegir> the agenda is on etherpad 14:02:31 <fdegir> #link https://etherpad.opnfv.org/p/xci-meetings 14:02:32 <joekidder> #info Joe Kidder 14:02:55 <fdegir> #topic Functest Healthcheck Status 14:03:04 <jmorgan1> #info Jack Morgan 14:03:25 <fdegir> as you might have noticed, we have functest healthcheck working on Ubuntu and openSUSE 14:03:41 <David_Orange> hi 14:03:41 <David_Orange> #info David Blaisonneau 14:03:58 <fdegir> Centos has issues which might be related to selinux etc but since centos is not voting yet, it's downprioritizes 14:04:37 <fdegir> #info Functest Healthcheck works fine on Ubuntu and openSUSE for os-nosdn-nofeature, os-odl-bgpvpn, and os-odl-sfc 14:05:00 <fdegir> the next topic is related to this one 14:05:17 <fdegir> #topic CI: Verify Jobs 14:05:49 <fdegir> #info verify jobs for external scenarios are enabled so we do deployment and healthcheck for the patches impacting those scenarios 14:06:19 <fdegir> #info these jobs are voting so if things fail, the change will get -1 from jenkins 14:07:12 <fdegir> any questions about these 2 topics? 14:07:55 <fdegir> moving to the next topic 14:07:59 <fdegir> #topic Baremetal Status 14:08:07 <fdegir> David_Orange: how are things? 14:08:53 <David_Orange> fdegir: it is a holidays week for me but last status on orange pods and lf pods were goofd 14:09:31 <fdegir> David_Orange: can you please elaborate good? 14:09:32 <fdegir> :) 14:09:46 <fdegir> David_Orange: do you mean you deployed on them successfully using pdf/idf or? 14:09:56 <David_Orange> the actual patch deploys ubuntu, on baremetal or VM, then a new patch configure the network depending on idf/pdf 14:10:35 <David_Orange> i deploy them with the pdf/idf structure that are locally on pods 14:10:54 <fdegir> ok so the provisioning and network config work 14:11:13 <David_Orange> but last review of lf pod4 idf is better and i have not many work to do to use the proposed idf 14:11:19 <David_Orange> yes 14:11:58 <fdegir> #info Provisioning and network configuration using PDF/IDF work on Orange and LF POD 14:12:05 <David_Orange> i add an issue on E/// pod2, but it was on network side, so i will check that next monday 14:12:44 <fdegir> ok 14:13:10 <David_Orange> the result of the last patch is an OS deployed with a configured network, (bridged or not) and a detailled inventory that can be used for next step 14:13:17 <fdegir> a question about the changes you have on gerrit 14:13:38 <David_Orange> this avoid the VIM to parse too much the PDF/IDF 14:13:44 <David_Orange> yes 14:13:57 <fdegir> I'm sure they all are important/needed but if you can tell us which ones are more important than others, we can look at those 14:14:21 <fdegir> this is first to go I suppose: https://gerrit.opnfv.org/gerrit/#/c/51659/ 14:15:27 <David_Orange> as i said i will wait for the approval of this patch to update the others, so yes you can say it is important 14:15:39 <OPNFV-Gerrit-Bot> Merged pharos: LF-POD4 PDF/IDF https://gerrit.opnfv.org/gerrit/51659 14:15:44 <fdegir> David_Orange: ^ 14:16:00 <David_Orange> but the other are https://gerrit.opnfv.org/gerrit/#/c/49067/ that contains the PDF/IDF for deployment in VM 14:16:11 <David_Orange> great :) 14:16:34 <David_Orange> and also a few python script that help for pdf/idf 14:17:13 <fdegir> ok, we will look at 49067 since you already made provisioning/config work 14:17:15 <David_Orange> the 2 others are https://gerrit.opnfv.org/gerrit/#/c/49069/11 for VM creation (opnfv_host + other if needed) 14:17:31 <David_Orange> https://gerrit.opnfv.org/gerrit/#/c/50361/7 for bifrost 14:17:51 <fdegir> ok, let me info them in order so people know which one to look at first 14:17:52 <David_Orange> the others are more light 14:18:20 <fdegir> #info The patch provisions the nodes using PDF/IDF is https://gerrit.opnfv.org/gerrit/#/c/49067/ and needs to be looked in first 14:18:24 <David_Orange> they other are all related to 49067 14:18:49 <fdegir> #info VM creation is done by https://gerrit.opnfv.org/gerrit/#/c/49069/ 14:19:11 <fdegir> #info Bifrost parts are done by https://gerrit.opnfv.org/gerrit/#/c/50361/ 14:19:15 <David_Orange> 52997 is for network config 14:19:42 <fdegir> #info Network config is done by https://gerrit.opnfv.org/gerrit/#/c/52997 14:19:49 <David_Orange> all thoses patch are done to deploy ubuntu, but are prepared for other OS 14:20:36 <fdegir> #info Patches are done for Ubuntu and they are prepared for other distros which require some work 14:20:58 <David_Orange> people that want to add centos or suse can ping me to find quickly where to add os specific data 14:21:22 <David_Orange> and last info, i start working on moving OSA to PDF/IDF 14:21:45 <fdegir> #info The work to support PDF/IDF by OSA is starting 14:22:15 <David_Orange> but i will wait for an approval on previous patches before publishing it 14:22:32 <fdegir> thanks David_Orange, that would be good 14:22:52 <fdegir> the changes are pretty big even though they are split so it might take some time for people to properly review them 14:23:17 <David_Orange> this patch will have a deep impact on the code and we need to talk on that. 14:23:28 <fdegir> I believe the order we have above would make things a bit easier since we know which ones to look at first now 14:23:38 <David_Orange> fdegir: sure 14:24:16 <fdegir> David_Orange: thanks for the update 14:24:21 <fdegir> moving to the next topic 14:24:22 <David_Orange> fdegir: i understand, but we have to be pragmatic, the code is not perfect, but it proves it works 14:25:09 <fdegir> let's discuss things on patches directly so we can find a good/sustainable way forward 14:25:39 <fdegir> #topic Scenario/Feature Status 14:25:51 <fdegir> anyone has any updates about the scenarios/features? 14:26:02 <hw_wutianwei> I am trying to enable the kubernetes dashboard 14:26:08 <hw_wutianwei> https://gerrit.opnfv.org/gerrit/#/c/52003/ 14:26:22 <fdegir> #info hw_wutianwei is working on enabling kubernetes dashboard for k8-nosdn-nofeature 14:26:28 <fdegir> #link https://gerrit.opnfv.org/gerrit/#/c/52003/ 14:26:29 <hw_wutianwei> the patch was passed on centos verify but failed on ubuntu. 14:27:20 <fdegir> hw_wutianwei: there are few things to note about the failure 14:28:13 <fdegir> bifrost is currently broken due to libvirt-python issue on ubuntu which electrocucaracha found 14:28:25 <fdegir> this should fix it: https://gerrit.opnfv.org/gerrit/#/c/53239/ 14:28:41 <hw_wutianwei> fdegir: yep, I have the same issue 14:28:56 <hw_wutianwei> fdegir, electrocucaracha: thanks 14:29:03 <electrocucaracha> the patch that I submitted to the libvirt community is waiting for the moderator 14:29:12 <hwoarang> a different one has been merged 14:29:16 <hwoarang> ah sorry, that was for bifrost 14:29:42 <fdegir> opensuse job should have exited with no failure so there seems to be a problem with jjb scripts 14:29:46 <fdegir> I'll look into it 14:30:53 <fdegir> and centos passing is good since it shows your patch is good 14:31:06 <fdegir> we can reverify it once bifrost issue is resolved 14:31:32 <hw_wutianwei> I will rebase that patch, and recheck 14:31:34 <fdegir> and as a side note, I am working on enabling generic scenario verify jobs as well so you will not need to hardcode things in vars files to get the right scenario tested 14:31:50 <hwoarang> fdegir: opensuse exited with failure so that's correct 14:32:02 <fdegir> hwoarang: it shouldn't have come to that point at all 14:32:08 <hwoarang> why not 14:32:21 <hwoarang> ah you mean the job shouldn't start because it's not supported 14:32:21 <hwoarang> ok 14:32:51 <fdegir> hwoarang: should have been something like this 14:32:52 <fdegir> hwoarang: https://build.opnfv.org/ci/job/xci-verify-centos-deploy-virtual-master/876/console 14:33:00 <hwoarang> yeah ok 14:33:06 <hwoarang> my bad 14:33:13 <fdegir> but this is probably due to not properly supporting generic scenarios in verify jobs 14:33:34 <fdegir> anything else hw_wutianwei ? 14:33:40 <hw_wutianwei> fdegir: once your work finished, It is good for us to verify 14:33:50 <hw_wutianwei> fdegir: that's all thanks. 14:34:05 <fdegir> hw_wutianwei: I'll use k8-nosdn-nofeature as my test scenario since it is way faster than os scenarios 14:34:14 <fdegir> thanks hw_wutianwei 14:34:52 <fdegir> I suppose the business is as usual for sfc and bgpvpn so skipping those 14:35:11 <fdegir> and moving to the kolla topic 14:35:24 <fdegir> #topic Kolla in XCI 14:35:54 <fdegir> as I said at the beginning of the meeting, I hoped to have more people to have a discussion 14:36:06 <electrocucaracha> I think this link has to be shared https://etherpad.opnfv.org/p/xci-kolla 14:36:09 <fdegir> but we need to go ahead with this no matter what 14:36:43 <fdegir> first I want to ask if anyone has any questions/concerns regarding Kolla in XCI? 14:37:49 <fdegir> i have few 14:38:10 <fdegir> so, did you come to an agreement regarding which kolla you would want to see in XCI? 14:38:27 <fdegir> is it upstream kolla or CableLabs/Orange/another? 14:38:39 <David_Orange> they are all made on upstream 14:39:13 <fdegir> David_Orange: that's right but for example CableLabs has a wrapper around upstream kolla 14:39:32 <fdegir> David_Orange: and Orange one is similar, it has a wrapper 14:39:33 <David_Orange> the orange one is an upstream version, with old PDF/IDF based config, and designate 14:40:07 <fdegir> so, which one we are talking about? 14:40:40 <sylvain_orange> the "wrapper" part is just to configure kolla actually so we can't really talk about "wrapper" for orange one 14:40:40 <David_Orange> in my opinion we have to work on a merge of all those version 14:41:01 <David_Orange> and follow as much as possible the kolla documentation. 14:41:28 <fdegir> sylvain_orange: ok 14:41:46 <David_Orange> following the steps in the documentation is also a way of validation the kolla work 14:41:59 <David_Orange> electrocucaracha: what do you think about that ? 14:42:03 <fdegir> my comment for Orange is based on the statement from Etherpad 14:42:04 <fdegir> "The Orange version is a wrapper around the upstream version, using PDF/IDF (maybe not the latest version) for deployement on several baremetal pods" 14:42:27 <fdegir> so there are differences between how the upstream kolla is consumed by you and CableLabs 14:42:33 <sylvain_orange> no problem, it was just a clarification on my side :) 14:42:37 <electrocucaracha> David_Orange: there are few things that are not fully documented but yes, that should be the idea 14:42:54 <David_Orange> fdegir: it may be my undestanding of the english word wrapper that is not good 14:43:00 <sylvain_orange> and I agree with david_orange, a merge is OK 14:43:17 <fdegir> and to be honest, we will have another way of consuming kolla in xci which is the result of the merge 14:43:22 <fdegir> which might be fine 14:43:35 <David_Orange> fdegir: we are using updtream code, without modifying it, just generating config and running commands as kolla doc specify it 14:43:58 <fdegir> so the answer to my question is there will only be one kolla and you will all be happy with it 14:44:14 <fdegir> did you have chance to discuss this with Randy and Steve? 14:44:18 <fdegir> the merger I mean 14:44:31 <David_Orange> in orange side yes 14:45:23 <electrocucaracha> in our previous talk, Randy was ok with that 14:45:32 <David_Orange> we had a talk with randyl a few weeks ago, he was ok 14:46:49 <fdegir> #info The Kolla that is proposed to be brought into XCI will be based on upstream Kolla and will also include the result of the merged wrappers from CableLabs Kolla and Orange Kolla 14:46:57 <fdegir> I hope this summarizes the intention 14:47:13 <electrocucaracha> +1 14:47:19 <David_Orange> we agree that this kolla version shall support PDF/IDF ? 14:47:26 <David_Orange> +1 14:47:36 <fdegir> David_Orange: yes - that's related to the other questions I have on the etherpad 14:48:04 <fdegir> as you see on xci-meetings etherpad, we have 3 requirements on Kolla 14:48:23 <fdegir> the most important one is staying in the framework set by XCI 14:48:51 <fdegir> which includes PDF/IDF among other things like having proper CI, following the previously set examples and practices, and so on 14:49:03 <fdegir> you know these so I don't want to repeat them 14:49:35 <fdegir> another important need is to have enough people caring kolla 14:49:55 <fdegir> since the reviews for kolla should properly be done 14:50:12 <David_Orange> if so, i would like every one to review https://gerrit.opnfv.org/gerrit/#/c/52997/2, and specificaly templates/vim_host_vars.yml.j2 and templates/vim_inventory.j2 14:50:42 <fdegir> as you can guess, we all are currently busy with what we are working on and expecting people to do more might not be fair and possible 14:51:02 <fdegir> so if you have your colleagues who want to help bringing Kolla in, ask them to join 14:51:25 <electrocucaracha> agree, do you have plans to add more jenkins jobs? 14:51:42 <fdegir> electrocucaracha: that's partly related to the 3rd requirement 14:51:45 <electrocucaracha> fdegir: ^ maybe that can facilitate somehow the extra work 14:51:55 <fdegir> hardware resources will be bottleneck if kolla comes in 14:52:08 <fdegir> initially we can support it but over time, what we have will not be enough 14:52:22 <fdegir> we only have 2 baremetal nodes and 6 vPODs 14:52:30 <fdegir> if you take multi-distro support into account 14:52:48 <fdegir> we are already short of resources but haven't hit the limit yet due to not having everything ready 14:53:39 <fdegir> electrocucaracha: so, we will have jenkins jobs but how fast they can be run remains to be seen 14:53:42 <David_Orange> fdegir: as it can be a requirement for ONAP in orange side, i will ask ressources on our side 14:53:53 <fdegir> David_Orange: that would be good 14:54:11 <fdegir> David_Orange: also, not just hardware resources - you will need human resources as well 14:54:25 <fdegir> to summarize the expectations, here is the list 14:54:28 <jmorgan1> this is a strange dynamic i think 14:54:39 <fdegir> jmorgan1: what do you mean? 14:54:44 <jmorgan1> asking for a feature project to bring in its own hardware 14:54:52 <jmorgan1> feature of xci 14:54:59 <fdegir> jmorgan1: this is not a feature project - it is a brand new installer 14:55:02 <jmorgan1> i understand the idea as resources are limited 14:55:17 <David_Orange> if it give us te opportunity to drop our internal kolla to have a pseudo reference for ONAP, we will work on that 14:55:39 <fdegir> jmorgan1: this is exact same requirement we ask(ed) installer projects 14:55:41 <fdegir> like armada 14:55:47 <jmorgan1> fdegir: ok, i see 14:55:49 <fdegir> so I am surprised to see your reaction 14:56:16 <fdegir> same with human part as well - when that project was presented, i was one of the people asking the question 14:56:28 <fdegir> so I am not putting a totally new requirement 14:56:42 <fdegir> the only new thing here is staying in XCI framework and I think that's fair 14:56:54 <fdegir> if not, then kolla is free to ask for regular installer project creation 14:56:55 <jmorgan1> i just seeing this as a limitation as xci scalles 14:56:55 <jmorgan1> scales 14:56:55 <jmorgan1> but i understand 14:57:15 <fdegir> so, the 3 things we ask are 14:57:20 <fdegir> #info Staying in XCI Framework 14:57:27 <fdegir> #info Bringing people to support Kolla in XCI 14:57:39 <fdegir> #info Bringing hardware to support Kolla in XCI 14:57:51 <fdegir> and the last 2 really depends on how much interest kolla gets 14:58:01 <fdegir> if none, than they are not needed 14:58:24 <fdegir> we have 2 minutes left and I want to have voting with everyone who is in the meeting now so we record the decision 14:59:25 <fdegir> #vote Do XCI Team agree having Kolla in XCI by meeting the requirements expected by XCI? 15:00:05 <fdegir> i don't know if this is the right way to start voting but please vote 15:00:13 <fdegir> it will be in raw logs anyways 15:00:27 <David_Orange> dont know :) 15:00:30 <David_Orange> #vote +1 15:00:38 <electrocucaracha> #vote +1 15:00:44 <jmorgan1> #vote +1 15:01:11 <sylvain_orange> #vote +1 15:01:15 <fdegir> #vote 0 15:01:38 <fdegir> hw_wutianwei: hwoarang: waiting for you two :) 15:01:44 <hwoarang> #vote 0 15:01:45 <hw_wutianwei> #vote 0 15:02:07 <fdegir> it was startvote but anyways 15:02:09 <fdegir> this passes 15:02:22 <electrocucaracha> \o/ 15:02:25 <fdegir> #info XCI Team agrees to have Kolla as part of XCI 15:02:42 <fdegir> looking forward to it 15:02:46 <David_Orange> great 15:02:52 <fdegir> and thank you all for joining today and talk to you next week 15:02:54 <fdegir> #endmeeting