14:00:09 <fdegir> #startmeeting Cross Community CI
14:00:09 <collabot`> Meeting started Wed Feb 14 14:00:09 2018 UTC.  The chair is fdegir. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:09 <collabot`> Useful Commands: #action #agreed #help #info #idea #link #topic.
14:00:09 <collabot`> The meeting name has been set to 'cross_community_ci'
14:00:36 <fdegir> anyone is around for a quick meeting?
14:01:09 <fdegir> we have few people on vacation so we can perhaps even end the meeting in few meetings
14:01:17 <fdegir> minutes*
14:01:23 <david_Orange> #info David Blaisonneau
14:01:28 <david_Orange> fdegir, Hi
14:01:33 <fdegir> hi david_Orange
14:01:47 <electrocucaracha> #info Victor Morales
14:01:56 <mardim> #info Dimitrios Markou
14:02:32 <peri> #info Periyasamy Palanisamy
14:03:03 <fdegir> the first item is
14:03:08 <joekidder_> #info Joe Kidder
14:03:08 <fdegir> #topic Upstream Ironic Issue
14:03:30 <fdegir> #info Due to a change in ironic, bifrost got broken which resulted loss of few days
14:03:44 <pkaralis> #info Panagiotis Karalis
14:03:52 <fdegir> #info The patch got merged and the issue should be resolved now
14:03:54 <fdegir> #link https://review.openstack.org/#/c/542206/
14:04:00 <fdegir> anyone tried xci lately?
14:04:05 <fdegir> like yesterday or today?
14:04:32 <david_Orange> no
14:04:41 <fdegir> I see hwoarang's patch passed CI so the problem looks resolved
14:05:02 <mbuil> #info Manuel Buil
14:05:25 <fdegir> #info At the same time, hwoarang has patches to pin ironic, ironic-inspector, ironic-client, and ironic-client-inspector
14:05:38 <fdegir> #link https://gerrit.opnfv.org/gerrit/#/q/topic:ironic-source-install
14:05:39 <peri> fdegir: i have tried it today with hwoarang patches and it works
14:05:47 <fdegir> peri: good to hear
14:06:01 <peri> looks like we need to export PYTHONIOENCODING=UTF-8 now
14:06:05 <fdegir> #info Once the CI passes, I'll submit those patches
14:06:20 <fdegir> peri: where?
14:06:41 <peri> while running xci-deploy.sh
14:06:55 <fdegir> peri: but CI passes without it, isn't it?
14:06:56 <peri> otherwise i saw it failing
14:07:22 <peri> not sure, this would be my local setup problem
14:07:34 <fdegir> peri: see this https://gerrit.opnfv.org/gerrit/#/c/52091/
14:07:52 <fdegir> the verify jobs passed - ignore centos
14:08:02 <fdegir> moving to next hot issue
14:08:08 <fdegir> #topic Functest Status
14:08:29 <fdegir> #info Ironic issue further delayed Functest integration and we didn't have chance to finish it
14:08:44 <fdegir> #info Work will continue once the xci bifrost patches got merged
14:09:01 <fdegir> and the next topic is for david_Orange and mbuil
14:09:05 <fdegir> #topic Baremetal Status
14:09:23 <fdegir> david_Orange: mbuil: ^
14:09:32 <david_Orange> i tested both pod availables
14:09:40 <david_Orange> i have different issues
14:10:29 <david_Orange> ericsson-pod2: i had to add local nat rules inside the jumphost but i can go to the node provisioning
14:10:48 <david_Orange> provisioning=enroll+deploy
14:10:57 <fdegir> david_Orange: are you getting needed support from Diangfeng about ericsson-pod2?
14:11:38 <david_Orange> now i can not see why os are not well installed
14:12:20 <david_Orange> fdegir, i had a talk with him about accessing ipmi (=> nat rules) but now i can not access the servers console to check what is wrong
14:12:44 <david_Orange> i used the serial on lan to check ipmi+pxe is ok, but i can not see after
14:13:11 <david_Orange> and java remote desktop over an ssh connection is not possible
14:13:29 <david_Orange> not desktop, console
14:13:49 <david_Orange> so i am wait for his help on that point
14:13:50 <fdegir> david_Orange: I just pmed him and he says he will contact you on irc
14:13:58 <david_Orange> fdegir, thanks
14:14:23 <fdegir> how about lf-pod4?
14:14:54 <david_Orange> on lf-pod4, i can go the deployment of the opnfv-host, but it fail due to the presence of a MAAS answering on the network used by the PXE
14:15:02 <fdegir> maas?
14:15:15 <david_Orange> MetalAsAService (canonical product)
14:15:22 <fdegir> yes
14:15:25 <david_Orange> a dhcp/pxe server
14:15:27 <fdegir> but my question was how
14:15:39 <fdegir> aren't those pods isolated from each other?
14:15:47 <david_Orange> i dont know, i opened a ticket
14:15:58 <fdegir> fuel/mcp uses maas on lf-pod2 so that might be it
14:16:02 <fdegir> AlexAvadanii: ^
14:16:32 <fdegir> david_Orange: AlexAvadanii can probably help out with this
14:16:43 <david_Orange> fdegir, this is possible, pods are not clean and i will probably have more side effect of previous installers
14:17:19 <fdegir> david_Orange: please add me and AlexAvadanii to the ticket once you get a response
14:18:03 <david_Orange> during debug time i add a new feature to reuse the previous bifrost image, it will make us gain many minutes
14:18:28 <fdegir> yep
14:18:34 <fdegir> you made it configurable I suppose
14:18:38 <david_Orange> and i am working on a role to configure network from pdf
14:18:41 <fdegir> we have similar thing already if I'm not mistaken
14:19:02 <david_Orange> fdegir, if the is no image and the param it build it
14:19:23 <david_Orange> (remove "and the param")
14:19:37 <fdegir> david_Orange: please look at CLEAN_DIB_IMAGES
14:19:42 <david_Orange> everything is functionnal on orange pod1 and pod2
14:19:59 <david_Orange> fdegir, ok, thx
14:20:03 <fdegir> david_Orange: https://gerrit.opnfv.org/gerrit/gitweb?p=releng-xci.git;a=blob;f=xci/config/user-vars;hb=HEAD#l48
14:20:09 <fdegir> good
14:20:36 <fdegir> I'll follow up lf-pod4/ericsson-pod2 issues with you and LF/Dianfeng to get them sorted out
14:20:57 <fdegir> david_Orange: anything else?
14:21:16 <david_Orange> great, thanks this is blocking for going further
14:21:21 <david_Orange> fdegir, nothing more
14:21:25 <fdegir> thanks david_Orange
14:21:33 <fdegir> #topic Scenario/Feature Status: k8-nosdn-nofeature
14:21:48 <fdegir> hw_wutianwei_ is on new year's holiday so he is not with us today
14:22:02 <fdegir> #info hw_wutianwei_ is working on getting the scenario running on centos
14:22:13 <fdegir> #topic Scenario/Feature Status: Congress in XCI
14:22:31 <fdegir> #info The change integrating congress to osa got merged today after 100+ patches
14:22:53 <fdegir> so I want to thank Taseer for his patience and perseverance for not giving up
14:23:30 <fdegir> #info Next step is to get congress installed as part of XCI by adding it to openstack_services.yml
14:23:49 <fdegir> #topic Scenario/Feature Status: os-odl-sfc
14:23:54 <fdegir> mbuil: mardim: ^
14:23:58 <fdegir> any update?
14:24:21 <mbuil> mbuil: Nothing from OPNFV-SFC from my side. I am still working on the integration of master-ODL into XCI
14:24:39 <fdegir> #info Integration of ODL from master is in progress
14:25:12 <fdegir> mardim: I suppose you are busy with other stuff and didn't have chance/time to look ad odl autorelease for deb?
14:25:43 <mardim> fdegir, exactly too overloaded for now :)
14:25:55 <fdegir> mardim: ok :)
14:26:00 <fdegir> thanks mbuil mardim
14:26:02 <mardim> But I will try to push some commits int the next days
14:26:17 <mardim> in*
14:26:28 <fdegir> #info Work on fixing ODL autorelease for deb will progress coming days
14:26:31 <fdegir> #topic Scenario/Feature Status: os-odl-bgpvpn
14:26:36 <fdegir> peri: your turn
14:26:50 <peri> no major updated on this scenario
14:27:15 <peri> zuul gate job is still unstable in osa neutron rep. i have communicated it with hwoarang
14:27:42 <peri> I hope he will look into it
14:28:12 <fdegir> ok
14:28:23 <fdegir> thanks peri
14:28:26 <peri> and still some more work needs to be done on https://review.openstack.org/#/c/523907/
14:28:55 <peri> that's all from me
14:28:59 <fdegir> ok
14:29:02 <fdegir> #topic Bifrost/OSA Shabump
14:29:59 <fdegir> #info The patches bumping shas of bifrost/osa will be merged once they pass CI
14:30:06 <fdegir> so please be ready to rebase your work
14:30:17 <fdegir> #link https://gerrit.opnfv.org/gerrit/#/c/52091/
14:30:29 <mbuil> fdegir: yes sir
14:30:32 <fdegir> #link https://gerrit.opnfv.org/gerrit/#/c/50367/
14:31:01 <fdegir> #topic OpenStack Helm
14:31:01 <electrocucaracha> nice
14:31:26 <fdegir> #info There will be conversations with OSH PTL and probably with the rest of OSH team during OpenStack PTG in Dublin
14:31:41 <fdegir> who is attending to PTG and who wants to join the conversation??
14:32:32 <fdegir> it seems noone but I'll pull hwoarang into it with me hopefully and will keep you updated about anything
14:32:48 <fdegir> #topic AOB
14:32:54 <fdegir> anyone wants to bring anything up?
14:33:23 <electrocucaracha> regarding he patches for pinning ironic services
14:33:28 <fdegir> yes
14:33:57 <electrocucaracha> is this only for ironic? or that includes the image creator?
14:34:53 <electrocucaracha> I mean diskimage-builder project
14:35:28 <fdegir> according to commit message, dib is not pinned
14:35:29 <fdegir> bifrost: Add ability to pin Ironic-{,client,inspector,inspector-client} SHAs
14:35:40 <fdegir> you can perhaps propose a change and we see
14:36:25 <electrocucaracha> sure, it's better to have others that affects the process as well
14:36:40 <fdegir> +1
14:37:10 <fdegir> I suppose that's all
14:37:21 <fdegir> so thank you all for joining and talk to you next week!
14:37:22 <electrocucaracha> and I have few arguments for Kolla proposal
14:37:32 <fdegir> oops
14:37:35 <fdegir> kolla again...
14:37:36 <fdegir> :)
14:37:39 <electrocucaracha> I'm not sure if david_Orange  has other things
14:37:53 <electrocucaracha> Basically, OpenStack Kolla offers the possibility to install OpenStack services thru Docker Containers. One benefit of this approach is that the Kolla containers that were created and tested by the community can be used in any production environment, this prevents any alteration during the deployment process. The Kolla community had done an effort to reduce the number of steps for deploying its services
14:37:53 <electrocucaracha> resulting in a simple configuration process.
14:38:44 <fdegir> electrocucaracha: david_Orange: can you put them to https://etherpad.opnfv.org/p/xci-kolla if they're not there already
14:39:20 <fdegir> electrocucaracha: david_Orange: and we spend at least 20 minutes for it in one of the meetings when everyone is back
14:39:22 <david_Orange> fdegir, electrocucaracha, ok i will do that
14:39:45 <electrocucaracha> fdegir: sounds fair
14:40:11 <mbuil> electrocucaracha: if we use Kolla, don't we lose the ability to pin on specific SHAs of openstack components?
14:40:26 <fdegir> electrocucaracha: we already have some groundwork done to support additional installers so we just need an open discussion followed by a voting of meeting participants
14:40:33 <electrocucaracha> mbuil: no really, there is a way to pin them
14:40:45 <fdegir> electrocucaracha: I think mbuil brings a valid point
14:41:14 <fdegir> electrocucaracha: if prebuilt containers are proposed to be used, you are pinning container versions not shas on source code level
14:41:17 <david_Orange> mbuil, we can pin on kolla-ansible side but also pin the docker images used
14:41:44 <mbuil> electrocucaracha: for example for neutron, is there an docker image created per commit?
14:42:09 <david_Orange> no, it is one by day IF we use official repo
14:42:11 <fdegir> or image have to be built to use different sha
14:42:22 <david_Orange> but we can build our own
14:42:29 <mbuil> david_Orange: ok thanks
14:42:36 <fdegir> david_Orange: and here goes one advantage of kolla :)
14:43:12 <fdegir> but, kolla and any other thing for that matter has possibility to support different use cases
14:43:40 <fdegir> so what we are looking here is not to find something same as osa but rather what kind of use cases people have and how we can support those
14:43:54 <fdegir> osa can be used for what we are using it for now
14:44:02 <fdegir> and kolla for some other purposes
14:44:09 <david_Orange> one of the advantage of using upstream images, for images not impacted by a patch is to gain building time
14:44:23 <mbuil> fdegir: aaah I see, thanks for the explanation
14:44:25 <fdegir> for example what randyl talked about during opnfv plugfest; pipelines for test projects
14:44:34 <electrocucaracha> the other thing that I was thinking about supporting different installers is not relaying in a single community for deploying Openstack
14:44:56 <fdegir> so, please try to avoid doing direct comparison and instead focus on what kolla could support doing
14:45:28 <fdegir> I think these are all important points which we need to talk during the meeting when everyone is here
14:45:42 <fdegir> mbuil: please add your questions/comments to etherpad so we don't forget them
14:46:15 <mbuil> fdegir: all right! I hope stupid questions are allowed :)
14:46:21 <fdegir> all the questions are welcomed
14:46:34 <fdegir> this is something we all need to vote on in few weeks
14:47:04 <fdegir> I'll send a mail to tech-discuss before that meeting to give heads up to the team so people can read it beforehand
14:47:11 <fdegir> and join the discussion well-informed
14:47:19 * electrocucaracha fdegir you read my mind
14:47:25 <fdegir> :)
14:47:38 <fdegir> now it is time to end meeting
14:47:43 <fdegir> thanks again!
14:47:49 <fdegir> #endmeeting