13:00:27 <fdegir> #startmeeting Cross Community CI 13:00:27 <collabot> Meeting started Wed Aug 30 13:00:27 2017 UTC. The chair is fdegir. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:00:27 <collabot> Useful Commands: #action #agreed #help #info #idea #link #topic. 13:00:27 <collabot> The meeting name has been set to 'cross_community_ci' 13:00:33 <fdegir> #topic Rollcall 13:00:42 <fdegir> please type in your name if you are joining the meeting 13:01:10 <yolanda> #yolanda 13:01:10 <David_Orange> #info David Blaisonneau 13:01:17 <yolanda> #info Yolanda 13:01:18 <hw_wutianwei> #info Tianwei 13:01:24 <mbuil> #info Manuel Buil 13:01:29 <georgk> #info Georg Kunz 13:01:54 <epalper> #info epalper Peri 13:02:01 <mbuil> yolanda: sorry, I was having lunch, let's talk after the meeting about your patch and openrc :P 13:02:08 <thaj> #info Thaj 13:02:19 <fdegir> lets start 13:02:20 <fdegir> #topic General XCI Framework status 13:02:41 <fdegir> the patch yolanda sent is open for reviews 13:02:50 <fdegir> anyone tried that? 13:02:55 <fdegir> mbuil perhaps... 13:03:08 <mardim> #info Dimitris Markou 13:03:39 <fdegir> #info The patch to generate and use self signed certificates for OSA is open 13:03:45 <fdegir> #link https://gerrit.opnfv.org/gerrit/#/c/39435/ 13:04:01 <mbuil> mbuil: yes! The SSL problem I was seeing when triggering functest tests is gone when adding the insecure variable to openrc, however, I think that should be automatically included in openrc when I read Yolanda's patch so I think there might be a bug there 13:04:26 <fdegir> mbuil: yolanda: do we still need to use OS_INSECURE=true if we use yolanda's patch? 13:04:36 <yolanda> fdegir, yes, because we use self signed certs 13:04:43 <fdegir> ok 13:04:59 <yolanda> a next step could be use certs from an external provider we trust, or build our own CA and trust it 13:05:05 <fdegir> #info OS_INSECURE=true should be set in openrc while using self signed certificates 13:05:10 <yolanda> but i think it can go on next steps 13:05:23 <hw_wutianwei> yolanda: can we export the certs in openrc? 13:05:26 <fdegir> #info Next step could be to use certs from an external provider we trust, or build our own CA and trust it but what we have is good enough now 13:05:44 <mbuil> yolanda, fdegir: but as far as I understood the openrc in the utility container should be this one ==> https://hastebin.com/ekukikurac.bash or? 13:05:48 <uli-k> #info Uli 13:05:58 <yolanda> mbuil yes 13:06:04 <yolanda> if it's not like that, there is a bug 13:06:32 <mbuil> yolanda: I am getting this one ==> https://hastebin.com/osudabuvan.bash 13:06:37 <fdegir> since mbuil is alrady trying yolanda's patch I action mbuil 13:07:01 <mbuil> yolanda, fdegir: I thought that with those aliases, OS_INSECURE=true is not needed anymore 13:07:06 <yolanda> mbuil so the openrc is on the opnfv host and on the build host. Can you see if both have the same content? 13:07:27 <yolanda> mbuil yep, that's wrong content for openrc, i expect the aliases to be there 13:07:59 <mbuil> yolanda: ok, let's discuss it after the meeting :) 13:08:11 <fdegir> #action mbuil to report findings on the patch and update README.rst with necessary information 13:08:15 <fdegir> thanks, moving on 13:08:24 <fdegir> David_Orange: any update? 13:08:31 <David_Orange> fdegir: yes 13:08:35 <fdegir> #info The PDF for lf-pod4 has been merged 13:09:02 <fdegir> #link https://gerrit.opnfv.org/gerrit/#/c/38283/ 13:09:06 <fdegir> please go ahead David_Orange 13:09:06 <David_Orange> #info Baremetal install is still under debug: i have an issue linking openstack router to extnet 13:09:32 <David_Orange> #info VM to VM is OK, VM to router is OK, but router to ext gateway is failing 13:10:02 <David_Orange> #info i am checking topology and hosts config, but i can not found where it is failing 13:10:48 <David_Orange> #info as i am on an interface with multi vlan and multi bridge, i am deploying with a new network topology with a single vlan for br-vlan 13:11:25 <hw_wutianwei> David_Orange: do you have the topology? 13:11:51 <hw_wutianwei> David_Orange: can you send to us? so we can check. 13:11:54 <David_Orange> #info yes, i can share it, maybe after the meeting 13:12:12 <fdegir> thanks hw_wutianwei! 13:12:25 <David_Orange> #info thanks hw_wutianwei 13:12:34 <fdegir> David_Orange: I've been having issues with ha after latest bump 13:12:43 <fdegir> David_Orange: what do you deploy? 13:12:47 <David_Orange> #info working in parallel on pdf, and moving actual j2 files to this new config source 13:13:02 <hw_wutianwei> fdegir: the issue about keepalived, I ask in irc openstack-ansible channel, they have fixed this problem. and I am testing 13:13:17 <David_Orange> fdegir: ha, but it has never worked, and the install was 1 month ago 13:13:36 <mbuil> David_Orange: can you check if you are using --provider:network_type=flat when creating the public network (or something similar)? 13:13:54 <David_Orange> fdegir: hw_wutianwei: in my install the keepalive between routers were working 13:14:03 <mbuil> David_Orange: when using OVS + ODL, that is needed, otherwise we have the same problem you describe 13:14:10 <fdegir> David_Orange: ok, for me it was failing there 13:14:39 <David_Orange> mbuil: as i just change the infra topology, i can not check now. 13:14:52 <hw_wutianwei> fdegir: me too 13:15:10 <David_Orange> mbuil: fdegir: i will reinstall with last topo version and check that 13:15:26 <fdegir> David_Orange: so we have some progress and challenges bt you have hw_wutianwei and mbuil there to help you as needed 13:15:34 <fdegir> so moving on to keepalived issue 13:15:51 <fdegir> #info After bumping to pike dated august 15th, keepalived playbook started failing 13:15:56 <fdegir> #link https://hastebin.com/oyigeqoqax.md 13:16:02 <David_Orange> #info i will check with mbuil and hw_wutianwei, thanks to them, feel free to send me your ssh keys if you want access 13:16:14 <fdegir> #info hw_wutianwei reported the issue to osa and they fixed it 13:16:21 <fdegir> hw_wutianwei: do you have link to the fix? 13:16:29 <hw_wutianwei> keepalived is caused by the python version 13:16:50 <fdegir> #info keepalived issue is caused by the python version 13:17:00 <fdegir> #info because of this, ha flavor is disabled at the moment 13:17:50 <fdegir> moving on... 13:17:50 <hw_wutianwei> 下午2:41:54 - cloudnull: […] If you have py3 13:17:50 <hw_wutianwei> 下午2:42:05 - cloudnull: i wonder if you have an old version of the keepalived role ? 13:17:50 <hw_wutianwei> 下午2:42:06 - cloudnull: https://github.com/evrardjp/ansible-keepalived/commit/2553ddffd948f3e8e8196cd017d27716b396554a 13:17:50 <hw_wutianwei> 下午2:42:21 - cloudnull: I got a change in that role a while back to address the py3 issue cloudnull: https://github.com/openstack/openstack-ansible/blob/master/ansible-role-requirements.yml#L29-L32 13:18:28 <fdegir> thx hw_wutianwei, I'll probably ping you later, perhaps tomorrow to get more details 13:18:45 <hw_wutianwei> fdegir: ok 13:18:55 <fdegir> #info We finalized migration to releng-xci repo and settling the structure there 13:19:10 <fdegir> #topic Scenario/Feature status 13:19:24 <fdegir> #info os-nosdn-nofeature works 13:19:36 <fdegir> epalper: what about os-odl-nofeature? 13:19:49 <epalper> reviews are in: https://gerrit.opnfv.org/gerrit/#/c/39239/, https://review.openstack.org/#/c/496583/ and https://review.openstack.org/#/c/496583/ 13:20:04 <fdegir> #info Patches for odl are under review 13:20:10 <epalper> testing is in progress. 13:20:11 <fdegir> #link https://gerrit.opnfv.org/gerrit/#/c/39239/ 13:20:18 <fdegir> #link https://gerrit.opnfv.org/gerrit/#/c/496583/ 13:20:27 <fdegir> #info Testing is in progress 13:20:37 <fdegir> epalper: you are working against master, isn't it? 13:20:43 <fdegir> not ocata I hope... 13:20:58 <epalper> #info yes 13:21:08 <fdegir> #info The work with odl is done against master 13:21:19 <fdegir> thanks epalper! 13:21:29 <fdegir> mbuil: your turn 13:21:46 <mbuil> epalper, fdegir: as far as I know, ODL Nitrogen does not work correctly with Pike yet, so don't get paranoid if tests fail :P 13:22:03 <fdegir> mbuil: I won't but you should :) 13:22:25 <fdegir> mbuil: how sfc is doing? 13:22:35 <mbuil> I am waiting for reviews to my two patches ==> https://gerrit.opnfv.org/gerrit/#/c/39517/ https://gerrit.opnfv.org/gerrit/#/c/39511/ 13:22:50 <fdegir> #info SFC patches are waiting for reviews 13:22:56 <fdegir> #link https://gerrit.opnfv.org/gerrit/#/c/39517/ 13:23:02 <fdegir> #link https://gerrit.opnfv.org/gerrit/#/c/39511/ 13:23:04 <mbuil> Those patches complete the os-odl-sfc scenario 13:23:19 <fdegir> #info The patches complete the os-odl-sfc scenario 13:23:20 <mbuil> Then, I am waiting for reviews on this patch ==> https://gerrit.opnfv.org/gerrit/#/c/40595/ 13:23:24 <fdegir> how is the testing going on? 13:23:39 <mbuil> That patch is required to make functest working with XCI 13:23:43 <fdegir> #info The patch to functest is also open 13:23:49 <fdegir> #link https://gerrit.opnfv.org/gerrit/#/c/40595/ 13:23:58 <fdegir> you tried it manually, isn't it? 13:24:12 <fdegir> and what was the result? 13:24:26 * fdegir heard rumours that it looked good... 13:24:29 <mbuil> The testing is completed. Try this private branch which has releng-xci master + Yolanda's patch + those two patches. Deployment is successful 13:24:29 <mbuil> v 13:24:33 <mbuil> https://github.com/manuelbuil/releng-xci/tree/integration-test-certificates 13:25:14 <epalper> #info we also need to get this review to be merged for ODL clustering to work: https://git.opendaylight.org/gerrit/#/c/61945/ 13:25:37 <mbuil> fdegir, epalper: Something interesting for epalper. The current ODL role does not work with ODL Nitrogen. This patch needs to be merged ==> https://git.opendaylight.org/gerrit/#/c/62323/ 13:25:45 <fdegir> I see dfarrell07 is there so I'm sure it will be done soon 13:26:27 <fdegir> #info For odl role to work with nitrogen, the patch needs to go in https://git.opendaylight.org/gerrit/#/c/62323/ 13:26:28 <mbuil> epalper: what is your email, I can add you as reviewer? 13:26:29 <fdegir> dfarrell07: ^ 13:27:14 <mardim> fdegir, : I am the owner of this patch https://git.opendaylight.org/gerrit/#/c/61945/ 13:27:17 <fdegir> while we are on sfc, any update regarding ovs nsh? 13:27:18 <dfarrell07> mbuil, fdegir <blunt honestly> getting packages shipping for Nitrogen and Oxygen is a higher prio than anything related to ansible-odl 13:27:41 <mardim> fdegir, I want to update it but I din't find the tim 13:27:46 <mardim> time* 13:28:03 <fdegir> dfarrell07: we consider this as warning 13:28:04 <dfarrell07> few more TODOs to get Oxygen working for frankbrockners 13:28:10 <fdegir> dfarrell07: but any help is appreciated :) 13:28:32 <epalper> mbuil: periyasamy.palanisamy@ericsson.com 13:28:52 <mardim> fdegir: So ovs-nsh updates 13:28:55 <fdegir> mardim: you mean ovs nsh 13:29:01 <fdegir> yes, please go ahead 13:29:26 <mardim> So as we told to the latest hangout that we had with you guys 13:29:49 <mardim> mbuil integrate my private neutron role to the xci relneg 13:30:05 <mardim> and we are not gonna push upstream the changes for the ovs-nsh 13:30:22 <fdegir> #info mbuild integrated the private role for ovs nsh to neutron 13:30:22 <mardim> because we are gonna have support for the nsh in OVS 2.8 right ? 13:30:43 <mardim> So I think the part for ovs-nsh is done right ? 13:30:48 <mardim> what do you think 13:30:49 <mardim> ? 13:30:56 <fdegir> mardim: if you think it's not worth the effort and assuming ovs 2.8 is soon, it sounds logical 13:31:31 <mardim> Yes I think that for now we can go with the releng-xci as we have it now 13:31:40 <fdegir> mardim: whenever we get ovs with what you need, we bump whatever is needed to take that in 13:31:52 <fdegir> beauty of chasing master 13:32:05 <mardim> fdegir, Can tyou explain that please 13:32:11 <mardim> I din't understand it 13:32:15 <mardim> sorry 13:32:27 <fdegir> mardim: I mean we do not have any constraints when it comes to what to bring in from upstream 13:32:55 <fdegir> mardim: as XCI needs to be on the latest of whatever we bring in from upstream 13:32:59 <mardim> fdegir, So you think that we can close thos jira tickets or ? 13:33:07 <fdegir> mardim: not yet 13:33:07 <mardim> https://jira.opnfv.org/browse/RELENG-228 13:33:18 <mardim> https://jira.opnfv.org/browse/SFC-96 13:33:22 <fdegir> mardim: we can discuss this later after the meeting 13:33:30 <mardim> fdegir sure 13:33:35 <fdegir> thanks mardim 13:33:48 <fdegir> I think scenario/feature topic is done 13:33:53 <fdegir> moving to testing 13:33:57 <fdegir> #topic Testing 13:34:11 <fdegir> #info Trials with functest alpine based containers started 13:34:23 <fdegir> #info It will be integrated to osa-periodic jobs as first step 13:34:48 <fdegir> mbuil: since you updated the progress you had with testing, I skip sfc tests here 13:35:21 <fdegir> #info snaps enabled patches verification for their patches against a pre-deployed pike based system 13:35:41 <fdegir> #info We have a real user and we have to take good care of them 13:36:00 <fdegir> #info See the job https://build.opnfv.org/ci/view/snaps/job/snaps-verify-master/ 13:36:22 <fdegir> #info The information about this is available on https://wiki.opnfv.org/display/INF/Improving+CI+for+Test+Projects#ImprovingCIforTestProjects-SNAPSPoC 13:36:46 <fdegir> and the task about stress test support is on David_Orange if I'm not mistaken 13:37:00 <fdegir> and it depends on what David_Orange already is working on; baremetal 13:37:19 <David_Orange> fdegir: you're not 13:37:34 <fdegir> #info the task about stress test support is on David_Orange and it depends on/related to what he is already doing with baremetal 13:37:40 <fdegir> thx David_Orange ! 13:37:49 <David_Orange> fdegir: you're welcome 13:38:01 <fdegir> next topic is 13:38:05 <fdegir> #topic Distro Support 13:38:29 <fdegir> #info The last I heard from hwoarang was that he was thinking of finishing SUSE support by last friday 13:38:52 <fdegir> #info He started publishing patches for XCI multi-distro support so things are progressing there as well 13:39:04 <fdegir> but we had issues with bifrost on centos 13:39:08 <fdegir> yolanda: any updates? 13:39:23 <yolanda> fdegir, i could fix the log publish from vms, in centos and on suse 13:39:36 <yolanda> but we are still hitting issues on centos jobs , related with selinux 13:40:00 <yolanda> still under investigation 13:40:05 <fdegir> this is the patch you are trying, isn't it? 13:40:05 <fdegir> https://review.openstack.org/#/c/497814/ 13:40:26 <yolanda> yep, but that's not a proper fix although it works :) 13:40:45 <fdegir> #info We are still hitting issues with centos jobs for bifrost and this is related to selinux 13:40:46 <yolanda> there seems to be a problem with serving images via nginx, because they don't have the right selinux type 13:41:06 <fdegir> #info There seems to be a problem with serving images via nginx, because they don't have the right selinux type 13:41:29 <ttallgren> I have been working on CentOS also but I seem to have a different problem 13:41:36 <fdegir> yolanda: I have no doubts about it being resolved since you are on top of it 13:41:49 <yolanda> fdegir, but as always, i could not focus on it 13:42:06 <fdegir> yolanda: I know but we will try our best 13:42:18 <fdegir> ttallgren: do you have logs? 13:42:27 <ttallgren> Yes 13:42:44 <fdegir> ttallgren: you can maybe share it so we can take a look at it after the meeting 13:43:05 <ttallgren> ipmitool is not able to reboot the opnfv VM, for some reason 13:43:29 <yolanda> mm, i have not arrived that long, as our jobs fail on provision the vms 13:44:48 <fdegir> ok, moving on to get through the agenda 13:45:06 <fdegir> yolanda: one last question: did we switch to centos minimal? 13:45:51 <yolanda> mm, i don't remember about it 13:46:58 <fdegir> yolanda: it seems we did for bifrost jobs 13:47:32 <fdegir> #info We switched to centos-minimal for bifrost-verify jobs 13:47:35 <fdegir> #link https://gerrit.opnfv.org/gerrit/gitweb?p=releng.git;a=blob;f=jjb/xci/bifrost-verify-jobs.yml;h=ef604fcfc7c0e550ac199e5d6d3385effbdd7e31;hb=HEAD#l29 13:47:45 <fdegir> #info ttallgren is having trouble with rebooting the opnfv VM 13:47:54 <yolanda> ah yep, i was looking at some logs from my patch now 13:47:59 <fdegir> next topic 13:48:09 <fdegir> #topic Extending CI 13:48:11 <yolanda> we may run with setenforce 0 because it passes there 13:48:25 <fdegir> yolanda: can you send a quick patch there? 13:48:33 <fdegir> I mean to releng? 13:48:49 <yolanda> for what? 13:48:57 <fdegir> yolanda: setenforce 0 thing you said 13:49:01 <yolanda> centos jobs in releng work 13:49:09 <yolanda> but i talk about the bifrost jobs themselves 13:49:17 <fdegir> ok 13:49:20 <ttallgren> bifrost is setting setenforce to 0 13:50:04 <fdegir> #info We started setting up periodic jobs 13:50:23 <fdegir> #info We will have jobs for bifrost and osa to test master to find a working version to pin 13:50:45 <fdegir> #info Skeleton for osa-periodic job is merged and execute shell part is in progress 13:51:03 <fdegir> and the last topic 13:51:07 <fdegir> #topic AoB 13:51:26 <fdegir> #info We presented the progress and short term plans to TSC yesterday 13:51:45 <fdegir> #info Presentation is available on https://wiki.opnfv.org/download/attachments/8687635/XCI_TSC_Report_August2017.pdf?api=v2 13:51:53 <fdegir> my feeling was good 13:52:14 <fdegir> at least we didn't get same questions again which I call a progress 13:52:25 <ttallgren> Did you discuss how this will affect the release model? 13:52:46 <fdegir> ttallgren: we haven't since I'm still trying to address dmcbride's questions 13:52:54 <ttallgren> Ok 13:53:05 <fdegir> ttallgren: and we intend to share the etherpad in a day or two 13:53:23 <fdegir> ttallgren: taking your comment into account regarding CD friendly release model 13:53:33 <fdegir> drawing an ascii art... 13:54:23 <fdegir> #info Next step is to clarify the impacts on release model and have a proposal regarding how it can be adjusted 13:55:01 <fdegir> #info We also need to document guides regarding how projects can be onboarded to XCI and get feedback from it 13:55:26 <fdegir> #info Documentation for XCI also started and will be on docs.opnfv.org under infrastructure 13:55:33 <fdegir> and one last announcement 13:55:54 <fdegir> #info Linux Foundation published XCI Solution Brief as well: https://www.opnfv.org/wp-content/uploads/sites/12/2017/08/OPNFV_SolutionBrief_XCI_080617.pdf 13:56:03 <ttallgren> Congratulations! 13:56:10 <fdegir> thanks ttallgren! 13:56:12 <yolanda> nice! 13:56:22 <mardim> great yes :) 13:56:32 <fdegir> thanks to you all 13:56:48 <David_Orange> great :) 13:56:52 <fdegir> I think we have great progress in such a short time and I believe we will have even more things 13:57:19 <fdegir> so I want to thank all of you for good work 13:57:26 <fdegir> and people expect more, just so you know... 13:57:48 <fdegir> closing the meeting if noone wants to bring anything else? 13:58:04 <fdegir> #endmeeting