08:00:52 #startmeeting Functest weekly meeting 28 Aug. 2018 08:00:52 Meeting started Tue Aug 28 08:00:52 2018 UTC. The chair is ollivier. Information about MeetBot at http://wiki.debian.org/MeetBot. 08:00:52 Useful Commands: #action #agreed #help #info #idea #link #topic. 08:00:52 The meeting name has been set to 'functest_weekly_meeting_28_aug__2018' 08:01:00 #info Cédric Ollivier 08:01:06 #info Juha Kosonen 08:01:13 #info Delia Popescu 08:01:13 #info Viktor Tikkanen 08:01:30 #info Cristina Pauna 08:02:11 #topic Action point follow-up 08:02:48 Few JIRA tickets were closed 08:02:53 #link https://jira.opnfv.org/browse/FUNCTEST-1003 08:03:07 #link https://jira.opnfv.org/browse/FUNCTEST-1006 08:03:14 #link https://jira.opnfv.org/browse/FUNCTEST-1007 08:03:22 #link https://jira.opnfv.org/browse/ARMBAND-393 08:04:15 Shaker arm64 image seems fine. We may double check the endpoint reachability. (Fuel side) 08:04:47 ollivier: I saw your mail, I am looking into it 08:05:12 great. thank you. 08:05:25 Still on-going: 08:05:33 - to advertise the new framework to OPNFV Features et all 08:05:41 - to check the release plan of OPNFV Features integrated in Functest 08:05:51 - to add a README dedicated to the new jjbs 08:06:10 I will send the different emails as soon as possible. 08:07:00 Parallel testing is working very well and the gating job is very useful :). 08:07:30 I managed to test for healthcheck and it was working fine, I hope today I will test the other testcases too 08:07:46 parallel testing I mean 08:08:37 I'm facing with floating ip issues when I run all 3 (fraser, master, ollivier) healtcheck testcases in //. But I think it's due to the deployment. 08:08:59 It would be great if you could confirm or not the possible issues. 08:09:10 I will look into it 08:09:25 maybe we have a quota limit for floating ips 08:09:55 Not here. Duration increases a lot in that case. Floating ips become ACTIVE after more than 3 minutes. 08:10:12 the pool is quite big. 08:10:25 ok, I will let you know what I get 08:10:46 Great. You simply have to edit globals.yaml to switch repo. 08:10:58 yes, I know, thank you 08:11:04 #topic Functest 08:11:16 #All actions to release expect updating docs have be done 08:11:30 #info All actions to release expect updating docs have be done 08:11:40 heat_ims is now working fine in Orange Onap Openlab 08:11:49 #link https://gerrit.opnfv.org/gerrit/#/c/61429/ 08:12:06 VIO should be supported now 08:12:17 #link https://gerrit.opnfv.org/gerrit/#/c/61431/ 08:12:26 #link https://jira.opnfv.org/browse/FUNCTEST-993 08:12:34 #link https://jira.opnfv.org/browse/FUNCTEST-994 08:13:17 Now endusers can convert all images to vmdk in that case. The config are patched if DEPLOY_SCENARIO matches vio. 08:14:09 I hope we will get feedbacks soon. I locally tested here and it seems fine (KVM supports vmdk) 08:15:36 #info tempest 08:15:44 #undo 08:15:44 Removing item from minutes: 08:15:50 #info tempest-full is closed to pass in Orange Onap Openlab (6 subtests in failure) 08:15:59 #link https://gerrit.opnfv.org/gerrit/61217 08:16:07 #link https://gerrit.opnfv.org/gerrit/#/c/61219/ 08:16:15 we may create an admin user for tempest too else we may reach quotas when creating router or network 08:17:06 Even if tempest allocates new users, we run few operations before (tenant network, router, etc...) 08:17:42 juhak: Do you agree to do so as for snaps and rally? 08:18:12 ollivier: ok 08:18:18 juhak: thank you 08:18:47 we should stop filtering testcases per installer (rally and tempest). 08:19:22 juhak: you already proposed yourself for this one :) 08:19:41 yes, I can take this 08:19:44 thank you 08:20:21 Then Functest won't have any installer logic. That's great. 08:20:37 it would be useful to allow running tempest-based testcases via specific users already created. 08:20:56 It's more or less needed to bypass Contrail RBAC. 08:21:33 It may be a simple flag into testcases.yaml. 08:21:46 (or env var) 08:22:17 I will have a look if nobody is interested by that action. 08:23:35 About rally, there is still an open question about integrating tasks from Neutron or other OpenStack services? 08:24:45 The point is simply to match the current openstack gates as much as possible. 08:24:46 are we planning to replace rally_sanity/full with those ones? 08:25:20 It's not my main idea. Rally full is useful even if it takes time. 08:26:56 I think Functest should be at least equal to OpenStack gates. From the time being, we simply differ regarding Neutron (+Cinder and Glance) tasks as you highlighted. 08:27:31 OpenStack CI cases probably do not support offline testing 08:28:38 That's a good point. As the possible hardcoded values or missing Neutron features (Neutron Trunk). 08:29:03 yes, that's right 08:29:28 If we can't integrate them it not a big deal. Our functional scope is already good and we simply know where we differ from OS. 08:30:49 If we can, integrating that new testcases are not really issues if we switch to parallel testing in gates. 08:32:43 CI cases as a such use same test scenarios as we are using, test parameters differs 08:34:45 num of users, tenants, concurrency etc 08:36:28 Yes and OpenStack installers differ from devstack deployments (extensions, nova configurations, etc...). I'm not convinced that we can add all of them. 08:38:27 that's true 08:38:43 I think it makes sens to check that opportunity even if we can't integrate them at the end. 08:39:59 agree, in general it would nice to conform OpenStack CI 08:40:37 would be 08:41:01 Still open. 08:41:22 #info a new testcase "intel-nfv-ci" is proposed (work in progress) 08:41:33 #link https://jira.opnfv.org/browse/FUNCTEST-1004 08:41:41 #link https://gerrit.opnfv.org/gerrit/#/c/61151/ 08:42:15 I think this change depends on updates in OpenStack. I haven't run it yet. 08:43:06 #topic Functest Kubernetes 08:43:15 #info arm64 containers are produced as well 08:43:56 #info new jjbs are being published (same model) 08:44:01 #link https://gerrit.opnfv.org/gerrit/61371 08:44:10 #link https://gerrit.opnfv.org/gerrit/61393 08:45:08 I think they have to be updated but I do first finish my setup. 08:46:21 #topic AoB 08:47:11 I think we MUST create branches as soon as Rocky is released. Else I do create another full Docker repo to gate OpenStack master. 08:48:32 functest:lastest <=> OpenStack Queens, ollivier:latest <=> OpenStack master ... 08:49:11 As soon as Rocky is released ollivier repo has to be synced with OpenStack Master or OpenStack Rocky 08:50:06 I think we should create that branches. Technically speaking Functest is already as stable to be released for Gambia and Hunter. 08:51:21 It would induce maybe more cherry-picks as discussed last weeks but at least it eases maintaining all the containers asked by OPNFV. 08:51:26 Agree? 08:52:17 Regarding a full release, only docs have to be updated. 08:52:49 no objections from my side... 08:54:29 Ok. I will add this point to the next release meeting. 08:54:52 I say we may leave just another week to see more results, then I'm ok 08:55:56 a branch is not a tag. And Yes, we have to wait a little bit because few patches are key for that release (tempest...) 08:56:09 ok 08:57:02 Docs have to be updated as well. 08:57:13 I will be part of ONS. 08:57:33 I dont' know who will be there. 08:58:43 I won't be there this time 08:59:41 neither do I 08:59:50 me neither 09:00:46 noted. 09:01:40 I will send an email to find a new Functest schedules as we discussed earlier. Valentin would like to join us. 09:01:57 That's all from side. Anything else? 09:02:07 not from me 09:02:44 no 09:03:41 Thank you very much. Great week! 09:03:47 #endmeeting