08:01:22 #startmeeting Functest weekly meeting August 22nd 08:01:22 Meeting started Tue Aug 22 08:01:22 2017 UTC. The chair is morgan_orange. Information about MeetBot at http://wiki.debian.org/MeetBot. 08:01:22 Useful Commands: #action #agreed #help #info #idea #link #topic. 08:01:22 The meeting name has been set to 'functest_weekly_meeting_august_22nd' 08:01:26 #topic call role 08:01:48 #info Juha Kosonen 08:01:54 #chair ollivier LindaWang juhak serena-zte 08:01:54 Current chairs: LindaWang juhak morgan_orange ollivier serena-zte 08:02:16 #info SerenaFeng 08:02:18 #info agenda for today: https://wiki.opnfv.org/display/functest/Functest+5.+Meeting 08:02:20 #info Shuya(OOL) 08:02:22 #info Linda Wang 08:02:31 #topic Action point follow-up 08:02:37 #info no pending action points 08:02:45 #topic Euphrates Status 08:02:52 #info mail sent last week to provide status on Master jenkins (Ubuntu and Alpine view) 08:02:58 #info Ubuntu view summary 08:03:05 #info odl test FAIl due to keystone v3 08:03:17 #info Apex: First scenarios OK (nofeature-nosdn, barometer), snaps_healthcheck FAIL with several scenarios 08:03:28 #info Compass: problem on baremetal seems to be solved (https://build.opnfv.org/ci/view/functest/job/functest-compass-baremetal-daily-master/1405/consoleFull), issue with refstack on several scenarios, issue with domino 08:03:37 #info Daisy: healthcheck and smoke Ok, Domino Fail 08:03:46 #info Fuel@x86: promise, domino, doctor FAIL, issue on smoke tests only with ovs scenario 08:03:54 #info Fuel@aarch64: no CI run (but there was an exception for armband) 08:04:03 #info Joid: issues on smoke with nosdn-nofeature and openbaton scenario, problem with healthcheck on lxd scenario 08:04:32 summary of the summary...we starte getting scenario OK (healthcheck & smoke) 08:05:28 most of the features are FAIL for the moment (and still integration work ahead: sdnvpn, sfc, calipso, vyosvrouter, openbaton test cases,...) 08:05:42 anything you want to add/comment on the status 08:06:43 I think we could exclude the features (doctor, promise, domino) from reporting at the beginning to focus on healthcheck and smoke 08:07:05 doctor is disabled, promise and domino are not 08:07:30 Hi Morgan, has the testcase cloudify_ims supported https? I saw it failed in compass. 08:08:25 LindaWang: no idea, I will ask valentin. I am pretty sure it successfully tested it on compass/Danube 2.0 (Orange communty Pod) 08:08:44 #action morgan_orange ask valentin on https support for cloudify_ims test case 08:08:45 #info Cédric 08:08:51 Ok. sounds good. 08:09:00 it also failed in daisy, we are debugging the reason 08:09:24 #info mail sent to ONOS team to get details for packaging/integration => if no answer this week case will be disabled => and if not integrated in MS6 => ref will be cleaned 08:09:45 We should install/configure dnsmasq for cloudify_ims in Alpine containers. 08:10:10 ollivier: I planned a dedicated topic on alpine status :) 08:10:17 I guess onos will not be updated in future. 08:10:23 morgan_orange: perfect 08:10:40 I do not see onos scenario right now 08:10:43 disable/remove? 08:11:09 yes, they will be obsolete 08:11:21 I think we could adopt the approach mentioned for onos 08:11:28 then I would prefer to remove the code we are hosting 08:11:29 and generalize to any third party 08:12:19 if there is no reply to Morgan's email about onos, that indicates there will be no updates from onos team. 08:12:21 MS6 is planned on the 25th if August 08:12:30 if no news => we remove 08:13:11 Do we disable right now or do we wait a little bit longer? There is a hook required for SFC. mail has been sent but no answer. I would prefer to remove it. 08:13:39 sfc is more active than onos I think 08:13:55 to remove the hook. https://git.opnfv.org/functest/tree/docker/Dockerfile#n99 08:13:58 we can disable right now 08:14:05 #agree 08:14:19 as we warn onos long time ago, we can remove on MS6 08:14:38 for SFC i can take the point to ask them and give them 1 more week but if no news => we remove 08:14:54 Jose will be back, he was involved in SFC he could decide .. :) 08:15:03 I think we should remove it. At least Tim's rozet patch should be updated 08:15:09 if required 08:15:28 Alpine containers didn't rely on it. 08:15:30 the hook is only related to onos, or is it also related to odl? 08:15:44 no. A patch version of an OpenSTack project 08:15:49 tacker I think 08:16:04 ollivier: got it. 08:16:34 #action morgan_orange contact sfc project to get the status and indicate we will disable first then remove beginning of September if no action 08:16:39 for odl /keystone v3: what shall we do, odl is now FAIL with all odl related scenario whatever the installer 08:17:41 boucherv: there was a question from you from Linda 08:17:45 I have no answer for ODL team via JIRA. I will contact them directly. I prefer first finish vnf. 08:18:14 #info no feedback from odl (upstream) on the approach for the odl suite 08:18:16 Is keystone v3 also not supported by odl_netvirt? 08:18:32 #action ollivier contact them directly (once vnf/cloudify_ims is completed) 08:18:47 #undo 08:18:47 Removing item from minutes: 08:19:01 #action ollivier contact odl team directly (once vnf/cloudify_ims is completed) 08:19:15 I do think so. I think their robot files rely on OS_AUTH_URL instead of hardcoding v2 08:19:32 It seems that odl_netvirt includes all tests of odl. 08:20:33 Yes. If all installers install netvirt I don't see the point to run obsolete basic tests. But we don't receive a clear answer. 08:21:13 yep wait for odl feedback, we can see through the answers on the mail thread that odl integration is not clear..different versions, different modules,... 08:21:22 LindaWang: What is your question? 08:21:51 not possible to clarify for Euphrates, but could be an action for RM to do that for next versions 08:22:12 boucherv: see above 08:22:17 boucherv: Has https been supported by cloudify_ims? 08:22:20 I think it's quite simple for all installers to give the features installed 08:22:53 ollivier: I have checked that with compass team, netvirt is installed. 08:23:18 Even if non netvirt scenarii. 08:23:30 LindaWang: HTTPS for openstack ? 08:23:38 boucherv: yes 08:24:07 boucherv: We can talk about that later :) 08:24:07 as we put in place a requirement management, something similar shall be done at the release level...I remember that lots of people asks if there was any command in opnfv to get the list of features/version of components 08:24:14 LindaWang: Yes! WIth insecure it's work 08:24:30 LindaWang: The issue is now the quota not the https 08:24:49 LindaWang: I work on a patch for quota but also snaps integration in vnf_base 08:24:53 #action morgan_orange contact D.McBride to report the need of a better control of the onboadred features/versions ..we coudl add Python version 08:25:32 boucherv: Steven has commited a patch about quota updation. 08:26:19 LindaWang: Yes, I know, I created a patch with the snaps quota https://gerrit.opnfv.org/gerrit/#/c/39441/ 08:26:29 LindaWang: But, I have to test it 08:26:31 Ok for the overall Ubuntu status, can we move to third party integration, or do we move to alpine status first? 08:26:38 Yes. Add doctor must be checked. (they must call python3 instead of python) 08:26:56 ok so integration.. 08:27:06 #third party integration 08:27:06 Parser is not ready too. 08:27:16 #info doctor must be checked. (they must call python3 instead of python) 08:27:23 #info Parser is not ready 08:27:31 do we have all the required information for parser? 08:27:41 Parser's PTL is working on that 08:27:58 he said he will modify according to Functest's requirements 08:28:18 ? 08:28:34 The patch has been ready to be merged for 1 month 08:28:46 which patch? 08:28:50 doctor 08:29:03 multithreading issue. Sorry 08:29:34 #info morgan_orange contacted ryota to merge doctor patch (ready since July) no feedback so far 08:29:36 Do you know why they fork several openstack project? 08:29:39 I think Doctor's PTL will do the negotiation 08:29:57 #info Parser PTL working on adaptations to meet Functest's requirements 08:30:07 we also got a request from the calipso project (dedicated scenario) 08:30:17 they have functional tests they would like to run 08:30:24 I sent them a mail yesterday 08:31:32 the contact is abashinos AT gmail.com 08:31:50 I send him some links (example, abstraction,..) 08:32:01 calipso? 08:32:22 I cannot find the wiki page of that project 08:32:41 it is the new name of the vina project 08:32:50 #link https://wiki.opnfv.org/display/vina/Calipso+Project+Proposal 08:32:57 serena-zte: I could help for the parser packaging if required. I would like to simply understand why they forked the official projects (obsolete?). These are required by other projects too. 08:33:01 okey, got it, thank you 08:33:21 #link https://build.opnfv.org/ci/job/calipso-verify-master/ 08:33:39 I guess they add some new features 08:34:17 #action serena-zte clarify with Parser PTL why they forked the official projects => conenction Parser PTL to ollivier 08:35:21 #info B.Souville indicated that he successfully tested Promise on a community lab, currently FAIL on all the CI runs 08:35:49 #info first test on sdnvpn => question from tomsou_ yesterday 08:35:57 They also hacks their files. 08:36:35 ollivier, yes, I talked with them why they do that before, they said they have to made some enhancement for OPNFV 08:36:50 I will let them talk to you directly. 08:37:04 #topic Alpine testing status 08:37:11 They did. I wrote the modifications in a flase JIRA ticket on functest 08:37:19 #link https://wiki.opnfv.org/display/functest/Running+new+Functest+containers 08:38:13 #info functest-vnf and functest-components now exposed in OPNFV docker hub 08:38:23 #link https://hub.docker.com/u/opnfv 08:38:45 No. they are dummy 08:38:58 ah ok 08:39:40 so use ollivier/functest-vnf, ollivier/functest-components for the moment 08:40:00 we may notice that Narinder made some test with functest-healthcheck and give a positive feedback 08:40:02 We have to test ollivier/functest-xxx for the last 3 containers. I sent two emails to Aric and Trevor for creating the automated builds. 08:40:17 #info We have to test ollivier/functest-xxx for the last 3 containers. I sent two emails to Aric and Trevor for creating the automated builds. 08:40:39 #info positive feedback from Narinder with functest-healthcheck 08:41:42 Regarding the global status: we are finishing them as it remains to test functest-vnf. 08:41:47 #info functest-vnf: installation cloudify (VNFM) + IMS OK, struggling with ruby tests => installation/configuration install/configure dnsmasq for cloudify_ims 08:42:05 Yes. musl differs from glibc 08:42:08 Do we need to test opnfv/functest-xxx or ollivier/functest-xxx? are they the same? 08:42:15 feature project shall also test ...thei feature... 08:42:53 They should be. But I'm not sure they do handle well the dependencies between core and the others. I sent an email about that too. 08:43:33 LindaWang: not yet for healthcheck/smoke = same, but not for components, features and vnf (pending request to LF) 08:44:25 But I don't think we could face with that bug (if functest-vnf is built before functest-core). I would propose to follow the wiki and to test opnfv as much as possible. 08:44:49 #info currently in CI: Alpine test only on Huawei virtual (E/// virtual pod is down, wait for Jose come back) 08:44:52 #link https://build.opnfv.org/ci/view/functest/job/functest-alpine-compass-huawei-virtual5-daily-master/ 08:45:01 badly I can't see their Docker settings. But we can bet they are equal 08:45:28 When opnfv/functest-vnf and opnfv/functest-components are ready, I will help test them in huawei-virtual5 08:46:07 it seems there are still issues with rally sanity /refstack on huawei virtual 5 08:46:34 Yes, rally_sanity failed complaining segmentation fault(core dumped). 08:46:37 Thank you. You could start with ollivier (eg vnf) if you agree. It's equal or better. 08:46:56 Still no idea how to fix that. 08:47:16 good thing is that is is reproducible... 08:47:17 Could you please test with ollivier/... ? 08:47:36 on Orange compass community pod, it was OK .. 08:48:11 juhak: are you able to test alpine container in Nokia labs? 08:48:52 #topic Functest API status 08:48:53 yes, I can test on apex 08:49:00 #undo 08:49:00 Removing item from minutes: 08:49:18 #action juhak test alpine containers on Nokia/apex community labs 08:49:31 just follow the wiki page.. 08:49:38 #topic Functest API status 08:49:57 #info Linda submitted lots of patches 08:50:27 any comment/update? 08:50:50 I suppose you will need also help to test it.. 08:50:51 I think orchestra vnfs don't work. 08:50:58 Basic framework and functions have been finished: functest env prepare; get testcases lists; run a test case 08:51:40 ollivier: we can contact Michael, they are not working in traditionnal CI as well. 08:51:51 #aciton morgan_orange contact orchestra regarding orchestra_* test cases 08:51:55 And nginx and uwsgi will be leveraged later. 08:51:56 morgan_orange: sounds great. Thank you. 08:51:59 #undo 08:51:59 Removing item from minutes: 08:52:07 #action morgan_orange contact orchestra regarding orchestra_* test cases 08:52:46 LindaWang: ok thanks 08:53:08 #topic Test API monitoring? 08:53:59 #info test API is no relatively critical, we faced some unstability over the last months 08:54:14 #info for ex last success Apex run were not able to push the results..:) 08:54:16 #info serena-zte: submitted some fix + watchdog mechanism 08:54:37 I add this topic => shall we add a monitoring at the lF level on the API? 08:54:41 Rohit submit a new patchset, but still some problems 08:54:51 they are currently monitoring JIRA... 08:55:12 the only think is that when LF IT team monitor they restric the access to the machine... 08:55:45 what do you mean by 'restrict access to the machine' 08:55:57 does it mean that we can not access it anymore? 08:56:02 access only to LF people 08:56:25 possible access but more controlled...today lots of people can have access to testresults.opnfv.org 08:56:40 we host the test API, the reporting and maybe some stuff I ingore... 08:56:46 yes 08:57:01 not realistic for Euphrates but an option to consider for F... 08:57:20 yes, we can consider it after TestAPI is a little bit stable 08:57:29 ok 08:57:41 currently, it is quite unstable, some times we need to fix the problem by ourselves 08:57:56 esp for few days ago :( 08:58:25 we can consider it an option for F 08:58:38 #info we could ask LF to ensure monitoring of test API (critical) for F - restricted access, isolated function. for the moment facing unstability and need manual restart/fix 08:58:42 #topic AoB 08:58:49 when almost all the features(web portal, testapi client) are finished 08:58:50 any topic you want to add/share 08:58:59 yes. ARM 08:59:07 ah yes good point 08:59:14 Alpine too? 08:59:43 #info ARM dockers - Several patches submitted 08:59:45 They also asked for a patch which shouldn't be merged as is. 09:00:05 #info dockerization shall respect good practices we put in places 09:00:26 Do I fix Jose's proposal? 09:01:20 https://gerrit.opnfv.org/gerrit/#/c/37761/ 09:01:34 if you have time..yes..Jose will be back on the 4th of Septemeber I think, armband project would like to go forward 09:01:40 I disagree that we patch staff in prepare_env 09:01:49 +1 09:01:55 that is part of the best practice... 09:02:00 They want to test for this week! 09:02:14 no IF aarch 64 in prepare_env ... 09:02:22 patch_file() 09:03:31 #depo(ENEA) 09:03:47 I am working on trying to get ARM arch being read 09:04:04 see update_config_file() or install_rally 09:05:13 ok it is already 11:05 09:05:34 I suggest depo and ollivier you continue on the irc chan 09:05:55 #endmeeting