08:01:12 <morgan_orange> #startmeeting Functest weekly meeting 29th of August
08:01:12 <collabot> Meeting started Tue Aug 29 08:01:12 2017 UTC.  The chair is morgan_orange. Information about MeetBot at http://wiki.debian.org/MeetBot.
08:01:12 <collabot> Useful Commands: #action #agreed #help #info #idea #link #topic.
08:01:12 <collabot> The meeting name has been set to 'functest_weekly_meeting_29th_of_august'
08:01:18 <morgan_orange> #topic call role
08:01:23 <morgan_orange> #info Morgan Richomme
08:01:38 <LindaWang> #info Linda Wang
08:01:58 <juhak> #info Juha Kosonen
08:02:15 <Debayan> #info Debayan Chaudhuri
08:02:19 <Shuya_ool> #info Shuya(OOL)
08:02:28 <morgan_orange> Hi Debayan welcome!
08:02:47 <OPNFV-Gerrit-Bot> valentin boucher proposed functest: More support of snaps for VNF  https://gerrit.opnfv.org/gerrit/39861
08:02:48 <morgan_orange> hi Shuya_ool ..we can see that a vnf topic is planned today :)
08:03:05 <morgan_orange> #info agenda for today: https://wiki.opnfv.org/display/functest/Functest+5.+Meeting
08:03:16 <boucherv> #info Valentin Boucher
08:03:17 <morgan_orange> let's start with the action points
08:03:30 <morgan_orange> #topic action points follow-up
08:03:38 <morgan_orange> #link http://ircbot.wl.linuxfoundation.org/meetings/opnfv-functest/2017/opnfv-functest.2017-08-22-08.01.html
08:03:46 <morgan_orange> #info AP1: morgan_orange ask valentin on https support for cloudify_ims test case
08:03:50 <morgan_orange> #info done during the meeting
08:03:57 <morgan_orange> #info AP2: morgan_orange contact sfc project to get the status and indicate we will disable first then remove beginning of September if no action
08:04:01 <Shuya_ool> morgan_orange:Hi Thankyou for helping
08:04:11 <morgan_orange> #info done, several patches done on sfc (get rid of sh script for tacker installation, functest shall be clean now)
08:04:19 <morgan_orange> #info AP3: morgan_orange contact D.McBride to report the need of a better control of the onboadred features/versions ..we coudl add Python version
08:05:04 <morgan_orange> #info done: mail thread in progress...not always easy to make understand that Functest mission is not to fix dependency problems of all the projects :)
08:05:06 <ollivier> not fully.
08:05:33 <morgan_orange> #info AP4: serena-zte clarify with Parser PTL why they forked the official projects => conenction Parser PTL to ollivier
08:05:41 <morgan_orange> no news? serena-zte?
08:06:06 <serena-zte> he is here now
08:06:12 <ollivier> sfc has to host the tacker utils (still in progress)
08:06:34 <morgan_orange> #info ollivier: sfc has to host the tacker utils (still in progress) - regarding AP2
08:06:54 <serena-zte> and he will be here soon
08:07:36 <morgan_orange> ok we can discuss parser integration in next slot
08:07:56 <shangxdy> Hi, i'm from parser team
08:08:05 <morgan_orange> #info parser discussion to be done in next topic
08:08:25 <morgan_orange> hi shangxdy let's finish the action point and discuss feature projects integration in next topic
08:08:30 <morgan_orange> shangxdy: welcome
08:08:39 <morgan_orange> #info AP5: juhak test alpine containers on Nokia/apex community labs
08:08:44 <juhak> #info done: healthcheck ok, smoke nok; tempest, rally and refstack fails
08:08:57 <shangxdy> I learn from serena that there some wonders in funectest about parser.
08:09:02 <morgan_orange> juhak: I assume that it is not linked to alpine => same results with ubuntu docker
08:09:17 <juhak> yes, that's right
08:09:19 <shangxdy> ok
08:09:32 <morgan_orange> shangxdy: yes we sent a mail to the contributors long time ago and did not get any feedback on our question
08:09:51 <morgan_orange> #info errors not linked to alpine (same error with ubuntu container)
08:09:54 <morgan_orange> juhak: it is on apex?
08:10:04 <juhak> yes, apex
08:10:06 <morgan_orange> Danube?
08:10:06 <shangxdy> Sorry,may be i missed the email.
08:10:17 <juhak> Danube 2.0
08:10:54 <morgan_orange> #info AP6: morgan_orange contact orchestra regarding orchestra_* test cases
08:11:02 <morgan_orange> #info not done will redo today, some issue with port 8082 on some PODs but last run on daisy shows that openbaton is properly installed
08:11:08 <morgan_orange> #action morgan_orange contact orchestra regarding orchestra_* test cases
08:11:20 <morgan_orange> anything you want to add regarding the action points?
08:12:08 <morgan_orange> #topic Euphrates status
08:13:02 <morgan_orange> #info some issues with the downloading of the images last week + issue with access to cirrios web site yesterday but everything seems OK now
08:13:29 <LindaWang> download.cirros-cloud.net is available now
08:13:42 <LindaWang> so maybe we should revert the url
08:13:48 <ollivier> Then we shoud revert the patch to use the official repo
08:13:52 <ollivier> Yes.
08:14:09 <LindaWang> ollivier: I will do that
08:14:13 <morgan_orange> LindaWang: shall we revert, plan a backup with artifact? trozet reported that we should log the fact that download failed
08:14:18 <ollivier> LindaWang: thank you
08:14:53 <ollivier> The issue is on jjobs which mute logs.
08:14:54 <morgan_orange> #action LindaWang revert from artifact to cirios web site (patch done yesterday to run CI despite cirrios web site weas no more answering)
08:15:20 <LindaWang> Another issue is each time in CI job, all the images will be downloaded
08:15:20 <morgan_orange> ollivier: we have some logs...we indicate that we start downloading :)
08:15:32 <LindaWang> It takes such long time
08:16:05 <morgan_orange> that is question for releng/helpdesk
08:16:29 <morgan_orange> #link http://testresults.opnfv.org/reporting2/display/master/functest/status-apex.html
08:16:32 <LindaWang> I guess  in CI the dir holding such images will be cleaned after each job is finished
08:16:43 <morgan_orange> #link http://testresults.opnfv.org/reporting2/display/master/functest/status-compass.html
08:16:50 <morgan_orange> #link http://testresults.opnfv.org/reporting2/display/master/functest/status-daisy.html
08:17:02 <morgan_orange> #link http://testresults.opnfv.org/reporting2/display/master/functest/status-fuel@x86.html
08:17:08 <morgan_orange> #link http://testresults.opnfv.org/reporting2/display/master/functest/status-fuel@aarch64.html
08:17:15 <morgan_orange> #link http://testresults.opnfv.org/reporting2/display/master/functest/status-joid.html
08:17:43 <morgan_orange> LindaWang: yes but it  could make sense to ask them if there is not a way to avoid that...
08:18:24 <ollivier> morgan_orange: I don't think it's printed by download.sh but by jjobs instead. I am checking
08:18:37 <morgan_orange> #action morgan_orange contact releng to see if we could not avoid cleaning downloaded images after each job (cache mechanism to save time on CI pipeline)
08:18:42 <morgan_orange> ollivier: yes it is the job
08:19:04 <ollivier> morgan_orange: the jjobs should add least print &2
08:19:14 <morgan_orange> https://git.opnfv.org/releng/tree/jjb/functest/set-functest-env.sh
08:19:39 <morgan_orange> L33 then we run download_images.sh
08:19:56 <morgan_orange> but if issue we just report fail without precising that it is due to the download
08:20:09 <ollivier> Our fault. if ${redirect} is /dev/null, everything is lost (see 2>&1)
08:20:33 <morgan_orange> #action ollivier patch set-functest-env.sh to indicate download failure issue
08:21:02 <morgan_orange> regarding the different scenarios for Euphrates
08:21:33 <morgan_orange> #info still early issues with apex scenarios: apparently trozetdetected a bug yesterday (discussion with Steven on api_check)
08:22:06 <morgan_orange> #info compass still issues on rally/refstack
08:22:24 <morgan_orange> #info verbose traces from tempest/rally/refstack in the CI
08:22:47 <LindaWang> The issue about rally in compass is about ceilometer.
08:23:09 <morgan_orange> #info LindaWang: issue about rally in compass is about ceilometer
08:23:10 <LindaWang> Compass is trying to fix that.
08:23:25 <morgan_orange> this suite has been added this summer by Martin
08:23:52 <morgan_orange> the tests need some adaptations on the service side
08:24:42 <morgan_orange> #info Daisy: stable results still some issue with Rally but could be similar than with compass
08:24:43 <LindaWang> yes, i have tested it in compass deployment locally with ceilometer available.  Several tests in ceilometer have passed.
08:25:01 <serena-zte> yes
08:25:13 <serena-zte> currently ceilometer is not support in daisy
08:25:27 <serena-zte> daisy team is working on the integration with ceilometer
08:25:34 <morgan_orange> #info for all the test cases: odl fail => we know why , domino => to be investigated but could be excluded from reporting
08:25:47 <ollivier> https://gerrit.opnfv.org/gerrit/#/c/40563/2/jjb/functest/set-functest-env.sh
08:27:08 <morgan_orange> #info fuel@x86 some progress on healthcheck/smoke: Doctor (disabled), Promise and Domino to be investigated/excluded from reporting
08:27:40 <morgan_orange> #info no run on aarch64: new docker generated last week, still some bug to be fixed
08:28:30 <morgan_orange> #info joid lxd scenario => issue with heat tests (investigation Narinder) Tempest/Rally/Refstack issue (like in Danube) + Promise/Domino to be excluded like Fuel
08:28:44 <morgan_orange> any comment?
08:29:11 <morgan_orange> #action morgan_orange disable Promise, Domino, Doctor from reporting (focus on healtcheck and smoke...)
08:29:30 <morgan_orange> you can see that no HA and noHA cenarios are distinguised in the reporting page
08:29:55 <morgan_orange> #topic VNF and Feature integration for Euphrates
08:30:05 <morgan_orange> #info vyos_vrouter patch merged today
08:30:10 <morgan_orange> thanks Shuya_ool
08:30:15 <morgan_orange> wait for CI run
08:30:40 <morgan_orange> as a reminder VNFs are enabled on the supposed most stable scenario for nosdn-nofeature
08:30:48 <ollivier> the container has been rebuilt.
08:30:54 <morgan_orange> Shuya_ool: in your patch you restricted the run to Fuel, any reason?
08:31:01 <morgan_orange> the other VNFs are tried on all the installers
08:31:32 <morgan_orange> and today the best chance to see a run is compass or daisy
08:31:38 <Shuya_ool> morgan_orange: because only test with fuel.
08:32:04 <morgan_orange> ok but is there any dependency to fuel or is it just because you tested it with fuel (as it was the only config you have)
08:32:22 <morgan_orange> if there is no dependency, I would suggest to remove it and try on compass/daisy/ ...
08:32:48 <morgan_orange> #info vEPC test tests in progress on community PODs
08:33:06 <morgan_orange> Debayan: maybe this case is new for several people, could you write a short describtion and the current status?
08:33:48 <Debayan> Sure Morgan - good morning to everyone ...
08:34:10 <OPNFV-Gerrit-Bot> valentin boucher proposed functest: Put vnf_test pass if more than 80% of vnf test result are OK  https://gerrit.opnfv.org/gerrit/40017
08:34:35 <Shuya_ool> morgan_orange: vrouter testcase is no dependency to fuel.
08:34:54 <Debayan> We have written a set of tests for OAI EPC using Juju as the deployment tool and ABot as the test driver
08:35:25 <Debayan> This is now integrated into Functest and we are testing it on the OPNFV Community Pod(Pod 3)
08:35:34 <morgan_orange> #action Shuya_ool remove dependency on fuel in testcases.yaml to try to run the testcase with all the installers
08:35:54 <morgan_orange> POD3 is Compass/Danube 3.0 hosted in Orange
08:36:06 <morgan_orange> #info Debayan We have written a set of tests for OAI EPC using Juju as the deployment tool and ABot as the test driver
08:36:21 <morgan_orange> #info Debayan  This is now integrated into Functest and we are testing it on the OPNFV Community Pod(Pod 3)
08:36:22 <Debayan> The testing is going fine and we should be ready to provide the updated code for review in a couple of days.
08:36:43 <morgan_orange> #info Debayan testing is going fine and we should be ready to provide the updated code for review in a couple of days.
08:36:50 <morgan_orange> sounds great first vEPC...
08:37:00 <ollivier> Does it require special requirements (we have to integrate it in Alpine as well)
08:37:29 <Debayan> No special requirements for vEPC testing ...
08:37:37 <morgan_orange> in theory MS6 was on the 25th of October (evidence of testing in CI) but I think we can be a little be flexible
08:37:50 <morgan_orange> I think we will see the requirements with the patch :)
08:37:51 <Debayan> 25th August I suppose ?
08:37:58 <morgan_orange> oops yes
08:38:23 <Debayan> Morgan - one question from my end ...
08:38:36 <ollivier> Yes. But as we integrate nodejs, ruby, python3, why not adding go...
08:38:58 <Debayan> We now have the interim code checked into our own public repo(forked off OPNFV Functest)
08:39:00 <morgan_orange> https://wiki.opnfv.org/display/SWREL/Euphrates?preview=/8689511/12388304/OPNFV%20Release%20%2522Euphrates%2522%20r9.pdf
08:39:12 <Debayan> Could we provide that for review ?
08:39:55 <ollivier> Just for reading. We must use gerrit (pull/request forbidden).
08:40:11 <morgan_orange> I think it would more convenient to put your code on OPNFV repo
08:40:38 <Debayan> OK thx Morgan/Ollivier for clarification ..
08:40:42 <morgan_orange> for reading it can be OK but it will be quicker to directly push your code in OPNFv repo and review it though gerrit
08:40:56 <morgan_orange> #action Debayan put the code for review on OPNFV repo
08:41:06 <morgan_orange> and it will be good for your stats :)
08:41:24 <ollivier> not sure. or create 2000 patchsets for 1 change
08:42:35 <morgan_orange> or create 100 Jiras...no just kidding..I BTW had to make come cleaning on the Jira..there were still Danube 2.0 open..so no worry if you saw some Jira movement recently
08:42:52 <morgan_orange> let's discuss a bit of parser then
08:43:05 <ollivier> Yes.
08:44:28 <shangxdy> Currently, parser team are working on the parser package setup
08:44:52 <shangxdy> I think we will finish it this week
08:45:11 <ollivier> We would like to understand why you're forking OpenStack project.
08:45:49 <ollivier> Other testcases (which we integrate) require the offical ones.
08:45:50 <shangxdy> Because there four subprojects in parser, so it may costs some time to solve it
08:46:43 <shangxdy> olivier, because nfv parser is  based on openstack tosca-parser for some specific features
08:46:49 <ollivier> We can't install the offical ones and your forks. mainly because You simply renames the python packages not the namespaces.
08:47:22 <ollivier> Yes  then why not add a dependency? Why did you need to fork them?
08:47:54 <shangxdy> Currently which project will use openstack toscaparser?
08:48:35 <ollivier> Domino ? I can check. But without any need, we should rely on the official packages.
08:49:00 <shangxdy> Because there many modifications in nfv parser
08:50:32 <shangxdy> We will submit most of the modification in nfv parser to openstack, but it is delayed often.
08:51:50 <morgan_orange> as it is already 10:50 I suggest to continue offline. shangxdy Thanks to Cedric we are now managing dependencies. When integrating 3rd party we need to get the feedback as upstream first shall be the rules. It is not always possible but it is hard to guess and if we want to keep consistant we need to take care of that.
08:51:52 <shangxdy> Additionaly, there are some feature in nfv pareser won't submitted to openstack because the features are not common requierments from the openstack side.
08:51:54 <ollivier> Yes but I'm pretty sure we can install both heat-translator  tosca-parser
08:51:58 <ollivier> we cannot
08:52:26 <morgan_orange> #action ollivier shangxdy clarify Parser requirement for functest by mail
08:52:54 <morgan_orange> #topic Functest API
08:53:02 <morgan_orange> LindaWang: anything you want to add
08:53:06 <morgan_orange> i saw a new patch
08:53:15 <shangxdy> Why does it is necessary to install both heat-translator  tosca-parser?
08:53:19 <LindaWang> Basic functions have finished.
08:53:30 <ollivier> because it's required by another projects.
08:53:45 <LindaWang> But we should launch the Restapi once the container is run
08:53:50 <morgan_orange> #info basic functions are finished for Functest API
08:53:56 <shangxdy> Which project are using heat-translator  tosca-parser?
08:53:59 <morgan_orange> #info pending patch to be reviewed
08:54:11 <morgan_orange> #link https://gerrit.opnfv.org/gerrit/#/c/40343/
08:54:19 <morgan_orange> shangxdy: domino
08:54:34 <LindaWang> Either nginx and uwsgi is leveraged or simply add the CMD ["functest_restapi"] in the Dockerfile
08:54:54 <ollivier> Yes domino
08:55:23 <ollivier> I would prefer not to override cmd.
08:56:03 <shangxdy> For functionality, nfv parser is compatible with openstack toscaparser except for some new features in nfv.
08:56:11 <ollivier> By default we should prepare/run the default config.
08:56:13 <LindaWang> Yes, it will infulence alpine based image, right?
08:56:28 <ollivier> yes at least for Alpine.
08:56:36 <morgan_orange> ok good point
08:56:39 <morgan_orange> and good transition
08:56:50 <morgan_orange> #topic Alpine XCI
08:57:10 <shangxdy> May be  it's a solution for parser to run virtual env
08:57:12 <morgan_orange> #info several pending patch to use functest-healtcheck for XCI and to run functest on XCI
08:57:53 <shangxdy> Can functest support that?
08:58:08 <ollivier> You should first fix your requirements.txt
08:58:16 <ollivier> tox is not enough
08:59:13 <morgan_orange> #info alpine run on Huawei virtual 5 say Success but logs are strange https://build.opnfv.org/ci/view/functest/job/functest-alpine-compass-huawei-virtual5-daily-master/lastSuccessfulBuild/console
08:59:26 <morgan_orange> #undo
08:59:26 <collabot> Removing item from minutes: <MeetBot.ircmeeting.items.Info object at 0x3554c50>
08:59:49 <morgan_orange> run in progress https://build.opnfv.org/ci/view/functest/job/functest-alpine-compass-huawei-virtual5-daily-master/23/console
08:59:52 <ollivier> bug in POD: gsutil?
09:00:03 <LindaWang> gsutil is not installed yet
09:00:14 <morgan_orange> gsutil is used to push results to artifact
09:00:31 <shangxdy> In addition, in the next release parser will provide microservice, then all will be ok, but in release E, i suggest to keep the current sate.
09:00:34 <morgan_orange> #topic internship
09:00:43 <morgan_orange> #info several candidates for all our internship proposals
09:00:45 <LindaWang> Tianwei is helping install and config gsutil
09:00:47 <LindaWang> https://build.opnfv.org/ci/view/functest/job/functest-alpine-compass-huawei-virtual5-daily-master/23/console
09:00:56 <morgan_orange> #info Jose is back on the 31st... we will need to answer...
09:00:57 <ollivier> shangxdy: Functest is not desinged to fix parser. Let's discuss offline how we can proceed
09:01:13 <ollivier> or atfer the meeting
09:02:03 <shangxdy> Currently, parser is the last testcase, so it won't affect other projects.
09:02:03 <morgan_orange> #action morgan_orange prepare list of students/ internships
09:02:27 <shangxdy> ollivier, ok
09:02:27 <morgan_orange> shangxdy: the framework is generic we cannot think like that ... :)
09:02:55 <ollivier> Normally if you create a new package, you should use a related namespaces. Here you simply rename the name not the contents
09:03:01 <morgan_orange> in French we could say "après moi le déluge" that can be translated "after me, i do not care"
09:03:07 <morgan_orange> #topic AoB
09:03:13 <morgan_orange> any topic you want to raise for today
09:03:19 <morgan_orange> #info jose will be back on the 31st
09:04:21 <morgan_orange> #info demo valentin on vIMS functionnal + stress tests (hopefully alpine integration will be finalized) next Thursday during the weekly testing meeting
09:04:42 <morgan_orange> #info this demo is based on the work done during plugfest and during summer...
09:05:16 <morgan_orange> ok thanks for joining, still lots of troubleshooting/integration ahead
09:05:22 <morgan_orange> have a good week
09:05:24 <morgan_orange> #endmeeting