10:01:21 <morgan_orange> #startmeeting functest weekly oct 27th 10:01:21 <collabot> Meeting started Tue Oct 27 10:01:21 2015 UTC. The chair is morgan_orange. Information about MeetBot at http://wiki.debian.org/MeetBot. 10:01:21 <collabot> Useful Commands: #action #agreed #help #info #idea #link #topic. 10:01:21 <collabot> The meeting name has been set to 'functest_weekly_oct_27th' 10:01:28 <morgan_orange> #info Morgan Richomme 10:01:34 <juhak> #info Juha Kosonen 10:01:40 <jose_lausuch> #info Jose Lausuch 10:01:49 <Valentin_Orange> #info Valentin Boucher 10:01:50 <jose_lausuch> I hope we have more participants than last week (only me) 10:01:51 <jose_lausuch> :) 10:02:32 <morgan_orange> I saw that (read the minutes) 10:02:45 <morgan_orange> #info agenda: https://wiki.opnfv.org/functest_meeting 10:02:57 <morgan_orange> #topic agenda bashing and action points 10:03:43 <morgan_orange> #info action point still to be completed (presentation for the summit) 10:03:53 <jose_lausuch> yes 10:04:00 <morgan_orange> #info pres started, i will send an updtae today or tomorrow (deadline 30th) 10:04:02 <jose_lausuch> i have a question here 10:04:24 <jose_lausuch> it has to be presented in pdf and 19:9 10:04:38 <jose_lausuch> is the javascript thing already usable? 10:05:07 <morgan_orange> it has not been merged and i am not sure it will 10:05:16 <morgan_orange> v=but I think it can be used if we want.. 10:05:34 <viktor_nokia> #info Viktor Tikkanen 10:05:37 <jose_lausuch> I was thinking of doing it for now in powerpoint... since I got the template to use 10:05:46 <morgan_orange> what do you prefer ppt template or reveal format? 10:05:58 <jose_lausuch> I prefer reveal 10:06:04 <jose_lausuch> but its about time to do it 10:06:15 <jose_lausuch> maybe it takes longer in reveal? Since we have to get used 10:06:19 <jose_lausuch> and also adapt the tempates 10:06:21 <jose_lausuch> icons, and so on 10:06:29 <morgan_orange> it is pretty quick in reveal 10:06:40 <morgan_orange> the longest part is to get the picture you want to display 10:06:53 <morgan_orange> in html or markdown it is very quick to do 10:06:54 <jose_lausuch> it has to be converted to pdf inn the end 10:07:04 <morgan_orange> yes and reveal is not perfect for that 10:07:16 <morgan_orange> but we could also host the pres on a web server 10:07:19 <jose_lausuch> I saw :) 10:07:27 <morgan_orange> ok let's powerpointize for the moment 10:07:41 <jose_lausuch> I will use ppoint for functest in depth 10:07:48 <jose_lausuch> maybe we can try the other presentation we have in reveal 10:07:54 <morgan_orange> ok 10:08:05 <morgan_orange> I already started the reveal one on the dashboard 10:08:08 <jose_lausuch> ok 10:08:18 <morgan_orange> #topic Sprint2 10:08:38 <morgan_orange> #link https://jira.opnfv.org/secure/RapidBoard.jspa?rapidView=59&useStoredSettings=true 10:08:58 <morgan_orange> #info 9 days remaining (basically the end of this sprint will take place during the design summit 10:09:02 <jose_lausuch> pbandzi_: ping 10:09:08 <pbandzi_> pong 10:09:22 <jose_lausuch> hi 10:09:23 <jose_lausuch> https://jira.opnfv.org/browse/FUNCTEST-80 10:09:26 <jose_lausuch> do you agree with this? 10:09:32 <jose_lausuch> can you have a look later? 10:09:40 <morgan_orange> are you all Ok with the tasks assigned to you? 10:10:29 <jose_lausuch> yes 10:10:42 <jose_lausuch> there are a lot of tasks still not started 10:10:42 <pbandzi_> jose_lausuch: I'll look at it 10:10:46 <jose_lausuch> thanks pbandzi_ 10:10:47 <morgan_orange> juhak: viktor_nokia pbandzi_ quinglong 10:10:50 <jose_lausuch> viktor_nokia: https://jira.opnfv.org/browse/FUNCTEST-72 10:11:00 <viktor_nokia> from my side there is no progress with FUNCTEST-72 (refactor tempest) 10:11:02 <jose_lausuch> can we discuss this later? 10:11:06 <viktor_nokia> OK 10:11:27 <jose_lausuch> morgan_orange: can we still create epics? 10:11:33 <morgan_orange> yes 10:11:34 <jose_lausuch> there are some tasks that are not included in an epic 10:11:42 <jose_lausuch> for example there is no epic for docker 10:11:44 <morgan_orange> and I must rename them 10:11:53 <morgan_orange> FUNCTEST_X is not very explicit 10:11:59 <juhak> no progess with FUNCTEST-71 either 10:11:59 <jose_lausuch> :) 10:12:00 <morgan_orange> but we can create new ones 10:12:31 <morgan_orange> juhak: any progress/question regarding for https://jira.opnfv.org/browse/FUNCTEST-71 ? 10:13:07 <juhak> no progress yet 10:13:14 <morgan_orange> pbandzi_: still Ok to look at the robot based testcases we could import from ODL? 10:13:25 <morgan_orange> I just opened the github pages 10:13:33 <morgan_orange> there are quiete a lot of tests 10:13:52 <morgan_orange> #action morgan_orange rename epics to be more explicit 10:14:19 <pbandzi_> morgan_orange: still ok - i have it on my todo list :) 10:14:22 <morgan_orange> #info Jose added new tasks (dockerization + running functest offline) 10:14:23 <jose_lausuch> I can give you a hand with that morgan_orange 10:15:35 <morgan_orange> ok cool pbandzi_, do not hesitate (as you are also buzy by the management of LF infra) of we need additional resources to address this selection (from ODL project) 10:15:47 <jose_lausuch> pbandzi_: I will create a task to take the IPs and ports from the config_functest.yaml file 10:16:02 <jose_lausuch> #info docker wiki updated https://wiki.opnfv.org/functest_docker 10:16:13 <morgan_orange> last week Quinglong said that he was able to run the tests on Huawei platform 10:16:21 <morgan_orange> so regarding the board 10:16:26 <jose_lausuch> #info possible now to run only a set of tests [vping,odl,tempest,rally] instead of all of them 10:16:28 <morgan_orange> I am confident for this sprint 10:16:57 <morgan_orange> #info on vIMS soon (working with Valentin_Orange today), we were able to deploy cloudify and met a timeout issue 10:17:52 <morgan_orange> any question on Sprint 2? 10:17:57 <jose_lausuch> morgan_orange: can you info the latests news about the dashboard ? 10:18:05 <jose_lausuch> discussionswith marion 10:18:07 <morgan_orange> yes 10:18:17 <morgan_orange> #info test collection API has been updated 10:18:51 <morgan_orange> #info a new method called dashboard has been added 10:19:18 <morgan_orange> #info this method retrieved "dashboard ready" data and will be used by LF web team in charge of the test dashboard 10:20:00 <morgan_orange> #info for the moment we defined 2 testcases "dashboard ready" 10:20:01 <morgan_orange> #link http://213.77.62.197/dashboard?project=functest&case=Tempest 10:20:07 <morgan_orange> #link http://213.77.62.197/dashboard?project=functest&case=vPing 10:20:33 <morgan_orange> #info for each testcase you can use filter (installer, version, pod, time ) 10:20:40 <jose_lausuch> great 10:20:59 <morgan_orange> #link http://213.77.62.197/dashboard?project=functest&case=vPing&installer=fuel&version=master&pod=opnfv-jump-2 10:21:25 <morgan_orange> #info we store the results only once (the raw results of the tests) 10:21:53 <morgan_orange> #info then in releng through a method called functest2Dashboard.py we post processed the raw results to provide dashboard ready graphs 10:22:07 <jose_lausuch> ok 10:22:12 <morgan_orange> #link https://wiki.opnfv.org/collection_of_test_results 10:22:35 <jose_lausuch> it is well explained 10:22:40 <jose_lausuch> thanks for that wiki 10:22:47 <morgan_orange> pbandzi_: for odl you may use the APi to push the results of the tests as they are today 10:23:14 <pbandzi_> ok 10:23:18 <morgan_orange> then I could add a method in functest2Dashboard that will graph duration of the test = f(time) and nb test run/nb test failed = f(time) 10:23:19 <morgan_orange> ... 10:23:39 <morgan_orange> #info Guy Rodrigue worked on an automatic doc (swagger like) 10:23:46 <morgan_orange> because wiki are difficult to maintain 10:24:04 <morgan_orange> #info we planned to work with Valentin_Orange to provide suwh results for vIMS testcase 10:24:39 <morgan_orange> #info LF foundation announced a version of the dashboard on the 9th of November 10:25:11 <morgan_orange> #info presentation planned during the summit 10:25:37 <morgan_orange> #info yardstick contributor shall also work on its yardstick2Dashboard.py file 10:25:45 <morgan_orange> ok with the dashboard? 10:26:12 <jose_lausuch> yes 10:26:16 <morgan_orange> there are still work (authentication, migration to target DB ...) but it is already possible to play with the APi and get dashboard ready info 10:26:17 <jose_lausuch> good work there 10:26:27 <jose_lausuch> fair enough 10:26:34 <jose_lausuch> I had an idea 10:26:41 <jose_lausuch> that we could use, all the test projects 10:26:45 <jose_lausuch> that report results 10:27:06 <jose_lausuch> when calling the script to run test X, we could add a flag to tell the script to report the results to the DB 10:27:12 <jose_lausuch> default=no 10:27:35 <morgan_orange> i think that you already implemented it? through the -r option when running tempest of vPing? 10:27:37 <jose_lausuch> we need to add it as an option 10:27:39 <morgan_orange> you mean a global option? 10:27:45 <jose_lausuch> for all the scripts 10:28:01 <morgan_orange> ok yes we could add such parameters 10:28:13 <jose_lausuch> in functest it is easy, but all should take it into account 10:28:18 <jose_lausuch> yardstick also 10:28:21 <morgan_orange> ok with default =no as theoretically the results shall be pushed from Pharos labs... 10:28:39 <morgan_orange> good transition to next point 10:29:13 <jose_lausuch> ok 10:29:15 <morgan_orange> #action jose_lausuch add a flag to tell the script to report the results to the DB and tell other test projects (info to share during the summit) 10:29:32 <jose_lausuch> :) 10:29:33 <morgan_orange> #topic Running Functest in CI in new platforms 10:30:06 <morgan_orange> I am confident for Sprint 2, but I feel some concern for post Sprint 2 and especially with the integration of the feature tests and the CI 10:30:21 <morgan_orange> so far we run automatically our tests only on LF POD2 (now fuel Master) 10:30:31 <morgan_orange> BTW there is an issue regarding CI 10:31:00 <morgan_orange> https://build.opnfv.org/ci/view/functest/job/functest-fuel-master/ 10:31:20 <morgan_orange> #action morgan_orange jose_lausuch look why CI is now failing on LF POD 2 (manual run of the container is OK) 10:31:28 <jose_lausuch> some problems with docker 10:31:32 <jose_lausuch> I'll take a look 10:31:37 <jose_lausuch> start.sh does not exist any more 10:31:42 <jose_lausuch> I will report it as a bug 10:31:50 <morgan_orange> I think it is important to ask Pharos resources 10:32:24 <morgan_orange> we need at least as many labs as installers assuming that their configuration may change (SDN controllers, specific OVS or KVM tuning to support feature tests) 10:32:53 <morgan_orange> fdegir: and trevor are working hard to provide a consisten view of the available resources 10:33:12 <jose_lausuch> thats inmportant 10:33:16 <morgan_orange> but we should now be able to run functest on at least 3 different PODs 10:33:34 <morgan_orange> (when cash on the installation like yesterday) => no more possible to work 10:34:00 <morgan_orange> regarding the candidates labs I see only (for the moment) Ericsson, Intel and Huawei 10:34:25 <morgan_orange> Huwaei we may ask Quinglong (as, as far as I understand, he is already performing tests for compass/onos) 10:34:46 <morgan_orange> #action morgan_orange ask Quinglong to deploy full functest on huawei lab trhough jenkins 10:34:59 <morgan_orange> jose_lausuch: could you on E/// side? 10:35:13 <jose_lausuch> I will ask for resources 10:35:38 <morgan_orange> #action morgan_orange ask trevor for Resources on intel for apex 10:35:55 <morgan_orange> if we could have already Apex, fuel, compass, only joid will miss 10:36:06 <morgan_orange> we are installing a joid based solution in one orange Pods 10:36:18 <morgan_orange> then we will cover all the installer, which will be a good point... 10:36:25 <jose_lausuch> es 10:36:26 <jose_lausuch> yes 10:36:42 <morgan_orange> #action jose_lausuch ask for resources (run functest one one of the E/// pod) 10:37:40 <morgan_orange> and for the change of config, for the moment it is fuzzy for me. I assumed we have default combination compass/Onos, fuel/ODL ...but it is not fully clear 10:38:03 <jose_lausuch> maybe fuel will support o.contrail soon 10:38:14 <morgan_orange> ok 10:38:34 <morgan_orange> any question on the CI, needs for PODs.. 10:39:27 <morgan_orange> #topic Design Summit and OPNFV Summit 10:39:28 <jose_lausuch> nop 10:39:48 <jose_lausuch> possible jam on tuesday :p 10:40:13 <morgan_orange> #info functest will be active. 1h tutorial from josé, presentation of the test collection API, and several presentations during the Summit 10:40:41 <morgan_orange> #info according to the dependencies, I think we need to meet at least 5 other projects 10:40:50 <morgan_orange> #link https://wiki.opnfv.org/feature_test_project_matrix 10:41:09 <jose_lausuch> yes 10:41:10 <morgan_orange> #info some high level Jira have been created, but they will surely not cover the needs 10:41:14 <jose_lausuch> there are project sessions 10:41:19 <morgan_orange> yes 10:41:20 <jose_lausuch> breakout sessions 10:41:23 <jose_lausuch> we could use them 10:41:36 <morgan_orange> yes 10:41:47 <morgan_orange> I think neither viktor_nokia nor juhak will attend 10:41:54 <morgan_orange> pbandzi_: will you be there? 10:41:54 <viktor_nokia> no 10:41:57 <jose_lausuch> pitty 10:41:58 <juhak> nope 10:42:22 <jose_lausuch> pbandzi_: you? 10:42:32 <pbandzi_> morgan_orange : i ll be on summit on November 10:42:36 <jose_lausuch> great 10:42:37 <morgan_orange> Valentin_Orange: and Guy Rodrigue will not be there as well 10:42:48 <pbandzi_> not sure if you talk about that :) 10:42:52 <morgan_orange> ok Quinglong may also be there 10:43:40 <morgan_orange> so we will have to discuss with Copper, Doctor, ONOS, OVNO (Open Contrail), PolicyTest and Promise during the sessions to precise our Jira and see if we can help or not 10:44:11 <morgan_orange> #action morgan_orange prepare the Summit, contact PTL to discuss functest implication in Doctor, Copper, ONOS, OVNO, Policytest and promise 10:44:26 <morgan_orange> do you see othe stuff we must discuss during the Summit? 10:44:36 <morgan_orange> Open Mano? 10:45:02 <jose_lausuch> maybe 10:45:03 <viktor_nokia> BTW, our Doctor expert (Tomi Juvonen) will attend the summit 10:45:05 <morgan_orange> I know that China Mobile is interested to have a session on orrchestrator (Open O) 10:45:19 <jose_lausuch> but... we got some insights already 2 weeks ago in dusseldorf 10:45:24 <morgan_orange> #info Tomi Juvonen (Doctor) will attend the summit 10:45:35 <jose_lausuch> about OpenMano 10:45:36 <jose_lausuch> hehe 10:45:50 <jose_lausuch> we can discuss it with ChinaMobile.. 10:45:59 <morgan_orange> my view regarding orchestrator (for the moment), Functest is using orchestrator to ease the run of tests 10:46:39 <morgan_orange> we are not integrating any orchestrator to OPNFV (it is out of scope for the moment) 10:46:50 <morgan_orange> just use them as tool (like Cloudify for vIMS testcase) 10:47:28 <morgan_orange> for B-Release we did not plan to add new testcases (Open Mano with pktgen VNF, Open X with testcase Y) 10:47:48 <morgan_orange> but we may start an exploring discussion on orchestrator during the Summit 10:47:59 <jose_lausuch> yes 10:48:00 <morgan_orange> what is your view on the different orchestrators? 10:48:20 <jose_lausuch> same as you 10:48:27 <jose_lausuch> since they are out of scope for release B 10:48:31 <morgan_orange> #info functest not dealing with orchestrator, just using some to ease the launch of tests 10:48:35 <jose_lausuch> we use in our own for the needed tests 10:48:48 <jose_lausuch> for the future we might have to handle it differently 10:49:34 <morgan_orange> #info ok for an exploring discussion during the Summit, see if for > C release, functest could propose one (still to ease the functional testing not as part as OPNFV) 10:50:19 <morgan_orange> juhak: viktor_nokia pbandzi_ j Valentin_Orange ..feel free to share your view on that (mailing thread) 10:50:29 <viktor_nokia> ok 10:50:37 <juhak> will do 10:51:00 <morgan_orange> so jose_lausuch SFC and BGP VPN will be managed by yardstick (my understanding) 10:51:52 <morgan_orange> #info Summit, jam session organization in progress...feel free to join :) 10:52:17 <jose_lausuch> ok 10:52:21 <morgan_orange> #topic Possible expectations for Nokia open lab 10:52:30 <viktor_nokia> a couple of words regarding (future) nokia open lab, should I info? 10:52:36 <morgan_orange> yes 10:52:50 <viktor_nokia> #info Nokia is seriously considering establishment of an open lab 10:52:58 <viktor_nokia> #info for Functest (and possibly some other OPNFV projects) testing purposes 10:53:10 <viktor_nokia> #info If you have any expectations regarding HW, network topology, access methods etc. please contact e.g. me or Juha 10:53:28 <viktor_nokia> in other words, we are trying to figure out what's the best way how we can help OPNFV community :) 10:53:35 <jose_lausuch> sounds good 10:53:37 <morgan_orange> good news 10:53:39 <morgan_orange> it will be connected to CI? 10:53:42 <viktor_nokia> or should we contact pharos guys next?... 10:53:56 <jose_lausuch> I think you should contact them, Trevor for example 10:54:03 <morgan_orange> and fdegir 10:54:03 <viktor_nokia> ok 10:54:08 <jose_lausuch> it would be nice if its connected to CI, as for LF labs 10:54:10 <jose_lausuch> and orange ones 10:54:15 <viktor_nokia> morgan_orange: not sure yet, possibly 10:54:52 <jose_lausuch> for HW specifications you can contact Aric maybe? 10:54:55 <viktor_nokia> we can discuss this (CI connection)later 10:54:58 <jose_lausuch> any hw in mind? 10:55:18 <jose_lausuch> hp, cisco,dell blades/servers? 10:55:33 <viktor_nokia> possibly HP or Nokia's own 10:55:45 <jose_lausuch> ok 10:55:57 <juhak> is there any info or guideline about the process for connecting a lab to CI? 10:56:12 <morgan_orange> OK I think in mid term it is important to have this connection to CI and to keep as transparent as possible on the results. It is not easy... same issues internally but at the end it will be a great added value for all the community 10:56:27 <morgan_orange> juhak: oh yes there is a good wiki page for that 10:57:38 <morgan_orange> I will find it (connection slow at the moment) and send it to you 10:57:38 <juhak> ok, good 10:57:49 <morgan_orange> #action morgan_orange send procedure for CI connectivy to juhak 10:58:11 <morgan_orange> pitty that you could not join the Summit 10:58:19 <viktor_nokia> :( 10:58:49 <morgan_orange> jose and myself are looking for a crazy finnish drummer or bas player or singer ... 10:59:04 <juhak> heh... 10:59:05 <jose_lausuch> hehe 10:59:07 <jose_lausuch> yes 10:59:31 <viktor_nokia> hard rock hallelujah 10:59:38 <morgan_orange> #topic AoB 10:59:49 <juhak> still waiting for OPNFV LF lab access 10:59:49 <morgan_orange> any other info to share? 11:00:01 <jose_lausuch> https://etherpad.opnfv.org/p/jam_session_summit 11:00:04 <morgan_orange> I think this time slot is more convenient for most of the contributors 11:00:19 <viktor_nokia> definitely... 11:00:51 <morgan_orange> ok let's keep the uS next week and if no US contributors we will use only the Asian/European one slot 11:01:11 <jose_lausuch> ok 11:01:20 <jose_lausuch> last week was sad 11:01:27 <jose_lausuch> It was like a monologue 11:01:35 <jose_lausuch> hehe 11:01:35 <morgan_orange> juhak: I saw some mails around, they want to setup a procedure but i can try to ask one more time 11:01:55 <morgan_orange> #action morgan_orange checks why LF access still not possibfor juhak (N weeks..) 11:02:12 <juhak> ok, thanks 11:02:21 <morgan_orange> if no more question we can stop now 11:02:29 <jose_lausuch> yes, its time 11:02:49 <morgan_orange> see you next week and enjoy your Jiras 11:02:52 <morgan_orange> #endmeeting