08:01:16 <jose_lausuch> #startmeeting Functest weekly meeting November 15th 2016
08:01:16 <collabot`> Meeting started Tue Nov 15 08:01:16 2016 UTC.  The chair is jose_lausuch. Information about MeetBot at http://wiki.debian.org/MeetBot.
08:01:16 <collabot`> Useful Commands: #action #agreed #help #info #idea #link #topic.
08:01:16 <collabot`> The meeting name has been set to 'functest_weekly_meeting_november_15th_2016'
08:01:20 <HelenYao> jose_lausuch: hey
08:01:28 <jose_lausuch> #chair morgan_orange, HelenYao, SerenaFeng
08:01:28 <collabot`> Current chairs: HelenYao SerenaFeng jose_lausuch morgan_orange
08:01:33 <jose_lausuch> #topic role call
08:01:36 <jose_lausuch> #info Jose Lausuch
08:01:39 <morgan_orange> #info Morgan Richomme
08:01:39 <CristinaPauna> #info CristinaPauna (ENEA)
08:01:40 <SerenaFeng> #info SerenaFeng
08:01:42 <juhak> #info Juha Kosonen
08:01:43 <jose_lausuch> pound in your names
08:01:49 <HelenYao> #info Helen Yao
08:01:53 <jose_lausuch> HelenYao: hey :)
08:01:55 <hideyasu> #info hideyasu_ool
08:02:03 <hujie> #info Jie Hu (escalator team)
08:02:06 <jose_lausuch> #info agenda for today here: https://wiki.opnfv.org/display/functest/Functest+Meeting#FunctestMeeting-15/11(8UTC)
08:02:30 <jose_lausuch> #info previous minutes: http://ircbot.wl.linuxfoundation.org/meetings/opnfv-functest/2016/opnfv-functest.2016-11-08-08.01.html
08:03:06 <jose_lausuch> #chair juhak
08:03:06 <collabot`> Current chairs: HelenYao SerenaFeng jose_lausuch juhak morgan_orange
08:03:09 <jose_lausuch> #topic Action point follow-up
08:03:41 <jose_lausuch> #info action: HelenYao check what are the versions of the OS python clients related to a stable version of OS (starting with Mitaka)
08:03:57 <jose_lausuch> I think it is related to this other one
08:03:57 <jose_lausuch> #info action: HelenYao detect with lines in the Dockerfile are altering the OS python clients (i.e. tempest, rally, kingbird, )
08:04:09 <jose_lausuch> #info we will discuss it later, but I created this page showing that info: https://wiki.opnfv.org/display/functest/OpenStack+python+clients
08:04:22 <jose_lausuch> we have a topic for this discussion
08:04:33 <morgan_orange> #info AP harcoded version of ruby in docker: I contacted valentin, it shall be possible to remove ref to hardcoded ruby version even if it is direcly inherited from Metaswitch test suite, so to be tested...but as we are not running vims a lot (planned for weekly job only), we can remove the dependecy
08:05:02 <jose_lausuch> morgan_orange: ok, can I action you to remove it?
08:05:11 <morgan_orange> #info patch for the new structure submitted and merged, master branch is now using the new structure
08:05:13 <morgan_orange> jose_lausuch: yes
08:05:28 <jose_lausuch> #action morgan_orange remove ref to hardcoded ruby version
08:05:43 <morgan_orange> #info note it was just a basic structure without any changes in the path management and/or file renaming
08:06:03 <jose_lausuch> but that was the task about
08:06:13 <jose_lausuch> after some troubles, we have it running in CI
08:06:15 <jose_lausuch> so good job
08:06:24 <jose_lausuch> #info action: HelenYao SerenaFeng morgan_orange and all check docker slicing proposal and provide comments
08:06:43 <HelenYao> i left the comment before the meeting
08:06:50 <jose_lausuch> well, I have an idea for an alternative to docker remote api: docker compose
08:07:10 <morgan_orange> sorry not a lot of comment from my side, I just discussed offline with Cedric which was thinking to use something that looks better than Ubuntu for what we want to achieve
08:07:12 <jose_lausuch> we could run containers in parallel to run parallel test
08:07:29 <jose_lausuch> #action HelenYao jose_lausuch SerenaFeng check  if things could work with docker compose
08:07:49 <jose_lausuch> morgan_orange: yes, he proposed another OS
08:07:54 <jose_lausuch> is cedric in the office today?
08:08:06 <morgan_orange> jose_lausuch: the thing with // testing -especially for smoke - if it works you saved time, if it fails you have to find the root cause
08:08:18 <HelenYao> if we want to deploy the container at the same time, docker compose is one solution
08:08:33 <HelenYao> docker remote api could be used to deploy container on demand
08:08:34 <jose_lausuch> you can create dependencies between containers as well
08:08:40 <jose_lausuch> for some tests, it makes sense parallelization
08:08:43 <jose_lausuch> for others it doesnt
08:08:45 <morgan_orange> jose_lausuch: yes - I planned to have lunch with him today. Unfortunately he was assigned with internal project at least until beginning of december
08:08:50 <jose_lausuch> for the blocking tests, we can setup dependencies
08:09:00 <jose_lausuch> for the other (tempest/rally/fetures) we can run in parallel
08:09:07 <jose_lausuch> but we need to change the openstack cleanup logic
08:09:13 <jose_lausuch> otherwise we will wipe everything
08:09:19 <jose_lausuch> oh, he is here
08:09:21 <jose_lausuch> ollivier: hi
08:09:24 <morgan_orange> for example on tempest we moved to serial , it fixed some issues that were not easy to troubleshoot
08:09:34 <ollivier> jose_lausuch: hello
08:10:01 <HelenYao> is the decision whether docker slicing is to be done or not settled?
08:10:01 <jose_lausuch> ollivier: I have proposed to have a look at docker compose to see if we could benefit from the slicing, instead of remote api
08:10:12 <jose_lausuch> ollivier: what OS did you propose for the container instead of ubuntu?
08:10:41 <jose_lausuch> HelenYao: it is not settled, it's a difficult decission I think, and we need to have a prototype and demo it to see if it works as expected
08:10:43 <HelenYao> if yes, I don't need to investigate virtualenv further and there is a patch about virtualenv if you are interested
08:10:55 <ollivier> jose_lausuch: Alpine? https://hub.docker.com/_/alpine/
08:11:01 <jose_lausuch> HelenYao: virtualenv for rally could make sense though
08:11:07 <jose_lausuch> ollivier: what do we gain with Alpine?
08:11:26 <ollivier> It's built around musl libc
08:11:34 <ollivier> space
08:11:38 <jose_lausuch> ok
08:11:43 <jose_lausuch> we can try
08:11:46 <jose_lausuch> to save some MB
08:11:57 <jose_lausuch> ubuntu base image takes around 200 MB I think
08:12:07 <ollivier> yes there is no rush but we should have a look too it
08:12:07 <morgan_orange> according to the page Alpine 16MB versus Ubuntu 232MB
08:12:21 <morgan_orange> but need to see if we can use it for our needs...
08:12:23 <ollivier> we should save space for every package
08:12:31 <jose_lausuch> would that change a lot the way we install things?
08:12:44 <ollivier> if it stop installing docs ...
08:12:53 <jose_lausuch> docs?
08:13:24 <jose_lausuch> #info ollivier proposes to use Alpine https://hub.docker.com/_/alpine/  (~16MB base image) instead of ubuntu (~200MB base image)
08:13:36 <ollivier> when you install a deb, there are other files than libs and binaries. For instance docs locales
08:13:46 <jose_lausuch> ok
08:14:05 <jose_lausuch> #action jose_lausuch HelenYao have a look at building the docker image with Alpine
08:14:13 <jose_lausuch> ollivier: also you if you have time :)
08:14:15 <jose_lausuch> ok
08:14:21 <jose_lausuch> we didn't start a topic for this
08:14:30 <jose_lausuch> let's start with the first topic
08:14:31 <jose_lausuch> #topic New JIRA assignments
08:14:55 <jose_lausuch> I've been cleaning up JIRA
08:14:55 <ollivier> jose_lausuch: I would like to work it. but I can't niw.
08:15:01 <jose_lausuch> and creating the sprints
08:15:04 <jose_lausuch> ollivier: ok, no problem
08:15:27 <jose_lausuch> #link https://jira.opnfv.org/browse/FUNCTEST-558?filter=11115
08:15:43 <jose_lausuch> I created an epic for the intern projects we have
08:15:44 <HelenYao> ollivier: good to know Alpine, it's a new way of looking at docker slicing
08:16:00 <jose_lausuch> security groups, api evolution, unit tests
08:16:12 <jose_lausuch> and by the way
08:16:15 <jose_lausuch> in this EPIC https://jira.opnfv.org/browse/FUNCTEST-541
08:16:33 <jose_lausuch> you will find all the tasks related to all the adaptation to TestCaseBase that ollivier did
08:16:49 <jose_lausuch> this is some work to do, so feel free to pickup whatever you feel like :)
08:17:04 <jose_lausuch> #action all: contribute to the testcase adaptation EPIC: https://jira.opnfv.org/browse/FUNCTEST-541
08:17:11 <morgan_orange> I may ad an Epic on the reporting evolution or I can use the CI evolution Epic
08:17:20 <jose_lausuch> I tihkn if everyone of us contribute a bit with this epic, we can get it done soon
08:17:53 <jose_lausuch> morgan_orange: ok
08:18:04 <jose_lausuch> as you wish
08:18:10 <jose_lausuch> #link https://wiki.opnfv.org/display/functest/_Functest-Danube+release+schedule
08:18:19 <jose_lausuch> I created 7 sprints, starting from first of this month
08:18:24 <morgan_orange> yep that is for adopting what ollivier did and illustrated for ODL, makes sense to adapt our existing test cases
08:18:33 <jose_lausuch> so the first sprint is finishing this week
08:18:51 <jose_lausuch> I did this way to finish the 7th sprint on the release date
08:18:54 <jose_lausuch> plans might change
08:18:58 <morgan_orange> it would also make sense to create a Vnfbase or something like that to offer an abstraction for VNF onboarding as we have feature onboarding now
08:18:58 <jose_lausuch> but I think it's ok for now
08:19:04 <jose_lausuch> take 1 minute to go though the wiki
08:19:07 <jose_lausuch> and the calendar
08:19:41 <jose_lausuch> morgan_orange: that task is already created
08:19:43 <morgan_orange> great JIRA job!
08:20:14 <morgan_orange> jose_lausuch: I can see, this is a real Epic with lots of tasks...
08:20:16 <jose_lausuch> morgan_orange: https://jira.opnfv.org/browse/FUNCTEST-535
08:20:23 <HelenYao> jose_lausuch: I have to say, you are good at wiki page presentation
08:20:42 <jose_lausuch> hehe Thanks!!
08:20:42 <jose_lausuch> :)
08:20:46 <morgan_orange> he is good at everything...
08:21:00 <jose_lausuch> come on, don't make me blush :p
08:21:04 <SerenaFeng> +1
08:21:21 <HelenYao> +2
08:21:30 <jose_lausuch> thanks :)
08:21:47 <HelenYao> although I don't have the real power to +2 :-P
08:21:49 <jose_lausuch> so, whenever you start with a task, put the current sprint
08:21:52 <morgan_orange> so shall we share the load on current test case adaptation
08:22:05 <jose_lausuch> https://jira.opnfv.org/secure/RapidBoard.jspa?rapidView=59&quickFilter=125
08:22:09 <jose_lausuch> this is a view of the current sprint
08:22:26 <jose_lausuch> so, it would be nice that ever one fixes the sprint in your current tasks
08:22:30 <jose_lausuch> to have a nice view here :)
08:22:46 <jose_lausuch> morgan_orange: yes, we can start with vping/tempest/rally
08:22:55 <jose_lausuch> I can volunteer for tempest
08:22:55 <HelenYao> it's awesome to have kanban, hooray
08:22:58 <morgan_orange> for https://jira.opnfv.org/browse/FUNCTEST-336, it will be the core of the internship
08:23:00 <jose_lausuch> juhak:  would be you ok with rally?
08:23:05 <morgan_orange> adding unit tests in functest utils...
08:23:09 <SerenaFeng> I will work on vping soom
08:23:22 <morgan_orange> I suppose i could take doctor, promise and domino
08:23:35 <SerenaFeng> and parser
08:23:50 <HelenYao> I can work on whatever left
08:23:57 <juhak> jose_lausuch: yes, sure
08:24:21 <jose_lausuch> #info SerenaFeng will work on vping adaptation https://jira.opnfv.org/browse/FUNCTEST-540
08:24:41 <jose_lausuch> #info juhak will work on rally adaptation https://jira.opnfv.org/browse/FUNCTEST-543
08:25:00 <jose_lausuch> #info jose_lausuch will work on tempest adaptation https://jira.opnfv.org/browse/FUNCTEST-542
08:25:02 <morgan_orange> BTW (other topic) on Colorado, all the CI runs are failing due to Rally ...
08:25:26 <morgan_orange> #info morgan_orange will work on doctor/copper/domino/copper
08:25:48 <jose_lausuch> morgan_orange: right, but at least we don't have problems with the functest env prepare
08:25:49 <jose_lausuch> https://build.opnfv.org/ci/view/functest/job/functest-fuel-baremetal-daily-colorado/580/console
08:25:52 <juhak> morgan_orange: didn't know that, I'll take a look
08:26:18 <jose_lausuch> juhak: it seems an error with some code: File "/home/opnfv/repos/functest/testcases/OpenStack/rally/run_rally-cert.py", line 243, in get_output     duration = line.split(': ')[1]
08:26:26 <jose_lausuch> but apparently it happens only in stable branch!
08:26:29 <jose_lausuch> that's weird
08:26:30 <morgan_orange> juhak: probably a config issue, issue introduced yesterday and similar whatever installer/scenario
08:26:42 <morgan_orange> but right only on stable
08:26:49 <ollivier> we should test line.split(': ') before [1]
08:27:10 <morgan_orange> #action juhak troobleshoot new errors on rally launch on stable branch
08:27:16 <jose_lausuch> #action look at the existing error in Rally test for Colorado jobs
08:27:19 <hideyasu> morgan, if you ok, i try doctor.
08:27:19 <jose_lausuch> #undo
08:27:19 <collabot`> Removing item from minutes: <MeetBot.ircmeeting.items.Action object at 0x1d12c10>
08:27:20 <jose_lausuch> :)
08:27:38 <jose_lausuch> hideyasu: that would be great, morgan_orange do you agree?
08:27:41 <ollivier> and space should be remove !
08:27:42 <morgan_orange> sure
08:27:53 <hideyasu> ok ,
08:28:07 <morgan_orange> #info hideyasu will work on doctor adaptation
08:28:22 <morgan_orange> #info morgan will work on promise/domino/copper
08:28:50 <SerenaFeng> #info SerenaFeng will work on Parser adaptation
08:28:55 <morgan_orange> BTW hideyasu we will have to work on the abstraction for vnf onboarding, we probably need to do it together - probably in next Sprint
08:29:07 <jose_lausuch> I assigned those tasks to you
08:29:09 <morgan_orange> with integration of your vrouter suite + OAI
08:29:30 <jose_lausuch> SerenaFeng: I need to create a task for parser
08:29:53 <morgan_orange> jose_lausuch: if you agree I closed https://jira.opnfv.org/browse/FUNCTEST-336 and the internship will create new tasks related to the unit test Epic
08:30:03 <morgan_orange> I close... not closed yet
08:30:05 <SerenaFeng> I can create it after the meeting
08:30:08 <jose_lausuch> SerenaFeng: done https://jira.opnfv.org/browse/FUNCTEST-565
08:30:13 <jose_lausuch> ok
08:30:17 <SerenaFeng> ok, you ar so fast
08:30:20 <jose_lausuch> morgan_orange: sounds good
08:30:38 <jose_lausuch> SerenaFeng: I trained a lot with my left hand, so I am like 2 persons in 1 now :p
08:31:04 <HelenYao> jose_lausuch: you are a talent for multi-tasking
08:31:09 <jose_lausuch> morgan_orange: maybe we can move the tasks to the other epic about unit tests
08:31:11 <SerenaFeng> Superman
08:31:34 <jose_lausuch> HelenYao: well, that doesn't work so good, doing 2 things at the same time is tough... only women can do that!
08:31:35 <jose_lausuch> :D
08:31:53 <jose_lausuch> ok, moving on
08:31:54 <jose_lausuch> #topic OpenStack client versions follow up
08:31:58 <HelenYao> interesting thoery
08:32:17 <jose_lausuch> #info this wiki created to keep track of the discussion: https://wiki.opnfv.org/display/functest/OpenStack+python+clients
08:32:22 <jose_lausuch> #info there is a table that shows the client versions according to the openstack release versions.
08:32:53 <HelenYao> it's great to have the clear picture
08:33:00 <jose_lausuch> #info created this patch to uplift the client versions corresponding to newton https://gerrit.opnfv.org/gerrit/#/c/24237/
08:33:26 <jose_lausuch> I took a while to complete the table looking at the release notes, but this way it is much clearer to me what versions we should use
08:33:27 <jose_lausuch> however
08:33:33 <jose_lausuch> when we install our requirements.pip
08:33:47 <jose_lausuch> they are overwritten by other commands after that one
08:33:51 <jose_lausuch> for example, rally
08:33:57 <jose_lausuch> snaps
08:34:08 <jose_lausuch> but snaps is using Newton clients, so it should be fine
08:34:13 <HelenYao> what's our plan for Newton support? is upgrading client enough for it or we will have further plan? I sent a mail about the proposal
08:34:32 <jose_lausuch> and most of the requirements don't use exact versions, they use '<='    '>=' instead
08:34:45 <jose_lausuch> HelenYao: no, I think we need to touch the code as well
08:34:51 <HelenYao> i am running virtualenv and this problem will be resolved
08:34:54 <jose_lausuch> HelenYao: yes, I neeed to read it
08:35:18 <jose_lausuch> #info the requirements dependencies can be resolved in some cases with virtual envs, like the rally case
08:35:31 <HelenYao> i tested in vitualenv and all requirements even the toughest kingbird does not cause chaos
08:35:39 <jose_lausuch> #info HelenYao working on venv
08:35:50 <jose_lausuch> that sounds good
08:35:57 <jose_lausuch> I think it's our way forward for now
08:36:23 <jose_lausuch> there is another discussion ongoing in the community for Danube: should we release OpenStack Ocata or Pike ?
08:36:31 <jose_lausuch> I will have to vote for Functest
08:36:36 <jose_lausuch> I need your opinion :)
08:36:39 <HelenYao> even Pike?
08:37:01 <jose_lausuch> sorry no
08:37:04 <jose_lausuch> Newton or Ocata
08:37:17 <jose_lausuch> the question is: shall we support Newton or Ocata? Or both ?
08:37:33 <HelenYao> do we have enough resource to take the rush?
08:37:45 <jose_lausuch> #info discussion in the community about Danube OpenStack release to support: Newton vs Ocata or both
08:37:45 <morgan_orange> is it a question for Functest? All the test projects will have the same questions...
08:37:57 <jose_lausuch> no, it's a question for al the ptl's ...
08:37:58 <HelenYao> supporting two version sounds aggressive
08:38:16 <jose_lausuch> do you know guys how stable Ocata is nowadays?
08:38:31 <jose_lausuch> I would vote Newton only, but I',m not sure
08:38:44 <hujie> David is asking the same question...
08:38:45 <HelenYao> i think Newton is not stable enough, not to mention Ocata
08:38:59 <jose_lausuch> hujie: yes, that question comes directly from David :)
08:39:12 <jose_lausuch> probably in March Newton will be stable enough
08:39:19 <jose_lausuch> shall I vote for Newton only?
08:39:28 <jose_lausuch> it's what I had in mind ...
08:39:36 <morgan_orange> +1
08:39:36 <jose_lausuch> ollivier: ?
08:39:43 <HelenYao> I would vote for Newton only +1
08:39:47 <jose_lausuch> ok
08:39:50 <jose_lausuch> SerenaFeng: ?
08:39:53 <jose_lausuch> juhak: ?
08:39:57 <SerenaFeng> Newton +1
08:40:02 <hujie> N +1
08:40:05 <juhak> newton +1
08:40:08 <jose_lausuch> ok
08:40:10 <jose_lausuch> thanks guys :)
08:40:36 <morgan_orange> #Agreed Functest will support only Newton in Danube
08:40:49 <jose_lausuch> #info Functest will vote for OpenStack Newton only support for Danube release
08:41:01 <HelenYao> Ocata is not released yet, it's very risky to support it as the schedule would be very tight
08:41:07 <jose_lausuch> morgan_orange:  you are always faster  :)
08:41:24 <jose_lausuch> releasing with Ocata would mean moving the milestones
08:41:47 <jose_lausuch> this is the message from David:
08:41:47 <jose_lausuch> OpenStack Ocata releases approximately one month before the planned release of OPNFV Danube.  In order to integrate Ocata with Danube, we would need to slip the scheduled release date for Danube at least one month.  Alternatively, we could release OPNFV Danube, per the existing schedule, using OpenStack Newton, then skip Ocata altogether.  For the OPNFV E-release, we would plan to intersect the OpenStack Pike release.  Another thing to consi
08:41:49 <HelenYao> it will impact all opnfv projects
08:42:05 <jose_lausuch> ok
08:42:13 <jose_lausuch> anything else do we need to do for this topic?
08:42:18 <jose_lausuch> we have APs
08:42:21 <jose_lausuch> and work to do
08:42:26 <jose_lausuch> let's see if we fix that soon
08:42:33 <jose_lausuch> and combine it with the docker slicing if it makes sense...
08:42:49 <jose_lausuch> moving to the next topic
08:42:55 <jose_lausuch> #info Answer to bitergia
08:43:07 <jose_lausuch> morgan_orange: do you wan to drive this topic?
08:43:33 <morgan_orange> #info after meeting in Barcelona, Jesus sent us some proposals
08:43:33 <jose_lausuch> hujie: we will talk about escalator after this
08:43:48 <hujie> no problem
08:43:51 <jose_lausuch> #topic  Answer to bitergia
08:44:02 <morgan_orange> #info it is up to us to say what we do and do not do
08:44:20 <jose_lausuch> can you summarize in 1 sentence the 2 proposals?
08:44:57 <jose_lausuch> 1) let me copy them from the email
08:45:04 <morgan_orange> #info bitergia can just work from the Test API (do everything) / let us do the Elasticsearch consolidation / let us do the kibana definition
08:45:09 <jose_lausuch> ok, sorry
08:45:22 <morgan_orange> SerenaFeng:  has the best view on this
08:45:35 <ollivier> jose_lausuch: hard to vote. We could switch at least to the OpenStack current stable.
08:45:41 <jose_lausuch> We can start by putting the results Dashboard on bitergia as Jesus suggested
08:45:48 <morgan_orange> regarding our load, I would suggest to ask them to consume the API
08:46:28 <morgan_orange> Whatever we choose, what bitergia is doing is always open source
08:46:37 <morgan_orange> so we will not lose the controls of the dashboard
08:46:37 <HelenYao> will cost be a consideration point :-)
08:46:41 <morgan_orange> yes
08:46:51 <morgan_orange> bitergia has already a contrcat with linux Foundation
08:47:01 <morgan_orange> so these aspects have also to be taken into account
08:47:19 <jose_lausuch> do we have to consider that outselves?
08:47:20 <HelenYao> the more they work, the more we have to pay
08:47:41 <jose_lausuch> do we have to justify our decission?
08:47:58 <HelenYao> the most important thing is, how much control we want for the issue
08:47:59 <morgan_orange> HelenYao: sure but they are very skilled and if they take 1h when we need 1 day...assuming that there is already a contract...
08:48:16 <morgan_orange> I think we hshould first focuse on what we want technically and ask Ray if he approves it
08:48:41 <jose_lausuch> and what do we want technically? :)
08:49:01 <HelenYao> yeah, from technical point of view
08:49:32 <SerenaFeng> in my opinion we already make the decision during the summit that we let Bitergia do all the work
08:49:58 <SerenaFeng> and we make the requirements
08:50:07 <SerenaFeng> is it not the case?
08:50:07 <morgan_orange> not all the work...we do the API that they consume + the requirements
08:50:24 <jose_lausuch> the api is our task and responsability, right?
08:50:30 <jose_lausuch> we should maintain it
08:50:35 <morgan_orange> yes
08:50:42 <HelenYao> by handing over all stuff to bitergia, the communication effort will rise enormously if we want to update the data. my cents are, we can keep control of the data and let them work on visualization as they are the experts
08:50:44 <SerenaFeng> yes, testspi is our responsibility
08:50:58 <SerenaFeng> as to the visualization, they take almost all the work
08:51:03 <morgan_orange> so we should answer to jesus mail and indicate that our option is that they start consuming the API to display the results
08:51:28 <SerenaFeng> yes, I think Jesus's opinion make sense
08:51:31 <jose_lausuch> #info  answer to bitergia: our option is that they start consuming the API to display the results
08:51:51 <jose_lausuch> #info we keep responsability and maintain the API
08:51:54 <SerenaFeng> we can start with implementing dashboards similar in functionality to the current ones
08:52:08 <jose_lausuch> yes
08:52:11 <jose_lausuch> ok
08:52:14 <SerenaFeng> they can start with implementing dashboards similar in functionality to the current ones
08:52:14 <SerenaFeng> [16:52:20]  <jose_lausuch>	 yes
08:52:21 <jose_lausuch> morgan_orange: can I action you on the answer?
08:52:29 <SerenaFeng> sorry wrong type
08:52:36 <jose_lausuch> or SerenaFeng?
08:52:37 <morgan_orange> #action morgan_orange answer bitergia
08:52:44 <jose_lausuch> ok thanks
08:52:54 <jose_lausuch> let's move to the last topic before we finish
08:53:02 <jose_lausuch> #topic Escalator testcases integration
08:53:03 <hujie> thanks
08:53:11 <jose_lausuch> #info Escalator team wants to participate the Danube release and provide son functional tests
08:53:17 <jose_lausuch> #info They want support upgrading all OPNFV environments deployed by various Installers
08:53:21 <jose_lausuch> #info To avoid re-inventing wheels, Escalator try to reuse the function of the installer.
08:53:28 <jose_lausuch> #info The very first steps are reusing Daisy4NFV installer for experimental
08:53:37 <jose_lausuch> hujie: please share the info you want
08:53:45 <jose_lausuch> so that we have an undestanding of what you need
08:53:53 <hujie> Hi Functesters!
08:54:14 <HelenYao> hey
08:54:27 <hujie> We have shared this idea in https://wiki.opnfv.org/download/attachments/5046455/Escalator%20Refreshing%20your%20OPNFV%20environment%20with%20less%20troubles%20v4.pptx?version=1&modificationDate=1466671406000&api=v2
08:54:46 <jose_lausuch> #info liknk https://wiki.opnfv.org/download/attachments/5046455/Escalator%20Refreshing%20your%20OPNFV%20environment%20with%20less%20troubles%20v4.pptx?version=1&modificationDate=1466671406000&api=v2
08:54:52 <jose_lausuch> #undo
08:54:52 <collabot`> Removing item from minutes: <MeetBot.ircmeeting.items.Info object at 0x2192f90>
08:54:55 <jose_lausuch> #link https://wiki.opnfv.org/download/attachments/5046455/Escalator%20Refreshing%20your%20OPNFV%20environment%20with%20less%20troubles%20v4.pptx?version=1&modificationDate=1466671406000&api=v2
08:55:27 <jose_lausuch> hujie: what kind of testing would you need?
08:55:36 <jose_lausuch> an upgrade is not the usual thing we do :)
08:55:37 <hujie> Escalator will use the deploy function to replace the version files on nodes.
08:56:40 <hujie> To aviod service outage, Escalator will try to communicate with VIM to move VMs from the nodes.
08:57:21 <jose_lausuch> so, it's a test case about VM migration?
08:57:45 <hujie> Yes
08:58:07 <jose_lausuch> hujie: is there a specific scenario about it?
08:58:13 <hujie> 1.Installing OpenStack via Daisy4NFV on scenarios "os-nosdn-nofeature-ha"
08:58:13 <hujie> 2.Query the version information of target nodes and VMs running on them
08:58:13 <hujie> 3.Live-migrating VMs from target compute nodes to spare nodes
08:58:38 <jose_lausuch> is Daisy a new installer different from Fuel/Apex/Compass/Joid ?
08:58:56 <morgan_orange> ok my undestanding is that 1,5: run Functest on fresh install + test case that will be helpful to check that we kept probablyt the data
08:59:07 <jose_lausuch> hujie: normally the deployments have 3 controllers and 2 computes. Are 2 computes ok for your tests?
08:59:14 <morgan_orange> 3.5 re run functest, check it is still OK and see if data are still there
08:59:26 <hujie> Escalator will submit the description of the testcase to Functest before next Monday
08:59:46 <jose_lausuch> #info 1.Installing OpenStack via Daisy4NFV on scenarios "os-nosdn-nofeature-ha"
08:59:52 <jose_lausuch> #info 2.Query the version information of target nodes and VMs running on them
09:00:02 <jose_lausuch> #info 3.Live-migrating VMs from target compute nodes to spare nodes
09:00:17 <jose_lausuch> #info Escalator will submit the description of the testcase to Functest before next Monday
09:00:22 <hujie> I think 2 computer nodes can be HA.
09:00:45 <jose_lausuch> hujie:  the HA deployments in CI are with 2 computes
09:01:01 <SerenaFeng> jose_lausuch Daisy is a new installer
09:01:08 <jose_lausuch> SerenaFeng: ok, thanks
09:01:22 <hujie> It is OK, we will find a solution in this configuration.
09:01:28 <jose_lausuch> ok, we will wait for the proposal and provide comments to the team
09:01:31 <morgan_orange> fro Openstack which upgrade will it manage? from Newton to ocata?...
09:01:32 <jose_lausuch> thanks
09:01:38 <morgan_orange> if so we do not support Ocata client :)
09:01:49 <jose_lausuch> it should be backwards compatible ...
09:01:52 <hujie> Newton, I think.
09:01:54 <jose_lausuch> but you are right
09:02:01 <morgan_orange> so Liberty to Newton?
09:02:10 <morgan_orange> oops
09:02:12 <jose_lausuch> Mitaka to Newton
09:02:37 <morgan_orange> yep even if skipping one version was also a pain points reported by the EUAG
09:03:06 <hujie> We plan to support OpenStack upgrade, and others as well.
09:03:08 <HelenYao> a side question, what is EUAG
09:03:09 <morgan_orange> so Functest will be used to test before and after migration + a new test case specific to escalator to garantee the persistance of data?
09:03:19 <morgan_orange> HelenYao:  End User Advisory group
09:03:27 <HelenYao> morgan_orange: thx
09:03:36 <jose_lausuch> that changes the way we execute things
09:03:39 <jose_lausuch> HelenYao: https://www.opnfv.org/about/opnfv-end-user-advisory-group
09:03:44 <jose_lausuch> nice picture Morgan
09:04:18 <hujie> Do you guys have more questions on Escalator?
09:04:18 <morgan_orange> I have to drop
09:04:24 <jose_lausuch> time is over
09:04:26 <jose_lausuch> thank you all
09:04:33 <hujie> Thanks....
09:04:35 <jose_lausuch> hujie: no, we will follow up by email when you have the proposal to us
09:04:45 <hujie> OK.
09:04:45 <jose_lausuch> #endmeeting