16:00:28 <trevor_intel> #startmeeting OPNFV Pharos
16:00:28 <collabot`> Meeting started Wed Dec 16 16:00:28 2015 UTC.  The chair is trevor_intel. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:28 <collabot`> Useful Commands: #action #agreed #help #info #idea #link #topic.
16:00:28 <collabot`> The meeting name has been set to 'opnfv_pharos'
16:00:36 <frankbrockners> #info Frank Brockners
16:00:39 <trevor_intel> #info Trevor Cooper
16:00:48 <fdegir> #info Fatih Degirmenci
16:00:55 <frankbrockners> IRC only or do we have a GTM?
16:01:10 <fdegir> irc
16:01:22 <frankbrockners> cool - thx
16:02:12 <trevor_intel> Agenda topics 1) CI/labs 2) LF 3) Templates 4) MAAS
16:02:27 <trevor_intel> Whate else or instead?
16:02:35 <fdegir> looks good to me
16:02:46 <StuartMackie> #info StuartMackie
16:03:25 <trevor_intel> #topic CI resources for Brahmaputra
16:03:36 <fdegir> trevor_intel: should I give a quick status
16:03:44 <trevor_intel> fdegir: please
16:03:56 <fdegir> #info CI Lab/POD Status Update
16:04:17 <fdegir> #info LF POD1: Waiting for Apex to move virtual deploys to Intel Machine from Intel POD4
16:04:27 <trevor_intel> #info     Lab resources for Release B  https://wiki.opnfv.org/pharos_rls_b_labs
16:04:47 <fdegir> #info LF POD1: Once it is done, pbandzi will convert it to a full POD which will be used for Apex CI, running stable
16:05:16 <fdegir> please note that after some thinking and talking with Tim, we moved Apex CI to LF POD1
16:05:27 <fdegir> updated the page https://wiki.opnfv.org/brahmaputra_testing_page
16:05:36 <fdegir> and will update the page https://wiki.opnfv.org/pharos_rls_b_labs#status_of_production_ci_resources_for_release_b
16:06:01 <fdegir> #info LF POD1: Not operational yet
16:06:15 <fdegir> #info LF POD2: Operational and used by Fuel CI.
16:06:39 <fdegir> #info Intel POD2: Operational and used by Apex development and project work/connected to Jenkins
16:06:45 <trevor_intel> fdegir: Plan for LF POD 1 has chnaged from Apex latest to Apex stable?
16:06:52 <fdegir> trevor_intel: that's correct
16:06:58 <fdegir> the reason is to keep LF access limited
16:07:04 <trevor_intel> fdegir: Ok I will update the Wiki
16:07:15 <fdegir> and LF POD has higher bandwidth which is needed by Apex
16:07:24 <trevor_intel> fdegir: That means INtel POD 2 is now Apex latest?
16:07:32 <fdegir> trevor_intel: that's correct
16:07:46 <fdegir> and also used for development work by apex
16:07:56 <fdegir> when it is not used by CI
16:07:57 <frankbrockners> fdegir: are there plans for using the recently bought servers in LF lab for virtual deploys?
16:08:07 <fdegir> frankbrockners: we have plans :)
16:08:11 <fdegir> frankbrockners: we lack hands
16:08:44 <trevor_intel> fdegir: a poet too?!? :)
16:08:50 <fdegir> :)
16:09:20 <fdegir> #info Intel POD3: Operational - connected to Jenkins and will be used by vsperf.
16:10:02 <frankbrockners> fdegir: what would the missing hands do if they were there?
16:10:04 <fdegir> #info Intel POD4: Operational - 1 fuel virtual deploy, 1 apex virtual deploy, 1 ubuntu build server (majority of project builds), 1 centos build server (ovsnfv builds), 1 free
16:10:30 <fdegir> frankbrockners: I lack all the details
16:10:38 <fdegir> frankbrockners: first, we need to fix pod1
16:10:51 <fdegir> frankbrockners: and then look at new blades to see how we can use them
16:10:59 <fdegir> frankbrockners: fix the os, packages etc
16:11:01 <frankbrockners> ok - if it a matter of configuration - let me know
16:11:12 <frankbrockners> peter already installed base os on some of the servers
16:11:17 <fdegir> frankbrockners: will do - was planning to do that yesterday
16:11:24 <frankbrockners> thx
16:11:26 <fdegir> frankbrockners: but we must first get apex guys out of lf
16:11:34 <fdegir> you will hear from me
16:11:54 <fdegir> #info Intel POD5: Operational (kind of) - Joid CI
16:12:10 <fdegir> #info Intel POD5 lacks jumphost so joid only deploys 3 controllers and 1 compute
16:12:18 <trevor_intel> fdegir: Jack is working on Intel jumphosts POD 5, 6, 8 ... will be up EOD today
16:12:20 <fdegir> #info Intel POD6: Waiting for jumphost
16:12:28 <fdegir> #info Intel POD8: Waiting for jumphost
16:12:39 <fdegir> #info Huawei SC POD1: Operational - Compass CI
16:12:44 <fdegir> thanks trevor_intel
16:12:59 <fdegir> #info Ericsson POD2: Operational - Fuel POD config is ongoing
16:13:35 <fdegir> did I miss any?
16:13:51 <trevor_intel> fdegir: I think that is all the CI stuff
16:14:01 <fdegir> good
16:15:17 <fdegir> just to record it again
16:15:49 <fdegir> #info LF POD1 (apex), LF POD2 (fuel), Intel POD5 (joid), Huawei POD1 (compass) for release testing purposes
16:16:02 <fdegir> meaning hands off, limited access, totally dedicated to CI
16:16:06 <trevor_intel> Also here https://wiki.opnfv.org/doku.php?id=pharos_rls_b_labs
16:16:34 <trevor_intel> Those are CI stable resources
16:16:57 <fdegir> #info Intel POD6 (joid), Intel POD8 (Huawei), Intel POD2 (Apex), Ericsson POD2 (fuel) for project work which is used for development if necessary
16:17:10 <fdegir> yes
16:17:27 <trevor_intel> Those are CI latest / test + dev resources
16:17:30 <fdegir> yes
16:17:41 <fdegir> since I'm still getting questions
16:17:45 <fdegir> just to record in mom
16:17:55 <trevor_intel> fdegir: Other than jump servers no other hw needs now?
16:18:07 <fdegir> trevor_intel: that's right
16:18:27 <fdegir> as I mentioned before, we should focus on making them operational
16:18:40 <fdegir> any other resources offered by community or exist in LF need to wait
16:18:56 <trevor_intel> fdegir: Any other help needed to get operational?
16:19:14 <fdegir> I think I can take care of the basic CI config
16:19:30 <fdegir> but when installer or other project teams need help
16:19:35 <fdegir> they might come to your way
16:19:47 <fdegir> or to LF
16:19:49 <trevor_intel> fdegir: ok
16:20:28 <trevor_intel> fdegir: what about the 2nd Jenkins instance?
16:20:57 <trevor_intel> for test
16:20:59 <fdegir> not there yet
16:21:12 <trevor_intel> But still teh plan, correct?
16:21:27 <fdegir> I think we passed the point
16:21:37 <fdegir> it is late already
16:21:53 <fdegir> teams need to learn stuff on production instance
16:21:56 <trevor_intel> fdegir: Do you mean too late for rls B?
16:22:01 <fdegir> yes
16:22:03 <fdegir> we can have it
16:22:12 <fdegir> but who will support teams to try stuff out is the question
16:22:25 <fdegir> the ones who know jenkins stuff is totally focused on release work
16:22:46 <trevor_intel> fdegir: ok I can see the issue
16:22:54 <fdegir> it is priorities as you know
16:23:31 <trevor_intel> I have a question re. mangement of the labs ... or it can wait
16:23:42 <fdegir> no need to wait :)
16:24:22 <trevor_intel> the issue of using a separate Jira project for lab operations
16:24:38 <fdegir> yes
16:25:08 <trevor_intel> Wouln't it be cleaner to seperate Pharos dev with lan operations?
16:25:30 <fdegir> but then it won't be in pharos backlog
16:25:34 <trevor_intel> Not sure how to label ops stuff it is all in one project?
16:25:40 <fdegir> making it disconnected from pharos
16:25:53 <trevor_intel> Don't want it disconnected
16:25:57 <fdegir> so it might not be simpler to see how much work needs to be done by pharos
16:26:16 <fdegir> if we have a separate project, it might result in 2 different backlogs
16:27:02 <fdegir> and more cross-project referencing
16:27:07 <trevor_intel> I am ok to keep it together but we need a scheme for labeling ops tasks/issues/labs/etc
16:27:26 <fdegir> I think what Jack suggested is good
16:27:29 <fdegir> component?
16:27:45 <trevor_intel> component = a particular lab
16:27:50 <fdegir> yep
16:28:03 <fdegir> I am not jira master but try to pull in Mark
16:28:21 <fdegir> he can perhaps say what he thinks
16:28:53 <trevor_intel> Ok I think we shoudl take this off-line from this meeting
16:29:01 <fdegir> yes
16:29:16 <fdegir> we can bug him and see what he thinks
16:29:28 <trevor_intel> fdegir: ok
16:29:29 <fdegir> offline I mean
16:29:43 <trevor_intel> change topic?
16:29:49 <trevor_intel> templates?
16:29:55 <fdegir> yes
16:30:01 <fdegir> frankbrockners is with us as well
16:30:09 <trevor_intel> #topic Config Files
16:31:16 <fdegir> frankbrockners: ping?
16:31:30 <frankbrockners> fdegir: yes - I'm here
16:31:38 <trevor_intel> Progress on config files?
16:31:44 <trevor_intel> What is your take?
16:32:17 <frankbrockners> #info Starting points for common config files have been defined
16:32:20 <frankbrockners> #link https://wiki.opnfv.org/genesis/configuration-files-discussion#common_config_files
16:32:53 <frankbrockners> #info will wait till the end of this week - hopefully we have comments/changes from everyone who cares in by then
16:33:12 <frankbrockners> #info can take it to a vote and close early next week in Genesis meeting (Monday)
16:33:39 <trevor_intel> frankbrockners: sounds good
16:33:40 <frankbrockners> #info IMHO deploy and inventory are quite far along
16:34:03 <frankbrockners> #info network config is still evolving
16:34:33 <frankbrockners> #info hope that by next week we can define "scenarios" using the common config files
16:35:06 <fdegir> frankbrockners: just a quick question regarding scenarios
16:35:06 <frankbrockners> (this is at least my hope) - let's see
16:35:08 <trevor_intel> frankbrockners: For now we said all test complexity will be handled by Jenkins jobs, correct?
16:35:24 <fdegir> will all the scenarios include a controller?
16:35:33 <fdegir> any of the controllers
16:35:50 <frankbrockners> trevor_intel: yes - testing figures out which tests to run
16:36:04 <fdegir> or do we have any possibility of having a scenario where there is no controller?
16:36:17 <frankbrockners> personally I'm not sure how this would work on any but the simple default scenarios
16:36:26 <frankbrockners> but I'm happy to be convinced otherwise
16:36:38 <trevor_intel> fdegir: why this question (I assumed there will be a controller)?
16:36:48 <frankbrockners> fdegir: IMHO all the scenarios should include a controller
16:37:02 <fdegir> trevor_intel: it will impact how we will handle combinations
16:37:03 <frankbrockners> otherwise we'd test openstack - which openstack already does
16:37:29 <fdegir> the structuring of jenkins jobs, the parameters, how those parameters will be evaluated and used by installers and test projects
16:37:30 <trevor_intel> fdegir: I know but why would there NOT be a controller?
16:37:45 <fdegir> trevor_intel: I have no reason, just asking
16:37:53 <frankbrockners> we could have a setup with pure neutron networking
16:38:00 <frankbrockners> but this isn't very useful for NFV
16:38:10 <frankbrockners> nor is this a new case to be run
16:38:18 <fdegir> I plan to have 1 job per installer per controller
16:38:19 <frankbrockners> openstack already tests openstack only cases
16:38:32 <fdegir> and then this job will have extra parameter to specify the feature
16:38:47 <fdegir> both controller and feature will be exposed to installers and test projects
16:38:59 <frankbrockners> fdegir - can you handle multiple versions of e.g. ODL that way
16:39:00 <frankbrockners> ?
16:39:02 <fdegir> and they will take care of the details in there
16:39:11 <fdegir> frankbrockners: we can
16:39:16 <frankbrockners> cool - thx
16:39:28 <fdegir> frankbrockners: but we need to know what else should be handled
16:39:33 <fdegir> is it just for odl or?
16:39:44 <fdegir> since jenkins is not that clever
16:40:00 <frankbrockners> that is a question that I don't have an answer for
16:40:03 <fdegir> btw, this is not a question I want answer
16:40:06 <fdegir> :)
16:40:13 <frankbrockners> onos and ocontrail just come on board
16:40:36 <frankbrockners> this is why I started the matrix_ https://wiki.opnfv.org/brahmaputra_testing_page#test_scenarios_for_brahmaputra
16:40:59 <frankbrockners> if we don't find owners for onos or ocontrail scenarios - then let's remove them from the map
16:41:07 <fdegir> ok
16:41:12 <fdegir> joid doesn't have virtual as of yet
16:41:19 <fdegir> or at least when I asked last week
16:41:22 <fdegir> anyway
16:41:47 <frankbrockners> key thing is to get visibility into what people want to pull together
16:42:08 <fdegir> yes
16:42:14 <frankbrockners> question: should I ask the "scenario" question as a standalone question on tech-discuss?
16:42:48 <fdegir> do you mean
16:42:49 <StuartMackie> Hey there - we just set up a pod in the Juniper NJ lab and connected to Jenkins for JOID/OpenContrail initially, and would expect to test other installers there too.
16:42:51 <fdegir> which scenarios people would like to have
16:42:58 <fdegir> which scenarios people believe we can have
16:43:01 <fdegir> or
16:43:10 <frankbrockners> yes - like to have AND willing to pull together and support
16:43:11 <fdegir> which scenarios we should focus - same as the above
16:43:34 <frankbrockners> "like" or "+1" is not enough - we need someone to step up and do the work
16:43:41 <fdegir> it would be good to see the response
16:43:50 <fdegir> someone might say we want this but the others say not possible
16:43:55 <fdegir> lack of this and that
16:43:59 <frankbrockners> let me push out that email after our meeting here
16:44:04 <fdegir> ok
16:44:16 <fdegir> thx StuartMackie
16:44:46 <fdegir> so we should some kind of outcome for config stuff by next week
16:44:55 <fdegir> (have)
16:45:05 <frankbrockners> fdegir - I share that "hope"
16:45:22 <fdegir> I always hope for the best but expect the worst
16:45:30 <frankbrockners> :-)
16:46:13 <trevor_intel> StuartMackie: will NJ lab be available as a community lab?
16:46:44 <StuartMackie> I guess so, I am not clear on what that exactly means
16:47:28 <StuartMackie> In particular - how is time scheduled on community labs?
16:48:00 <trevor_intel> StuartMackie: means 1) Pharos compliant AND 2) part of the OPNFV lab resource plan
16:48:21 <trevor_intel> StuartMackie: depneds on how it is used
16:48:53 <trevor_intel> StuartMackie: Can be for dev or CI
16:49:23 <trevor_intel> StuartMackie: See here https://wiki.opnfv.org/pharos_rls_b_labs
16:49:54 <trevor_intel> narindergupta: ping
16:50:01 <narindergupta> trevor_intel, pong
16:50:32 <trevor_intel> narindergupta: do you haev any updates re. MAAS this week?
16:50:55 <narindergupta> trevor_intel, need help on creating jobs for MAAS
16:51:04 <trevor_intel> #topic MAAS
16:52:00 <narindergupta> trevor_intel, i never worked on jenkins before so need help there. MAAS part and API i can handle though.
16:52:15 <trevor_intel> fdegir: I am thinking people are focussedon release and will be difficult to get help for this now? What do you think?
16:52:29 <fdegir> narindergupta: I think you can forget the jenkins stuff for creating the script
16:52:37 <fdegir> trevor_intel: that's right
16:52:54 <fdegir> narindergupta: as long as you script, we can put it in a jenkins job and see what happens
16:53:21 <fdegir> narindergupta: the crucial thing here is
16:53:32 <narindergupta> trevor_intel, ok i can script where script input will be MAAS IP API key and which release of ubuntu or cent os needs to be installed
16:53:33 <fdegir> narindergupta: where that script is executed on jenkins
16:53:49 <narindergupta> that can execute in linux foundation lab
16:53:51 <fdegir> narindergupta: meaning that it should be able to use the api on the machine
16:53:59 <fdegir> narindergupta: then that should be no problem
16:54:13 <fdegir> narindergupta: how will it work?
16:54:20 <narindergupta> MAAS URL and API key can be called form anywhere
16:54:31 <fdegir> narindergupta: can I say; give me a centos jumphost with Centos7 and these packages installed
16:55:12 <fdegir> narindergupta: or give me a single machine with this OS and packages where I can run virtual deployment or build
16:55:15 <narindergupta> fdegir, for cents answer is yes and for package you need to give me predefined list to modify the image to do that
16:55:21 <fdegir> narindergupta: they're basically same
16:55:30 <fdegir> narindergupta: ok
16:55:42 <narindergupta> so i can run installation of package after initial base os install both should work
16:55:43 <fdegir> narindergupta: then if you can create a script and we just get a simple build machine booked to start with
16:56:03 <fdegir> narindergupta: I can pass you the packages and the OS (can start with ubuntu 14.04)
16:56:13 <fdegir> narindergupta: I book that using the script from jenkins
16:56:31 <narindergupta> fdegir, correct
16:56:33 <fdegir> once it is provided by maas, I do hello world and then build something
16:56:40 <fdegir> and we move to virt deployment of something
16:56:48 <fdegir> and then to jumphost/pod stuff
16:57:03 <narindergupta> fdegir, i am confused now
16:57:14 <narindergupta> fdegir, what do you mean by virtue deployment?
16:57:39 <fdegir> like I want a machine where I can run virtual deployment of installer X
16:58:00 <fdegir> which means the cpu of the machine should have certain features
16:58:08 <fdegir> certain packages are installed
16:58:10 <narindergupta> fdegir, ok yes MAAS can give you a machine to run your scripts
16:58:11 <fdegir> like libvirt
16:58:15 <fdegir> yep
16:58:32 <fdegir> so, when you have a script and machines I can book
16:58:34 <narindergupta> fdegir, MAAS will install the OS
16:58:45 <fdegir> that's good too
16:58:57 <fdegir> I can quickly create a jenkins job to try stuff out
16:59:05 <fdegir> last 2 minutes
16:59:09 <narindergupta> fdegir, early next week you can have it
16:59:18 <fdegir> good
16:59:32 <narindergupta> fdegir, you want shell script or python?
16:59:38 <fdegir> doesn't matter
17:00:01 <fdegir> we can ask aric to set venv on a host in LF
17:00:17 <narindergupta> fdegir, ok sounds good i will work on it where you can say give me machine in intel lab with particular OS installed
17:00:31 <fdegir> ok
17:00:42 <fdegir> we start with most basic and build on top
17:00:51 <narindergupta> fdegir, sure sounds good to me
17:01:43 <trevor_intel> I will end meeting now unless other topics?
17:02:03 <trevor_intel> going ...
17:02:18 <trevor_intel> thanks all for attending
17:02:24 <trevor_intel> #endmeeting