14:36:08 <frankbrockners> #startmeeting OPNFV Brahmaputra release daily scrum
14:36:08 <collabot> Meeting started Mon Feb  8 14:36:08 2016 UTC.  The chair is frankbrockners. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:36:08 <collabot> Useful Commands: #action #agreed #help #info #idea #link #topic.
14:36:08 <collabot> The meeting name has been set to 'opnfv_brahmaputra_release_daily_scrum'
14:36:11 <collabot> debrascott: Error: Can't start another meeting, one is in progress.  Use #endmeeting first.
14:36:14 <ashyoung> #info Ashlee Young
14:36:19 <radez> #info Dan Radez
14:36:21 <anac1> #info Ana Cunha
14:36:24 <frankbrockners> #info Frank Brockners
14:36:27 <bin_> #info Bin Hu
14:36:30 <morgan_orange> #info Morgan richomme
14:36:30 <frankbrockners> #chair debrascott
14:36:30 <collabot> Current chairs: debrascott frankbrockners
14:36:31 <debrascott> Thank you Frank, I had to reset all network
14:36:47 <joekidder> #info Joe Kidder
14:36:53 <debrascott> #topick roll call
14:36:54 <trevor_intel> #info Trevor Cooper
14:36:55 <narindergupta> #info Narinder Gupta
14:37:24 <frankbrockners> while we're getting settled here... - can we start with the ODL Be question from Friday?
14:37:28 <debrascott> #info same agenda as Friday https://wiki.opnfv.org/releases/brahmaputra/minutes
14:37:46 <debrascott> Frankbrockners sure.
14:37:50 <frankbrockners> cool
14:37:59 <debrascott> I have a question about ODL upgrade
14:38:11 <debrascott> Do they provide easy path from Li to Be?
14:38:15 <JonasB> #info Jonas Bjurel
14:38:19 <frankbrockners> From what I've seen Compass/Apex/Fuel are all in favor of making ODL Be the default
14:38:44 <frankbrockners> narindergupta - you needed to check on feasibility for Joid, did you?
14:39:01 <narindergupta> frankbrockners: yes i have to
14:39:18 <frankbrockners> narindergupta: when do you expect to know?
14:39:21 <narindergupta> frankbrockners: as i am  not odl expert needs to look for any configuration changes
14:39:41 <narindergupta> frankbrockners: i am working on it in intel pod6 but need help on it as well.
14:40:00 <narindergupta> frankbrockners: i send a query last week on same.
14:40:07 <bin_> ODL Be is good for IPv6
14:40:42 <narindergupta> frankbrockners: as i understand this is still not release and builds available on daily basis
14:40:58 <narindergupta> frankbrockners: https://nexus.opendaylight.org/content/repositories/opendaylight.snapshot/org/opendaylight/integration/distribution-karaf/0.4.0-SNAPSHOT/distribution-karaf-0.4.0-20160208.142515-3986.tar.gz i will try to use
14:41:01 <radez> narindergupta: there is an RC out now
14:41:23 <narindergupta> radez: may i know the location of distribution?
14:41:24 <debrascott> does anyone know about upgrade path for ODL between Li & Be? If Joid unable to support Be with install could a workaround be Joid install Li and upgrade to Be?
14:41:32 <frankbrockners> narindergupta's question was on which modules to load (see email): "Currently we enable feature odl-ovsdb-openstack  in Li if Be needs more feature needs to be enabled then we need to know.  "
14:41:40 <narindergupta> correct
14:41:43 <frankbrockners> anyone feel to help with guidance?
14:41:50 <radez> narindergupta: we get ours from the centos nfv sig: http://cbs.centos.org/repos/nfv7-opendaylight-4-testing/x86_64/os/Packages/
14:41:57 <frankbrockners> should be the same across installers anyway
14:42:03 <radez> but that atleast shows that RC2 is out :)
14:42:25 <radez> dfarrell07: ^^ where can we get the tarball for RC2 Be?
14:42:27 <frankbrockners> ODL still does not build dbkg..
14:42:44 <narindergupta> correct i need the already build tar ball
14:43:18 <radez> narindergupta: let's see if dfarrell07 is listening, if not now I'll get the location from him for you after the meeting
14:43:31 <narindergupta> radez: thanks
14:43:33 <frankbrockners> apart from the tarball - radez: do you have a list of features that get installed for the different scenarios with  ODL Be?
14:43:48 <narindergupta> radez: frankbrockner yeah we need to know that too
14:44:20 <frankbrockners> or JonasB - do you have that list?
14:44:25 <narindergupta> frankbrockners: there was difference from helium to lithum
14:44:30 <frankbrockners> ashyoung might have that list as well
14:44:42 <radez> let me look in our code base, we use the puppet module which does most of it
14:45:03 <frankbrockners> thanks radez - we can hook up in the Genesis call later today
14:45:06 <narindergupta> radez: ok thanks let me know i will work on integration with joid
14:45:07 <radez> ack
14:45:28 <ashyoung> I don't know it off hand, but will look into it today
14:45:33 <frankbrockners> assuming we can get the charm adapted for ODL Be, are there any other issues we know of that projects have with ODL Be as default?
14:46:00 <ashyoung> no negative responses came in
14:46:16 <frankbrockners> morgan_orange? any tests that are specific to ODL Li?
14:46:18 <debrascott> ashyoung: +1 that’s what I saw too but it was over weekend
14:46:53 <frankbrockners> pbandzi: Do you know of any tests for ODL which are ODL Li specific and that would behave differently on ODL Be?
14:47:01 <morgan_orange> frankbrockners: no ODL test suite is very limited (just northbound API) so no big risk I think
14:47:25 <morgan_orange> and if some tests have been run with Be, the ODL tests are always fine
14:47:28 <frankbrockners> morgan_orange: ODL Be might actually do better than Li
14:47:29 <ashyoung> Should we repost the question to the thread due to the weekend?
14:47:35 <JonasB> nearindergupta: Have a look at the Fuelplugin for ODL - that should guide you: https://review.openstack.org/#/admin/projects/openstack/fuel-plugin-opendaylight
14:47:42 <frankbrockners> ashyoung - go ahead
14:47:47 <ashyoung> doing it now
14:48:00 <frankbrockners> ashyoung - thanks
14:48:05 <frankbrockners> let's see what comes back
14:48:22 <radez> narindergupta: yea we use the default features here: https://github.com/dfarrell07/puppet-opendaylight/blob/master/manifests/params.pp and we add to it the openstack feature here: https://github.com/radez/apex/blob/stable/build/opnfv-tripleo-heat-templates.patch#L907
14:48:38 <radez> and there's a sfc feature to if you need that?
14:48:42 <debrascott> OK, assuming that Be is agreed to by TSC, we still have to get a good view of all tests results by scenario & feature
14:49:10 <debrascott> Morgan_Orange is your table on the wiki up to date?
14:49:14 <frankbrockners> ... ok - let's hash things out further in the Genesis meeting - debrascott - thanks for the 15min - and sorry for the delay
14:49:25 <JonasB> All fuel results from the week-end was with Be.
14:49:27 <narindergupta> radez: thanks i will have a look
14:49:33 <debrascott> frankbrockners: no problem
14:49:44 <debrascott> JonasB; thank you good to know
14:49:58 <morgan_orange> JonasB OK so it answers to your question frankbrockners => no regression due to Be seen on Fuel scenario over the week end
14:50:43 <morgan_orange> debrascott: I tried to update the table (as I know some people like it) but I prefer the dynamic dashboard
14:51:16 <morgan_orange> I created new view per installer and scenario http://testresults.opnfv.org/proto/brahma/index-<installer>.html with installer = apex, compass, fuel, joid
14:51:30 <morgan_orange> we can see
14:51:56 <morgan_orange> from a Functest perspective: compass scenario are almost all ready for release
14:52:12 <morgan_orange> Good stability of fuel tests over the week end
14:52:24 <debrascott> morgan_orange: thank you. This is important to digest the whole matrix so we can suggest what the minimum set is to release and what might need to be later
14:52:26 <morgan_orange> Main issue reduce Tempest errors
14:52:40 <ashyoung> email sent
14:53:05 <JonasB> morgan_orange: 404 not found
14:53:23 <debrascott> morgan_orange: are you dealing with multiple errors with Tempest?
14:53:25 <morgan_orange> http://testresults.opnfv.org/proto/brahma/index-fuel.html
14:53:29 <morgan_orange> ?
14:53:34 <frankbrockners> http://testresults.opnfv.org/proto/brahma/index-fuel.html works for me
14:53:34 <morgan_orange> http://testresults.opnfv.org/proto/brahma/index-compass.html
14:53:41 <morgan_orange> http://testresults.opnfv.org/proto/brahma/index-apex.html
14:53:46 <morgan_orange> http://testresults.opnfv.org/proto/brahma/index-joid.html
14:54:13 <debrascott> morgan_orange: how complex is the troubleshooting for Tempest issues? multiple, intermittent, consistent but no direct cause associated….etc?
14:54:23 <morgan_orange> Tempest is running 210 tests covering all the main APis/modules of OpenSTack
14:54:29 <morgan_orange> with compass wr have 1 errors
14:54:43 <morgan_orange> with fuel it is more about 20-30 with some that are related
14:55:08 <morgan_orange> we master the custom list of tests
14:55:30 <morgan_orange> but as compass is almost 100% OK, we shall try to have the same results with the different installers
14:55:47 <morgan_orange> 2 types of issues
14:56:05 <morgan_orange> on fuel it seems very reproducible, so fixing some config should possible
14:56:28 <morgan_orange> on apex, we are facing a strange things, the number of executed tests vary from 1 run to another without error logs
14:56:51 <morgan_orange> a workaround  has been suggested by Rally upstream community, we need to test it, but it affects only POD1
14:57:51 <morgan_orange> #info integration of doctor initiated on apex scenario
14:58:10 <morgan_orange> #info only ovno (ocl) is not done so far regarding what we planned for B Release
14:58:18 <debrascott> Morgan_orange: thank you, helps me to understand & discuss better when asked
14:58:26 <debrascott> only 2 min left.
14:58:39 <frankbrockners> folks... the TSC will ask for timelines tomorrow..
14:58:59 <debrascott> anac1 do you have a summary page available for yardstick?
14:59:10 <frankbrockners> so - do we have an idea on a draft schedule?
14:59:10 <ashyoung> I propose we also take up the Li to Be issue at TSC
14:59:24 <frankbrockners> ashyoung: agreed
14:59:27 <anac1> debrascott: no
15:00:08 <debrascott> anac1: if I create a wiki similar to Morgan’s can you populate with test results before TSC tomorrow? They will want this view to help drive decisions
15:00:09 <JonasB> Feb 26?
15:00:23 <anac1> radez: we are using Li workaround for Apex, please let me know when Be will be available and we'll remove it
15:00:49 <radez> anac1: will do, the patch is verifying now and should merge today when it passes
15:01:00 <anac1> radez: ty
15:01:02 <debrascott> JonasB: the questions posed last week were Mar 8 and then see if anything might be ready to pull in before that or if we should set a date later than Mar 8
15:01:10 <frankbrockners> JonasB... that is an option - real question is whether https://etherpad.opnfv.org/p/steps_to_brahmaputra is the latest, or whether we could pull in the schedule based on recent enhancements
15:01:37 <frankbrockners> where does Mar/8 come from?
15:02:19 <debrascott> Frankbrockners: that was the addition of more days suggested on the etherpad, Chris took the latter date rather than start with Feb 26
15:02:22 <JonasB> Have never heard of anything beyond Feb?
15:03:16 <frankbrockners> Etherpad say Mar/3 right now (due to added padding to accomodate testing)
15:03:59 <frankbrockners> morgan_orange, anac1 - what are the thoughts on final timelines? By when do you thing we can go through the final sequence of 4 runs per scenario?
15:04:00 <debrascott> There was additional padding requested, went out further than Mar 3. I can’t get my browsers working right now or I could answer better
15:04:27 <debrascott> They had board meeting last week and this was discussed at that level
15:04:27 <frankbrockners> would that be by Feb/16 or potentially earlier?
15:04:59 <debrascott> If we can pull in from Mar 8, wonderful but let’s set a date we are really sure we can make
15:05:01 <morgan_orange> frankbrockners: main concern is on joid/scenario (troubleshooting on Intel POD5 planned this week) confident for others..
15:05:53 <debrascott> There was also padding put in to accomodate longer time to write final test results
15:07:11 <anac1> frankbrockners: yardstick is working on bug fixes, as now we can run some scenarios - Li is unstable, nosdn should be ok to start eh 4 runs
15:07:36 <frankbrockners> debrascott - could you create an updated view of the schedule for all of us and send out?
15:08:00 <debrascott> frankbrockners: will try to consolidate, and send out, yet
15:08:03 <debrascott> *yes
15:08:10 <frankbrockners> anac1 - thanks. do you think 1 more week is enough to get things under control
15:08:12 <frankbrockners> ?
15:09:12 <anac1> frankbrockners: as Compass people are on vacation, i'd suggest 10 days. we can focus on fuel, apex, joid this week and some fixes for compass that is on yarsdtick side
15:09:22 <frankbrockners> and nosdn should be tablestakes... I would assume that plain vanilla ostack deployments work :-)
15:09:35 <frankbrockners> 10 business days?
15:10:08 <frankbrockners> that would mean start final release runs on Feb/22?
15:12:16 <frankbrockners> anac1: still there?
15:13:59 <debrascott> frankbrockners: we need to continue this discussion today but meeting is 15 over
15:14:09 <anac1> frankbrockners: partly... we can start with nosdn, let's see how thie runs tonight will go
15:14:43 <frankbrockners> ok... let's see whether we can get a bit better picture just prior to the TSC tomorrow
15:15:10 <debrascott> anac1: TSC will meet tomorrow and must decide on a date, please send out status no matter what time it is US time.
15:15:23 <anac1> debrascott: ok
15:15:36 <debrascott> anac1: ty
15:16:22 <debrascott> #endmeeting