rdo_meeting_-_2017-02-15
LOGS
15:01:03 <jpena> #startmeeting RDO meeting - 2017-02-15
15:01:03 <zodbot> Meeting started Wed Feb 15 15:01:03 2017 UTC.  The chair is jpena. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:03 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:01:03 <zodbot> The meeting name has been set to 'rdo_meeting_-_2017-02-15'
15:01:04 <openstack> Meeting started Wed Feb 15 15:01:03 2017 UTC and is due to finish in 60 minutes.  The chair is jpena. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:06 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:01:08 <eck`> number80: ah ok, i had no idea :)
15:01:08 <openstack> The meeting name has been set to 'rdo_meeting___2017_02_15'
15:01:10 <jpena> #topic roll call
15:01:12 <leifmadsen> o/
15:01:14 <tosky> o/
15:01:22 <amoralej> o/
15:01:30 <jpena> #chair leifmadsen tosky amoralej
15:01:30 <zodbot> Current chairs: amoralej jpena leifmadsen tosky
15:01:30 <openstack> Current chairs: amoralej jpena leifmadsen tosky
15:01:33 <rbowen> o/
15:01:34 <number80> o/
15:01:42 <jpena> #chair rbowen number80
15:01:42 <zodbot> Current chairs: amoralej jpena leifmadsen number80 rbowen tosky
15:01:43 <openstack> Current chairs: amoralej jpena leifmadsen number80 rbowen tosky
15:02:40 <chandankumar> \o/
15:03:19 <jpena> #chair chandankumar
15:03:19 <zodbot> Current chairs: amoralej chandankumar jpena leifmadsen number80 rbowen tosky
15:03:19 <openstack> Current chairs: amoralej chandankumar jpena leifmadsen number80 rbowen tosky
15:03:35 <social> anyone hitting type=AVC msg=audit(1487168176.211:89): avc:  denied  { read } for  pid=13482 comm="ip6tables.init" name="modprobe.d" dev="sda3" ino=8409209 scontext=system_u:system_r:iptables_t:s0 tcontext=unconfined_u:object_r:modules_conf_t:s0 tclass=dir
15:04:10 <jschluet> o/
15:04:20 <openstackgerrit> Evgeny Antyshev proposed openstack/packstack master: Ability to provide demo image's properties  https://review.openstack.org/434334
15:04:23 <jpena> #chair jschluet
15:04:23 <zodbot> Current chairs: amoralej chandankumar jpena jschluet leifmadsen number80 rbowen tosky
15:04:24 <openstack> Current chairs: amoralej chandankumar jpena jschluet leifmadsen number80 rbowen tosky
15:04:34 <jpena> let's start with the agenda
15:04:41 <jpena> #topic Ocata release checklist
15:05:49 <jpena> rbowen: ^^
15:05:50 <rbowen> As usual, we are drafting a blog post about the Ocata RDO release.
15:06:01 <rbowen> That is at
15:06:02 <rbowen> #link https://etherpad.openstack.org/p/rdp-ocata
15:06:18 <rbowen> The things we try to do around a release are described at
15:06:20 <rbowen> #link https://www.rdoproject.org/rdo/release-checklist/
15:06:34 <rbowen> What I need at this point is some life to that draft
15:06:41 <rbowen> Some bullet points of exciting things in this release.
15:06:53 <rbowen> What I have now is pretty much a "hey, we released something" and it's kind of boring.
15:07:05 <rbowen> So if there's things that you worked on that are interesting, please add bullet points to that draft.
15:08:07 <rbowen> Note that we're not taking credit for things that happened upstream, but it's worth mentioning a few of them, so people care about trying it out.
15:08:22 <rbowen> Thanks. </eom>
15:09:04 <jpena> #action everyone have a look at https://etherpad.openstack.org/p/rdp-ocata and contribute
15:09:45 <amoralej> rbowen, i have the list of new packages introduced in ocata
15:09:59 <rbowen> amoralej: Thanks.
15:10:22 <jpena> next topic?
15:10:29 <jpena> although related to this one
15:10:31 <rbowen> Yes, I'm done.
15:10:40 <jpena> #topic Review release engineering steps for the next week https://review.rdoproject.org/etherpad/p/ocata-release-preparation
15:10:50 <jpena> amoralej, I think you added this one
15:10:54 <amoralej> ok
15:11:12 <rdogerrit> Merged rdoinfo: Remove under-review tag for python-networking-sfc  https://review.rdoproject.org/r/5188
15:11:12 <amoralej> at the beginning of the etherpad you have the list of todo items
15:11:25 <amoralej> as, finish under-review packages
15:11:47 <amoralej> we need to create the ci jobs for ocata CBS repos
15:11:52 <amoralej> if someone can take it
15:12:00 <amoralej> it should be pretty easy, i think
15:12:12 <chandankumar> amoralej: what about creating a new list of new packages introduced in ocata release
15:12:20 <chandankumar> in the same package list
15:12:30 <chandankumar> s/new/
15:12:45 <number80> amoralej: which kind of jobs?
15:12:52 <number80> we're supposed to have them already
15:12:59 <amoralej> https://ci.centos.org/view/rdo/view/cbs-promote-pipelines/
15:13:05 <amoralej> i don't see there
15:13:18 <jpena> amoralej: oh, promotion jobs
15:13:19 <amoralej> maybe it's just the view
15:13:49 <amoralej> the ones we used to test official repos
15:13:59 <amoralej> buildlogs.c.o and mirror.c.o
15:15:12 <number80> AFAIK, they are a WIP
15:15:25 <amoralej> we need those to test the stable packages and to create publish images
15:15:41 <amoralej> lemme check
15:16:10 <number80> Yes for images, but they are not used for promotion as they're broken
15:16:51 <amoralej> https://review.rdoproject.org/r/#/c/4503/ is for RDO Trunk but not for cloudsig repos
15:17:03 <number80> We use weirdo jobs to test stable packaging
15:17:29 <amoralej> and oooq also number80
15:17:46 <number80> not really
15:17:53 <amoralej> is not https://ci.centos.org/view/rdo/view/cbs-promote-pipelines/job/rdo-promote-cloudsig-testing-newton/ used?
15:18:06 <amoralej> or you mean for GA?
15:18:07 <number80> not for tagging testing to stable
15:18:13 <trown> o/
15:18:41 <amoralej> anyway, we need to create weirdo-testing-ocata
15:18:49 <number80> amoralej: we tag automatically upstream GA to stable
15:18:54 <number80> Yes, I agree
15:19:13 <number80> dmsimard: ^
15:19:32 <dmsimard> yeah
15:19:39 <dmsimard> is the repo self contained and good to go yet ?
15:19:46 <number80> not yet
15:19:47 <amoralej> not yet dmsimard
15:20:08 <amoralej> but it'd be great if we can get the pipeline created as soon as possible
15:20:10 <dmsimard> ok, well, not just weirdo jobs anywya
15:20:12 <dmsimard> anyway*
15:20:15 <dmsimard> we need oooq things as well
15:20:30 <amoralej> we may be able to test some scenarios, as packstack pretty soon
15:21:33 <dmsimard> We should branch stable/ocata in packstack.
15:21:44 <amoralej> dmsimard, do you think we could get the pipelines for cloudsig this week?
15:21:52 <amoralej> yeah dmsimard, it's in the list too
15:22:02 <amoralej> i think we can branch and tag as-is
15:22:28 <dmsimard> amoralej: afaict for weirdo it's just to add one line with "ocata" in it, but we need to create a pipeline with oooq in it -- trown, anything special required for oooq ?
15:22:53 <trown> dmsimard: just need to make sure there is an ocata repo config that matches up
15:23:05 <number80> well, there is a proposal about release criteria but I wait for apevec/mburns to provide final draft
15:23:41 <number80> It will define which set of jobs will be used for releasing RDO
15:23:57 <openstackgerrit> Evgeny Antyshev proposed openstack/packstack master: Ability to customize flavors in Tempest  https://review.openstack.org/423216
15:24:04 <dmsimard> I guess there has to be something tripleo in there :)
15:24:21 <number80> #action number80 request kbsingh to create ocata stable repositories
15:24:44 <number80> Right, this is something which should be done ahead of time if we want to be ready
15:25:42 <trown> I can make sure there is repo config for tripleo quickstart
15:25:51 <dmsimard> trown: <3
15:25:56 <trown> #action trown make sure there is ocata relase config for quickstart
15:25:57 <amoralej> good trown
15:26:39 <amoralej> dmsimard, will you take care of adding the jobs to rdo-infra/ci-config?
15:26:47 <amoralej> i can send the weirdo part
15:26:53 <amoralej> but not that sure about oooq
15:27:06 <rdogerrit> David Moreau Simard created rdo-infra/ci-config: Add ocata weirdo-testing job  https://review.rdoproject.org/r/5277
15:27:10 <dmsimard> amoralej: ^ should do it, one line ;)
15:27:16 <amoralej> you were fast :)
15:27:23 <dmsimard> however we need to create the actual pipeline
15:27:46 <dmsimard> this is more or less copy/pasting https://github.com/rdo-infra/ci-config/blob/master/jenkins/jobs/promote-cloudsig-testing-newton.yml
15:28:13 <dmsimard> I can send a review for that once I get the ack from trown that it should be okay
15:28:21 <amoralej> ok
15:28:40 <amoralej> #action dmsimard to send cloudsig-testing-ocata once oooq is ready
15:28:56 <amoralej> I'd say, let's create for stable too
15:29:06 <dmsimard> that one won't work until stable is GA
15:29:12 <amoralej> not only -testing
15:29:13 <dmsimard> weirdo will fail horribly
15:29:26 <dmsimard> we can create it once stable is ga.
15:29:46 <amoralej> yes but, as we'll tag direct to -release in GA we'll probalby have to test there
15:29:49 <amoralej> ok
15:29:58 <amoralej> about CDN synchronization
15:30:08 <amoralej> do we need something to get the script ready?
15:30:11 <amoralej> jpena ^
15:30:16 <dmsimard> nothing for release, no
15:30:16 <amoralej> for ocata
15:30:30 <dmsimard> when number80 and apevec tag things for release, it's done automatically but kbsingh needs to sign
15:30:39 <jpena> for stable/ocata trunk, we need some current-passed-ci repo
15:30:56 <jpena> otherwise the CDN sync script won't work
15:31:09 <amoralej> we don't have current-passed-ci pipeline yet
15:31:16 <dmsimard> right, we did the current-tripleo one but current-passed-ci isn't there
15:31:19 <dmsimard> I'll take that
15:31:38 <dmsimard> #action dmsimard to create current-passed-ci pipeline for ocata
15:31:39 * number80 not sure to understand
15:31:57 <number80> Cloud SIG repo is completely managed by CentOS team
15:32:23 <amoralej> yes, number80, it just came to my mind the synchronization of rdo-trunk repos to buildlogs
15:32:31 <amoralej> not Cloud SIG
15:32:37 <number80> ah right
15:33:53 <amoralej> about packages, was still thinking about what to do with release-trailings
15:34:07 <amoralej> for tripleo, i think there may be a RC1 before GA
15:34:25 <amoralej> but probably not for neutron plugins
15:34:41 <amoralej> he was proposing cross-tagging newton release if they don't tag
15:34:57 <number80> I'd just wait until they're ready
15:35:20 <apevec> why, we are not blocking GA no them?
15:36:05 * number80 not sure if we're speaking about upstream or downstream GA
15:36:30 <apevec> number80, re. release criteria - I've https://trello.com/c/B25DJFZx/572-rdo-propose-an-evolution-of-dod but that will be for Pike
15:36:36 <number80> ack
15:36:52 <apevec> for Ocata we are on DoD from last year i.e. current-ci promotion
15:37:19 <apevec> = weirdo puppet/packstack + oooq non-ha/ha
15:37:43 <apevec> number80, not sure what you mean by downstream ?
15:37:51 <dmsimard> apevec: current-ci promotion is for trunk, this is for cbs
15:37:51 <apevec> but here we are discussing RDO
15:37:53 <number80> ok, so for components that are release blockers, I'd suggest to use snapshots from latest passing RDO trunk CI
15:38:02 <dmsimard> so the equivalent would be the -testing pipeline
15:38:06 <apevec> dmsimard, I mean equivalent
15:38:08 <dmsimard> right
15:38:35 <number80> apevec: I was not sure to understand the sentence, just forget it, you already answered my question :)
15:38:55 <number80> anything that is not release blocker and not ready by GA, should wait
15:38:57 <apevec> number80, so snaps might be a problem when upstream versioning is broken
15:39:02 <apevec> i.e. master < stable/newton
15:39:02 <number80> e.g networking-* mess
15:39:12 <apevec> hence I propose to just retagg latest newton
15:39:17 <apevec> to avoid update issues
15:39:40 <apevec> other opt is to just not ship them, but not sure what that might break
15:39:55 <amoralej> we used to install neutron plugins in ooo
15:39:59 <amoralej> in overcloud images
15:40:01 <number80> apevec: if they're release blockers, ack, not release blockers, I am not convinced (e.g networking)
15:40:04 <amoralej> we disabled some of them
15:40:12 <amoralej> but i'm not sure about current status
15:40:29 <amoralej> trown ^ do you know what neutron plugins do we install in overcloud?
15:41:19 * number80 thinks that they are likely not tested
15:41:21 <amoralej> so apevec, number80, anything not required for DoD criteria is considered not release blocker?
15:41:36 <apevec> right
15:41:43 <amoralej> yeah, but if they are installed we need to provide some installable packages
15:42:05 <number80> Yes, but that's testing by ooq image build job
15:42:32 <apevec> amoralej, ok, then action is to make a list what is included/used in ooo
15:42:36 <amoralej> yeah, so that's a good point
15:42:40 <amoralej> yes i will
15:43:06 <amoralej> #action amoralej create a list of required packages for oooq jobs
15:43:09 <apevec> and ship those as snaps from passed-ci, other's -> noship until they have ocata release
15:43:17 <trown> amoralej: not sure
15:43:27 <amoralej> i'll check trown, no problem
15:44:28 <amoralej> same for other packages?
15:44:36 <amoralej> tacker is not branched or tagged yet
15:45:33 <amoralej> and i think is not tested in our pipeline
15:45:49 <apevec> noship
15:46:05 <amoralej> ok
15:46:30 <number80> for tacker, I'd check w/ dradez post-release, it could be a blocker for OPNFV
15:47:06 <amoralej> yeah, i was looking for him yesterday but he didn't connect
15:47:10 <amoralej> i'll send him a mail
15:48:51 <amoralej> i think that's it, regarding ocata release
15:49:09 <jpena> next topic, then?
15:49:19 <amoralej> sorry, just a last one
15:49:42 <amoralej> we should make sure we have enough slaves to rebuild with GA tag
15:49:45 <amoralej> in review.r.o
15:50:35 <amoralej> i'd propose to let DLRN-rpmbuild and cbs jobs run on any node
15:50:49 <amoralej> bare centos or weirdo nodes
15:51:16 <amoralej> wdyt?
15:51:51 <jpena> it should work on both, but what's the advantage? bare-centos can run on any of our two clouds
15:52:10 <amoralej> last week there was periods where we had a bunch of empty weirdo ones and few bare-centos
15:52:12 <amoralej> and long queue
15:52:18 <apevec> we could get temp more trystack nodes
15:52:20 <amoralej> i'm not sure if that's fixed
15:52:31 <apevec> dmsimard, ^
15:52:58 <jpena> I think that was related to the change we made to use bare-centos nodes, I saw some mismatch between what nodepool thought it was requested, and what zuul was actually requesting
15:52:59 <dmsimard> we made it so we could run the koji jobs on bare-centos
15:53:07 <dmsimard> apevec, number80: right ?
15:53:25 <dmsimard> we fixed by adding centos-packager and createrepo or something
15:53:58 <dmsimard> so dlrn-rpmbuild and cbs jobs can run on bare-centos-7 nodes which are available on both rcip-dev and trystack
15:54:20 <dmsimard> FWIW we got greenlighted some special case usage of more trystack capacity, if we're in a pinch, I'll be able to add it
15:54:47 <amoralej> how many bare-centos-7 nodes can we get?
15:54:49 <number80> yes
15:55:00 <dmsimard> we can get probably about 20 more
15:55:45 <amoralej> so, iiuc, there is a global quota on each cloud and a minimal number of nodes of each type on each cloud ?
15:55:48 <dmsimard> brings us to a theoritical max of about ~35 concurrent bare-centos-7
15:56:07 <amoralej> ok, that should be ok
15:56:21 <jpena> to be time-conscious, we're 4 minutes before the end of the hour and have one pending topic
15:56:36 <amoralej> yeah, ok, i'm done
15:56:38 <dmsimard> #action dmsimard to prep additional trystack capacity for bare-centos-7
15:56:46 <dmsimard> #undo
15:56:50 <dmsimard> #action dmsimard to prep additional temporary trystack capacity for bare-centos-7
15:56:54 <dmsimard> keyword temporary :)
15:56:58 <jpena> cool
15:57:00 <jpena> let's move on
15:57:01 <dmsimard> hm, I might not be chair
15:57:15 <jpena> #action dmsimard to prep additional temporary trystack capacity for bare-centos-7
15:57:20 <jpena> I think zodbot is quiet
15:57:28 <dmsimard> ah
15:57:34 <jpena> #chair     PTG - if you want to talk about what you did in Ocata, and you'll be at the PTG, please contact rbowen@redhat.com to schedule your interview
15:57:34 <zodbot> Current chairs: - Ocata PTG about amoralej and at be chandankumar contact did if in interview jpena jschluet leifmadsen number80 please rbowen rbowen@redhat.com schedule talk the to tosky want what you you'll your
15:57:36 <openstack> Warning: Nick not in channel: PTG
15:57:36 <rbowen> This is the last time I'll bug you about this. If you'll be at the PTG, and want to talk for 10-20 minutes about the work you did on Ocata, please ping me to schedule your interview. I have roughly 12 appointments open, with 6 of those tentatively spoken for.
15:57:37 <openstack> Warning: Nick not in channel: -
15:57:38 <openstack> Warning: Nick not in channel: if
15:57:39 <openstack> Warning: Nick not in channel: you
15:57:40 <openstack> Warning: Nick not in channel: want
15:57:41 <openstack> Warning: Nick not in channel: to
15:57:42 <openstack> Warning: Nick not in channel: talk
15:57:43 <openstack> Warning: Nick not in channel: about
15:57:43 <dmsimard> uh oh
15:57:45 <openstack> Warning: Nick not in channel: what
15:57:46 <openstack> Warning: Nick not in channel: did
15:57:46 <jpena> ouch
15:57:47 <rbowen> What the heck
15:57:52 <jpena> #topic PTG - if you want to talk about what you did in Ocata, and you'll be at the PTG, please contact rbowen@redhat.com to schedule your interview
15:57:56 <rbowen> Um
15:57:57 <rbowen> ok,
15:58:09 <rbowen> well, as the agenda item says, I'll be doing interviews at the PTG
15:58:15 <rbowen> about what you did in Ocata.
15:58:34 <rbowen> These will be used at future conferences, on various blog sites, and at OpenStack Summit.
15:58:51 <rbowen> I have just a few appointments remaining, so if you like to talk, ping me.
15:58:58 <rbowen> THanks to all of you who have already signed up.
15:59:02 <rbowen> </eom>
15:59:14 <jpena> thanks rbowen
15:59:21 <jpena> #topic chair for next meeting
15:59:25 <jpena> any volunteers?
16:00:45 <amoralej> ok, i can take it
16:00:56 <jpena> #action amoralej to chair the next meeting
16:00:59 <jpena> thanks amoralej ;)
16:01:31 <jpena> we're over the hour, let's close the meeting and then have any open floor chat
16:01:34 <jpena> #endmeeting