rdo_meeting_-_2016-09-14
LOGS
15:00:48 <chandankumar> #startmeeting RDO meeting - 2016-09-14
15:00:48 <zodbot> Meeting started Wed Sep 14 15:00:48 2016 UTC.  The chair is chandankumar. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:48 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:00:48 <zodbot> The meeting name has been set to 'rdo_meeting_-_2016-09-14'
15:00:48 <openstack> Meeting started Wed Sep 14 15:00:48 2016 UTC and is due to finish in 60 minutes.  The chair is chandankumar. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:49 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:50 <apevec> yes
15:00:51 <openstack> The meeting name has been set to 'rdo_meeting___2016_09_14'
15:00:57 <chandankumar> #topic roll call
15:00:59 <chandankumar> \o/
15:01:09 <apevec> bot clash!
15:01:12 <jpena> o/
15:01:20 <amoralej> o/
15:01:21 <dmsimard> \o
15:01:26 <hrybacki> o/
15:01:28 <DrBacchus> o/
15:01:36 <flepied> o/
15:01:37 <number80> damn, I'm not an op currently
15:01:50 <chandankumar> #chair apevec jpena amoralej dmsimard DrBacchus flepied hrybacki number80
15:01:50 <zodbot> Current chairs: DrBacchus amoralej apevec chandankumar dmsimard flepied hrybacki jpena number80
15:01:57 <trown> o/
15:02:02 <chandankumar> #chair trown
15:02:02 <zodbot> Current chairs: DrBacchus amoralej apevec chandankumar dmsimard flepied hrybacki jpena number80 trown
15:02:35 <chandankumar> jruzicka: meeting time
15:02:39 <jruzicka> o/
15:02:54 <misc> nan mais ça reshi
15:03:01 <chandankumar> #chair jruzicka
15:03:01 <zodbot> Current chairs: DrBacchus amoralej apevec chandankumar dmsimard flepied hrybacki jpena jruzicka number80 trown
15:03:07 <misc> argh
15:03:08 <misc> hi
15:03:12 <jrist> o/
15:03:18 <chandankumar> #chair jrist
15:03:18 <zodbot> Current chairs: DrBacchus amoralej apevec chandankumar dmsimard flepied hrybacki jpena jrist jruzicka number80 trown
15:03:26 <chandankumar> so starting with first topic
15:03:37 <chandankumar> #topic Infrastructure concerns
15:03:44 <chandankumar> Duck: please go ahead.
15:04:00 <Duck> quack
15:04:02 <chandankumar> #chair Duck
15:04:03 <zodbot> Current chairs: DrBacchus Duck amoralej apevec chandankumar dmsimard flepied hrybacki jpena jrist jruzicka number80 trown
15:04:06 <coolsvap> o/
15:04:13 <chandankumar> #chair coolsvap
15:04:13 <zodbot> Current chairs: DrBacchus Duck amoralej apevec chandankumar coolsvap dmsimard flepied hrybacki jpena jrist jruzicka number80 trown
15:04:51 <jschlueter> o/
15:04:56 <chandankumar> #chair jschlueter
15:04:56 <zodbot> Current chairs: DrBacchus Duck amoralej apevec chandankumar coolsvap dmsimard flepied hrybacki jpena jrist jruzicka jschlueter number80 trown
15:05:49 <chandankumar> Duck has some infrastructure concerns for for seperate mailing lists and dedicated meeting.
15:06:30 <rdogerrit> Merged openstack/UcsSdk-distgit: Use pypi.io for Source0 URL  http://review.rdoproject.org/r/2284
15:06:43 <Duck> well, more generally, on how to help on the infra
15:06:54 <apevec> [infra] tag on rdo-list would not be good enough?
15:06:59 <Duck> and having a proper channel to discuss would be nice
15:07:14 <Duck> apevec: it is not used, as I know of
15:07:15 <apevec> let's keep all on #rdo
15:07:28 <dmsimard> I think this was brought up before, I think it's appropriate to use rdo-list (with [infra] tag perhaps) and it's fine to keep discussions in #rdo as far as I'm concerned. Don't be shy :)
15:07:31 <apevec> yeah, let's start using it to help easier filtering
15:07:44 <Duck> when the certificate expiration happened, twice, there was nothing on the ML
15:07:59 <apevec> there sensu checks
15:08:11 <Duck> I'm fine with it but it does NOT work
15:08:12 <apevec> you mean would should send email alerts from those?
15:08:19 <dmsimard> Duck: there actually was an email on the ML :)
15:08:25 <Duck> also this meeting never has any infra topics
15:08:36 <dmsimard> https://www.redhat.com/archives/rdo-list/2016-August/msg00237.html
15:08:41 <Duck> which would mean everything is perfect but… I tend to not believe it
15:08:44 <apevec> we have one today about CBS outage :)
15:08:52 <DrBacchus> It didn't use the specified tag.
15:09:10 <Duck> dmsimard: without the tag => filtered out
15:09:17 <dmsimard> Duck: the meeting tends to be regarding topics of discussions, not updates on the status of the infrastructure
15:09:22 <Duck> so there is de facto no mail to wrn
15:09:23 <DrBacchus> The difficulty being that Duck and misc are responsible for infra across many projects, and are unlikely to see these messages unless they are appropriately flagged.
15:09:27 <apevec> shall we document list of "official" tags ?
15:09:52 <dmsimard> apevec: it's hard to enforce, even on openstack-dev there are untagged emails :)
15:10:06 <apevec> yeah, but we can try
15:10:08 <misc> the concern is also that infrastructure is all over the place
15:10:19 <misc> and we didn't followed up on the rdo-infra plan :/
15:10:40 <dmsimard> misc: did you mention you were coming to Barcelona ?
15:10:54 <dmsimard> It could be a great opportunity to have work sessions
15:10:54 <misc> dmsimard: yep
15:11:01 <misc> dmsimard: and yep
15:11:12 <DrBacchus> While it's understood that people will spin up stuff as needed, we want to ensure that those services are supported long-term, and that requires access, documentation, and automation.
15:11:15 <Duck> so with rbowen we began to list the resources, where they are, and who is responsible for
15:11:18 <misc> but I think duck and I have different concerns :)
15:11:27 <apevec> if folks arrive early on Monday, let's plan rdo infra meetup then?
15:11:31 <Duck> would be nice to ensure it is complete
15:11:43 <Duck> and there is no SPOF
15:11:46 <dmsimard> Duck, DrBacchus: I believe jpena has an inventory somewhere
15:11:50 <DrBacchus> That doc is here: https://docs.google.com/spreadsheets/u/1/d/1cpTHMixEiaqHA-qR-ol2uO76USYna08WrTvUJPCxHSI/edit#gid=0
15:11:51 <dmsimard> of the hosts
15:11:52 <Duck> and a todolist of things so we could pick some
15:11:55 <Duck> etc etc
15:11:57 <DrBacchus> Oh, ok, well, we should merge those things, then.
15:12:03 <dmsimard> DrBacchus: that document isn't public
15:12:33 <DrBacchus> Yeah, at the moment it's invite-only but I can move that to somewhere public.
15:12:43 <dmsimard> DrBacchus: It's ok I'll request access
15:12:51 <DrBacchus> There's also https://www.rdoproject.org/infra/ which is probably not up to date either.
15:12:55 <misc> another concern is also that we have no idea on who do what :)
15:13:24 <apevec> let's document that all in a single doc
15:13:25 <dmsimard> misc: it's a fair assumption that jpena and I are responsable/accountable for most of the infrastructure
15:13:30 <Duck> so there's a few things OSAS is involved into, like web stuff
15:13:44 <Duck> the goal is to have more things public and more people involved too
15:13:59 <Duck> so no take over, no blocking people, etc
15:14:14 <rdogerrit> hguemar created openstack/neutron-lib-distgit: Sync requirements  http://review.rdoproject.org/r/2292
15:14:59 <apevec> ok, problem statement is clear, let's get actions
15:15:21 <rdogerrit> hguemar created openstack/neutron-lib-distgit: Create rpm-mitaka branch  http://review.rdoproject.org/r/2293
15:15:22 <chandankumar> feel free to add actions against yourself. :-)
15:15:27 <DrBacchus> So, action items are 1) update https://www.rdoproject.org/infra/ with all the services, and who is responsible for each
15:15:39 <dmsimard> apevec: I'm +1 for a work session on monday in BCN
15:15:51 <number80> +1
15:16:00 <trown> #chair rbowen
15:16:00 <zodbot> Current chairs: DrBacchus Duck amoralej apevec chandankumar coolsvap dmsimard flepied hrybacki jpena jrist jruzicka jschlueter number80 rbowen trown
15:16:06 <dmsimard> It should be productive and put us all on the same page
15:16:07 <rbowen> That seems a reasonable place to start, as that is the biggest hole at the moment, it seems?
15:16:12 <Duck> BCN?
15:16:17 <rbowen> barcelona
15:16:18 <misc> barcelona
15:16:18 <number80> Barcelona IATA code
15:16:22 <Duck> ho
15:16:34 <misc> (cause RH use IATA code for whatever historical reason)
15:16:34 <Duck> so misc you'll be my representative then
15:16:39 <chandankumar> rbowen: will i add that action item to you?
15:16:44 <jpena> +1 to work session, but I'm arriving to BCN around 11am
15:16:47 <flepied> we meet at the airport ? ;-)
15:16:51 <Duck> misc: with all my concerns too
15:17:05 <misc> Duck: sure
15:17:06 <rbowen> chandankumar: Sure, I can add the info that I have. Also need jpena to crosscheck and update.
15:17:07 <dmsimard> rbowen, Duck, misc: Are we comfortable with starting with that ? BCN is in a month.
15:17:10 <number80> flepied: just redhatter internal slang :)
15:17:14 <misc> dmsimard: I am ok
15:17:39 <rbowen> #action rbowen to update https://www.rdoproject.org/infra/ with services and responsibilities, then hand off to jpena for crosscheck.
15:17:46 <apevec> we can work on documents in the meantime
15:18:13 <jpena> rbowen: ack
15:18:13 <chandankumar> next topic?
15:18:16 <Duck> dmsimard: well, probably better to have some live to work better. May be able to join via Blue Jeans.
15:18:21 <imcsk8> o/
15:18:26 <Duck> yes, let's clarify the doc
15:18:28 <chandankumar> #chair imcsk8
15:18:28 <zodbot> Current chairs: DrBacchus Duck amoralej apevec chandankumar coolsvap dmsimard flepied hrybacki imcsk8 jpena jrist jruzicka jschlueter number80 rbowen trown
15:18:34 <dmsimard> Duck: yeah sure, we can have a hangouts or bluejeans
15:19:04 <Duck> dmsimard: could you do an invitation when exact timing is determined?
15:19:06 <dmsimard> #action Schedule a work session for RDO Infrastructure sometime, somewhere monday afternoon
15:19:06 <misc> dmsimard: we likely need to go to the office, no ?
15:19:16 <misc> cause I do not trust network to work :)
15:19:19 <dmsimard> misc: ah, there is an office in bcn? We could do it there
15:19:27 <dmsimard> how far is it from the conference center ?
15:19:29 <jpena> dmsimard: yep, there's an office
15:19:40 <amoralej> a brand new one :)
15:19:42 <Duck> misc: you need to see, battlle with, and drink with people live
15:19:43 <dmsimard> #undo
15:19:43 <zodbot> Removing item from minutes: ACTION by dmsimard at 15:19:06 : Schedule a work session for RDO Infrastructure sometime, somewhere monday afternoon
15:19:45 <number80> misc: ack for office, but we need to take care of unbadged people who wants to join us
15:19:50 <rbowen> There may also be space/time at the event venue. I need to check with Jmadriaga on that. Unless we prefer the office.
15:20:00 <dmsimard> #action Schedule a work session for RDO Infrastructure sometime at the BCN Red Hat office
15:20:07 <apevec> rbowen, I'd prefer conf
15:20:13 <dmsimard> bah
15:20:14 <misc> dmsimard: not too far
15:20:16 <number80> rbowen: that'd be fantastic if we get space at conference site
15:20:20 <dmsimard> #undo
15:20:20 <zodbot> Removing item from minutes: ACTION by dmsimard at 15:20:00 : Schedule a work session for RDO Infrastructure sometime at the BCN Red Hat office
15:20:32 <rbowen> ok, I'll ask and see what there is. Should be easier Monday than actually during the event.
15:20:32 <dmsimard> #action Schedule a work session for RDO Infrastructure somewhere monday afternoon
15:20:44 <apevec> 3rd time is a charm
15:21:08 <apevec> move on?
15:21:15 <chandankumar> #topic phase out rpm-xxx branches
15:21:24 <chandankumar> number80: you?
15:22:12 <apevec> rpm-RELEASE to be precise
15:22:14 <Duck> rbowen: thanks
15:22:15 <apevec> rpm-master stays :)
15:22:39 <number80> yes
15:22:57 <apevec> related card is https://trello.com/c/URAtrhLU/86-automate-stable-packages-releases
15:23:06 <number80> as soon as I'll finish branching newton, I'll submit a review to drop rpm-RELEASE branches
15:23:11 <apevec> we're starting this with Newton
15:23:17 <number80> Yes
15:23:24 <number80> no rpm-newton
15:23:37 <jpena> are we removing rpm-mitaka and earlier?
15:23:43 <apevec> ideally yes
15:23:49 <apevec> do you see issues with migration?
15:23:59 <number80> jpena: it will need testing to avoid storming DLRN
15:24:03 <number80> apevec: ^
15:24:09 <number80> hash are differents
15:24:15 <jpena> that requires DLRN/puppet-dlrn support, just need to know in advance to update DLRN workers and fix any errors
15:24:41 <number80> actually if keeping rpm-mitaka is preferred, that's an option
15:24:43 <jpena> we have a bit of backlog
15:24:45 <apevec> number80, can't avoid that, we'll have to let full rebuild
15:25:12 <apevec> but yeah if we don't have DLRN resource to process that quickly better leave mitaka
15:25:19 <number80> #info do not merge any rdoinfo review removing rpm-RELEASE branches before checking w/ jpena
15:25:30 <number80> please :)
15:25:51 <rdogerrit> trown created rdoinfo: Add DIB and python-tripleoclient to tripleo UC exceptions  http://review.rdoproject.org/r/2294
15:25:52 <jpena> number80: my plan is to update DLRN/puppet-dlrn on the instance when we create the Ocata workers
15:26:07 <jpena> that's next Friday or Monday next week
15:26:14 <jpena> (I guess)
15:26:21 <number80> jpena: ack, I'll submit review and will let you merge it
15:26:21 <apevec> yes, RC1 is branching point
15:26:45 <number80> yeah newton branching status
15:26:51 <number80> clients are all done
15:26:53 <apevec> #link https://trello.com/c/zT4BghsV/366-branching-newton-rdo
15:27:18 <number80> services are starting today, I identified few ones that could be done in // by someone else ^
15:27:28 <number80> (I'll add more later)
15:27:29 <amoralej> i'd like to help on that
15:27:36 <number80> \o/
15:27:37 <apevec> yes, me and amoralej will join
15:27:48 <amoralej> but i'll need some additional permissions in review.r.o
15:28:01 <apevec> just need to split the checklist among us
15:28:09 <apevec> amoralej, I'll look at that after meeting
15:28:21 <amoralej> ok apevec, let's sync after the meeting
15:28:27 <apevec> #action apevec to check branching permissions for amoralej
15:28:43 <number80> as for mitaka, I'm checking that in parallel of newton branching
15:29:12 <number80> so as follow-up, I have some fedora bundled updates to do + mass-tagging
15:29:25 <number80> mass-tagging will happen after running weirdo generic jobs
15:29:37 <imcsk8> apevec, amoralej i would like to help with branchin too
15:29:59 <apevec> imcsk8, ack - let's first see how initial split goes
15:30:12 <apevec> we might have so other followup tasks besides branching
15:30:19 <number80> yes
15:30:43 <Duck> number80: btw, I live in Mitaka :-)
15:30:52 <Duck> number80: Mitaka-shi
15:30:54 <imcsk8> apevec: ok, i'll be standing by
15:30:55 <apevec> #info  number80 is checking Mitaka in parallel of newton branching
15:31:06 <chandankumar> next topic?
15:31:10 <number80> yes
15:31:16 <chandankumar> #topic python-shade packaging
15:31:30 <rdogerrit> Merged openstack/glance_store-distgit: Use tarballs.o.o for Source0 URL  http://review.rdoproject.org/r/2277
15:31:54 <chandankumar> I think it goes under  release independet repo rdo-tools
15:32:07 <apevec> yes, it was proposed to put it in EPEL
15:32:24 <apevec> but I sugget to develop it first in new rdo-tools (coming soon) repo
15:32:44 <apevec> which was discussed on rdo-list and one of the previous meetings
15:32:51 <chandankumar> apevec: yes,
15:33:05 <chandankumar> up to what time new rdo-tools repo is coming?
15:33:08 <number80> apevec: can you take care of repo creation w/ CentOS?
15:33:13 <number80> chandankumar: asap
15:33:17 <apevec> it's lower prio due to newton wrapup
15:33:25 <apevec> but I'll start the process w/ tags in CBS
15:33:38 <apevec> #action apevec to request CBS tags for rdo-tools repo
15:33:58 <apevec> related, I've already requested Ocata CBS tags
15:34:20 <apevec> so we'll start with ocata-testing for deps on Trunk master
15:34:24 <apevec> jpena, ^ FYI
15:34:38 <jpena> apevec: ack, the puppet-dlrn update for that is already updated
15:34:59 <chandankumar> Anything more on this, if not then next topic?
15:35:07 <apevec> next
15:35:14 <chandankumar> #topic [Ci-users] [Infra] - Major outage for several services
15:35:22 <number80> thank
15:35:24 <chandankumar> https://lists.centos.org/pipermail/ci-users/2016-September/000392.html
15:35:29 * Arrfab is here
15:35:31 <chandankumar> please read it
15:35:41 <chandankumar> Arrfab: dmsimard please go ahead
15:35:44 <apevec> so D-Day is Monday Oct 10
15:36:03 <apevec> plan is to have Newton GA repo pushed before that
15:36:14 <chandankumar> #chair Arrfab
15:36:14 <zodbot> Current chairs: Arrfab DrBacchus Duck amoralej apevec chandankumar coolsvap dmsimard flepied hrybacki imcsk8 jpena jrist jruzicka jschlueter number80 rbowen trown
15:36:16 <dmsimard> What about cycle-trailing ?
15:36:17 <apevec> Arrfab, just to confirm buildlogs.c.o and mirror.c.o are not affected ?
15:36:40 <chandankumar> Outage for: CBS, ci.centos.org, internal DLRN build server
15:36:40 <apevec> dmsimard, trown is optimistic 3o will be ready Oct 6
15:36:43 <Arrfab> apevec: they aren't
15:36:59 <trown> EmilienM: is ^ totally off base?
15:37:02 <dmsimard> that is quite optimistic :)
15:37:18 <dmsimard> considering puppet-openstack typically doesn't release the same day as GA
15:37:22 <Arrfab> apevec: so last chance to see pkgs landing on either buildlogs or mirror.centos.or is monday 10th in the morning (CET time)
15:37:27 <apevec> remember that number80's goal from Austin meetup is to release RDO minutes after upstream announcement :)
15:37:39 <apevec> Arrfab, ack
15:37:56 <Arrfab> dmsimard: true that you mentioned some "sub-components" that will probalby only appear in the days after GA
15:37:57 <EmilienM> dmsimard: we do
15:38:03 <Arrfab> dmsimard: nothing new about this ?
15:38:11 <EmilienM> we released Mitaka on GA and will release newton on GA
15:38:12 <jschlueter> ack on not available Oct 10
15:38:17 <dmsimard> EmilienM: you do ? I guess I'm thinking of something else
15:38:18 <EmilienM> only TripleO might slip a bit
15:38:24 <jpena> I understand the internal DLRN build server is affected?
15:38:27 <dmsimard> jpena: yes
15:38:28 <EmilienM> we have max 2 weeks to release final
15:38:43 <jpena> so no builds during these 2 days unless we have a fallback
15:38:49 <dmsimard> jpena: it doesn't last two days
15:39:01 <dmsimard> jpena: We're looking at an optimistic 8 hours of downtime on the 10th
15:39:02 <number80> Yeah, I like M-I style goals :D
15:39:13 <trown> dmsimard: I thought it was 48
15:39:21 <dmsimard> trown: you probably read 4-8
15:39:25 <dmsimard> right quaid ^ ?
15:39:30 <Arrfab> dmsimard: I wouldn't count on such a small downtime though (even if we want to minimize that)
15:39:41 <jpena> The mail says "the announced downtime will be probably ~48h."
15:39:43 <apevec> it's 48h
15:39:57 <Arrfab> trown: correct : that's the downtime announced by RH IT to move and power back the racks
15:40:04 <dmsimard> Arrfab: hrm, I think I saw quaid mention 4-8 hours. Must have misread that :(
15:40:21 <jpena> so, should we kick a temp build server just in case we have an emergency for centos-newton?
15:40:37 <dmsimard> jpena: either that or expedite the nodepool setup
15:41:10 <Arrfab> jpena: if that's critical for you/rdo, wondering if we want to postpone the migration (but all the RH people will be angry about this, again)
15:41:19 <apevec> dmsimard, jpena yeah, good opportunity to move to new arch :)
15:41:33 <apevec> Arrfab, no, let do it
15:41:37 <jpena> Arrfab, no need to postpone
15:41:43 <apevec> there's never good time
15:41:55 <apevec> Ocata is already starting
15:42:01 <Arrfab> ok, as said, we'll work hard with all the involved people to get cbs.centos.org back online as first service
15:42:01 <dmsimard> jpena: I think if we make the nodepool idea a priority it's realistic to have it live before oct 10th, I don't believe there is a lot of work involved -- most of the pre-reqs are already in place ?
15:42:02 <apevec> CD is 24x7x365
15:42:15 <Arrfab> and CI being the second one (in parallel)
15:42:34 <apevec> Arrfab, ack for that prio list
15:42:36 <rdogerrit> trown proposed rdoinfo: Add DIB to tripleo UC exceptions  http://review.rdoproject.org/r/2294
15:42:49 <Arrfab> dmsimard: just curious : what's that "nodepool" thing ?
15:43:08 <dmsimard> Arrfab: we want to decentralize the package building to help us scale and build things faster
15:43:32 <quaid> dmsimard: I was saying what I'd wanted not what we had, sorry for the confusion
15:43:35 <jpena> dmsimard: my only concern is where we can start with it (and how we can connect the nodepool instances with the "central" DLRN server), but yes, let's go for it
15:43:50 <dmsimard> Arrfab: as we package more and more projects, and there are more and more commits (and we package every commit), one single build server is no longer enough
15:43:56 <Arrfab> dmsimard: oh, ok, but so those other pkgs will never appear on the centos infra either ...
15:44:15 <dmsimard> Arrfab: nodepool is what we use in review.rdoproject.org to dispatch builds onto ephemeral VMs (same thing as upstream openstack-infra)
15:44:45 <Arrfab> dmsimard: so a kind of "pre-build" before submitting a pkg to cbs.centos.org ?
15:44:49 <apevec> Arrfab, those are Trunk rpms which we push to http://buildlogs.centos.org/centos/7/cloud/x86_64/rdo-trunk-master-tested/
15:45:03 <apevec> it's build on each upstream commit
15:45:14 <apevec> push to buildlogs is after passing CI
15:45:28 <Arrfab> apevec: yeah, but that I mean is that if you built on something else (external builders), that will not appear on buildlogs either
15:45:41 <Arrfab> apevec: *ack*
15:45:52 <apevec> that's a challenge that jpena and dmsimard  are solving
15:46:23 <chandankumar> action items here?
15:46:32 <Arrfab> jpena, dmsimard : let's discuss that when you have a minute
15:46:50 <dmsimard> #action dmsimard and jpena to start working towards moving dlrn package building to review.rdoproject.org (nodepool) on post-commit upstream pipeline
15:46:57 <Arrfab> from my side, I so consider that we keep the announced date for HW migration
15:47:06 <jpena> yes
15:47:31 <chandankumar> next topic?
15:47:32 <jpena> worst case, if the package building is not ready when we're close to the deadline, we can spawn a temp builder
15:48:12 <apevec> yes, have a planB
15:48:28 <Arrfab> jpena: well, as said, up2you, but don't expect those pkgs to land on centos infra then (as long as everybody is aware, that's fine)
15:48:58 <jpena> Arrfab: understood. After the DLRN instance is back, we can just copy stuff back to it
15:50:30 <apevec> let's move on, we have few more topic
15:50:35 <chandankumar> #topic exceptions for upper-constraints based repo for tripleo gated projects
15:50:40 <chandankumar> trown: please go ahead
15:50:41 <trown> this is mine
15:50:53 <apevec> please do not use master-head !
15:50:54 <trown> this is pretty much resolved in the etherpad
15:51:12 <trown> I did not realize there was a list of exceptions for tripleo to upper-constraints
15:51:14 <apevec> we just need a list which projects should be unpinned
15:51:17 <trown> put up a patch to add DIB http://review.rdoproject.org/r/2294
15:51:22 <trown> EOL
15:51:32 <chandankumar> trown: so fast!
15:51:32 <apevec> cool, +2
15:51:52 <chandankumar> moving to next item if no action item?
15:51:58 <trown> yep
15:52:12 <chandankumar> #topic What should we improve for test days?
15:52:17 <chandankumar> rbowen: please go ahead
15:52:35 <chandankumar> some suggestions from etherpad:
15:52:37 <rbowen> Well, I didn't create this topic, but, yeah, it's something that we need to do better.
15:52:46 <rbowen> There are some good suggestions in there.
15:52:49 <chandankumar> * Test setup matrix cleanup
15:53:00 <chandankumar> * More detailed, beginner-friendly instructions for testing. (Test day as product preview for potential users.)
15:53:08 <rbowen> I would like test day to bringing potential new users to us, and be a way for people to see what's coming in the future.
15:53:09 <apevec> color says it was flepied
15:53:11 <chandankumar> * ask other teams to provide use cases they want to validate
15:53:18 <rbowen> That's hard at the moment, because the instructions assume you already know a lot.
15:53:27 <chandankumar> * only on single etherpad with instructions and registering people to use cases
15:53:28 <flepied> yes it was me as a feedback from doing the last ones
15:53:36 <rbowen> This is perhaps something that pkovar and the other docs folks can help with? Or QE?
15:53:50 <imcsk8> chandankumar: +1 for registering people
15:53:51 <rbowen> I think people show up, but they're not sure how to get started.
15:54:01 <trown> we also kind of need new folks to take on these actions, as just getting stuff testable consumes much of existing pre-testday effort
15:54:03 <imcsk8> would test scripts help?
15:54:15 <rbowen> Test scripts could be a useful thing, sure.
15:54:24 <chandankumar> imcsk8: +1 for different test setup
15:54:41 <rbowen> trown: Yeah, folks spend most of test day just standing it up, and then we're out of time.
15:54:54 <chandankumar> what about including tempest run on each of the setup?
15:55:47 <imcsk8> i think test days are more about user experience
15:56:30 <chandankumar> Any action items from here?
15:56:34 <flepied> imcsk8: agreed. we should target to reproduce what a typical user will experiment when installing RDO
15:56:40 <imcsk8> there could be more promotion for test days amond the community
15:56:56 * chandankumar reminds we have 4 mins left.
15:57:06 <rbowen> Perhaps we can take this back to the mailing list, then.
15:57:11 <chandankumar> rbowen: +1
15:57:16 <amoralej> but then, it'd be useful to let users provide real use cases instead of follow some of them provided by us
15:57:16 <apevec> +1
15:57:18 <imcsk8> +1
15:57:18 <rbowen> We have a lot of actions, but unclear who will actually do them.'
15:57:20 <rdogerrit> hguemar proposed openstack/neutron-distgit: Add %upstream_version fallback macros  http://review.rdoproject.org/r/2291
15:57:21 <rdogerrit> hguemar created openstack/neutron-distgit: Sync requirements  http://review.rdoproject.org/r/2295
15:57:40 <chandankumar> rbowen: will you do that?
15:57:45 <rbowen> Yes, I will.
15:57:57 <rbowen> #action rbowen to raise topic of improving test day experience on rdo-list
15:58:05 <chandankumar> #topic Announcements
15:58:20 <chandankumar> #info Demos/Q&A at RDO booth at OpenStack Summit Barcelona - signup at https://etherpad.openstack.org/p/rdo-barcelona-summit-booth
15:58:34 <chandankumar> #info We are planning an RDO Community Meetup, in conjunction with the Ceph community, at OpenStack Summit. Tuesday night, October 25th, 5pm-8pm. Agenda is at https://etherpad.openstack.org/p/rdo-barcelona-meetup-schedule
15:58:45 <chandankumar> #info Newton RC1 test day September 22, 23 - Details coming soon to https://www.rdoproject.org/testday/
15:59:16 <chandankumar> #info New tempest rpms are available now here https://trunk.rdoproject.org/centos7-master/consistent/ Feel free to try out
15:59:37 <chandankumar> #info openstack-sahara-tests is available in RDO now https://review.rdoproject.org/r/#/q/topic:add-sahara-tests
15:59:51 <chandankumar> #topic chair for next meeting
16:00:01 <imcsk8> i can do it
16:00:01 <chandankumar> Anyone up for that?
16:00:13 <chandankumar> #action imcsk8 will chair for next meeting
16:00:19 <chandankumar> #topic open floor
16:00:41 <amoralej> question about promotion job in CI
16:00:44 <chandankumar> apevec: please have a look on python-heat-tests-tempest package https://review.rdoproject.org/r/#/c/1980/
16:01:07 <amoralej> scheduling is now every 12 hours instead of 4 hours
16:01:18 <chandankumar> i have tested it, works fine.
16:01:20 <trown> amoralej: only for stable pipelines
16:01:33 <amoralej> it's 12 hours for master now
16:01:33 * chandankumar reminds we are in last minute.
16:01:39 <amoralej> if i checked it right
16:02:01 <flepied> could we put it back to 4 hours?
16:02:14 <amoralej> yeah, please
16:02:20 <trown> amoralej: hmm you are right... not sure when that changed... it should definitely be 4
16:02:34 <chandankumar> amoralej: will i end the meeting, we are running out of time?
16:02:45 <amoralej> ok, thanks trown
16:02:46 <trown> chandankumar: ya go for it, we can discuss after if needed
16:02:50 <chandankumar> #endmeeting