15:05:39 <amoralej> #startmeeting RDO meeting - 2017-12-20
15:05:40 <openstack> Meeting started Wed Dec 20 15:05:39 2017 UTC and is due to finish in 60 minutes.  The chair is amoralej. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:05:41 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:05:43 <openstack> The meeting name has been set to 'rdo_meeting___2017_12_20'
15:05:52 <amoralej> #topic rollcall
15:05:57 <jpena> o/
15:06:01 <mary_grace> \o
15:06:02 <ykarel> o/
15:06:10 <amoralej> #chair dmsimard jpena ykarel mary_grace
15:06:11 <openstack> Current chairs: amoralej dmsimard jpena mary_grace ykarel
15:06:49 <rbowen> o/
15:06:59 <amoralej> #chair rbowen
15:07:00 <openstack> Current chairs: amoralej dmsimard jpena mary_grace rbowen ykarel
15:07:14 <amoralej> let's start with the first topic
15:07:30 <amoralej> #topic CBS certificate renewal on Dec, 31/Jan, 1
15:08:04 <amoralej> I'm not sure who wrote the note "I'll be taking care of that"
15:08:09 <amoralej> apevec?
15:08:10 <dmsimard> number80 ^
15:08:23 <amoralej> number80 is on pto i think
15:08:30 <dmsimard> oh
15:08:42 <dmsimard> we had been using number80's cert
15:08:49 <amoralej> yes
15:08:49 <dmsimard> so I'm not sure how someone else can take care of that
15:08:51 <apevec> it needs to be number80
15:08:54 <dmsimard> unless we use someone else's cert
15:09:07 <apevec> or yeah we change cert
15:09:23 <dmsimard> apevec: it's either him or you :P
15:09:30 <dmsimard> no one else has build privileges in cloud sig I think
15:09:31 <amoralej> #chair apevec
15:09:31 <openstack> Current chairs: amoralej apevec dmsimard jpena mary_grace rbowen ykarel
15:09:36 <amoralej> i have cert also
15:09:39 <apevec> so iirc it was Haikel who wrote that
15:09:46 <apevec> iiuc*
15:10:01 <apevec> so I'd assume he is taking care
15:10:05 <amoralej> so the note means he will take care while in pto?
15:10:12 <amoralej> oook
15:10:22 <apevec> yeah, that's not good and we need to fix that :)
15:10:42 <apevec> there is discussion ongoing about creating SIG bot cert
15:10:51 <amoralej> apevec, any progress on the discussion?
15:10:57 <amoralej> i saw the mails about it
15:11:07 <amoralej> but no conclusion, right?
15:11:23 <apevec> yeah no progress yet
15:11:25 <dmsimard> apevec: why do we need to wait until the certificate expires ? can't we renew it before expiration and just switch it ?
15:11:49 <apevec> that's also good question
15:11:57 <apevec> cert can be regen at any time
15:12:07 <amoralej> yeah, that's what i was thinking, maybe i could install mine as secret
15:12:29 <amoralej> there is anything special whith the change of year?
15:12:36 <dmsimard> well I just mean that number80 can renew it at any time
15:12:45 <amoralej> number80 is on pto
15:12:46 <amoralej> :)
15:12:53 <dmsimard> ok, let me rephrase that
15:12:55 <amoralej> let's see if we can take care
15:13:00 <dmsimard> number80 should have been able to renew it at any time
15:13:02 <rdogerrit> Merged openstack/networking-l2gw-distgit rpm-master: Workaround due to Tempest plugin split  https://review.rdoproject.org/r/11039
15:13:09 <dmsimard> amoralej: do you have build privileges on cbs ?
15:13:12 <amoralej> dmsimard, i got it now, i think so
15:13:12 <amoralej> yes
15:13:18 <amoralej> for cloudsig
15:13:34 <dmsimard> ok, I think the real solution is to have a "service" account like paas sig has
15:13:40 <dmsimard> but it might be short notice to do that
15:13:46 <dmsimard> we can reach out to #centos-devel
15:13:58 <amoralej> dmsimard, there has been some conversations about it
15:14:19 <amoralej> does paas sig have a service accunt?
15:14:22 <dmsimard> yeah
15:14:22 <amoralej> account?
15:14:47 <dmsimard> amoralej: see for example http://cbs.centos.org/koji/packageinfo?packageID=3175
15:14:49 <dmsimard> "paas"
15:14:57 <amoralej> i thought there is not solution for service account in CBS
15:15:00 <dmsimard> it's used by their CI afaik
15:15:02 <dmsimard> amoralej: there is
15:15:20 <dmsimard> amoralej: they tend to stray away from that in *brew* because of a policy where each package must be reviewed by a human
15:15:25 <dmsimard> or something like that
15:15:30 <amoralej> then, i don't understand well the discussion about the generic account some time ago in centos-devel
15:15:38 <apevec> yeah paas is example I brought up,
15:15:44 <apevec> but unsure how they got it
15:15:45 <apevec> it'
15:15:53 <apevec> s not documented in official SIG guide
15:16:10 <amoralej> maybe it's just a regular account
15:16:13 <amoralej> with regular expiration
15:16:16 <amoralej> but with paas name
15:16:16 <dmsimard> ok i'll take the action to chat with them
15:16:20 <dmsimard> maybe
15:16:28 <dmsimard> #action dmsimard to reach out to #centos-devel about getting a service account like paas
15:16:48 <apevec> it was topic on centos contributor meetup at cern
15:17:04 <apevec> where conclusion was yes and TBD
15:17:25 <amoralej> ok, let's see what centos guys tell us
15:17:36 <amoralej> maybe there is a chance to get it before end of year?
15:17:49 <apevec> amoralej, I'd go with the plan to put yours today
15:17:53 <apevec> and see how that goes
15:17:59 <amoralej> ok
15:18:02 <dmsimard> I'm asking now, let's catch up after the meeting
15:18:08 <apevec> i.e. do we know all places where is it used
15:18:21 <number80> dropping by, reason to do it on Jan, 1 is that it's easy to remember
15:18:23 <amoralej> #action amoralej to install new certificate for CBS task
15:18:46 <amoralej> number80 but it's holiday!, not a good day for anything :)
15:19:03 <number80> random date means that nobody knows but the person who set it up when to renew it
15:19:16 <amoralej> that's also true
15:19:19 <number80> best date to do that, nobody is impacted :)
15:19:33 <amoralej> just the one that has to do it
15:19:37 <amoralej> that's not fair
15:19:47 <amoralej> :)
15:20:11 <amoralej> let's see if we can get the service account
15:20:34 <number80> There's already one created, pending validations from KB, we also need email alias
15:20:51 <number80> apevec: I think it's a manager thing to get a team alias ^
15:21:29 <dmsimard> KB is on PTO already so if we need him that means it'll be after the holidays
15:22:02 <number80> Yeah, he's the one who has to review that
15:22:40 <amoralej> number80, what if we install new cert today?, i guess we'll have service account before expiration
15:22:44 <amoralej> in 6 months
15:23:09 <number80> works for me, just wanted to notify this was under radar, anyway
15:23:18 <amoralej> yeah, thanks for that
15:23:34 <amoralej> ok, i'll follow up after the meeting
15:23:37 <amoralej> next topic
15:24:07 <amoralej> #topic Upgrading review.rdoproject.org to software factory to 2.7
15:24:10 <amoralej> dmsimard,
15:24:14 <amoralej> all yours
15:24:19 <dmsimard> ohai
15:24:28 <dmsimard> we have this thing on review.rdoproject.org, software factory
15:24:59 <dmsimard> we'd like to upgrade from 2.6 to 2.7 which will enable us to do things like zuul v3, nodepool v3 and generally allow us to update things and resolve issues
15:25:14 <dmsimard> the upgrade and the implementation of zuul v3 are two distinct things
15:25:33 <dmsimard> initially, we'll want to upgrade to 2.7 and then later, sometime after the holidays, enable zuul v3
15:25:47 <amoralej> we can have both in parallel?
15:26:08 <dmsimard> the definitive plan for the zuul v3 migration hasn't been discussed yet and it's something we'll want to talk about before we do it
15:26:31 <dmsimard> I've reached out and briefly discussed this with mrunge, for example -- opstools doesn't have a lot of knowledge about zuulv3 so it is likely we will need to help
15:26:58 <amoralej> ok
15:27:30 <dmsimard> and we had a meeting yesterday with the TripleO CI team (weshay, EmilienM, sshnaidm|afk, pabelanger, mwhahaha) to discuss what it means to migrate to zuulv3 in review.rdo (see summary here: http://lists.openstack.org/pipermail/openstack-dev/2017-December/125735.html )
15:27:42 <weshay> 0/
15:28:08 <dmsimard> for example, we agreed that the job configuration, roles and playbooks should live in openstack-infra/tripleo-ci
15:28:15 <amoralej> #chair weshay
15:28:15 <openstack> Current chairs: amoralej apevec dmsimard jpena mary_grace rbowen weshay ykarel
15:28:26 <dmsimard> so we can include things as-is from review.rdo with zuul v3
15:28:54 <dmsimard> the specifics/details have not been ironed out yet but we started talking about it so there's that
15:29:16 <amoralej> so the .zuul in openstack-infra/tripleo-ci will be used by review.o.o and review.r.o ?
15:29:33 <dmsimard> amoralej: yes, zuul v3 allows us to selectively include components from different projects
15:29:51 <dmsimard> we'll be including the configuration from openstack-infra/zuul-jobs, openstack-infra/openstack-zuul-jobs, etc.
15:30:26 <dmsimard> the general idea being to centralize everything in one place (tripleo-ci) so that it can be used across not just review.o.o and review.r.o but also further down the stream eventually
15:30:54 <amoralej> ok, sounds reasonable
15:31:05 <dmsimard> so, circling back to the upgrade of software factory from 2.6 to 2.7
15:31:28 <dmsimard> review.rdoproject.org hasn't been particularly productive recently so it's a shame we hadn't thought of that before, but we'd like to upgrade it before the holidays if we can
15:31:51 <dmsimard> that means either doing the upgrade tonight or tomorrow evening (to have tristanC present)
15:32:03 <dmsimard> otherwise it means we're pushing back the upgrade to january
15:32:19 <amoralej> for upgrade, iiuc, the options are to have parallel zuulv3 and v3 and automatic creation of legacy-* jobs as intermediary until all jobs are ported to native v3?
15:32:31 <amoralej> other options?
15:32:34 <dmsimard> amoralej: again, the upgrade will not enable zuulv3 immediately
15:32:53 <amoralej> yes, but the day it's enabled we'll have the problem
15:33:12 <dmsimard> yes, but this can be another meeting topic on it's own :D
15:33:18 <amoralej> that's the migration that i'm concerned about
15:33:19 <amoralej> ok
15:33:22 <amoralej> good for me
15:34:10 <dmsimard> Formal question: Would we be okay with updating review.rdoproject.org tonight at 23:00 UTC ? We'd plan for one hour window with fairly brief disturbances.
15:34:27 <amoralej> i'm ok
15:34:28 <dmsimard> apevec_, jpena, amoralej, weshay, rlandy|rover ^
15:34:29 <rbowen> +1
15:34:31 <amoralej> +1
15:34:38 <weshay> +1
15:34:45 <rlandy|rover> +1
15:34:46 <weshay> rlandy|rover, ^
15:34:48 <ykarel> +1
15:35:02 <dmsimard> ok, thanks
15:35:05 <jpena> just a second, I'm checking if the latest backup is there
15:35:20 <dmsimard> jpena: we would likely manually run a backup before running the upgrade
15:35:38 <dmsimard> #agreed review.rdoproject.org to be updated tonight december 20th at 23:00UTC
15:35:45 <dmsimard> #action dmsimard to send an email to notify about the upgrade
15:36:03 <jpena> anyway, the backup is there, so +1
15:36:09 <amoralej> ok
15:36:26 <amoralej> let's move to next topic
15:36:28 <chandankumar> \o/
15:36:37 <amoralej> #topic Update about status of infra
15:36:40 <amoralej> #chair chandankumar
15:36:41 <openstack> Current chairs: amoralej apevec chandankumar dmsimard jpena mary_grace rbowen weshay ykarel
15:36:46 <amoralej> i added that topic
15:37:06 <amoralej> dmsimard, jpena so what's the status of infra?, is rdo cloud migration finished?
15:37:22 <dmsimard> s/migration/upgrade/ ?
15:37:26 <amoralej> sorry
15:37:27 <amoralej> yeah
15:37:43 <amoralej> is it fully functional now?
15:37:55 <amoralej> and about nodepool issue?, it's finally fixed?
15:38:01 <jpena> the RDO Cloud controller upgrade was finished yesterday, and compute nodes were updated today. I think the final details are being ironed out (the rdo-nodepool tenant quota is not correct, but it's not hurting us)
15:38:25 <jpena> nodepool is not 100% fixed yet, we had another hang early today and we're investigating it
15:38:33 <amoralej> so we can expect it to work normally?
15:38:59 <amoralej> apart from the nodepool hangs, i mean
15:39:01 <jpena> yes, unless nodepool hangs (or there's a new issue, of course)
15:39:11 <dmsimard> so, something to be wary about
15:39:41 <dmsimard> it took me about 1hr+ last night to properly clean up everything that had leaked throughout the upgrade
15:40:09 <dmsimard> we had a large backlog of jobs in review.rdo and nodepool was started which actually killed RDO cloud
15:40:32 <dmsimard> there was 75+ heat stacks to clean up, 100's of networks/orphaned ports
15:40:37 <dmsimard> VMs in error everywhere
15:40:54 <amoralej> today i still had the impression that jobs were slow to dequeue
15:41:01 <amoralej> but maybe it's just a perception
15:41:07 <dmsimard> nodepool got stuck
15:41:18 <dmsimard> I'm not sure if it's been restarted yet, I haven't been involved in troubleshooting that (yet)
15:41:41 <amoralej> ok
15:41:45 <dmsimard> what I want to say, though, is to be careful if there is a large zuul backlog and we're starting or restarting nodepool
15:41:50 <dmsimard> because nodepool can and will kill RDO cloud
15:41:52 <amoralej> hopefully we can find the root cause soon
15:42:34 <amoralej> ok, i think we are done with this topic
15:42:42 <rlandy|rover> jpena: you have a test instance on the openstack-nodepool tenant  - are you still using that?
15:42:47 <amoralej> #topic Test day
15:42:54 <amoralej> rbowen, ^ is it yours?
15:42:58 <rbowen> I just wanted to thank everyone who worked so hard to put together the test cloud
15:43:03 <rbowen> and the people that showed up to test.
15:43:18 <rbowen> I expect that our turnout for this will improve as word gets out that we're doing it.
15:43:31 <rbowen> We had great help promoting it from all through the community.
15:43:44 <rbowen> For next time, we need to continue to improve our testing instructions/suggestions.
15:44:01 <rbowen> Anyways, thanks so much to everyone that helped with that.
15:44:06 <amoralej> yes, and to provide a better network option
15:44:30 <rbowen> Especially dmsimard jpena apevec_ and amoralej
15:44:45 <dmsimard> rbowen: actually we got some feedback from the openstack technical committee
15:44:47 <amoralej> but i think it proof it can be useful to involve new people in test day
15:44:54 <rbowen> dmsimard: Yeah? Tell us more.
15:44:56 <rdogerrit> Carlos Goncalves proposed openstack/octavia-dashboard-distgit rpm-master: Initial commit  https://review.rdoproject.org/r/11033
15:44:57 <dmsimard> I forgot how or why we ended up discussing it there
15:45:12 <dmsimard> but they thought it was a great idea
15:45:28 <dmsimard> said we should have advertised it more than that
15:45:32 <rbowen> If we could get that rolled into the Trystack project, and get Trystack rescued in the process, that would be cool too.
15:45:34 <dmsimard> not like we didn't try :/1
15:45:52 <rbowen> ok, we'll saturate the various openstack lists next time. :-)
15:46:35 <dmsimard> I'll formally reach out to them about the idea in general, maybe we could get more or less of an endorsement for the initiative
15:47:00 <dmsimard> #action dmsimard to discuss the test day cloud idea with the TC
15:47:37 <amoralej> anything else about this topic?
15:48:40 <amoralej> i guess, no, i'll move on
15:48:41 <amoralej> #topic Cancelling 27th Dec and 03rd Jan meeting due to holidays
15:48:56 <amoralej> last week we talked about cancelling on 27th
15:49:14 <chandankumar> on 3rd Jan i and ykarel is available
15:49:16 <amoralej> in my case i'm on PTO on january, 3rd
15:49:29 <chandankumar> I am not sure how many will be there?
15:50:02 <jpena> I'll still be on PTO on January 3
15:50:07 <amoralej> chandankumar, i think you can run the meeting on 3rd unless there are no topics
15:50:20 <amoralej> to avoid two weeks without meeting
15:50:26 <chandankumar> amoralej: ack sir
15:51:04 <rdobot> [sensu] NEW: master.monitoring.rdoproject.org - check-delorean-newton-current @ http://tinyurl.com/y8g8m46f |#| Last check execution was 10977 seconds ago
15:51:12 <amoralej> unless the rest of the teams prefers to cancel it
15:51:56 <amoralej> #agreed the meeting on december 27th will be cancelled
15:52:01 <chandankumar> #info 27th Dec, 2017 RDO meeting is cancelled, Happy Holidays :-)
15:52:10 <amoralej> #agreed the meeting on january on 3rd is kept
15:52:27 <amoralej> #action chandankumar will chair the RDO meeting on January the 3rd
15:52:47 <amoralej> #topic open floor
15:52:57 <chandankumar> amoralej: one more thing 26th Dec Office hour is also cancelled
15:53:02 <chandankumar> due to same
15:53:09 <amoralej> #undo
15:53:10 <openstack> Removing item from minutes: #topic open floor
15:53:37 <amoralej> #info the office hour on december 26th is cancelled
15:53:50 <chandankumar> amoralej: i will send the cancellation email to the list
15:53:58 <amoralej> thanks chandankumar
15:54:06 <amoralej> #topic open floor
15:54:23 <amoralej> any topic not in the agenda you'd like to bring?
15:54:46 <chandankumar> one question related to repoexplorer
15:55:09 <chandankumar> repoexplorer currently generates stats from git repository
15:55:32 <chandankumar> is there any plan to extend to include open reviews and bug as well as github pr and issues?
15:55:36 <chandankumar> jpena: ^^
15:56:06 <jpena> chandankumar: I'm not aware of plans to do that in the short term. fbo_ ^^ ?
15:56:22 <chandankumar> jpena: not short term but in roadmap
15:57:15 <jpena> I haven't seen anything in the backlog
15:57:29 <chandankumar> i will open an issue on that
15:57:55 <chandankumar> that's it
15:58:02 <chandankumar> from myside
15:58:09 <amoralej> ok
15:58:17 <amoralej> i think we can close the meeting
15:58:29 <amoralej> thank you everyone
15:58:48 <amoralej> Happy new year for all RDOers!!
15:59:10 <dmsimard> yay \o/
15:59:12 <chandankumar> Merry Christmas to all RDOers/Stackers and their families :-)
15:59:48 <amoralej> #endmeeting