15:01:14 <ykarel> #startmeeting RDO meeting - 2019-04-03
15:01:14 <openstack> Meeting started Wed Apr  3 15:01:14 2019 UTC and is due to finish in 60 minutes.  The chair is ykarel. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:15 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:01:18 <openstack> The meeting name has been set to 'rdo_meeting___2019_04_03'
15:01:31 <jpena> o/
15:01:38 <ykarel> @all add topic to agenda:  https://etherpad.openstack.org/p/RDO-Meeting
15:01:43 <ykarel> #topic roll call
15:01:51 <ykarel> #chair jpena mjturek fmount
15:01:52 <openstack> Current chairs: fmount jpena mjturek ykarel
15:02:15 <baha> o/
15:02:29 <ykarel> #chair baha
15:02:30 <openstack> Current chairs: baha fmount jpena mjturek ykarel
15:02:53 <amoralej> o/
15:03:12 <ykarel> #chair amoralej
15:03:13 <openstack> Current chairs: amoralej baha fmount jpena mjturek ykarel
15:04:18 <ykarel> ok let's start
15:04:27 <ykarel> #topic [fmount] Ceph Nautilus update
15:04:50 <fmount> so for ceph nautilus we made some progresses
15:05:15 <fmount> now we have scenario001-standalone green applying few hacks from ceph side
15:06:07 <fmount> as ykarel knows the first one (disabling v1 from ceph.conf) don't give us a clear view
15:06:50 <amoralej> fmount, but iiuc, that fixed the ceph part
15:07:02 <amoralej> then we also need to hack tripleo to wait for quorum, right?
15:07:04 <fmount> the ideal scenario is to have both v1+v2 enabled in ceph.conf for backward compatibility, but seeing cinder able to create a volume is good enough
15:07:26 <fmount> amoralej: yes, for the quorum we have a PR ready for ceph-ansible
15:07:29 <ykarel> fmount, v1+v2 both are enabled, right, it's just removed from conf
15:07:38 <ykarel> 6789+3300 both listening
15:07:43 <amoralej> mmm
15:07:51 <fmount> ykarel: I was checking it on our first env and I don't see 6789
15:08:09 <fmount> ykarel: I'm talking about the first env in which we created a volume using cinder
15:08:16 <amoralej> so, maybe he does not like syntax of host mon?
15:08:40 <fmount> amoralej: not sure
15:08:43 <amoralej> fmount, ok
15:08:51 <ykarel> fmount, amoralej https://logs.rdoproject.org/21/18721/28/check/rdoinfo-tripleo-stein-centos-7-scenario001-standalone/bb9468a/logs/undercloud/var/log/extra/netstat.txt.gz
15:08:55 <amoralej> did you report it to ceph cores?
15:08:58 <ykarel> both 6789 and 3300 open
15:09:00 <fultonj> o/
15:09:07 <amoralej> what's the next step?
15:09:09 <fmount> ykarel: I'll do this
15:09:13 <ykarel> #chair fultonj
15:09:14 <openstack> Current chairs: amoralej baha fmount fultonj jpena mjturek ykarel
15:09:31 <fmount> from my point of view next step is to PR ceph-ansible for the quorum hack
15:09:38 <fultonj> (fmount has been leading this from the ceph integration side, i'm just listening in)
15:09:56 <fmount> and investigate the mon host = [] config on ceph.conf with ceph-core people
15:10:01 <amoralej> ok
15:10:05 <fmount> and ceph-ansible people
15:10:06 <amoralej> sounds good
15:10:16 <fmount> but now things are more clear
15:10:34 <amoralej> i'd like to spend some time checking to add support for nautilus in puppet-ceph
15:10:36 <fmount> so despite we have some hacks the fix are not blockers
15:10:41 <tosky> amoralej: re sahara-tests; we may have a new release, but 0.8.0 is good enough for stein (not 0.7 for sure); if you want to use the latest snapshot, be my  guest
15:10:43 <amoralej> so that we can keep ceph in p-o-i
15:11:37 <fmount> for me that's all, fultonj  would you like to add any consideration?
15:11:46 <amoralej> i think getting it working with those hacks is a good progress
15:11:55 <fmount> ykarel: I'll report that log you've pasted ^^
15:11:57 <fultonj> +1, job did turn green
15:12:19 <ykarel> fmount, ack
15:12:38 <Duck> qauck o/
15:12:47 <ykarel> #chair Duck
15:12:48 <openstack> Current chairs: Duck amoralej baha fmount fultonj jpena mjturek ykarel
15:13:36 <ykarel> anything else in this topic? or we move
15:13:45 <fmount> for me we can move
15:13:49 <fultonj> move
15:13:52 <ykarel> ok let's have some action
15:13:56 <ykarel> #action fmount to investigate mon host config part involving ceph people and push a PR to ceph-ansible handle ceph quorum part
15:14:11 <ykarel> #action amoralej to look for changes in puppet-ceph for including nautilus
15:14:32 <ykarel> #topic Status of stein GA preparation
15:14:40 <ykarel> amoralej would you like to update ^^
15:14:58 <amoralej> yes
15:15:03 <ykarel> ack
15:15:18 <amoralej> so, we have builds for RC releases for services
15:15:44 <amoralej> main missing builds in CBS for stein are tripleo and kolla which are trailing release
15:15:51 <amoralej> and still some tempest plugins
15:16:01 <amoralej> which are still pending to release new tags
15:16:13 <amoralej> we are validating it with weirdo jobs
15:16:20 <amoralej> currently packstack scenarios are all passing
15:16:40 <amoralej> but we need some more tempest plugins for p-o-i
15:17:04 <amoralej> once we have all weirdo passing, we'll tag them in -testing
15:17:20 <amoralej> and keep building new updates until GA releases are published
15:17:53 <amoralej> for Stein GA, i think we should release with nautilus
15:18:31 <amoralej> so, having nautilus working in tripleo and p-o-i and get it released in StorageSIG i 'd say is a blocker for RDO Stein
15:19:07 <amoralej> i don't think we should update ceph release post-GA
15:19:24 <amoralej> so we need to get it working and released
15:19:42 <amoralej> i think that's it
15:19:44 <rdogerrit> Merged openstack/sahara-tests-distgit stein-rdo: openstack-sahara-tests-0.8.0-1  https://review.rdoproject.org/r/19922
15:19:47 <amoralej> well, i sent https://review.rdoproject.org/r/#/c/19909/
15:19:52 <amoralej> ykarel, ^
15:20:04 * ykarel looks
15:20:07 <amoralej> to add Stein to dashboards
15:20:21 <ykarel> okk we will be having a stein promotion soon btw
15:20:34 <ykarel> amoralej, so stein GA on 12th
15:20:46 <ykarel> you mean we need to get ceph nautilus ready before ^^
15:21:07 <amoralej> yep
15:21:25 <ykarel> fultonj, fmount what's your take on this deadline?
15:21:41 <amoralej> i think it's better to hold the release from RDO some days that release with luminous and switch later
15:23:04 <amoralej> in fact, i'd say upstream release will probably be on 10th-11th, they always release in wed-thu
15:23:05 <amoralej> iirc
15:23:08 * fultonj looks
15:23:22 <ykarel> hmm possibly they will do on Wednesday iirc
15:23:37 <ykarel> last time doing on Thursday they faced some issue
15:23:39 <amoralej> yeah, i think so too
15:23:40 <fultonj> ykarel: stein/nautilus must work togehter by the 12th right?
15:23:42 * fmount reading
15:24:06 <ykarel> fultonj, yes that would be ideal case
15:24:08 <amoralej> fultonj, fmount yes
15:24:12 <fultonj> ykarel: i'm expecting we get ceph-ansible patch so we can meet it
15:24:21 <fultonj> ykarel: i agree
15:24:25 <fultonj> dsavineau: ^ fyi
15:24:35 <amoralej> anyway, tripleo will probably be released later
15:24:41 <fultonj> fmount: is it ok with you?
15:24:46 <fmount> fultonj: ykarel amoralej after the patch if the issues we had are no blockers we can do this
15:24:49 <amoralej> so, from tripleo/ceph-ansible we have some more time
15:24:55 <fultonj> gfidente: ^ fyi
15:24:57 <fmount> fultonj: yes
15:25:13 <amoralej> but we'd need to get it officially released in mirror.c.o
15:25:32 <amoralej> so that we can make centos-openstack-release stein depend on ceph-nautilus
15:25:49 <fultonj> amoralej: so it's a little earlier than the 12th then?
15:26:05 <amoralej> fultonj, probably upstream will be on 10th
15:27:01 <amoralej> we like to release RDO asap after upstream, but we can wait a bit if needed
15:28:04 <fultonj> amoralej: thank you, let's make it happen and include nautilus
15:28:44 <fmount> fultonj: agree   ^
15:29:05 <amoralej> fultonj, as said before, i prefer to wait and include nautilus that release with luminous and switch to nautilus later
15:29:14 <amoralej> so yeah, i thing we are all in the same page
15:29:50 <fultonj> ++
15:30:17 <ykarel> ok so this sounds a plan: get RDO-stein released with nautilus, let's move on to next topic
15:30:26 <ykarel> #topic [mjturek | baha] ppc64le  container builds update
15:30:27 <fultonj> ceph-ansible patch --> ceph-ansible tag --> ceph-ansible rpm in centos storage sig --> this rdo job with the rpm
15:30:34 <mjturek> hey!
15:30:46 <amoralej> fultonj, yep, sounds good
15:30:57 <mjturek> so update is that we can build the containers using the upstream playbooks with buildah enabled
15:31:19 <ykarel> nice
15:31:27 <mjturek> next step is for me and baha to get it working reproducibly for cico
15:31:33 <mjturek> which seems doable
15:31:40 <mjturek> need to mock some zuul values
15:31:45 <mjturek> and make an iptables rules
15:31:49 <mjturek> rule chain*
15:31:57 <baha> We haven't run into any dealbreakers yet, just stuff that will take time to try out
15:32:11 <mjturek> yep, any questions? if not we're good to move on
15:32:24 <mjturek> oh!
15:32:33 <mjturek> except, we're not sure what to  compare against
15:32:40 <mjturek> I have a list of the containers that built here http://paste.openstack.org/show/748813/
15:33:37 <ykarel> i think https://github.com/openstack/tripleo-common/blob/master/container-images/overcloud_containers.yaml is the list
15:33:40 <ykarel> mjturek, ^^
15:33:54 <ykarel> and you can also check the jobs we run in tripleo promotion
15:34:08 <mjturek> cool! will compare that
15:34:08 <mjturek> thanks
15:34:32 <baha> Thanks ykarel!
15:34:59 <ykarel> mjturek, baha this is the job we run: https://review.rdoproject.org/zuul/builds?pipeline=openstack-periodic&job_name=periodic-tripleo-centos-7-master-containers-build-push
15:35:09 <ykarel> so you can check that also to compare
15:35:48 <mjturek> will do!
15:35:50 <ykarel> mjturek, baha so are you guys also testing the ppc build containers somewhere?
15:36:04 <ykarel> like some tripleo job
15:36:14 <mjturek> no we aren't
15:36:44 <ykarel> okk so there is a plan to do in future?
15:37:05 <mjturek> yes  for sure, but first goal is getting them published
15:37:21 <ykarel> okk good
15:37:39 <ykarel> so let's move to next topic
15:37:44 <ykarel> #topic Chair for next meeting
15:37:53 <ykarel> any volunteer?
15:37:54 <amoralej> running some tripleo-standalone on ppc with those would be nice
15:38:30 <amoralej> i will chair
15:38:41 <ykarel> #action amoralej to chair next meeting
15:39:10 <ykarel> #topic Open Discussion
15:39:17 <rdogerrit> Merged openstack/cloudkitty-distgit stein-rdo: Add flask and flask-restful dependencies  https://review.rdoproject.org/r/19926
15:39:19 <ykarel> so bring any topic now
15:40:06 <Vorrtex> I have one question... Why does this meeting "chair" anyone who joins, when in typical IRC/Openstack meetings, the chair (as denoted above by "who will chair the next meeting") is generally left to one individual who runs the meeting?
15:41:27 <amoralej> Vorrtex, by chair as in "Who will chair next meeting?" means who will run it
15:41:45 <amoralej> chairing everyone allows anyone to add actions to the meeting notes iirc
15:42:25 <amoralej> i remember this was discussed time ago, but tbh, i'm not sure if it's needed
15:42:45 <Vorrtex> I'm 80% sure anyone can add an action, but the people you add to the "#chair" list can do other meeting-specific actions, like changing topic and other things.
15:43:59 <amoralej> we probably do it just by habit then
15:44:19 <Vorrtex> ha ha, that's fine I suppose, just thought it was a strange practice and was curious why it was done ha ha.
15:44:34 <amoralej> we want anyone to be able to add actions, info and links
15:44:47 <amoralej> but only the one that runs open, close, change topics
15:47:26 <ykarel> +1 Okk so let's continue the habit :) if there is no harm else found the reason for not doing it
15:48:15 <amoralej> yep wfm
15:48:32 <ykarel> so any other thing to discuss, or let's close
15:49:52 <ykarel> okk so let's close, Thanks all
15:49:57 <ykarel> #endmeeting