15:00:14 #startmeeting openstack-helm 15:00:15 Meeting started Tue Apr 2 15:00:14 2019 UTC and is due to finish in 60 minutes. The chair is portdirect. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:16 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:19 The meeting name has been set to 'openstack_helm' 15:00:22 Agenda is here: https://etherpad.openstack.org/p/openstack-helm-meeting-2019-04-02 15:00:26 o/ 15:00:27 o/ 15:00:28 #topic rollcall 15:00:36 o/ 15:00:37 lets give it 5 mins for people to arrive 15:00:37 o/ 15:01:03 o/ 15:02:08 o/ 15:02:27 o/ 15:02:28 o/ 15:03:37 o/ 15:04:34 full house today :) 15:04:48 ok - lets go :) 15:05:02 #topic multi-os spec 15:05:14 so - this got merged recently 15:05:32 evrardjp: thanks for all your work on driving this, it looks solid 15:05:49 yeah evrardjp, great work there 15:06:12 whoops should have included the link: https://review.openstack.org/#/c/645142/ 15:06:24 thanks for all the comments 15:06:32 and thanks for merging this 15:06:35 I have a question 15:06:41 about the implementation details 15:07:53 some charts already have this kind of features, but not under a software key, do you prefer having a full swoop of all charts in one go, or should we continue the way we have now, and slowly migrate the charts to the new key? 15:08:03 I think we all agreed what the future state looks like 15:08:31 (it's to know how to create the backlog in storyboard) 15:08:44 (if such backlog is necessary) 15:09:11 i have found in the past that hitting one chart 1st as a spearhead makes sense 15:09:12 I would prefer just doing the right thing for the new changes, and adapt the existing when we find issues 15:09:23 often using either keystone, or heat for that 15:09:28 yeah 15:09:34 and then expanding out to nova/neutron 15:09:43 and then finally hitting all the rest in a single ps 15:09:45 I think we agree on that 15:09:55 oh ok 15:09:59 I see 15:10:04 sounds about right 15:10:11 thanks for the clarification 15:10:14 that's all I had 15:10:51 ok - also we can use the rocky images for suse/ubuntu for gating: https://cloud.docker.com/u/openstackhelm/repository/docker/openstackhelm/nova/tags 15:11:59 let see if we can get this closed out before the summit - it would be a great thing to get a pin in 15:12:39 anything else here? 15:12:55 it's not 100% fully working AFAIK -- but we'll include example values when they are working and the helm charts are changed 15:13:08 nothing else 15:14:36 #topic Office hours 15:14:52 so in the past mattmceuen had setup office hours for OSH 15:15:22 which was great, though attendance was low and as a result they faded out 15:15:36 a few people have asked for us to re-instate these 15:16:10 and I think we have grown to the point where they could add real value for people looking to get involved in OpenStack-Helm 15:16:18 and also seek support for issues 15:16:42 srwilkers / mattmceuen / lamt / alanmeadows / anticw: thoughts? 15:17:04 sounds good to me 15:17:06 The success will depend on attendence of all of us -- the presence of cores is paramount though 15:17:13 ++ ^ 15:18:17 it's kind of a chicken and egg problem -- if folks would take advantage of it I'm happy to be here :) 15:18:38 mattmceuen: i think we are the eggs here :D 15:19:32 ok - so are we up for trying this again, until the end of this cycle? 15:20:05 if so i think we need to operate with the expectation that at least one core would attend each office hour 15:20:46 im fine with seeing how it works until the end of the cycle 15:20:52 if so - I will send out a scheduling voter today to the ml, for a single hour per week slot 15:21:03 I am totally looking for more eggs 15:21:31 pretty sure others of my team would love too 15:21:55 * mattmceuen can only think of breakfast now 15:22:14 ok - I'll send out the mail 15:22:18 o/ dropping by and that sounds good. Also breakfast. 15:22:38 would be great to have a more informal meeting time as well 15:22:42 lamt: ! hey bud 15:22:43 o/ lamt 15:22:58 portdirect: I think that's kinda it 15:23:10 the informal meeting time is kinda the purpose of office hours 15:23:18 we just talk all together about things 15:23:20 yup 15:23:53 ok - so that kinda wraps up our agenda for today, other than reviews 15:23:58 I think it also helps on building a "community" as we'll be talking with each other more 15:23:59 #topic reviews 15:24:14 https://review.openstack.org/#/c/647493/ Add internal tenant id in conf (irc: LiangFang) 15:24:14 https://review.openstack.org/#/c/647747/ Libvirt: support per node overrices (irc: swebster) 15:24:36 overrices ... doh! 15:24:40 ^ can we get some eyes on these today 15:24:47 thanks 15:24:51 swebster: I am starving now 15:25:03 better than crunchy under-rices srwilkers :D 15:25:20 hah 15:25:23 itxaka, anything to ping? 15:25:28 o/ sorry for being late! 15:25:48 yes, the overrides patches for keystone, horizon and nova placement have been updated to be in sync with the multi-os spec 15:26:02 oh so keystone could totally deserve a review already 15:26:10 https://review.openstack.org/#/c/647403/ and https://review.openstack.org/#/c/645141/ and https://review.openstack.org/#/c/642067/ 15:26:16 thanks itxaka! 15:26:26 so any reviews are welcome, they should not neede any more changes...I hope! 15:26:26 woot! 15:27:00 itxaka: we probably want to get keystone in 1st, then will move on to the others (see the 1st topic in the meeting) 15:27:10 sounds good to me! 15:27:18 nice 15:27:21 also we should add some gates around these 15:27:58 I also wanted to bring attention to https://review.openstack.org/#/c/644821/ -> change the nova cache to dogpile.cache.memcached 15:27:58 yeah we will :) 15:28:26 though that can probably follow - as these just add the appropriate config for the multi-distro spec supporting what we use today 15:29:08 https://review.openstack.org/#/c/644821/4 lgmt 15:29:20 we found out that the default memcache_pool backend has issues under python3 for >= rocky so changing it to a different backend by default would be good until upstream fixes are available in oslo_cache for it 15:30:02 keystone is also using the dogpile.cache.memcached backend so its already tested in there and found to cause no issues 15:30:41 nice, thanks portdirect ! 15:31:20 i suppose i should chime in: https://review.openstack.org/#/c/648730/ 15:31:44 im hoping this is ready to go now, moves us to mariabackup 15:32:13 we had to move to rsync for a bit, as the migration from perconabackup to mariabackup was in flux 15:32:21 but this should close the loop on that 15:32:51 and also gives us an appropriate user for sst work, rather than overusing admin/root for this 15:33:49 ok to move to the parking lot? 15:34:02 for whatever this means, I suppose yes 15:34:13 #topic parking-lot 15:34:45 anything else we should be thinking about this week? 15:36:34 ok - lets give everyone some time back 15:36:38 thanks everyone 15:36:44 thanks all! o/ 15:36:51 thanks all, enjoy your breakfasts 15:37:17 #endmeeting