16:00:52 <gagehugo> #startmeeting openstack-helm
16:00:52 <openstack> Meeting started Tue Mar  3 16:00:52 2020 UTC and is due to finish in 60 minutes.  The chair is gagehugo. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:53 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:56 <openstack> The meeting name has been set to 'openstack_helm'
16:01:11 <gagehugo> #link https://etherpad.openstack.org/p/openstack-helm-meeting-2020-03-03 agenda
16:01:28 <lamt> o/
16:01:59 <mattmceuen> o/ !
16:02:36 <megheisler> o/
16:02:44 <portdirect> o/
16:02:52 <stevthedev> o/
16:03:33 <portdirect> hey everyone - gagehugo has stepped up to host these for the next couple of weeks
16:03:37 <dwalt> o/
16:04:27 <gagehugo> portdirect: I will do my best :)
16:04:51 <portdirect> I really owe him a huge debt of gratitude for doing so, ive been spread rally thin recently
16:04:53 <srwilkers> o/
16:05:21 <portdirect> gagehugo: you'll rock at it, and i think some of your exp from keystone et al will really help round here, thanks man
16:05:24 <srwilkers> portdirect: you're a master thin-man
16:05:44 <gagehugo> np!
16:05:53 <gagehugo> #topic Midcycle Virtual PTG - March
16:06:21 <gagehugo> So the plan was to have a virtual midcycle meeting (via webex) some time this month iirc
16:06:36 <portdirect> yup
16:06:48 <gagehugo> I will put something together today to gauge a timeframe that works for most people
16:07:28 <gagehugo> either an etherpad or a survey thing that we've used in the past for keystone
16:08:02 <gagehugo> Does anyone have any thoughts/opinions on this?
16:08:15 <stevthedev> Sounds good to me. What's on the agenda for this meeting?
16:08:52 <gagehugo> good question, I'll also figure that out
16:09:00 <gagehugo> and let people post things they want to discuss
16:10:26 <gagehugo> #topic open discussion
16:10:48 <portdirect> RE the virtual midcycle - as always docs are probably our biggest and most pressing pain point
16:10:59 <gagehugo> that's a good point
16:11:15 <songgongjun> hi everyone, I want to discuss the issue about ovrrides
16:11:43 <gagehugo> songgongjun o/ sure, what's up
16:12:02 <songgongjun> Regarding the patch of the overrides feature upgrade, patch1 (https://review.opendev.org/#/c/707775/) and patch2 (https://review.opendev.org/#/c/707788/) two patches depend on each other. Since only patch1 has been merged, openstack-helm may not deploy openstack normally. So, to reduce the impact of this issue, please merge patch2 as soon as
16:12:02 <songgongjun> possible.
16:12:26 <gagehugo> yeah we noticed that this morning
16:13:06 <gagehugo> thanks for bringing it up though!
16:13:20 <songgongjun> Can you merge as soon as possible
16:14:04 <songgongjun> In addition, in order to achieve overrides more parameters than conf section, I submitted a patch(https://review.opendev.org/#/c/701676/). Please help review, thank you!
16:14:23 <gagehugo> I will take a look
16:14:53 <portdirect> songgongjun: can you provide some examples of the over=rides you intend to use with the above
16:15:27 <portdirect> to merge things liek this, it would be good to be able to validate them - and so i'd really like to see that excersized in at gate if possible
16:15:30 <portdirect> *like
16:15:32 <srwilkers> i think examples are key here - the values overrides foo has always seemed like black magic in the past, and every time i've looked at the gotpl for it, i want to cry in a corner
16:17:30 <lamt> esp the daemonset gotpl
16:17:39 <srwilkers> yeah, that's the bit i meant - my bad
16:18:14 <lamt> I also think we should always test OSH against HTK changes in OSH-infra
16:18:15 <songgongjun> The work of enable dpdk in starlingx needs to achieve the overrides of parameters like images, tags, labels, and pods. This feature is implemented through the overrides of openstack-helm.
16:18:46 <portdirect> songgongjun: example of this would be great
16:19:09 <songgongjun> Thanks.
16:19:11 <portdirect> it would also let us anser the q of why we would want to do this as a single helm release, rather than multiples
16:19:36 <portdirect> over-riding pods is pretty dramatic :-O
16:19:46 <srwilkers> songgongjun: i didn't realize the goal was to override literally everything in values per daemonset
16:22:25 <songgongjun> Hi， srwilkers，I didn't understand what you meant, can you explain it
16:23:37 <portdirect> songgongjun: with the changes you have made, what is it possible to over-ride?
16:23:51 <portdirect> by the sounds of it the entire values tree?
16:24:09 <srwilkers> to portdirect's point, i don't understand why this couldn't be handled via multiple releases instead of adding the ability to override the entirety of the values trees via the daemonset overrides
16:24:16 <portdirect> but also - injecting deep-merging arbitary yaml over the rendered output?
16:25:43 <srwilkers> without some examples as requested that actually show this vetted out, this seems heavy handed
16:28:08 <songgongjun> OK, I will think about it, it will be updated in the commit MSG later
16:28:42 <portdirect> songgongjun: that would be great
16:29:03 <portdirect> its also worth noting that other users of osh are using dpdk in production deployments
16:29:30 <portdirect> so would be good to get the requirements that starlingx has that goes beyond what is supported currently
16:30:47 <songgongjun> OK, thank you for your suggestions, that's all my questions
16:31:12 <gagehugo> anyone else have anything they want to discuss this week?
16:35:34 <gagehugo> thanks everyone!  I'll get something put together for the midcycle today and send out a email
16:35:40 <gagehugo> #endmeeting