16:00:51 <gtema> #startmeeting sdk_osc
16:00:51 <openstack> Meeting started Thu Mar 18 16:00:51 2021 UTC and is due to finish in 60 minutes.  The chair is gtema. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:52 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:54 <openstack> The meeting name has been set to 'sdk_osc'
16:01:18 <gtema> diablo_rojo, diablo_rojo_phon, gouthamr, gtema, stephenfin, amotoki is anybody here for the meeting?
16:01:26 <diablo_rojo> o/
16:01:43 <diablo_rojo> I am here but I don't think I have anything I need to discuss?
16:02:02 <gtema> not really, perhaps vPTG planning
16:02:11 <diablo_rojo> Its mostly if any of the students need to discuss anything
16:02:17 <diablo_rojo> Oh thats a good idea
16:02:38 <diablo_rojo> Deadline will be before our next meeting
16:03:06 <diablo_rojo> I definitely think we should meet for a couple hours one or two days like last time.
16:03:10 <gtema> I opened ethercalc and my system froze ;-)
16:03:26 <gtema> I booked Wed 21 15-17
16:03:36 <gtema> but haven't really planned anything yet.
16:03:43 <diablo_rojo> gtema, yeah I botched making that link- I somehow managed to use, not our instance of ethercalc and so it likes to freeze.
16:03:46 <gtema> Is there a demand for PTG?
16:03:57 <diablo_rojo> gtema, maybe make a planning etherpad and circulate it?
16:04:23 <gtema> something I wanted to do over a week already, but haven't really managed
16:04:47 <diablo_rojo> Honestly, I think it would be cool to get all the students together, but also for you to do a braindump of like.. what all needs to be done so we can get that into storyboard so its easier for people to pick things up?
16:04:50 <gtema> what was the link I need to follow also for getting booked?
16:05:20 <gtema> I have one thing for the PTG - it's a R1.0 planning with all the challenges
16:05:27 <gtema> so there is something to be discussed
16:05:37 <diablo_rojo> https://openinfrafoundation.formstack.com/forms/april2021_vptg_survey
16:05:41 <gtema> and getting all the students is surely cool
16:05:42 <diablo_rojo> Apparently so
16:05:45 <diablo_rojo> lol
16:05:56 <gtema> thks
16:06:03 <diablo_rojo> No problem!
16:06:04 <stephenfin> o/
16:06:10 <diablo_rojo> Also, don't forget to register: https://april2021-ptg.eventbrite.com/
16:06:17 <diablo_rojo> Hello stephenfin :) Just chatting about the PTG
16:07:05 <gtema> diablo_rojo - filled the survey. Thanks, lost the link
16:07:27 <diablo_rojo> gtema, no worries :)
16:08:09 <gtema> I will try not to forget to send email with planning
16:09:40 <gtema> are there other interesting topics to discuss?
16:09:59 <diablo_rojo> gtema, if you don't by the middleish of next week I can try to remember.
16:10:03 <diablo_rojo> Nothing else from me?
16:10:14 <gtema> thks
16:10:31 <gtema> was it a question? "Nothing else from me?"
16:10:41 <stephenfin> I've one thing
16:10:45 <stephenfin> Releases
16:11:19 <stephenfin> I _think_ we're currently tied to trailing release or some such thing for OSC
16:11:28 <stephenfin> i.e. we have 'stable/{release}' branches
16:11:46 <gtema> ah yeah, this one is also completely lost in overload this week
16:11:56 <diablo_rojo> gtema, no question :) Sorry to be confusing.
16:12:25 <stephenfin> Do we want to look at changing that to independent?
16:12:37 <gtema> diablo_rojo - was more a joke ;-)
16:12:46 <diablo_rojo> I feel like that's reasonable stephenfin ?
16:12:51 <stephenfin> Rationale being that OSC isn't really tied to any particular OpenStack release. In theory it should support every release
16:13:05 <diablo_rojo> gtema, lol I am too tired to catch that sorry. Need more caffeine.
16:13:29 <gtema> I have no problems with that. But maybe TC have something in mind?
16:13:52 <stephenfin> Yeah, I'm not sure how we do it, but I can investigate
16:14:13 <amotoki> I think we are still in sync with our release to some extent I think.
16:14:16 <diablo_rojo> From a TC standpoint, I don't think we care? It's more the release team?
16:14:24 <stephenfin> I figure we can and should release any time we have enough "stuff" to justify it
16:14:28 * diablo_rojo digs through pile of hats for release hat
16:14:36 <stephenfin> amotoki: You think? How so?
16:15:20 <amotoki> stephenfin: I thoght we are discussing about _indepedent release, but I might misundertand the context.
16:15:52 <stephenfin> No, you understood correctly :)
16:15:55 <amotoki> we sometimes have things related to releases
16:16:12 <stephenfin> Most of our work now is less about keeping up with new changes, since APIs change less frequently than before. It's more about filling in gaps in old releases
16:16:13 <gtema> we can actually release client things for not yet released stuff
16:16:14 <amotoki> so stable branches can help us regarding backports.
16:16:36 <stephenfin> and you can use new OSC versions with old service versions
16:16:49 <stephenfin> Hmm, can we get stable branches that are based on releases instead?
16:16:58 <stephenfin> i.e. stable/5.0 ?
16:17:14 <stephenfin> Maybe that's too much work unless we're very careful about how often we release
16:17:14 <gtema> ugh, that's interesting
16:17:37 <amotoki> it is debatable. we cover most releases, so it is not a problem for most cases I think,
16:17:59 <gtema> I actually dislike heavily stable branches for SDK and OSC. They are made to be always backward compatible
16:18:04 <amotoki> so I don't have strong preference on this
16:18:24 <stephenfin> gtema: True, though we do occasionally do things like drop Python versions
16:18:32 <gtema> yes
16:18:53 <gtema> and we are also potentially caught in the service using dedicated version of SDK
16:18:54 <stephenfin> that's less of an issue for OSC though, since that's only used in the client environment
16:19:16 <stephenfin> yeah, more of an issues for sdk since that can be embedded in the service
16:19:19 <gtema> I would be careful here
16:19:23 <amotoki> yes for most cases. no stable branches sometimes disallow us to fix bugs in released versions
16:19:37 <amotoki> that's all I see as a minus point.
16:19:46 <amotoki> but that would be minor.
16:20:00 <stephenfin> let me think about it more and maybe discuss with release management folks
16:20:03 <gtema> recently I landed image command redesign (to rely on SDK) and some time later figured out that some other dependent client got broken by that
16:20:03 <amotoki> I mean big fixes
16:20:18 <gtema> I guess it was masakari
16:20:30 <stephenfin> It's not high priority. It just occurred to me that we might want to change things when gtema was working through patches ahead of the release
16:20:57 <stephenfin> amotoki: Yeah, that's a good point. I'll pay particular attention to this
16:21:19 <gtema> it clear stephenfin and I am also currently "bombed" by release team wishing a release of OSC
16:21:33 <stephenfin> do you need help?
16:21:47 <stephenfin> I've got some spare time at the moment since nova is in feature freeze
16:21:57 <stephenfin> not much but some :)
16:21:58 <gtema> well, it's generally just cur a release
16:22:05 <gtema> cut
16:22:39 <gtema> and I landed today few things, one self-approved with project cleanup, since around a year it lacks reviews, but people permanently ask for it
16:22:51 <gtema> so I guess we can cut release now
16:23:10 <stephenfin> Oh, yes, I saw that. I thought we were still waiting for the openstacksdk side of that to merge so I hadn't reviewed it. Sorry /o\
16:23:21 <amotoki> antoher way is to overcome it by cutting more releases from master branch independently from OpenStack release cycle
16:23:24 <gtema> it landed actually 1,5 year ago
16:23:46 <stephenfin> gtema: Ouch. Then I'm doubly sorry. I missed that
16:23:47 <amotoki> but we still need to take into account our OpenStack distributors like RHOSP
16:23:49 <diablo_rojo> I could probably help with that if you need too. I vaguely remember how to do it.
16:23:53 <stephenfin> fwiw, I reviewed this morning and it looks good to me
16:23:56 <gtema> there is continuous place for improvement and adding new services inside, but OSC side is now at least capable calling it
16:25:11 <stephenfin> amotoki: Also a good idea. fwiw, we're having discussions about how we can improve our OSC situation internally at the moment. We kind of ignore it now (as in there is no one person in charge of it) since most people are using OSC from their clients which aren't necessarily running the same stack as their servers
16:25:40 <stephenfin> so RHOSP shouldn't be a major concern in any changes to lifecycle, IMO
16:25:49 <stephenfin> I can't speak for any other distro, however
16:26:00 <gtema> dear RedHatters - clarify it ;-)
16:26:44 <gtema> I see lot of "distos" and cloud providers try to tie to the release
16:26:47 * mordred could never get anyone to care about OSC or SDK while I was there ...
16:26:56 <openstackgerrit> James Palmer proposed openstack/openstacksdk master: Add support for Resource Filters  https://review.opendev.org/c/openstack/openstacksdk/+/776271
16:26:57 <gtema> and always they consider SDK/OSC part of that
16:27:14 <amotoki> thanks, so perhaps it is better to keep our release on OSC/SDK from the mater continuously
16:27:30 <stephenfin> mordred: Horizon either. It's not classed as core, I guess
16:27:39 <stephenfin> even though it's what 90% of users see, I know
16:28:01 <amotoki> hehe, as my hat of horizon, horizon faield to support system-scope in Wallaby :-(
16:28:03 <stephenfin> gtema: If we're cutting a release, do we want to go through the final few "not in merge conflict" patches for OSC and see if any are good to go?
16:28:16 <stephenfin> Or should we just cut and punt those to the next release?
16:28:30 <gtema> stephenfin - yupp, let's do that
16:28:38 <stephenfin> Okay, I'll spend tomorrow at that so
16:28:45 <gtema> I mean the first one (time lag is killing the sense)
16:28:49 <stephenfin> oh
16:28:55 <stephenfin> okay yeah
16:29:27 <stephenfin> I see about 30 or so, and many more in merge conflict. Will see how many are ready to roll and defer the rest
16:29:34 <stephenfin> This will be a big release
16:29:50 <gtema> yupp, but moslty only covering gaps
16:30:08 <stephenfin> Yeah, true
16:30:17 <diablo_rojo> Even still, exciting to have progress on things
16:30:23 <stephenfin> indeed
16:30:51 <gtema> yeah, I sadly admit I failed to complete switch of compute things to use SDK
16:31:14 <gtema> started good but then lost the track
16:31:53 <gtema> and then figured out that for SDK R1 we also need lot of love in parallel
16:32:40 <gtema> anything else we want to discuss?
16:32:56 <stephenfin> not from me
16:33:26 <amotoki> nothing from me too
16:33:47 <amotoki> this timeslot per month works for me :)
16:33:58 <gtema> great
16:34:18 <gtema> okay, then I wish everybody a nice day/evening/sleep
16:34:30 <gtema> #endmeeting