14:00:12 <abhishekk> #startmeeting glance
14:00:13 <openstack> Meeting started Thu May 20 14:00:12 2021 UTC and is due to finish in 60 minutes.  The chair is abhishekk. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:14 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:16 <openstack> The meeting name has been set to 'glance'
14:00:18 <jokke> o/
14:00:18 <abhishekk> #topic roll call
14:00:22 <dansmith> o/
14:00:23 <abhishekk> #link https://etherpad.openstack.org/p/glance-team-meeting-agenda
14:00:30 <abhishekk> #chair rosmaita
14:00:31 <openstack> Current chairs: abhishekk rosmaita
14:00:35 <rosmaita> o/
14:00:56 <Steap> o/
14:01:01 <abhishekk> Just FYI, no power at my end since morning and laptop battery is only 10% now
14:01:18 <abhishekk> Lets start
14:01:28 <jokke> abhishekk: oof, lets be quick then
14:01:33 <abhishekk> #topic release/periodic jobs update
14:01:35 <abhishekk> yep
14:01:41 <abhishekk> M1 is next week
14:01:59 <abhishekk> we need to be quick on some specs
14:02:06 <abhishekk> 1. unified quota
14:02:11 <abhishekk> 2. Cache API
14:02:28 <abhishekk> Other than that we don't have much in hand for M1
14:02:33 <dansmith> well,
14:02:40 <dansmith> there needs to be a spec for the policy stuff at some point,
14:02:46 <abhishekk> Periodic jobs all green for this week
14:02:53 <dansmith> but there has been basically no feedback on the poc stuff
14:03:17 <abhishekk> Yes, I know, I have added it as a separate point for discussion
14:03:20 <dansmith> okay
14:03:22 <rosmaita> when is the spec freeze?
14:03:27 <abhishekk> M2
14:03:35 <rosmaita> ok
14:03:50 <abhishekk> Moving ahead
14:04:00 <abhishekk> #topic Implement glance-unified-quotas
14:04:10 <abhishekk> #link https://review.opendev.org/c/openstack/glance-specs/+/788037
14:04:29 <rosmaita> i am way behind on reviews, sorry about that
14:04:39 <abhishekk> jokke, raised some concerns, I guess we need to come to an agreement in this week
14:05:14 <rosmaita> i will commit to reviewing and commenting today
14:05:20 <abhishekk> thank you
14:05:51 <rosmaita> well, dont' thank me yet
14:05:53 <abhishekk> I still think, this will be better option to have than our existing quota
14:06:08 <jokke> I raised few specifics inline to the spec and few general concerns on the review level
14:06:10 <rosmaita> abhishekk: +++++++++++
14:06:29 <abhishekk> yep
14:07:03 <jokke> As this will be changing our API behaviour, it will be very difficult to justify refactoring it in the future, so I'd like a bit of brainstorming what we could do better to avoid digging ourselves in the hole
14:07:06 <abhishekk> I saw some inline answers/justifications from dansmith as well, so hopefully those will help to reduce your concerns
14:07:28 <dansmith> this does not change our api behavior, as noted in the spec,
14:07:29 <jokke> if the answer is there is nothing else we can do, then I guess it's better than nothing, but lets remember what ever we do, we're stuck with it
14:07:41 <dansmith> because we can get over limit for all these calls already due to the global limits
14:07:56 <abhishekk> I am also not sure how it will change API behavior?
14:08:01 <dansmith> and also as noted, we can further refine this, we are not "stuck"
14:08:35 <jokke> dansmith: so are you saying there is no new behaviour tested and enforced by tempest due to this work that is assumed to behave as implemented for the foreseeable future?
14:09:27 <dansmith> jokke: I'm saying what the user experiences is the same, since they could be returned an over limit code for any of these APIs already if the operator has set the global limits,
14:09:34 <dansmith> and the way out of those are the same (delete some stuff)
14:09:59 <dansmith> I haven't seen an API concern raised by you on the spec, so if you have one, please do it there so we can continue with abhishekk's battery
14:10:33 * abhishekk yep, only 8% now
14:10:49 <abhishekk> So, continue this discussion on the spec
14:10:56 <abhishekk> moving to next and important topic
14:10:56 <jokke> yeah, move on we can take this out of band
14:11:12 <abhishekk> #topic Policy refactoring (Need reviews ASAP)
14:11:21 <abhishekk> https://review.opendev.org/q/topic:%22policy-poc%22+(status:open%20OR%20status:merged)
14:11:42 <dansmith> there are a few refactors in the front of that set,
14:11:44 <dansmith> which are +2 by abhishekk, so some quick +Ws to get that set down to a smaller size would be great
14:11:58 <abhishekk> As promised dansmith has modified the modify_image policy injection
14:12:06 <abhishekk> those are sitting idle since 2 weeks now
14:12:07 <dansmith> but the last one is the chunky one we need to look at in terms of approach, and there are questions in some of the test changes which I think are things that were wrong before
14:12:27 <abhishekk> I guess we need some eyes there so that we can start with next step which is spec for this stuff
14:13:30 <abhishekk> I am really struggling to find why some property protection tests checking for conflict instead of forbidden
14:14:03 <dansmith> yeah, it's been weeks since I was looking at that (as noted) but I too think it was probably tests confirming actual behavior, not desired
14:14:03 <abhishekk> So either one glance from rosmaita or jokke will be more helpful there
14:14:31 <jokke> ack
14:14:50 <rosmaita> ok
14:14:56 <abhishekk> So, I would suggest we need to spend some time around this as this is most priority item for us in this cycle
14:15:12 <rosmaita> this is https://review.opendev.org/q/topic:%22policy-poc%22+(status:open%20OR%20status:merged)  ?
14:15:16 <abhishekk> as this will circle around secure RBAC stuff
14:15:21 <abhishekk> rosmaita, yes
14:15:30 <rosmaita> ok, sorry to be obtuse
14:15:46 <dansmith> rosmaita: I think you missed those convos, but this is a PoC for refactoring policy enforcement up to the api layer where we can make proper decisions based on the request
14:16:04 <rosmaita> \o/
14:16:09 <dansmith> rosmaita: like what has happened in other projects lately, and is required for us to be able to do things like system scope and actual flexible policies
14:16:20 <rosmaita> you are preaching to the choir
14:16:26 <abhishekk> ++
14:16:27 <dansmith> it will be a giant refactor, so we decided to PoC image update as the gnarly-est thing first
14:16:44 <rosmaita> thanks for the context
14:16:45 <dansmith> rosmaita: glad to hear it.. long overdue, AFAICT.. like quotas :)
14:16:51 <rosmaita> i will be really interested in looking at it
14:17:25 <abhishekk> that's it for today
14:17:32 <abhishekk> Moving to Open discussion
14:17:39 <abhishekk> #topic Open discussion
14:18:02 <dansmith> let me just say to jokke, that I think one change is import could not have had an overlimit before and can after this,
14:18:37 <jokke> just heads up, like agreed we've been looking into the caching API with abhishekk over this week. So hopefully updates and progress coming soon
14:18:45 <dansmith> but that is the only change I think, and even still, not a hole we're stuck in, since it mirrors other behavior and... it's not a strict behavior people have to depend on
14:18:56 <rosmaita> sounds like a lot is going on in glance this cycle!
14:19:18 <abhishekk> Just to update you all, I am struggling with post covid symptoms so not able to contribute with my full strengths yet
14:19:25 <rosmaita> i think a user should always be open to the possibility of getting overlimit
14:19:32 <dansmith> rosmaita: +1000
14:19:37 <rosmaita> abhishekk: please rest up!
14:19:43 <abhishekk> over to rosmaita now
14:19:48 <abhishekk> rosmaita, yep
14:19:54 <dansmith> abhishekk: rest up!
14:20:08 <abhishekk> yes
14:20:17 <jokke> abhishekk: yeah I haven't told that you yet today, take care of yourself buddy!
14:20:33 <abhishekk> thank you
14:20:39 <abhishekk> I will go through logs tomorrow
14:20:51 <abhishekk> thank you all
14:20:53 <Steap> yeah, health should always be the current cycle's top prio
14:21:06 <abhishekk> +1 Steap
14:21:32 <dansmith> we should #endmeeting while abhishekk can still do a clean shutdown
14:21:33 * abhishekk shutting down now, rosmaita will chair this now
14:21:55 <abhishekk> if nothing to discuss then I can do it
14:21:59 <jokke> go
14:22:04 <rosmaita> ok, from now on the official language of Glance will be Sweedish
14:22:05 <jokke> thanks all
14:22:13 <dansmith> heh
14:22:27 <abhishekk> ok, bye
14:22:37 <rosmaita> anything else for open discussion, or should we adjourn early?
14:23:05 <Steap> We could say it's "slut" if we insist on speaking Swedish
14:23:40 <dansmith> oh my
14:23:44 <dansmith> that word means something very different in english :)
14:23:49 <Steap> yep
14:24:00 <rosmaita> i guess maybe i don't want to conduct this meeting is Sweedish after all
14:24:09 <Steap> Written in big letters at the end of every movie
14:24:09 <dansmith> hah
14:24:19 <jokke> lat oss icke ga igenom
14:24:21 <dansmith> well, TIL
14:24:22 <Steap> was a bit of a surprise for French me as well
14:24:34 <Steap> jokke: are you actually fluent in Swedish as well as Finnish?
14:24:43 <rosmaita> Steap: what kind of movies are you watching?
14:24:52 <Steap> rosmaita: Swedish ones, duh
14:24:57 <jokke> Steap: hell not ... jeg talar icke svensk
14:25:22 <Steap> jokke: Jag heter Steap!
14:26:11 <rosmaita> Okej, låt oss avsluta det här mötet
14:26:31 <Steap> see, it's easy!
14:27:06 <jokke> rosmaita: I expect nothing less than you translate all our specs to swedish now :D
14:27:35 <rosmaita> well, i was really thinking we should rewrite glance in Go
14:27:43 <jokke> rosmaita: fortran
14:27:47 <rosmaita> något annat?
14:29:45 <rosmaita> låter som ingenting ... jag tror att vi kan skjuta upp
14:30:11 <Steap> fårtran?
14:30:55 <rosmaita> i think it's still FORTRAN, even in Sweedish
14:31:46 <Steap> Is "Sweedish" the language spoken in Sweden when people are smoking?
14:32:15 <rosmaita> i am really sorry that i started this
14:32:42 <rosmaita> OK, going ... going ...
14:33:03 <rosmaita> well, at least abhishekk should get a laugh
14:33:12 <rosmaita> #endmeeting