14:01:17 <jokke> #startmeeting glance
14:01:17 <openstack> Meeting started Thu Dec 10 14:01:17 2020 UTC and is due to finish in 60 minutes.  The chair is jokke. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:01:18 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:01:21 <openstack> The meeting name has been set to 'glance'
14:01:29 <jokke> #topic roll call
14:01:59 <jokke> Abhishek just lost power so will be minute or two when he tries to get on from mobile
14:02:03 <rosmaita> o/
14:02:03 <jokke> o/
14:02:14 <rosmaita> ok, i was wondering
14:03:39 <abhishekk> o/ from mobile
14:03:40 <Steap> o/
14:03:45 <jokke> #link https://etherpad.opendev.org/p/glance-team-meeting-agenda quite a few things on the agenda
14:03:50 <jokke> wb abhishekk
14:03:54 <jokke> #chair abhishekk
14:03:56 <openstack> Current chairs: abhishekk jokke
14:04:07 <abhishekk> will be difficult for me to switch over agenda, so jokke could you please continue chairing?
14:04:17 <jokke> sure
14:04:22 <abhishekk> thank you
14:04:34 <jokke> #topic release/periodic job updates
14:05:29 <abhishekk> just passed milestone 1
14:05:40 <jokke> so we seem to gave some serious gating issues still. Milestone 1 has not been tagged as of now
14:05:56 <abhishekk> yes
14:06:13 <abhishekk> those jobs are failing for othercouple of projects as well
14:06:20 <rosmaita> is it lower-constraints?
14:06:25 <abhishekk> yes
14:06:30 <abhishekk> and same error for docs job
14:06:54 <rosmaita> i had to update requirements and l-c for cinder, it's the new pip resolver
14:07:20 <abhishekk> then some tempest jobs are failing with unknown error and no logs were generated
14:07:21 <rosmaita> https://review.opendev.org/c/openstack/cinder/+/766085
14:07:25 <rosmaita> a bit of a PITA
14:07:31 <jokke> ok, I will look the cinder patch after the meeting and see if I can find the help for us from there
14:07:43 <abhishekk> jokke, rosmaita ++
14:08:03 <rosmaita> jokke: i did a pip freeze on a successful local unit test run, and used those versions
14:08:20 <rosmaita> but i don't remember why we even have a lower constraints job any more
14:08:37 <abhishekk> also we need to get Steap's patch once these issues resolved
14:08:40 <jokke> rosmaita: ah, so it's just dependency versions instead of something about formatting of them files
14:09:00 <rosmaita> yeah, pretty sure
14:10:16 <jokke> rosmaita: I remember one upon a time we had centralized dependency management where these things had to be done only on global requirements and bot would propose them down, but someone thought that rather than having unified requirements it's better to let everyone just swing it project by project
14:10:53 <rosmaita> :)
14:11:28 <abhishekk> i think we faced this issue earlier as well
14:11:35 <jokke> yup
14:11:47 <jokke> it's every few months
14:12:26 <abhishekk> hmm
14:12:41 <jokke> but we can move forward I'll ping ye guys if I can't figure it out myself looking what had to be taken in for cinder and with local runs
14:12:59 <rosmaita> ok, i will be around today (on PTO tomorrow)
14:13:06 <jokke> rosmaita: great, thanks
14:13:10 <abhishekk> ping me on telegram if I didn't returned back online due to power failure
14:13:29 <jokke> #topic milestone 2
14:13:34 <jokke> 5 weeks away
14:13:46 <jokke> abhishekk had some priorities laid down in the agenda
14:14:04 <abhishekk> yes, and we have couple of specs + RBAC stuff to address
14:14:29 <jokke> Mainly looking into the proxying those calls for Image Import and RBAC
14:14:57 <abhishekk> ++
14:15:13 <abhishekk> others are trivial and can be implemented quickly
14:16:07 <jokke> I will look into addressing any comments on the image import proxy spec and cluster awareness spec tomorrow if I got the gate unblocked by that
14:16:08 <rosmaita> i left comments on all the specs listed in the agenda
14:16:27 <abhishekk> rosmaita, ack, thank you
14:16:28 <rosmaita> was curious what others thought about the chunk download support for RBD
14:16:33 <jokke> rosmaita: great, thanks!
14:16:56 <rosmaita> jokke: don't thank me, made more work for you
14:17:06 <abhishekk> :P
14:18:25 <abhishekk> So next topic is review priorities where I listed some specs to be reviewed
14:18:35 <jokke> so I'm failing to see the usecase for that chunk download
14:18:50 <jokke> and if we do it, we really should do it for all stores, not just rbd
14:19:06 <jokke> oh sorry
14:19:06 <abhishekk> agree
14:19:07 <rosmaita> well, we do support the range header
14:19:19 <abhishekk> agree for use case
14:19:19 <jokke> #topic important spec reviews
14:19:20 <rosmaita> i remember Darini fixing a bug about that back in the osic days
14:19:57 <abhishekk> is it documented somewhere?
14:19:58 <rosmaita> anyway, i left a comment there that i'd like to hear more about how this will be tested
14:20:02 <abhishekk> about range header?
14:20:25 <rosmaita> i'll look for a reference
14:20:31 <jokke> rosmaita: ok, so clearly there is some holes on it and it's not tested ;)
14:20:37 <jokke> One of those again
14:22:47 <jokke> on top of those there's the discovery api for tasks and sqlalchemy-migrate deprecation on the list. Any marks on them before we move on?
14:23:29 <abhishekk> for task thing we will continue discussion on specs
14:23:53 <rosmaita> sqlalchemy-migrate is fine
14:24:08 <rosmaita> i'm not sold on using the discovery endpoint for tasks, though
14:24:20 <rosmaita> left a comment on the spec
14:24:48 <rosmaita> abhishekk: i added the commit hashes for range header support to the agenda
14:25:05 <abhishekk> rosmaita, ack, will have a look
14:25:25 <abhishekk> also I replied to your comment on tasks specs
14:25:55 <jokke> rosmaita: ok, I see your point. Maybe we need to have fresh new endpoint for probing the tasks
14:26:13 <abhishekk> I took shortcut :D
14:26:37 <rosmaita> or have multiple api-level policies on the tasks API
14:26:41 <jokke> the tasks api is just usage nightmare and we deprecated it for removal ages ago for good reason. I'd be very much against just opening it back up as it is
14:26:51 <rosmaita> that tasks-api-access thing was a last minute hack
14:27:03 <jokke> yeah
14:27:35 <rosmaita> well, you could maybe do something like v2/messages/{id}
14:27:46 <rosmaita> and under the hood, grab the task and reformat it for display
14:27:55 <abhishekk> ack, sounds good
14:28:18 <abhishekk> will update the specs accordinglyu
14:28:25 <jokke> ok, lets keep the discussion going on the spec and move on
14:28:54 <jokke> oh, od left
14:28:58 <jokke> #open discussion
14:29:05 <jokke> #topic open discussion
14:29:31 <jokke> rosmaita: looks like you added the chunk download refs on the agenda, thanks
14:29:33 <rosmaita> fungi helped me figure out how to fix the redirects
14:29:34 <rosmaita> https://review.opendev.org/c/openstack/glance-specs/+/766260
14:29:39 <rosmaita> for glance-specs
14:29:41 <abhishekk> great
14:29:46 <jokke> \\o \o/ o// o/7
14:30:01 <rosmaita> i didn't add any new ones, this just fixes the current ones (up to stein)
14:30:08 <jokke> rosmaita++ fungi++ that's great news, thanks!
14:30:18 <rosmaita> but i don't think there's any way to test it other than to merge it and see what happens
14:30:19 <abhishekk> will have a look and work for other redirects
14:30:28 <rosmaita> it worked for cinder, so that's some evidence
14:30:36 <abhishekk> cool
14:30:50 <fungi> well, it's an apache behavior, so the way to test it is to serve content from apache
14:31:13 <abhishekk> ack
14:31:20 <fungi> the apache configuration we use is public, if anyone wants it
14:31:37 <abhishekk> thank you fungi
14:31:42 <rosmaita> try this out: https://specs.openstack.org/openstack/cinder-specs/specs/ussuri/image-encryption.html
14:31:49 <abhishekk> will ping you if required
14:32:24 <abhishekk> rosmaita, thank you, will try once back online
14:32:35 <rosmaita> :)
14:33:15 <abhishekk> great, I don't have anything else to discuss at the moment
14:33:47 <jokke> I'm still trying to wrap my head around the rbac stuff and it's impact
14:34:14 <Steap> regarding RBAC
14:34:29 <Steap> Lance mentioned that migrating policies would be a "manual" process
14:34:43 <Steap> that worries me a tiny bit :)
14:35:56 <abhishekk> :D
14:36:11 <jokke> we just need to be super vigilant and make sure we do not expose anything unintended
14:36:33 <jokke> Which is my biggest worry on the amount of pressure this wor has to rush it through
14:36:56 <jokke> s/wor/work/
14:37:00 <abhishekk> testing, more testing and more and more testing is what we need to do in less time
14:37:57 <abhishekk> we should sign a MOU :D
14:38:25 <jokke> MOU?
14:39:02 <rosmaita> memorandum of understanding?
14:39:16 <abhishekk> just kidding, memo for understanding the regression
14:39:24 <jokke> ah
14:39:27 <jokke> ;)
14:40:20 <jokke> well if anyone has understanding of that and wants to write it down, I'd be very happy to read and comment :P
14:40:37 <abhishekk> haha, rosmaita is the perfect person for this
14:41:16 <rosmaita> i'm driving through a tunnel, couldn't catch that last comment
14:41:25 <jokke> Yeah I think rosmaita is pretty much the only one left who has even chance of having full understanding tf is going on there ;)
14:41:59 <jokke> rosmaita: I just said that if anyone has understanding and wants to write it down, I'd be more than happy to read and comment
14:42:15 <rosmaita> still can't hear you
14:42:21 <jokke> :P
14:42:22 <rosmaita> :)
14:42:47 <abhishekk> :P
14:43:05 <jokke> yeah I guess most of the people whe were around when that masterpiece was written are long gone. At least I've been exposed only in small bits to it
14:44:03 <abhishekk> :D
14:44:27 <jokke> but I've nothing else. We can take the 15 unless someone else has something still to bring up
14:44:37 <abhishekk> sounds good
14:45:00 <rosmaita> nothing from me
14:45:11 <jokke> ok, thanks all!
14:45:13 <abhishekk> thank you all
14:45:18 <rosmaita> bye!
14:45:19 <jokke> #endmeeting