14:00:48 <jokke_> #startmeeting glance
14:00:49 <openstack> Meeting started Thu Aug 16 14:00:48 2018 UTC and is due to finish in 60 minutes.  The chair is jokke_. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:50 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:52 <openstack> The meeting name has been set to 'glance'
14:00:58 <jokke_> #topic roll call
14:01:01 <abhishekk> o/
14:01:01 <jokke_> o/
14:01:25 <jokke_> #link https://etherpad.openstack.org/p/glance-team-meeting-agenda
14:01:48 <rosmaita> o/
14:02:59 <jokke_> lets get started ... I guess Sean will join us if he can
14:03:06 <jokke_> #topic updates
14:03:08 <rosmaita> ok, cool
14:03:18 <rosmaita> sorry i was late ... is it just us?
14:03:30 <jokke_> abhishekk: is here as well
14:03:38 <abhishekk> hey
14:03:44 <rosmaita> hello abhishekk !
14:03:59 <abhishekk> hey Brian!!
14:03:59 <rosmaita> welcome back, hope you are well-rested!
14:04:14 <abhishekk> yes, I am, thank you
14:04:14 <jokke_> so I've been looking the bugs logged recently (mainly around after RC-1 and I think we look pretty decent
14:04:23 <jokke_> might not have broken the world
14:04:30 <rosmaita> good to hear
14:04:32 <jokke_> indeed welcome back abhishekk
14:04:50 <abhishekk> jokke_, :D
14:04:56 <jokke_> #topic release updates
14:05:09 <abhishekk> thank you both for taking care of multi-stores
14:05:48 <rosmaita> hth
14:06:50 * smcginnis missed calendar reminder
14:06:58 <jokke_> smcginnis: \o
14:07:08 <rosmaita> you haven't missed much yet
14:07:11 <rosmaita> abhishekk is back!
14:07:15 <abhishekk> looks like we have full house now
14:07:48 <abhishekk> smcginnis, o/
14:08:10 <jokke_> rosmaita: stage is yours ref release updates
14:08:19 <rosmaita> ok
14:08:29 <smcginnis> Welcome back abhishekk
14:08:35 <rosmaita> so, we do need to do a RC-2
14:08:44 <rosmaita> not much in it atm
14:08:51 <abhishekk> smcginnis, :D
14:08:57 <rosmaita> #link http://git.openstack.org/cgit/openstack/glance/log/?h=stable/rocky
14:09:13 <rosmaita> everything after 17.0.0.0rc1 tag
14:09:26 <abhishekk> I guess we need to document hidden images changes
14:09:33 <rosmaita> so my question is, is there any reason to wait
14:09:35 <smcginnis> Looks like the release note and the quoted chars would both be good.
14:09:38 <smcginnis> https://github.com/openstack/glance/compare/17.0.0.0rc1...stable/rocky
14:09:50 <rosmaita> abhishekk i thought we had?
14:09:59 <smcginnis> I saw the backport proposed for tests, but I don't think that's critical for RC.
14:10:00 <rosmaita> is it missing from api-ref?
14:10:25 <abhishekk> it is
14:10:41 <rosmaita> ok, i think i was supposed to do that, i will get right on it
14:10:50 <rosmaita> (hidden images in api-ref, i mean)
14:10:54 <jokke_> that should be quick to gate
14:11:05 <jokke_> if it's just doc change
14:11:20 <rosmaita> yeah, but still needs to go through integrated queue to merge, i think
14:11:41 <abhishekk> rosmaita, I guess it will be quick
14:11:43 <rosmaita> i had a question about release notes
14:12:02 <rosmaita> no bugfixes are listed, do we want to add some?
14:12:08 <jokke_> smcginnis: yeah, I cherry-picked those func tests into stable ... I don't really care at which point they get merged as long as they do. I think it was good testing to keep track of
14:12:56 <smcginnis> jokke_: ++
14:13:08 <rosmaita> also, about release notes, i think the number of cores fix mentioned in the prelude was actually done in queens
14:13:13 <rosmaita> (at least, it's mentioned in the queens release notes)
14:13:25 <rosmaita> doesn't hurt to mention it again, though
14:14:20 <rosmaita> so, it looks like outstanding is: (1) api-ref update for hidden images, (2) release notes update (if we want one)
14:14:27 <jokke_> rosmaita: IIRC the number of cores was backported after the release ... it's indeed included at least in the latests queens but I didn't think it made to the actual queens release
14:14:48 <rosmaita> could be, i didn't look that closely
14:15:15 <jokke_> the bugs section in renos would not hurt. Did we do something mention worthy, anyone had a look?
14:15:34 <rosmaita> well, what made me think was the password bug we're including in rc-2
14:16:01 <abhishekk> yes, we can mention that
14:16:24 <jokke_> yeah
14:17:14 <rosmaita> apropos of nothing, "limit default workers to 8" was in 16.0.0.0rc2
14:17:52 <jokke_> yes just noticed
14:18:16 <jokke_> so could remove that bit when doing the bugs
14:18:50 <rosmaita> i don't see a lot of bugs, scrolling through the logs ... mostly stability changes and features
14:19:22 <jokke_> for some reason it was listed when I looked the git log between 16.0.0 and HEAD
14:19:55 <smcginnis> The password one is probably the only notable one I can recall. But if we don't mention it, I think that's OK. Not sure how many people have actually been impacted by that. At least to the point that we need to highlight it there.
14:20:19 <abhishekk> In rocky we mostly worked on new features
14:20:55 <rosmaita> ok, so how does the timing look?
14:21:08 <rosmaita> i can probably get the api-ref update done in an hour or so
14:21:09 * jokke_ is wondering how much of a technical dept we accumulated this cycle or if we really are that stable atm.
14:21:43 <smcginnis> We're just a bunch of stable folks. :)
14:21:45 <abhishekk> https://launchpad.net/bugs/1753964
14:21:45 <openstack> Launchpad bug 1753964 in Glance queens "Image remains in queued state for web-download when node_staging_uri end with "/"" [High,Fix committed] - Assigned to Abhishek Kekane (abhishek-kekane)
14:21:47 <jokke_> rosmaita: I can do the reno change ... shouldn't take long and they both will gate quickly
14:22:25 <rosmaita> ok, then let's try to get those through today and maybe i can propose the release later this afternoon
14:22:33 <rosmaita> ok, that's all for RC-2
14:22:41 <rosmaita> tips jobs:
14:22:48 <rosmaita> short story, everything looking good
14:23:08 <rosmaita> we do have a maintenance issue, though, because we don't want them running in the stable branches
14:23:21 <rosmaita> i talked to corvus in the infra channel yesterday, got some advice
14:23:33 <rosmaita> so here's a sample patch for glance_store:
14:23:41 <rosmaita> https://review.openstack.org/#/c/586334/
14:23:56 <rosmaita> i think we should merge it and then backport to stable/rocky to verify that it works
14:24:03 <rosmaita> i don't know how else to test it
14:24:18 <rosmaita> also, it was a bit more complicated than i thought
14:24:41 <smcginnis> Makes sense. (as far as zuul goes)
14:24:50 <rosmaita> if you are interested, i put a comment on PS1 explaining how it works
14:24:57 <rosmaita> (or at least my understanding)
14:25:11 <abhishekk> ack
14:25:47 <rosmaita> ok, that's all from me
14:27:09 <jokke_> #topic ptg planning
14:27:40 <jokke_> #link https://etherpad.openstack.org/p/stein-ptg-glance-planning
14:27:58 <jokke_> the etherpad I've been promising for ages
14:28:26 <rosmaita> btw, i haven't been approved for travel yet
14:28:33 <smcginnis> Added to https://wiki.openstack.org/wiki/PTG/Stein/Etherpads
14:28:43 <jokke_> first of all, I think the delayed spec freeze worked pretty well and we should do that again.
14:28:45 <smcginnis> rosmaita: Are you thinking it might be an issue?
14:28:57 <rosmaita> hard to say
14:29:04 <jokke_> it's getting close for sure
14:29:16 <jokke_> we're less than a month away
14:29:19 <rosmaita> yep
14:29:35 <abhishekk> 22 days to go
14:30:32 <jokke_> rosmaita: so regardless if you're gonna be there or not, your imput to the planning is more than welcome ... we see what we can do about it if you can't be there
14:30:42 <rosmaita> sure
14:31:00 <jokke_> first of all, we need to start thinking what we want to do next cycle
14:31:01 <abhishekk> but be there
14:31:05 <smcginnis> We should see if anyone can stream a hangout if needed.
14:31:08 <rosmaita> i'll put in a proposal for the location policies refactoring
14:31:34 <jokke_> so we have that, we have stabilizing the multi-store
14:31:55 <rosmaita> is anything arising from the edge comupting discussions?
14:32:23 <jokke_> I want to start using that for the tasks as well so we can get rid of the horrid way of using the file store for workspace and staging
14:32:31 <rosmaita> and isn't there some kind of issue with eventlet and py3?
14:33:06 <abhishekk> +1, this we need to handle on top priority
14:33:10 <jokke_> so Tuesday is Edge day. They have proposal for image stuff in there
14:33:24 <rosmaita> ok, cool, so we can use that time instead of ours
14:33:35 <jokke_> #link https://etherpad.openstack.org/p/EdgeComputingGroupPTG4
14:34:11 <jokke_> it might be that we should schedule spare slot for any continuation if we don't have time issue (which I don't think we will)
14:35:15 <jokke_> what's that eventlet py3 issue?
14:35:49 <rosmaita> i don't know, i sort of remember hearing that there's an issue about the eventlet-based server and py3
14:35:57 <abhishekk> I guess https is not working with py3 due to eventlet issue
14:36:00 <rosmaita> (i will be happy to be wrong)
14:36:27 <rosmaita> abhishekk you are right, that's it, and iirc it is unlikely to be fixed
14:37:12 <smcginnis> The keystone team has been doing a lot of good work this release moving to use flask. It's a big(ish) effort, but might be worth considering that if it solves some of our problems.
14:37:26 <abhishekk> rosmaita, yes
14:37:28 <jokke_> abhishekk: yeah, there is that, but I don't think it's super critical tbh. As discussed last time there is plenty of ways handle that ssl termination ... I'd be very interested to see some definite resolution for that 'though
14:38:35 <jokke_> smcginnis: I'd be more than happy to discuss that option as soon as we have more than 3-4 people working on Glance ;)
14:38:41 <abhishekk> jokke_, agree
14:38:46 <smcginnis> jokke_: Yeah...
14:40:21 <jokke_> But back to the PTG planning ... discuss internally what your companies would like to achieve, what you can and want to work on and propose sessions so we can make work plan for Stein
14:40:40 <abhishekk> ack
14:41:27 <jokke_> if we don't get stuffed with topics I'll schedule couple of strategically placed ad-hoc slots there for anything that might come up earlier on the week
14:41:44 <smcginnis> Something always comes up. :)
14:41:51 <jokke_> indeed
14:42:02 <jokke_> ok, moving on
14:42:06 <jokke_> #open discussion
14:42:12 <jokke_> #topic open discussion
14:42:18 <rosmaita> well, the mother of all security bugs became public this week
14:42:36 <rosmaita> (that's related to the image locations policies refactoring effort)
14:42:50 <jokke_> rosmaita: yup, that's the storm raising I mentioned last week
14:43:15 <rosmaita> #link https://bugs.launchpad.net/glance/+bug/1546507
14:43:15 <openstack> Launchpad bug 1546507 in Glance ocata "Regular user in non-default non-recommended configuration can delete any image file" [Critical,Confirmed] - Assigned to Mike Fedosin (mfedosin)
14:44:30 <jokke_> I think we almost decided to sack that whole bug last time in Denver (we had lengthy discussion about it at Fri)
14:45:19 <jokke_> it was one of those "We've been telling to you for long time not to expose locations to end users." things
14:45:31 <rosmaita> yes, i think the plan was to issue a general warning, but there's nothing we can do to really fix it short of rewriting glance in flask, maybe
14:46:05 <rosmaita> anyway, we should probably come up with something at the PTG
14:46:21 <rosmaita> at least the bug title isn't as alarmist as the original
14:46:48 <jokke_> well rewriting in <insert web framework here> won't solve the issue per se ... it's all folding back to the very faulty design of exposing those locations and their operations
14:48:24 <rosmaita> ok, i had one agenda item
14:49:18 <rosmaita> smcginnis can correct me, but i think because of the way reno works, if you fix a typo in a release note, the release note will be associated with that commit, and it will show up in the latest release unless you specifically add an exclusion
14:49:32 <rosmaita> so i am thinking the best thing is not to make corrections to these
14:49:33 <smcginnis> Yep, that is correct.
14:49:47 <jokke_> oh :O
14:49:49 <rosmaita> this is why i bring it up: https://review.openstack.org/#/c/574929/
14:49:54 <smcginnis> The recommended way to fix those is to go directly to the stable branch where they were for.
14:50:21 <jokke_> I thought reno is tracking them by the hex it generates when the file is created
14:51:00 <smcginnis> Hmm, maybe Doug fixed that, but I know it was the behavior that it matched on the branch where any commit modified it.
14:51:31 <dhellmann> rosmaita : we fixed that bug, so it should be fine to fix them
14:51:37 <jokke_> ok, maybe we should play it safe and just add it to the disallowed changes
14:51:39 <rosmaita> dhellmann : thanks!
14:51:44 <dhellmann> you want to make sure the changes are on the  right branch, though
14:51:49 <dhellmann> don't fix it on master and backport to ocata
14:51:56 <jokke_> "Review the renos when they were introduced or live with it"
14:51:58 <rosmaita> so fix on ocata only?
14:52:01 <dhellmann> right
14:52:06 <dhellmann> fix it where you want the change to show up
14:52:11 <rosmaita> ok, thanks
14:52:23 <jokke_> thanks dhellmann
14:52:25 <rosmaita> see, another reason to delete those old branches
14:52:30 <rosmaita> :)
14:52:35 <abhishekk> :D
14:54:03 <jokke_> ok about that then, should we clean up old releasenote files from master when ever the new branch is cut 'cause if we just fix them on old branches we'll end up with multiple versions of those files and someone will run their spell checkers against master and propose those changes forever
14:54:25 <rosmaita> dhellmann: ^^
14:55:02 <dhellmann> it's better to just leave all of that stuff in place and reject the patches you don't want
14:55:22 <rosmaita> ok, so maybe we update
14:55:28 <rosmaita> disallowed minor code changes
14:55:32 <rosmaita> and reject the patch
14:56:09 <jokke_> #info need to add release note modifications in master to disallowed minor changes list
14:57:00 <rosmaita> how do we feel about fixing in the stable branch?
14:57:40 <smcginnis> The the stable branch? :P
14:57:48 <rosmaita> :P
14:58:28 <jokke_> I really think https://youtu.be/AJQ3TM-p2QI?t=44s
14:58:43 <rosmaita> without looking, i bet that is "computer says no"
14:58:49 <jokke_> ;)
14:58:50 <dhellmann> you could probably automate a check to prevent changes to files on master if they exist on stable branches
14:58:52 <abhishekk> :P
14:59:37 <dhellmann> given the recent push by the i18n team to translate the release notes, eliminating typos seems like a good thing to support
14:59:39 <rosmaita> ok, we can tell that person to propose a patch implementing doug's suggestion, and then we'll remove the redundant "the" from the release note
14:59:58 <dhellmann> it's hard enough to translate our jargon without having to decipher typos :-)
15:00:31 <rosmaita> i'm pretty sure that was my typo, too
15:00:45 <dhellmann> or you could just ask them to move the fix to the right branch
15:01:07 <dhellmann> folks who fix typos are unlikely to be up to the level of automation I described
15:01:19 <rosmaita> you never know!
15:01:28 <rosmaita> i don't think we have the review bandwidth, though
15:01:38 <dhellmann> how many of these do you really get?
15:01:55 * jokke_ pings flaper for the -2 & abandon bot
15:02:13 <dhellmann> I feel like being too hostile to simple fixes is going to make the pipeline of potential new contributors very very shallow
15:02:26 <dhellmann> I started in openstack with documentation typo fixes
15:02:42 <jokke_> dhellmann: diffilcult to get shallower than the dry land currently is :D
15:02:50 <dhellmann> so let's not make it any worse, then
15:02:56 <jokke_> I don't think us being bit hostile will change that
15:03:16 <rosmaita> well, i just -2'd the patch and asked it to be proposed to stable/pike
15:03:23 <jokke_> well tbh. We've got couple of new names popping up this cycle
15:03:25 <rosmaita> and i did ask nicely
15:03:29 <rosmaita> :)
15:03:43 <rosmaita> and, btw, it definitely was my typo: https://github.com/openstack/glance/blame/master/releasenotes/notes/exp-emc-mig-fix-a7e28d547ac38f9e.yaml
15:03:46 <dhellmann> if someone's first interaction with our community is "we don't care about typos, either go away or build something you don't understand" then our reputation *outside* of our community won't be improved at all
15:04:01 <jokke_> dhellmann: that
15:04:08 <jokke_> and we're out of time
15:04:18 <jokke_> so thanks all!
15:04:27 <rosmaita> bye!
15:04:31 <dhellmann> part of the review team's responsibility is to guide new contributors, so I'm always a bit disappointed when this comes up
15:04:48 <abhishekk> thank you all
15:04:53 <jokke_> #endmeeting