21:00:19 <notmyname> #startmeeting swift
21:00:20 <openstack> Meeting started Wed Apr  5 21:00:19 2017 UTC and is due to finish in 60 minutes.  The chair is notmyname. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:00:21 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:00:24 <openstack> The meeting name has been set to 'swift'
21:00:26 <notmyname> who's here for the swift team meeting?
21:00:31 <patchbot> I'm here for the meeting!
21:00:41 <dmorita> o/
21:00:44 <mattoliverau> o/
21:00:49 <mathiasb> o/
21:00:51 <tdasilva> hi
21:00:53 <rledisez> hi o/
21:01:00 <kota_> o/
21:01:12 * SotK lurks
21:01:18 <notmyname> I fixed patchbot, so it can join us in meetings again
21:01:23 <tdasilva> has patchbot grown a sense of humor?
21:01:41 * patchbot is alive!
21:02:20 <mattoliverau> what it attempts to be humour
21:02:21 <notmyname> clayg: acoles: jrichli: here?
21:02:21 <acoles> hi
21:02:29 <notmyname> mattoliverau: no kidding
21:02:58 <clayg> no
21:03:06 <jrichli> yes sir :-)
21:03:09 <m_kazuhiro> o/
21:03:11 <torgomatic> .
21:03:13 <notmyname> a few things to go over this week
21:03:16 <notmyname> #link https://wiki.openstack.org/wiki/Meetings/Swift
21:03:23 <notmyname> #topic follow-up from last week
21:03:41 <notmyname> mattoliverau: anything new on stable branch test failures? anything else to do there?
21:04:03 <notmyname> I've landed all the backport patches, including yours
21:04:16 <mattoliverau> I haven't been as productive as I'd liked there this week. got sidetracked on one tho
21:04:22 <notmyname> mattoliverau: I think we're good on this, but did you find anythign else?
21:04:40 <mattoliverau> FAIL: test_GET (test.unit.proxy.test_server.TestAccountController)
21:05:12 <notmyname> that's not the same one you had a patch for, right?
21:05:32 <mattoliverau> fails every 100ish time, but also on master. It seems to be FakeConn or GETorHeadHandler loosing a request when trying to find qouum.
21:05:34 <mattoliverau> nope.
21:05:40 <notmyname> ok
21:05:45 <mattoliverau> but just kept showing up
21:05:47 <notmyname> mattoliverau: do you have a bug filed for it in launchpad?
21:05:53 <mattoliverau> so went ot test master
21:06:10 <mattoliverau> no, found it yesterday arvo, so wanted to see if it was a problem elsewhere.
21:06:23 <notmyname> ok
21:06:28 <mattoliverau> I'll put one up today.. I'm part way through investigating it tho
21:06:29 <notmyname> please file that when you get some info on it
21:06:32 <notmyname> thanks
21:06:48 <notmyname> next up...
21:07:07 <notmyname> patch 451524 has landed. approved by the tc yesterday
21:07:08 <patchbot> https://review.openstack.org/#/c/451524/ - governance - Use case analysis for Golang addition to Openstack (MERGED)
21:07:14 <notmyname> thank you tdasilva for writing that up
21:07:27 <mattoliverau> yeah nice work tdasilva
21:07:30 <jrichli> yay!
21:07:33 <notmyname> what this means is that the tc has formally approved swift using golang
21:07:35 <kota_> nice!
21:08:01 <notmyname> there's a lot of work to be done, of course, some of which is figuring out the logistics for -infra
21:08:07 <clayg> umm.. *is* that what i means?
21:08:18 <tdasilva> was asking the same question
21:09:26 <notmyname> according to the tc process (https://governance.openstack.org/tc/reference/new-language-requirements.html) yes
21:09:48 <notmyname> step one, approval of the need for a new language, has been done
21:09:50 <notmyname> After reviewing and agreeing on the need for a new language, the team driving this effort should start working on the second phase. The needs of a new language should not be questioned during the evaluation of the second phase, unless the team working on it decides the language is not needed anymore.
21:10:37 <notmyname> as a reminder to everyone, we don't want golang because it will be fun. we need it as part of solving some bigger problems facing every swift cluster
21:10:47 <notmyname> the high-level outline of that is at https://wiki.openstack.org/wiki/Swift/Fixing-rebalance-and-golang
21:11:30 <notmyname> so far, clayg has been digging into that work (the general rebalance problem)
21:11:42 <notmyname> but there is a *lot* to do
21:11:57 <notmyname> it is a great area for people to help out in
21:12:10 <notmyname> and one of the two main priorities for stuff to get done (the other being container sharding)
21:12:39 <notmyname> I'm happy the first step of the new TC process is over
21:12:47 <notmyname> any questions on that before we move on?
21:12:54 <clayg> notmyname: but now that swift has outlined our use-case - does that mean that openstack-client can start to merge golang patches as soon as common infrastructure and/or shared modules are developed?
21:13:22 <notmyname> clayg: my understanding is that it's been approved for swift. I don't know if that's a general approval or not
21:14:34 <tdasilva> clayg: IIUC, yes
21:15:51 <notmyname> but note that we are expected to help out with building that common infrastructure as part of getting golang into our codebase
21:16:03 <notmyname> next follow-up...
21:16:06 <notmyname> https://bugs.launchpad.net/swift/+bug/1675500
21:16:06 <openstack> Launchpad bug 1675500 in OpenStack Object Storage (swift) "Container/account disk drive fault results replication on all rest drives" [High,In progress] - Assigned to Pavel Kvasnička (pavel-kvasnicka)
21:16:06 <acoles> there's a second resolution needed before projects can release golang https://governance.openstack.org/tc/reference/new-language-requirements.html#guarantee-compatible-functionality-for-the-base-common-libraries
21:16:37 <clayg> notmyname: ok awesome!
21:16:53 <notmyname> mattoliverau: rledisez: acoles: last week the three of you said you'd look into that bug
21:16:55 <notmyname> any news?
21:17:19 <notmyname> kota_: you commented too :-)
21:17:21 <mattoliverau> Yeah, it's deinfitely a bug, and happens on all modes
21:17:35 <mattoliverau> no matter the number of replicas I mean
21:17:39 <kota_> yup
21:17:48 <notmyname> https://review.openstack.org/#/c/448480/ is proposed, and clayg has -1 for text failure(?)
21:17:49 <patchbot> patch 448480 - swift - Container drive error results double space usage o...
21:18:11 <mattoliverau> I wrote some probe tests and linked them in the patch for those who want to see.
21:18:13 <acoles> notmyname: sorry, I didn't yet, got distracted by reconstructor bugs
21:18:29 <notmyname> acoles: no worries. looks like kota_ mattoliverau and clayg have looked at it
21:19:04 <notmyname> clayg: oh sorry, not a test failure. needs more work to clean up the over-replicated data
21:19:04 <mattoliverau> Both clayg and I put some sample patches in the bug to stop the bleeding, the patch goes further. and I think that's where its stalled
21:19:10 <notmyname> good
21:19:15 <notmyname> what's the next step then?
21:19:32 <notmyname> what do we need to give to pavel? what do we need to get from him?
21:19:42 <mattoliverau> what... should the fix be. should we behave like object replication, and things sit on handoffs until its ready.
21:19:46 <notmyname> onovy: ^ FYI
21:19:46 <mattoliverau> *fixed
21:19:58 <mattoliverau> (too early for me to make sentences)
21:20:00 <mattoliverau> :P
21:20:02 <notmyname> heh
21:20:30 <mattoliverau> should handoffs only push to primaries.
21:21:50 <notmyname> mattoliverau: kota_: can you keep working with pavel this week on getting a viable patch?
21:21:51 <clayg> mattoliverau: are you sure "handoffs only push to primaries" is a thing object replication does?  Maybe it is... but even when they see a 507?  seems unlikely
21:22:31 <mattoliverau> clayg: I'm just going on comments in the patch, because I haven't coffeed :P
21:22:38 <kota_> notmyname: i'm available in this week but idk if pavel is
21:22:48 <notmyname> ok
21:23:13 <notmyname> also, don't be shy to push changes over what's proposed
21:23:17 <notmyname> that's ok
21:23:28 <kota_> as mattoliverau described, to fix the bug with minimum code, i think clayg's patch can be good
21:23:43 <mattoliverau> I think claygs question in patch was, there is a high prioriy bug, should we stop the bleeding then fix the rest when we have more time it iterate.
21:23:45 <kota_> and i'm waiting pavel's opinion at the patch now.
21:23:52 <notmyname> ok
21:24:21 <notmyname> anything more to talk about on that patch?
21:24:45 <notmyname> ok, moving on
21:24:51 <notmyname> to new things
21:24:56 <notmyname> #topic tc vision document
21:25:07 <notmyname> #link https://review.openstack.org/#/c/453262/
21:25:08 <patchbot> patch 453262 - governance - Draft technical committee vision for public feedback
21:25:12 <notmyname> pay attention to this
21:25:30 <notmyname> it was just proposed and is an attempt to frame the next few years of openstack
21:25:38 <notmyname> the tc is looking for comments
21:25:47 <notmyname> please read over it and leave comments
21:26:04 <notmyname> #topic migrating to storyboard
21:26:13 <notmyname> remember when we talked about storyboard?
21:26:20 <mattoliverau> yup
21:26:30 <mattoliverau> and we were surprised it was still a thing
21:26:44 <notmyname> what was that? I know at the PTG. maybe also Austin? Barca?
21:26:52 <notmyname> it's been talked about a few times
21:26:57 <notmyname> and it's come up again
21:27:07 <notmyname> is there someone here from the storyboard team?
21:27:10 <diablo_rojo> Yep
21:27:15 <notmyname> diablo_rojo: hi! welcome
21:27:22 <SotK> o/
21:27:23 <diablo_rojo> I think SotK is creeping too
21:27:34 <notmyname> diablo_rojo: so tell us more. you've talked about swift using storyboard. we've been interested in the past
21:27:42 <notmyname> but we don't know anything about it or how that works
21:28:03 <notmyname> welcome SotK
21:28:05 <diablo_rojo> So basically we have migrated a few projects- Monasca, refstack, and the interop wg- and they joined infra.
21:28:29 <diablo_rojo> Its a pretty simple process, we have scripts to migrate open items in lp over so you dont need to recreate everything from scratch.
21:28:57 <notmyname> it's a complete replacement of launchpad right?
21:28:59 <diablo_rojo> We were just wondering if you were interested/ if there was anything you may have noticed sb doesn't have that you would need to be able to use it.
21:29:04 <diablo_rojo> notmyname, correct
21:29:22 <diablo_rojo> I am actually working on making it a Queens release goal to have all projects migrated by the end of the release
21:29:36 * diablo_rojo was working on drafting that before this meeting :)
21:29:44 <notmyname> last we talked about it, we're interested, but we don't know much about storyboard. and therefore we also don't really understand what a migration means
21:29:57 <clayg> diablo_rojo: are any not projects that don't host code on git.openstack.org *using* storyboard today?
21:30:00 <notmyname> but it sounds like it will be a thing everyone does, so getting on the storyboard train early is good ;-)
21:30:21 <clayg> s/are any not projects/are any projects/g - ssory
21:30:23 <clayg> heh
21:30:32 <mattoliverau> How can we look at it and have a play with its features, is there a rest env, or should we go raise a bunch of feautures in refstack or something? :P
21:30:43 <diablo_rojo> clayg, Craton isn't an official project and they want to be part of the next wave of migrations- is that what you are talking about?
21:30:55 <SotK> mattoliverau: there is storyboard-dev.openstack.org for playing around with
21:31:01 <mattoliverau> nice :)
21:31:12 <diablo_rojo> mattoliverau, I have sent a few things out to the ml with blog posts and a bunch of links, let me get the archive link
21:31:23 <mattoliverau> diablo_rojo: ta
21:31:23 <notmyname> what does it do with existing links in existing commit messages?
21:31:30 <notmyname> do bug numbers stay the same?
21:32:01 * notmyname is sure you've already thought about and handled all this
21:32:12 <SotK> yeah, story IDs are the same as the bug number in LP
21:32:25 <diablo_rojo> mattoliverau, here's the ml thread: http://lists.openstack.org/pipermail/openstack-dev/2017-February/112282.html
21:32:39 <clayg> diablo_rojo: nope, I'm just curious to what extend OpenStack is building an island - sounds like to begin with it's standard "Open
21:32:42 <mattoliverau> #link  http://lists.openstack.org/pipermail/openstack-dev/2017-February/112282.html
21:33:10 <notmyname> ok, so if you ran migration scripts today, what do we do tomorrow? how does it "work" from a day-to-day perspective? we jsut go there instead of LP and that's it?
21:33:22 <clayg> ... well just that launchpad isn't anyone's favorite - but utlimately it's going to do things other FOSS projects currently get done with launchpad
21:33:27 <clayg> ... or github issues
21:33:43 <notmyname> clayg: it == storyboard?
21:33:52 <clayg> yeah
21:33:55 <diablo_rojo> notmyname, yep, you just stop using lp and file things in sb instead- also probably want to tell vendors about it too.
21:33:57 <SotK> I don't know what the plan is for existing commit messages, probably before all projects are merged it'd require manually fixing them to use the Story: #### syntax
21:34:45 * notmyname needs to go re read the linked storyboard docs
21:35:19 <notmyname> SotK: has any project that does vuln management migrated to it yet?
21:35:20 <clayg> mattoliverau: thanks for the link
21:35:21 <diablo_rojo> clayg, maybe fungi can shed some light on that
21:35:31 <clayg> notmyname: good question!
21:35:42 <mattoliverau> lol, it was diablo_rojo's :)
21:36:51 <SotK> notmyname: not yet afaik
21:37:11 <notmyname> but there's functionality for that already? private bugs with private code reviews?
21:37:26 <diablo_rojo> notmyname, that is in progress work actually.
21:37:29 <notmyname> has the vuln management team given feedback?
21:37:31 <diablo_rojo> SotK, its about half done?
21:37:32 <notmyname> ah ok
21:37:38 <acoles> diablo_rojo: is there a mechanism to prevent new items going to launchpad after a migration? a redirect?
21:37:41 <notmyname> so it might be wise to wait for that?
21:37:46 <SotK> clayg: I know of other storyboards that are deployed and used to varying extents
21:38:01 <clayg> SotK: that's *awesome*
21:38:04 <diablo_rojo> notmyname, yeah if that is something you need, we would wait to move you till after thats done.
21:38:18 <notmyname> diablo_rojo: yeah, definitely
21:38:40 <clayg> notmyname: our needs there aren't complex - the half they've implemented might be fine
21:38:45 <mattoliverau> notmyname: what do you mean, we don't have bugs, esp embargo ones :P
21:38:50 <diablo_rojo> notmyname, we will write you down on the list of projects that need that before they can move.
21:38:56 <tdasilva> what about using it for other things like trello board?
21:38:57 <SotK> notmyname: indeed, there are private stories (where the story and all related things like tasks + comments + history)
21:39:20 <SotK> the missing functionality is in automatically making vulnerabilities visible to the vmt
21:39:24 <notmyname> last question (for now, I think), is the migration story written up somewhere? is it in that email? like, we do #1, then you do #2, then next is #3...
21:39:47 <SotK> there is also the fact that our migration script currently doesn't migrate private bugs
21:39:50 <notmyname> SotK: hmm.. ok
21:40:31 <diablo_rojo> notmyname, No, its really just an agreement on a date and then we get an infra friend (usually fungi) to run the scripts and then its done. Not much you need to do but spread the word.
21:40:37 <notmyname> SotK: one would hope that the number of private bugs would be small enough that they could be moved by hand :-)
21:40:47 <SotK> notmyname: indeed :)
21:41:08 <notmyname> so what acoles asked... do we "turn off" LP after a migration?
21:41:12 <notmyname> is that possible?
21:41:24 <acoles> we need it to become readonly right?
21:41:29 <notmyname> right
21:41:32 <notmyname> (I think)
21:41:33 <SotK> correct I believe
21:41:39 <acoles> or does its entire history get migrated?
21:41:50 <SotK> (at least, I'm pretty sure that's what infra did to theirs way back when)
21:42:01 <diablo_rojo> acoles, not the history, just open items
21:42:01 <notmyname> which one?
21:42:04 <notmyname> ok
21:42:29 <notmyname> history stays in LP, LP becomes read-only, open things get migrated, and new stuff goes to storyboard
21:42:34 <acoles> so we'd want launchpad to stay there in readonly mode, somehow
21:42:47 <notmyname> which also solves the existing links problem
21:43:03 <clayg> looks just "gone" to me?  https://bugs.launchpad.net/openstack-ci
21:43:23 <acoles> notmyname: how does it solve links? closes-bug to an open bug in launchpad woul dneed to be changed?
21:43:41 <notmyname> acoles: meaning that existing links to LP would still work
21:43:50 <notmyname> ...but I'm not sure about what clayg just said
21:44:14 <SotK> clayg: I may be misremembering, I'm sure fungi can clarify at some point
21:44:29 <clayg> notmyname: i mean, it'd be nice to have an exmple of a project that used launchpad and now uses story board
21:44:39 <clayg> if so such project exists we could just "do it"
21:44:47 <notmyname> yeah
21:44:48 <diablo_rojo> clayg, Monasca is one you could talk to :)
21:44:49 <notmyname> SotK: diablo_rojo: this has been very helpful. thanks. what do we as a team need to consider next or do next?
21:45:07 <notmyname> talk to Monasca, it sounds like
21:45:10 <notmyname> what else?
21:45:26 <diablo_rojo> notmyname, I would say figure out if there is anything else you need as a project besides VMT things or if what is there currently works for you guys.
21:45:27 <clayg> so monasca still has bugs -> https://bugs.launchpad.net/monasca
21:45:41 <clayg> i can report a new one -> https://bugs.launchpad.net/monasca/+filebug
21:45:44 <tdasilva> would it make sense to use for other functionality like "trello board" type of functionality...just to get our feet wet with sb???
21:45:53 <SotK> notmyname: play around with storyboard and see if there is anything that jumps out as being likely to block your use of it
21:45:56 <tdasilva> before using it for bugs
21:46:05 <diablo_rojo> We can run a test migration to our sandbox to see if there are any other issues.
21:46:20 <clayg> new open high prio bugs still seems active on lp -> https://bugs.launchpad.net/monasca/+bug/1636508
21:46:20 <openstack> Launchpad bug 1636508 in Monasca "Smoke test fails and is not under ci or within openstack review process" [High,New]
21:46:31 <acoles> notmyname: yep, merged commit links ought to still link to something on launchpad, future patches use Story tags, and pending patches....we need to remember to update commit messages...manually?
21:46:35 <notmyname> last bugs in LP for monasca have been updated in the past few days
21:46:56 <diablo_rojo> acoles, yes that's accurate I believe
21:46:57 <notmyname> acoles: yes?
21:47:50 <notmyname> ok, I'd like to move on in this meeting, but thank you for the info
21:48:02 <acoles> notmyname: I thought I was agreeing with you...at least on a requirement if not a feature :)
21:48:02 <diablo_rojo> notmyname, thanks for your time :)
21:48:03 <notmyname> we'll look at it
21:48:11 <SotK> indeed, thanks :)
21:48:21 <notmyname> #topic forum topics
21:48:24 <clayg> diablo_rojo: SotK: thanks!
21:48:30 <notmyname> #link http://forumtopics.openstack.org
21:48:40 <notmyname> those are the proposals for the forum in atlanta
21:48:54 <notmyname> this morning I was asked to provide some more commentary on the swift talks
21:48:54 <kota_> boston?
21:49:00 <notmyname> boston
21:49:02 <notmyname> thanks kota_
21:49:47 <notmyname> my understanding is that the forum topics are things that couldn't be discussed elsewhere because of who's attending (ie ops)
21:50:09 <notmyname> and I was asked to provide that justification for the swift sessions
21:50:34 <notmyname> in that, those who are building the schedule and selecting the talks aren't familiar with swift so need some more context
21:50:43 <notmyname> please help me out on this
21:50:54 <notmyname> I think everyone should have the ability to leave comments on the topics
21:51:46 <notmyname> I'll be going through them as well, but if we all help out, we'll get through the swift ones faster
21:52:00 <mattoliverau> kk
21:52:04 <notmyname> the swift ones should all say "swift", but you can also just look for the ones I submitted
21:52:22 <clayg> tdasilva: I think monasca basically started using storyboard and they curently have some bugs tracked on lp and some tracked on storyboard - it's not an unreasonable "let's try this out pattern"
21:52:56 <notmyname> we submitted a lot of stuff that would have been sessions if the ptg/forum split would not have happened. and of course we look for a lot of ops feedback for a lot of these
21:53:22 <mattoliverau> so need to give them a ops spin ;)
21:53:47 <notmyname> but adding more context and more justification for being included will help make our boston summit more productive overall
21:54:04 <notmyname> any questions on that?
21:54:40 <notmyname> ok. here's a pre- "thank you for your help" :-)
21:54:44 <mattoliverau> not yet, will take a look at them and see if I have comments :)
21:54:51 <notmyname> #topic open discussion
21:54:55 <notmyname> next week's meeting
21:55:10 <notmyname> next week I'll be at the openstack leadership training thing in Ann Arbor
21:55:24 <notmyname> (you may have seen it mentioned on the ML)
21:55:32 <notmyname> so I won't be here to chair
21:55:47 <notmyname> is there another volunteer or shall we skip next week's meeting?
21:56:21 <mattoliverau> lets just assume to skip the next one, if something comes up we need to talk about, someone can volunteer :)
21:56:34 <notmyname> heh
21:56:48 <mattoliverau> So everyone check the adjenda to see if its on, but lets assume it's not
21:56:50 <patchbot> matt just wants to sleep late!
21:57:11 <mattoliverau> patchbot:  you can read my mind, what kind of updates have they done to you :P
21:57:16 <acoles> patchbot?
21:57:17 <kota_> lol
21:57:20 * jrichli thinks patchbot has gained a sense of humor
21:57:42 <acoles> or notmyname has gained an alter ego :)
21:58:05 <mattoliverau> all this time, there was no bot, it's just notmyname
21:58:16 <clayg> heh
21:58:18 <jrichli> lol
21:58:18 <notmyname> anything else to bring up this week? 2 min left
21:58:18 <acoles> mattoliverau: lol
21:58:54 <notmyname> ok, /fin
21:59:05 <notmyname> thank you for coming and for your work on swift, both code and community
21:59:13 <notmyname> you are what makes swift awesome
21:59:17 <notmyname> #endmeeting