21:00:02 <notmyname> #startmeeting swift
21:00:03 <openstack> Meeting started Wed Jul 19 21:00:02 2017 UTC and is due to finish in 60 minutes.  The chair is notmyname. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:00:04 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:00:06 <openstack> The meeting name has been set to 'swift'
21:00:10 <notmyname> who's here for the swift team meeting?
21:00:24 <kota_> hi
21:00:25 <m_kazuhiro> o/
21:00:26 <jrichli> o/
21:00:26 <mathiasb> o/
21:00:32 <tdasilva> o/
21:00:48 <timburke> o/
21:00:49 <torgomatic> hi
21:00:59 <acoles> hello
21:01:09 <notmyname> welcome everyone
21:01:14 <notmyname> meeting agenda is at
21:01:15 <notmyname> #link https://wiki.openstack.org/wiki/Meetings/Swift
21:01:28 <notmyname> #topic denver ptg
21:01:34 <notmyname> reminder that the PTG is coming up
21:01:44 <notmyname> you should have already registered for your ticket and hotel :-)
21:01:54 <notmyname> #link https://etherpad.openstack.org/p/swift-ptg-queens
21:02:08 <notmyname> that's the topic etherpad where we're collecting the things we want to discuss in person
21:02:48 <notmyname> please encourage everyone to be at the PTG, especially ops. getting the devs and ops together is crucial
21:03:35 <notmyname> if you're still planning your week, monday and tuesday will be for cross project sessions (scheduled) and any ad-hoc conversations we may have. wed-fri will have a reserved room for swift
21:03:37 <mattoliverau> o/ (sorry for being late)
21:03:55 <notmyname> are there any questions about the PTG that we can answer in here or that I can hunt down for you?
21:04:54 <notmyname> great! no questions
21:05:09 <notmyname> (please feel free to ask at any time in -swift if you do have questions)
21:05:16 <notmyname> #topic bug triage work
21:05:40 <notmyname> so we're one week in to this (or a little longer, depending on how you count). how are we doing?
21:05:47 <notmyname> #link https://etherpad.openstack.org/p/swift-bug-triage-list
21:06:24 <notmyname> I see that tdasilva had put his name on several and they're mostly crossed off now. great job!
21:07:01 <acoles> looks like tdasilva is the champ!
21:07:51 <tdasilva> honestly, i've been picking the low hanging fruit just to make progress
21:07:55 <tdasilva> cross stuff out
21:07:56 <notmyname> heh
21:08:03 <tdasilva> hopefully make the list more manageable
21:08:05 <notmyname> see! go do stuff now, and you get all the easy ones!
21:08:32 <tdasilva> i'd like to think there's lots of that
21:08:35 <notmyname> everyone else should go do bug triage today so tdasilva doesn't get all the low-hanging-fruit :-)
21:09:31 <notmyname> please put your names on bugs, go triage them, and cross them off after updating the status in LP (see the notes at the top of the etherpad)
21:09:41 <notmyname> (I'm speaking to myself, too)
21:10:19 <notmyname> (maybe literally since it doesn't seem like anyone else is talking in here) ;-)
21:10:47 <tdasilva> heh...I was playing with tags a big, but would like to get feedback and see what people think
21:11:09 <tdasilva> I created a few that maybe in the future can help us just to see where most bugs go?
21:11:34 <timburke> i wish the graphs on http://status.openstack.org/bugday/ went back more than 3 days...
21:11:35 <notmyname> instead of trying to evaluate a bug based on the title listed (to decide if you're going to sign up for it), just sign up and then take a look :-)
21:11:38 <notmyname> tdasilva: those look great
21:11:46 <acoles> tdasilva: I think the service tags could be useful
21:12:01 <notmyname> tdasilva: like a "bugweek" or something?! crazytalk
21:13:01 <tdasilva> notmyname: acoles and I thought of the same idea of maybe utilizing the first two days at PTG
21:13:04 <tdasilva> if time allows
21:13:13 <notmyname> that's a great idea
21:13:17 <timburke> i'm in :-)
21:13:41 <tdasilva> if we plan it maybe even folks not going to PTG can participate
21:14:12 <notmyname> what would be in the plan?
21:14:42 * jungleboyj walks in late
21:14:44 <tdasilva> I guess to start is to announce that we are going going to do that ;)
21:15:11 <tdasilva> then maybe get everybody to sign up for some number of bugs...idk never done this before... :D
21:15:18 <notmyname> :-)
21:15:47 <acoles> we'll need to be flexible to allow us to attend useful stuff that may be happening on those days, but I like the idea
21:15:57 <tdasilva> yep!
21:15:58 <notmyname> ok, I can imagine what that looks like (the announcement and organization)
21:16:02 <notmyname> yeah
21:17:08 <notmyname> I think we need to add a little bit of description to the top of the bug etherpad for people to come up to speed without having to have followed these meetings and the other IRC conversations
21:17:12 <notmyname> I can work on that
21:17:55 <tdasilva> sounds good
21:18:12 <notmyname> is there anything more that needs to be brought up in this meeting about bug triage work?
21:18:35 <tdasilva> thanks to Christian for getting that list up
21:18:41 <acoles> +1 ^^
21:18:49 <notmyname> yes! his work has been invaluable
21:19:14 <notmyname> #topic upcoming releases, priority reviews
21:19:34 <notmyname> as I've talked about recently, I'd like to do releases asap
21:19:39 <notmyname> #link https://wiki.openstack.org/wiki/Swift/PriorityReviews
21:19:52 <notmyname> the priority reviews page has some important things listed
21:20:14 <notmyname> https://review.openstack.org/#/c/448480 and https://review.openstack.org/#/c/478416/ and https://review.openstack.org/#/c/475038/ need reviews
21:20:15 <patchbot> patch 448480 - swift - DB replicator cleanup
21:20:16 <patchbot> patch 478416 - swift - Add multiple worker processes strategy to reconstr...
21:20:17 <patchbot> patch 475038 - python-swiftclient - Allow for uploads from standard input.
21:20:45 <notmyname> I believe timburke and acoles are looking at the ec reconstructor one (right?)
21:20:59 <acoles> notmyname: yes
21:21:26 <notmyname> one question I have is about acoles's composite replicated rings patch. where are we with it? could it get in? should it be added to the list?
21:22:00 <acoles> yo mean https://review.openstack.org/#/c/477000
21:22:01 <patchbot> patch 477000 - swift - Ring rebalance respects co-builders' last_part_moves
21:22:02 <notmyname> the stdin swiftclient patch is something that will really help a lot of use cases
21:22:31 <notmyname> acoles: maybe I do
21:22:52 <acoles> notmyname: I'd love to have 477000 in next release, kota_ has been looking at it and i did have clayg's +2
21:23:10 <mattoliverau> I've been looking at the db replication one, not as much as I'd like.. just want to test on a my SAIO a little. Bec comes back soon so will have more time, but if others want to look feel free :)
21:23:39 <notmyname> thanks mattoliverau
21:23:53 <kota_> it looks acoles update the patch since i reviewed, so i should circle back to there
21:23:55 <notmyname> I'd like to have a volunteer for the stdin swiftclient patch
21:24:04 <notmyname> kota_: thanks
21:24:42 <acoles> kota_: thanks, I'll try to catch up with you at end of your day if you have any queries
21:26:00 <notmyname> can someone volunteer to review https://review.openstack.org/#/c/475038/ please?
21:26:00 <patchbot> patch 475038 - python-swiftclient - Allow for uploads from standard input.
21:26:51 <jrichli> I'll try to carve out some time to take a look this week
21:26:55 <tdasilva> o/
21:27:20 <notmyname> thanks jrichli and tdasilva
21:27:39 <notmyname> that's the last patch needed for a swiftclient release (aside from the boilerplate authors/changelog update)
21:28:24 <notmyname> Ithink that covers the stuff on priority reviews
21:28:36 <notmyname> any questions/comments/concerns over upcoming releases?
21:29:16 <notmyname> note that the docs migration stuff is still in progress (most of the big stuff is done, but there's a little more to do yet)
21:29:42 <notmyname> oh, the swiftclient docs migration has a little bit to do, IIRC. that should be done before a release. i'll work on that
21:30:07 <notmyname> #topic barbica/castellan dependency
21:30:15 <notmyname> tdasilva: this is your topic. what's up?
21:30:23 <notmyname> https://review.openstack.org/#/c/364878/28
21:30:23 <patchbot> patch 364878 - swift - Retrieve encryption root secret from Barbican
21:30:52 <tdasilva> so i've been reviewing mathiasb patch and noticed that it depends on castellan, but the patch does not update requirements.txt
21:31:25 <tdasilva> talked to mathiasb a bit and it seems that is still up in the air on whether we want to add that dependency or not
21:31:42 <tdasilva> so I just wanted to hear from folks what they think are the pros and cons
21:32:07 <notmyname> what if we added the functionality without the requirements update (for now) and added the requirements update later, in part depending on it's usage
21:32:14 <notmyname> (yes I know there's a catch-22 there)
21:32:52 <acoles> tdasilva: is it similar scenario to keystonemiddleware which is also not in requirements.txt?
21:32:54 <jrichli> I think I had given mathiasb the advice to shy away from the dependency.  so as not to require that unless you used that middleware
21:33:03 <timburke> add it to [extras] in setup.cfg?
21:33:25 <tdasilva> timburke: hmm...interesting, had not heard of that before
21:33:29 <timburke> acoles: there's a slight difference, in that we don't *import* from keystonemiddleware
21:33:46 <tdasilva> and that's what i was thinking, how do we at least point users to these other dependencies that they might need
21:33:52 <tdasilva> extra-requirements.txt ??
21:33:58 <mathiasb> we've discussed it in person and in the patch comments, and imho the consensus has been not to add the dependencies at this time in requirements.txt
21:34:02 <timburke> then you can do things like `pip install swift[castellan]` or w/e
21:34:43 <timburke> we currently do something like that for swiftclient: https://github.com/openstack/python-swiftclient/blob/master/setup.cfg#L35-L37
21:35:14 <notmyname> I'm not very familiar with that mechanic, but it seems like a simple solution
21:35:28 <notmyname> mathiasb: tdasilva: what do you think?
21:35:30 <tdasilva> yep, then we can just document it
21:35:31 <timburke> https://www.python.org/dev/peps/pep-0426/#extras-optional-dependencies
21:35:38 <acoles> timburke: that's interesting, had not noticed that mechanism before
21:35:59 <mathiasb> i'm not familiat with that either, but i can definitely look into it
21:36:10 <notmyname> cool! a way forward!
21:36:25 <mathiasb> we do mention the requirements in the docs
21:36:41 <notmyname> tdasilva: mathiasb: were there other things to bring up on this topic here in the meeting?
21:37:01 <tdasilva> also, the patch already has a +2 and I've been reviewing and think it looks pretty good, but it would be nice to get other folks to review before I give my +2
21:37:13 <tdasilva> especially folks that have been playing with encryption before :D
21:37:41 <notmyname> jrichli: would you have any time to look?
21:38:00 <tdasilva> maybe we can have barbican support in Pike! :D
21:38:31 <notmyname> that would be really great!
21:38:34 <jrichli> notmyname: its on my list to review - i had started, but got interupted
21:38:42 <notmyname> thanks
21:39:02 <notmyname> tdasilva: but please don't let jrichli's review hold up your review. go ahead and put your vote down too :-)
21:39:32 <tdasilva> ok, i think we will need another patchset, but mostly around 'smaller' items like we just discussed and docs
21:39:34 <acoles> last time I looked it seemed in good shape
21:39:47 <tdasilva> code is very small
21:40:11 <notmyname> :-)
21:40:40 <notmyname> anything else on this patch?
21:40:50 <mathiasb> there's a related barbicanclient bug that tdasilva discovered that people actually running the code might bump into
21:41:05 <notmyname> should it block this patch?
21:41:13 <tdasilva> blocks testing :D
21:41:26 <notmyname> hmmm. that's not good
21:41:35 <mathiasb> tdasilva: do you have the link? typing on my phone atm :/
21:41:43 <tdasilva> yeah, you kinda have to go manually
21:41:44 <tdasilva> small change
21:41:45 <tdasilva> https://review.openstack.org/#/c/484926/2
21:41:46 <patchbot> patch 484926 - python-barbicanclient - fix default version
21:42:20 <tdasilva> i think barbican client had a big refactor a few days ago and it introduced a small regression
21:42:27 <notmyname> should the swift patch have a depends-on?
21:42:42 <tdasilva> the patch itself doesn't depend on it
21:42:54 <notmyname> oh. yeah, I was noticing that the gate passes
21:42:56 <tdasilva> it's mathiasb script to set the env up that depends on it
21:43:39 <tdasilva> btw: mathiasb script and Vagrantfile are very useful, thank you mathiasb!
21:44:12 <mathiasb> yw, anything to make testing and reviewing easier!
21:44:38 <jrichli> +1
21:45:03 <notmyname> do you have any concerns about landing our patch before the barbican client patch?
21:45:33 <notmyname> or we should land ours and encourage them to land theirs and everything is good?
21:45:38 <notmyname> (ie no defined order)
21:45:55 <tdasilva> notmyname: no, but honestly i don't think it will happend. the barbican patch is small and I'd say "high priority" for them. client is broken, can't use it
21:46:03 <notmyname> ok
21:46:12 <notmyname> so it's likely theirs will land first anyway :-)
21:46:26 <tdasilva> i *think* so
21:46:39 <jrichli> but i guess that means we can't successfully run mathias's script currently?
21:46:53 <notmyname> jrichli: you hear that? you're now in a race with the barbican team. go, go, go!
21:47:10 <jrichli> lol
21:47:25 <notmyname> ok, let' move on
21:47:30 <tdasilva> jrichli: no, it fails at the very end when setting up the barbican secret, but it is not hard to fix the client manually and then create the secret manually
21:47:30 <notmyname> #topic open discussion
21:47:40 <notmyname> I was just looking at the openstack release dates. looks like we need a swift release by the week of august 21. that's not too far away. in part meaning that if we cut a swift release now, we'll have a fast-follow for another right before pike
21:48:01 <notmyname> also that when we cut swiftclient, that will very very likely be what's in pike
21:48:17 <notmyname> does anyone else have anything to bring up in today's meeting?
21:48:48 <notmyname> reminder that the 0700 meeting is happening next week. thanks kota_ for chairing
21:48:56 <kota_> o/
21:49:06 <kota_> will do my best
21:49:15 <notmyname> then it will be fantastic
21:50:11 <notmyname> just as an FYI, I'll be out for the first two weeks of august on a family trip. I'll have limited time online
21:50:27 <notmyname> thank you for your work on swift this week!
21:50:34 <notmyname> thanks for coming to the meeting
21:50:37 <notmyname> #endmeeting