20:59:43 <notmyname> #startmeeting swift
20:59:44 <openstack> Meeting started Wed Aug  5 20:59:43 2015 UTC and is due to finish in 60 minutes.  The chair is notmyname. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:59:45 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
20:59:47 <openstack> The meeting name has been set to 'swift'
20:59:53 <notmyname> who's here for the swift meeting?
20:59:54 <wbhuber> \o/
20:59:55 <minwoob_> o/
20:59:57 <jrichli> hello
20:59:57 <hurricanerix> o/
20:59:58 <mattoliverau> o/
20:59:59 <ho> hi
20:59:59 <MooingLemur> 🐄
21:00:00 <acoles> here
21:00:06 <torgomatic> .
21:00:15 <ajiang> hi
21:00:22 <notmyname> MooingLemur: do you have lemur emoji too?
21:00:37 <MooingLemur> If there were one, I'd use it all the time.
21:00:43 <zaitcev> :lemur:
21:01:18 <aerwin> o/
21:01:18 <notmyname> not a very big agenda this week
21:01:20 <notmyname> #link https://wiki.openstack.org/wiki/Meetings/Swift
21:01:27 <kota_> hi
21:01:39 <notmyname> first up, hackathon next week
21:01:43 <notmyname> #topic hackathon
21:01:49 <notmyname> next week in austin
21:01:55 <notmyname> also, no IRC meeting next week
21:02:09 <notmyname> #link https://etherpad.openstack.org/p/swift-midcycle-aug-2015
21:02:20 <notmyname> that's a place if you want to collect some ideas before we get together
21:02:30 <mattoliverau> yay, hackathon
21:02:33 <notmyname> but we will start off as normal building a list of topics to talk about
21:02:43 <notmyname> mattoliverau: yay seeing everyone :-)
21:02:51 <mattoliverau> +100
21:03:03 <notmyname> so...any questions about next week? any prep needed?
21:03:22 <jrichli> I had a couple things that might help out
21:03:35 <jrichli> there is a map - the link is in the invite, but here it is:
21:03:38 <jrichli> map: https://www.google.com/maps/d/edit?mid=zyXNIjVDS-Us.kvQV8EWN67aU&usp=sharing
21:04:04 <jrichli> Look at it using the satellite view
21:04:17 <jrichli> The map shows you a break in the trees where you can walk from the hotel to IBM
21:04:27 <jrichli> and it shows you where to enter.
21:04:35 <notmyname> nice map
21:04:41 <jrichli> thx!
21:04:47 <jrichli> You will need to get a badge in the front and wait to be escorted to the location, unless you work for IBM in the US.
21:04:57 <tdasilva> cool! but don't see option for satellite view :(
21:05:05 <notmyname> tdasilva: down in the bottom left?
21:05:06 <jrichli> bottom left?
21:05:27 <jrichli> If you work for IBM in the US, then you can use your badge to enter building 908, where you will see signs posted about where to go.
21:05:30 <tdasilva> ah, firefox worked
21:06:12 <mattoliverau> cool, thanks jrichli
21:06:15 <notmyname> jrichli: anything else for us for next week?
21:06:31 <jrichli> i guess that is all.
21:06:48 <notmyname> great! I'm looking forward to it
21:06:58 <jrichli> me too
21:06:58 <notmyname> #topic other general stuff
21:07:14 <notmyname> the week after the swift hackathon is the ops midcycle
21:07:20 <notmyname> #link https://www.eventbrite.com/e/openstack-ops-mid-cycle-meetup-tickets-17703258924
21:07:45 <notmyname> I'll be there, so if you (or a coworker) is going, find me and let's make sure swift is meeting the needs of ops people
21:08:02 <notmyname> also, something came up with defcore at the end of last week
21:08:15 <notmyname> I want to bring it up (mention it) here, and then probably talk more about it next week
21:08:50 <notmyname> one thing we need to look in to is using testr (test repository) as the test runner in swift. it's what we already use in python-swiftclient
21:09:29 <notmyname> if we are able to do that, it means that the tests we maintain as part of swift are more likely to be used by defcore to measure what is allowed to be "openstack object storage"
21:09:36 <notmyname> which is a good thing
21:10:02 <notmyname> any other general announcements from anyone before we get into reviewing ongoing work?
21:10:49 <notmyname> ok :-)
21:11:05 <notmyname> #topic ongoing work (versioning middleware)
21:11:15 <tdasilva> yay! it's getting there
21:11:17 <notmyname> this is a big one because it's been going on a while is is blocking work
21:11:20 <notmyname> tdasilva: yay
21:11:23 <notmyname> tdasilva: what's the status?
21:11:30 <tdasilva> thanks to kota for finding bugs and helping fix them
21:11:34 <tdasilva> kota_: ^
21:11:45 <kota_> tdasilva: you're welcome
21:11:48 <notmyname> nice
21:12:00 <notmyname> #link https://review.openstack.org/#/c/134347/
21:12:06 <tdasilva> notmyname: i submitted a new patchset a while back, so it is ready for review
21:12:11 <notmyname> ok
21:12:20 <kota_> and now i'm reviewing
21:12:20 <notmyname> can we get some volunteers to look at it asap?
21:12:23 <notmyname> kota_: thanks
21:12:39 <mattoliverau> I'll look too
21:12:42 <notmyname> thanks mattoliverau
21:12:46 <tdasilva> thanks guys!
21:13:06 <notmyname> I'd love to have this merged this week so we can move on the copy middleware and unblock encryption (and the other cleanup work)
21:13:27 <tdasilva> yes, me too
21:13:28 <notmyname> tdasilva: anything else on this patch?
21:13:51 <tdasilva> no, please review and try not to get confused with the backwards compatibility code :P
21:13:58 <notmyname> thanks
21:14:00 <notmyname> #topic ongoing work (encryption)
21:14:08 <notmyname> jrichli: acoles: what's going on here?
21:14:16 <notmyname> #link https://etherpad.openstack.org/p/swift_encryption_issues
21:14:28 <jrichli> acoles has been working hard reviewing the patches.  thanks acoles!
21:14:38 <notmyname> great
21:14:49 <jrichli> I have re-work to do over the next few days.
21:14:52 <notmyname> jrichli: how are you on that etherpad and the issues raised there?
21:14:59 <acoles> also we have some topics to work through next week
21:15:09 <acoles> yes, the etherpad!
21:15:29 <jrichli> acoles and I talked a little about the issues, and mahatic added  notes about probetests.  but not a lot of movement there.
21:15:37 <notmyname> ok
21:15:53 <acoles> looks like conditional gets we have a way forward on etherpad
21:16:02 <notmyname> jrichli: acoles: is there something specific you'd like to see before the end of the week before the hackathon?
21:16:21 <jrichli> not on my end
21:16:31 <acoles> notmyname: i'd like to see myself complete review of all jrichli good work ;)
21:16:36 <notmyname> :-) ok
21:17:02 <notmyname> I think there should be great progress next week on encryption :-)
21:17:15 <jrichli> that would be good
21:17:16 <notmyname> #topic ongoing work (hummingbird)
21:17:27 <notmyname> redbo: dfg: hurricanerix: what's happening on your side of things?
21:17:50 <dfg> we're mostly working on replication
21:18:09 <hurricanerix> notmyname: I started looking over that testr stuff, which isn't the most fun thing ;)
21:18:15 <notmyname> hurricanerix: thanks
21:18:22 <notmyname> dfg: interesting. ok
21:18:45 <notmyname> do you need anything before next week?
21:18:54 <dfg> what we have is going really well so far but still tweaking
21:18:59 <dfg> notmyname: i don't think so
21:19:02 <notmyname> ok
21:19:28 <notmyname> have you had any opportunity to gather some numbers about it?
21:19:37 <notmyname> benchmark numbers
21:20:04 <dfg> um. i think we're still looking into that. hopefully we'll have something by next week
21:20:09 <notmyname> ok, great
21:20:11 <notmyname> thanks
21:20:24 <notmyname> anything else on that front for this meeting today?
21:20:39 <dfg> i added a section for direct to object tests for the hackathon
21:20:46 <dfg> in the therpad
21:20:58 <dfg> but other than that i don't thin so
21:21:04 <notmyname> yeah, I saw that. sounds great (and I agree with them being probe tests)!
21:21:16 <dfg> ya
21:21:26 <notmyname> #topic ongoing work (other)
21:21:41 <notmyname> what else are you (anyone) working on that you'd like to give an update on today?
21:22:12 <ho> I have one question. In OpenStack Installation guide (Kilo), default value for allow_account_management was changed from TRUE to FALSE. do you know the reason?
21:22:32 <notmyname> IIRC it was always false in our example configs
21:23:00 <notmyname> ho: false is a more secure option, so that's why our sample config have it that way
21:23:01 <ho> in keystone env, the value should be TRUE to have an ability to delete account.
21:23:17 <ho> i think
21:23:33 <mattoliverau> I'm debugging an initial version of shrinking in container sharding, hopeing to have it done by hackathon so we can talk about it.. and maybe some benchmarks.. but getting sick got in the way of that.. so crossing fingers I get some in time.
21:23:37 <notmyname> ho: many deployments might have only one (or a few) proxy servers with that on and those are firewalled off to only allow access from an auth system
21:23:53 <notmyname> mattoliverau: great
21:24:49 <notmyname> joeljwright: what's going on in swiftclient? anything that needs to be brought up?
21:25:15 <joeljwright> I've been digging through the bug backlog
21:25:23 <notmyname> great! thanks
21:25:25 <joeljwright> have closed a couple, confirmed a couple
21:25:29 <notmyname> (everyone should do that!)
21:25:32 <joeljwright> but I'm going to need help
21:25:44 <joeljwright> there are also quite a few patches that need some reviews
21:25:53 <ho> notmyname: thanks! in swift we have a lot of garbage. i will check it.
21:26:06 <wbhuber> joeljwright: beep me the ones you'd like to have my help with.
21:26:07 <joeljwright> the no negative feedback dashboard could do with looking at
21:26:18 <notmyname> https://review.openstack.org/#/q/status:open+project:openstack/python-swiftclient,n,z
21:26:26 <joeljwright> wbhuber: thanks
21:26:33 <joeljwright> really, I think there are plenty
21:26:43 <joeljwright> I'm assigning a couple to myself now
21:26:50 <joeljwright> but feel free to pitch in and grab some :)
21:27:35 <notmyname> https://goo.gl/uO4b7l are "new" bugs for swiftclient
21:27:51 <notmyname> bugs in a "new" status need to be looked at and confirmed, closed, or get more info
21:27:52 <joeljwright> I'll probably ping IRC next week while I know you guys are hacking to try and get some reviews...
21:27:59 <notmyname> joeljwright: great
21:28:04 <joeljwright> some of those are up to 4 years old!
21:28:48 <notmyname> yup :-(
21:28:51 <kota_> wow
21:29:55 <notmyname> next (last) up
21:29:56 <notmyname> #topic open discussion
21:30:18 <notmyname> is there anything else that needs to be brought up this week? patches? bugs? specs?
21:30:24 <joeljwright> who is the best person to talk to about the object expirer?
21:30:40 <notmyname> joeljwright: this is a great place to ask :-)
21:30:44 <dfg> oh- ya. we want to rewrite that thing.
21:30:53 <notmyname> dfg: are you working ont hat?
21:30:55 <joeljwright> one of the bugs reported relatively recently against the swiftclient is regarding x-delete-at and x-delete-after and segmented objects
21:30:56 <notmyname> *aren't
21:31:15 <MooingLemur> I'm having to split the load and run it on 4 hosts in our environment. :P
21:31:17 <dfg> um- we were starting to but one of our guys got pulled off. meant to put it on the hackthon todo
21:31:18 <MooingLemur> (object-expirer)
21:31:50 <joeljwright> this hits me too, so I'd be interested in helping with a patch, but it seems like the wrong solution to have a POST set an expiry header on all segemnts of an object…
21:32:10 <notmyname> dfg: ah, I was thinking for the reaper patch you just updated
21:32:50 <dfg> no. the obj expirere would be a pretty big change. i just forgot to add it
21:33:13 <dfg> to the hackathon thing
21:33:33 <notmyname> dfg: can you put it on the etherpad, please?
21:33:48 <dfg> anywya- sorry it wouldn't be any change to how swiftclient uses it. that just reminded me :)
21:34:20 <tdasilva> joeljwright: what's the bug number?
21:34:24 <joeljwright> well that was the point of asking - I think we need a discussion to work out what the 'right' solution here is
21:34:35 <joeljwright> https://bugs.launchpad.net/python-swiftclient/+bug/1478830
21:34:35 <openstack> Launchpad bug 1478830 in python-swiftclient "x-delete-at and x-delete-after do not apply to segments of large objects" [Undecided,New]
21:34:35 <uvirtbot> Launchpad bug 1478830 in python-swiftclient "x-delete-at and x-delete-after do not apply to segments of large objects" [Undecided,New]
21:34:36 <uvirtbot> Launchpad bug 1478830 in python-swiftclient "x-delete-at and x-delete-after do not apply to segments of large objects" [Undecided,New] https://launchpad.net/bugs/1478830
21:34:40 <joeljwright> reported against 2.5.0 so, recent
21:35:53 <notmyname> joeljwright: I agree (with needing a discussion)
21:35:58 <wbhuber> kota_: notmyname: can you elaborate on the global ec cluster improvement feature?
21:36:23 <kota_> wbhuber: ah yes
21:36:24 <wbhuber> notmyname: by looking at the etherpad, it'll be one topic to discuss at hackthon, right?
21:36:27 <notmyname> looks like kota_ just proposed a spec
21:36:35 <kota_> I added just yesterday
21:36:41 <MooingLemur> I was looking at that
21:36:43 <notmyname> wbhuber: the etherpad is goign to seed the topic list we create next week
21:36:51 <kota_> and pushed a new spec
21:36:58 <kota_> into swift-spec
21:37:05 <wbhuber> kota_: good we have ideas to feed to that improvement
21:37:15 <zaitcev> Oh god. Every day I keep postponing looking at specs.
21:37:22 <kota_> wbhuber: thanks
21:37:24 <notmyname> joeljwright: would you prefer to start/keep the discussion in the bug comments for now, or do you think we should push it to the mailing list, or schedule some meeting time, or somethign else?
21:37:40 <kota_> gerrit is here, https://review.openstack.org/#/c/209447/
21:37:53 <kota_> I wanna talk about it in detail in hackathon.
21:37:55 <wbhuber> kota_: just added myself a reviewer. =]
21:37:56 <joeljwright> I think the comments are ok for now
21:38:07 <kota_> wbhuber: nice!
21:38:15 <joeljwright> notmyname: we can always move the discussion to proposed patches or IRC later
21:38:17 <MooingLemur> the multiregion-EC would have rings as wide as the total replicas still, right?
21:38:23 <minwoob_> kota_: I'll add myself as well.
21:38:24 <notmyname> joeljwright: ok, great. thanks for bringing it up :-)
21:38:34 <MooingLemur> including all the extra ones you'd want for geographic dispersion
21:38:41 <kota_> minwoob_: :)
21:39:22 <wbhuber> kota_: i imagine both peluse and clayg would like to partake this spec eventually
21:39:45 <notmyname> oh yeah. there will be a lot of eyes on it, I'm sure ;-)
21:40:02 <notmyname> anything else to bring up this week, or shall we adjourn early?
21:40:03 <kota_> wbhuber: I hope so, let's work together :)
21:40:39 <wbhuber> kota_: you got it.
21:41:33 <notmyname> ok, that looks like it
21:41:38 <notmyname> thanks for coming and participating today.
21:41:42 <notmyname> and thanks for working on swift
21:41:44 <notmyname> #endmeeting