21:00:53 <notmyname> #startmeeting swift
21:00:53 <openstack> Meeting started Wed Oct  7 21:00:53 2015 UTC and is due to finish in 60 minutes.  The chair is notmyname. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:00:54 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:00:57 <openstack> The meeting name has been set to 'swift'
21:01:01 <notmyname> hello, everyone. who's here for the swift meeting?
21:01:05 <mattoliverau> o/
21:01:06 <wbhuber> o/
21:01:06 <minwoob> o/
21:01:07 <cschwede> o/
21:01:07 <jrichli> o/
21:01:09 <ho> o/
21:01:09 <hrou> o/
21:01:12 <tdasilva> hello
21:01:12 <m_kazuhiro> o\
21:01:18 <kota_> o/
21:01:18 <cutforth> o/
21:01:21 <joeljwright> hey
21:01:22 <peluse> yo
21:01:25 <blmartin> o/
21:01:42 <acoles> hi
21:01:51 <notmyname> welcome
21:01:54 <notmyname> #link https://wiki.openstack.org/wiki/Meetings/Swift
21:02:07 <notmyname> a few things added to the agenda. we'll get to those
21:02:09 <notmyname> but first...
21:02:15 <notmyname> #topic liberty release
21:02:24 <notmyname> good news, everyone! we've got a release!
21:02:30 <joeljwright> yay!
21:02:35 <notmyname> 2.5.0 is tagged
21:02:42 <kota_> great
21:03:08 <notmyname> unless someone finds a bug and comes up with a patch in ...oh... the next 24 hours, that's what we're using for liberty
21:03:24 <notmyname> thank you everyone for working on it
21:04:10 <notmyname> ...sorry, kid just walked in the room ;-)
21:04:23 <kota_> lol
21:04:24 <notmyname> so as a retrospective on this
21:04:32 <notmyname> during the liberty cycle we had 2 releases
21:04:36 <notmyname> that's lower than normal
21:04:48 <notmyname> the first one, 2.4.0, was in development for 4 months
21:04:55 <notmyname> then 2.5.0 was 1 month
21:05:04 <notmyname> so if you thought it felt rushed, it did :-)
21:05:27 <notmyname> lots of things conspired to make it a crunch, but I'll take blame for it.
21:05:52 <notmyname> I'd like to get better at figuring out how to pace things better and avoid a friday night at midnight review session
21:06:07 <notmyname> (or saturday am for some)
21:06:24 <notmyname> so ideas, and all that: more than welcome :-)
21:06:38 <notmyname> any thoughts or questions about the 2.5.0 release?
21:06:39 <mattoliverau> There should be no blame, things happen. All we can do its work together, so nice work all :)
21:08:06 <notmyname> I've intentionally not starred any patches yet since the release. go find something interesting or work on the stuff that needs just one more +2 ;-)
21:08:12 <torgomatic> you need a keyboard macro to type out "it's released when it's ready; time-based releases are awful"
21:08:18 <torgomatic> maybe bind it to a foot pedal for easy repetition
21:08:41 <mattoliverau> lol
21:08:43 <notmyname> (*or* the stuff that needs one more +2)
21:09:16 <notmyname> thanks again. I'm actually pretty happy with what's in it. there's some things we know we need to get done, but I'm really happy we solved or mitigated all of the major stuff
21:09:55 <notmyname> I'd like to spend some time this week on summit planning, but before we get to that, let's go to hrou and robefran for symlinks
21:09:58 <notmyname> #topic symlinks
21:10:09 <hrou> robefran, do the honors ? : )  Thanks notmyname.
21:10:10 <notmyname> robefran: hrou: you added this to the meeting. what's up?
21:10:21 <notmyname> #link https://etherpad.openstack.org/p/swift_symlinks
21:10:34 <robefran> https://review.openstack.org/#/c/232162/
21:10:55 <robefran> just sumbit patch with prelim symlink implementation and functests
21:11:47 <robefran> primary goal is to create func tests to define details of symlink behaviour
21:12:03 <robefran> prelim implementation is just to exercise the func tests
21:12:14 <notmyname> ok
21:12:16 <hrou> yep, more of just an FYI for folks and to share feedback, feel free to look at the etherpad where we have some open design questions (or the code of course).
21:12:21 <robefran> expect lots of changes, refactorization, improvements of the implementation
21:12:34 <notmyname> ok. are there any specific questions you'd like to discuss here today?
21:13:04 <robefran> nothing specific, we'll work through the etherpad, and the gerrit review
21:13:14 <robefran> just wanted to let people know that it is out there
21:13:19 <notmyname> great, thanks
21:13:26 <notmyname> I'm looking forward to it
21:13:36 <notmyname> let's move on then
21:13:40 <notmyname> #topic encryption
21:13:47 <mattoliverau> interesting, thanks for looking into it
21:13:49 <notmyname> jrichli: this one is you. take it away
21:13:53 <jrichli> ok, thx
21:14:13 <jrichli> we are busy working tasks on the trello board
21:14:24 <jrichli> one of them in particular is stuck until we get feedback on direction
21:14:27 <jrichli> Conditional GETs on https://etherpad.openstack.org/p/swift_encryption_issues
21:14:46 <notmyname> #link https://trello.com/b/63l5zQhq/swift-encryption
21:15:10 <notmyname> jrichli: that's a lot of words on the etherpad about conditional gets :-)
21:15:14 <jrichli> please feel free to add comments - leave your handle.  discuss in IRC, whatever.  hrou is heading that effort
21:15:14 <notmyname> what's the summmary?
21:15:21 <jrichli> hrou?
21:15:43 <hrou> Thanks jrichli, yep !  Actually acoles was looking to schedule some time to meet either way, but for other folks if you have any thoughts that'd be great
21:16:10 <jrichli> basically, it proposes changes to the encryption of etag
21:16:50 <jrichli> and this involves maybe deriving keys for the encryption of other things - like body and metadata so that the same key can be used without comprimising security
21:16:59 <jrichli> hrou - anything to add to that summary?
21:17:27 <hrou> notmyname, jrichli nope that's perfect, yea the "lots of words" : )  Is because it changes the crypto scheme somewhat, and more because there was a lot of back and forth about other options.
21:18:04 <acoles> it looked pretty promising when i first read a few weeks back.
21:18:38 <notmyname> sounds like we should all (re)read over that etherpad this week
21:18:59 <jrichli> that would be great.  There are also a couple other etherpads to read:
21:19:11 <notmyname> hrou: is there anything specific you need here today in the meeting?
21:19:15 <notmyname> jrichli: ^
21:19:22 <hrou> notmyname, acoles I wonder if we should cleanup the etherpad to make that "re-reading" easier for others, there's some stuff we can remove really at this point, though if you'd like to see the history ..
21:19:25 <hrou> notmyname, nope!
21:19:43 <acoles> hrou: looking...
21:19:53 <notmyname> hrou: yeah, that might help everyone (ie me) understand what the current thinking is
21:20:03 <mattoliverau> +1
21:20:13 <acoles> hrou: yes!
21:20:13 <notmyname> ok, me and mattoliverau ;-)
21:20:14 <hrou> notmyname, cool, let us crypto folks talk it over I think we can improve it : )  Make life easier on the rest.
21:20:17 <notmyname> thanks
21:20:22 <jrichli> sounds good.  So, there is https://etherpad.openstack.org/p/swift_production_keymaster_issues for Production Keymaster Discussion
21:20:40 <jrichli> since we haven't talked much about what a production keymaster would be like
21:20:53 <jrichli> hopefully we can nail down some decisions
21:21:15 <jrichli> an overview of what IBM has prototyped is at: https://etherpad.openstack.org/p/swift_production_keymaster_overview
21:21:33 <notmyname> great
21:21:50 <notmyname> if in whatever cleaned-up version you have, putting those up at the top would be helpful
21:22:04 <mattoliverau> phew /me got some reading
21:22:18 <jrichli> :-p
21:22:35 <hrou> : )
21:22:35 <jrichli> that's all I got.  thx!
21:22:47 <notmyname> jrichli: hrou: thanks for driving the encryption stuff
21:22:57 <jrichli> yw!
21:23:03 <notmyname> #topic summit prep
21:23:05 <hrou> thanks!
21:23:15 <notmyname> most of us (I think) will be together in tokyo in a few weeks
21:23:29 <notmyname> we haven't done much by way of planning yet
21:23:34 <notmyname> we need to get on that :-)
21:23:37 <notmyname> #link https://etherpad.openstack.org/p/tokyo-summit-swift
21:23:51 <notmyname> there's an etherpad where we can put together ideas of topics to discuss
21:24:07 <notmyname> here's what it's going to look like, I'm told, in the sessions
21:24:32 <notmyname> it will be similar to what we had in the vancouver workrooms, but this time there will be a screen as well
21:25:21 <mattoliverau> \o/
21:25:24 <notmyname> and learning from vancouver, I'd like to group similar topics together so we can be a little more flexible with the time and not have to mentally context switch so quickly
21:25:36 <mattoliverau> great idea
21:25:38 <notmyname> here's the schedule
21:25:40 <notmyname> #link https://docs.google.com/spreadsheets/d/1tpLN5emWhcMmSmkn8z_HuclcjnEPevP77BhdnFN9KCs/pubhtml?gid=5&single=true
21:26:02 <notmyname> we have 2 fishbowl sessions. one of which we'll us for ops feedback
21:26:12 <notmyname> so we'll have one more for a big audience
21:26:37 <notmyname> then we have 12 workroom sessions
21:26:48 <notmyname> and then all day friday as free-form
21:27:14 <notmyname> so basically, we've got all tay wednesday thursday and friday for swift stuff (except for the first 2 slots on thursday
21:27:33 <notmyname> also means that if you've got a talk you're giving or want to see, you'll likely miss something
21:28:07 <notmyname> on the etherpad we've got some ideas right now
21:28:13 <notmyname> or proposals, rather
21:28:24 <notmyname> and I've put down some ideas at the bottom
21:28:42 <notmyname> what I'd like to see is that next week we have enough to build a schedule
21:28:51 <notmyname> ie to sort them
21:29:28 <notmyname> joeljwright: I really want to spend some more time on swiftclient, so I'd love to work with you on a few topics there
21:29:43 <joeljwright> notmyname: unfortunately I can't make it :(
21:29:46 <notmyname> :-(
21:29:50 <notmyname> that is unfortunate
21:30:04 <notmyname> I'd still love your input :-)
21:30:17 <joeljwright> I'll do whatever I can
21:30:42 <mattoliverau> virtual joeljwright?
21:30:52 <notmyname> dfg_: one of my ideas was to talk about the hummingbird sync protocol and what's possible there to unify things and make everything better
21:31:22 <joeljwright> mattoliverau: :)
21:31:30 <notmyname> we should probably have some EC summary topics
21:32:41 <notmyname> if you have an idea, but you htink it might need more than one time slot, indicate that so we can schedule properly
21:32:55 <mattoliverau> I'm hoping to start some early sharding POC #3 benchmarking this/next week, so hopefuly that will go well. in any case there _should_ be something to talk about.
21:32:56 <notmyname> any thoughts or questions? will you be able to add some ideas to the list this week?
21:33:02 <dfg_> notmyname: oh ok. we'd prob have enough material to talk about if you wanted
21:33:02 <notmyname> mattoliverau: good!
21:33:16 <notmyname> dfg_: cool
21:33:45 <notmyname> basically, you know how we do the "what are we going to talk about" at the beginning of the hackathons? that's what we've got to do on the etherpad this week
21:34:04 <tdasilva> do we get stickers?
21:34:12 <mattoliverau> lol
21:34:28 <notmyname> no stickers. but there's a tshirt
21:34:35 <tdasilva> +1
21:34:36 <notmyname> http://d.not.mn/tshirt_hint.jpg
21:35:09 <dfg_> well i think some of the things we've seen with hummignbird have been cool and may be interesting even if you weren't running it. so if there's room for it then we could do one
21:35:09 <acoles> notmyname: c'mon the hint should be bigger each week :P
21:35:17 <notmyname> :-)
21:35:23 <jrichli> acoles +1
21:35:28 <notmyname> dfg_: yeah!
21:35:33 <jrichli> acoles: you have already seen it!
21:35:34 <mattoliverau> how mysterious and exciting :)
21:35:59 <notmyname> dfg_: the whole idea of "make sure we're all talking the same language on the backend" is a good one. that includes the replication protocol
21:36:26 <notmyname> dfg_: I'm excited about what' you've found and how it's working for you :-)
21:36:59 <notmyname> anything else from anyone on summit planning or prep?
21:38:27 <notmyname> ok then :-)
21:38:44 <notmyname> have a good week. let's talk about ranking the summit topics next week
21:38:48 <notmyname> thanks for working on swift
21:38:53 <notmyname> #endmeeting