21:04:29 <notmyname> #startmeeting swift
21:04:30 <openstack> Meeting started Wed Mar  9 21:04:29 2016 UTC and is due to finish in 60 minutes.  The chair is notmyname. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:04:31 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:04:33 <openstack> The meeting name has been set to 'swift'
21:04:37 <notmyname> swift team meeting time. who's here?
21:04:39 <mattoliverau> o/
21:04:41 <pdardeau> o/
21:04:44 <cutforth> o/
21:04:46 <jrichli> o/
21:04:47 <dmorita> o/
21:04:47 <ho_away> o/
21:04:48 <jlhinson> o/
21:04:48 <tdasilva> helo!
21:04:51 <minwoob> o/
21:04:54 <kota_> \o/
21:04:56 <awelleck> o/
21:05:01 <acoles> hi
21:05:13 <gmmaha> o/
21:05:14 <sgundur> hi
21:05:24 <notmyname> great to see everyone
21:05:41 <notmyname> I want to keep this week's meeting fairly short
21:05:54 <notmyname> #topic hackathon summary
21:06:00 <notmyname> last week was the hackathon in bristol
21:06:13 <notmyname> acoles: thank you, again, for hosting and all the work you put in
21:06:17 <notmyname> it was a fantastic week
21:06:24 <jrichli> +1
21:06:32 <ho_away> +1
21:06:41 <kota_> so great
21:06:49 <dmorita> It was nice to see you guys
21:06:57 <acoles> thanks, we were glad to host you all
21:06:58 <pdardeau> awesome work acoles!
21:07:07 <ntata> o/
21:07:15 <notmyname> I'd like to spend most of the meeting tim on this topic to hear from everyone (and share with those who weren't there)
21:07:29 <notmyname> what did you like? what can we improve? what are you happy got done? etc
21:07:33 <siva_krishnan> o/
21:08:19 <notmyname> personally, I found huge benefit in the conversations about the community itself. I've got several things to work on about how to track things, what to track, what *not* to track, etc. thank you for all your input
21:08:33 <notmyname> I'm also really happy that fast-post and the ops runbook docs landed
21:08:45 <mattoliverau> free baby sitting or birthing facilities for those of us who were home-ridden just in case :P
21:08:58 <mattoliverau> +1--
21:09:00 <mattoliverau> +100
21:09:02 <notmyname> mattoliverau: we'll work on that for next time :-)
21:09:10 <acoles> we missed mattoliverau
21:09:27 <siva_krishnan> :-D
21:09:50 <mmotiani> \o/
21:09:58 <acoles> notmyname: agree, discussions about community and process were good
21:10:03 <notmyname> what was another thing that happened last week that you enjoyed or found benefit in?
21:10:38 <acoles> also good to get consensus on container sync patches
21:10:59 <notmyname> yep
21:11:12 <pdardeau> as newcomer to community, insight/learning that's not possible via irc
21:11:25 <notmyname> I'll be bugging people about those in the days to come to see if we can get the container stuff into the next release
21:11:30 <pdardeau> and meeting people in person!
21:11:38 <jrichli> seems like we planned out the topics pretty well.  maybe the last couple days were harder to know what was being discussed when
21:11:58 * mattoliverau is looking forward to some summaries cause it sounds like he missed alot.
21:12:11 <kota_> jirichli: exactly
21:12:30 <notmyname> yeah, it did feel like a bunch of topics got piled intot he last day
21:13:14 <notmyname> mattoliverau: mostly we were in a room and talked. then we all went to a different room and talked and ate. then we went to a different room and talked and had beer
21:13:29 <jrichli> dont forget the bus rides!
21:13:35 <jrichli> good conversations there
21:13:41 <notmyname> right
21:13:46 <acoles> mattoliverau: and rarely we went to individual rooms and slept :)
21:14:00 <mattoliverau> notmyname: thanks :P Now I'm up to speed :P
21:14:10 <acoles> oh yeah, the bus could have been better :/
21:14:24 <mattoliverau> acoles: lol
21:14:28 <notmyname> so acoles provided us a bus to and from the HPE office. and we all rode it together instead of everyone coming and going at different times
21:14:48 <notmyname> the bus had the unintentional side effect of fostering a lot of good conversations it seems
21:14:50 <jrichli> acoles got a scare one day - was gonna order a bunch of taxis
21:15:10 <acoles> single replica bus is not robust
21:15:17 <mattoliverau> lol
21:15:30 <notmyname> here's one summary thing that happend
21:15:35 <mattoliverau> multiple taxis are eventually consistant :P
21:15:37 <notmyname> we spent a lot of time talking about long review times
21:15:42 <acoles> mattoliverau: lol
21:16:16 <notmyname> and the result is (1) prioritization of reviews is very imported instead of just number of patches reviewed. and (2) *good* reviews are very important to encourage
21:16:16 <mattoliverau> notmyname: so a long review on long reviews?
21:17:18 <acoles> did clayg write something up about good reviews? i saw something, on github maybe, is it on gerrit?
21:17:27 <notmyname> also another think I'll try to facilitate going forward are some different ways to communicate with each other like voice and/or video chat. probably around a specific topic
21:17:30 <notmyname> acoles: yep
21:17:35 <mattoliverau> yeah, the reviewer guidelines
21:17:41 <acoles> it looked great
21:17:42 <jrichli> sounded like we are gonna accept any form of communicating a design for now : you dont *have* to write a spec, right?
21:17:48 <notmyname> still in progress, but definitely a good idea to have
21:18:05 <notmyname> https://review.openstack.org/#/c/287616/
21:18:05 <patchbot> notmyname: patch 287616 - swift - Adding review guidelines
21:19:10 <notmyname> so, other things talked about include global cluster improvements (including container sync). we made good progress there I hear
21:19:17 <notmyname> anyone want to give a better summary on that?
21:19:59 <notmyname> not everyone at once ;-)
21:20:32 <kota_> ah yeah
21:20:33 <notmyname> I can give it a try
21:20:35 <tdasilva> basically what I got from that is that we will continue to spend time improving obth
21:20:43 <notmyname> ah, good. go kota_ and tdasilva :-)
21:20:43 <acoles> we reviewed use cases, pros/cons of container sync and global clusters, and agreed both were valuable features
21:20:47 <tdasilva> as they help with different use cases
21:21:03 <acoles> and both have some deficiencies which people are working on fixing
21:21:13 <mattoliverau> cool, +1 to that
21:21:17 <kota_> yup, and i think we need something like table for users which staff is suitable for you use case in upstream docs.
21:21:19 <tdasilva> understanding the pros/cons was very valuable to me
21:21:32 <tdasilva> kota_: +1
21:22:07 <kota_> i have a picture for the white board and am willing to write down in docs in the future.
21:22:15 <mattoliverau> yeah andy was really excited after that topic and started getting container sync into openstack-ansible
21:22:16 <notmyname> kota_: great :-)
21:22:25 <notmyname> we also talked about notifications, and that quickly went into the metadata search use case
21:22:35 <notmyname> and we talked with the searchlight team
21:23:11 <notmyname> the idea we came up with is something like container sync (its mechanisms to poll the container DBs) to directly feed the indexer instead of a message bus
21:23:14 <mattoliverau> great, seems to be hugh community interest in that one and I wasn't even there
21:23:16 <notmyname> (like rabbit)
21:23:49 <notmyname> joel and tim worked with alex on client docs and have some stuff in progress there
21:23:56 <notmyname> anything else to bring up?
21:24:06 <acoles> there was a session on RBAC and policy which I missed
21:24:18 <notmyname> right. me too :-(
21:24:22 <acoles> anyone here attend that?
21:24:27 <notmyname> ho_away: ?
21:24:40 <ho_away> acoles: we had a great explanation from donagh for it :-)
21:24:56 <acoles> ho_away: oic. i must ask him for a replay!
21:25:02 <tdasilva> thanks to jrichli and acoles on the the crypto status update
21:25:24 <tdasilva> I think we also got a good plan on how we see it landing on master
21:25:26 <notmyname> yeah! the goal for crypto is to have a reasonably good feature branch by the summit
21:25:37 <notmyname> (and merging shortly thereafter)
21:25:47 <notmyname> jrichli: acoles: do I remember correctly?
21:25:57 <jrichli> notmyname: yes! and review the COPY as mw patches
21:25:58 <acoles> notmyname: yes
21:26:16 <tdasilva> lol
21:26:20 <acoles> oh yeah, tdasilva gave great overview of whats happening with COPY moving to middleware
21:26:30 <notmyname> that hasn't landed yet? ;-)
21:26:43 <mattoliverau> lol, I was going to make the same joke :P
21:26:55 <acoles> i saw some more +1 on one patch, thanks reviewers!
21:27:05 <mattoliverau> tdasilva: typey typey
21:27:12 <jrichli> :-)
21:27:19 <notmyname> mattoliverau: I think it's on us reviewers now!
21:27:21 <tdasilva> mattoliverau: it's done...i'm just waiting for reviews :(
21:27:23 <ho_away> acoles: yeah, he gave us a good explanation why it is important (i also learned from it)
21:27:35 <mattoliverau> damn.. thats on me then :P
21:27:40 <tdasilva> lol
21:27:52 <acoles> mattoliverau: and clayg was badgering to get concurrent gets landed
21:27:59 <tdasilva> what's typey typey for "go review"
21:28:17 <acoles> starey starey
21:28:21 <notmyname> overall, to say again, it was a great week and very productive
21:28:34 <tdasilva> just in general...I liked the way we chose the topics to discuss in Austin better than this time...just my 2 cents...
21:28:38 <notmyname> thank you to everyone who attended, and I hope everyone else will be at the next one :-)
21:28:54 <notmyname> tdasilva: that's good to know. I'll have to buy some more dot stickers ;-)
21:29:19 <tdasilva> notmyname: hehehe...i think letting people rank what's important and focusing on that was nice
21:29:30 <tdasilva> i guess it goes with the whole rocks and sands anology
21:29:44 <notmyname> yeah, I think I liked that better too.
21:30:00 <notmyname> ok, in the interest of time, I want to move on to some general stuff
21:30:11 <notmyname> #topic general summary of things
21:30:19 <notmyname> (how's that for a generic topic)
21:30:30 <mattoliverau> nice
21:30:40 <notmyname> I dont' want to discuss the per-release vs per-commit docs for swiftclient this week
21:30:51 <notmyname> we just cut a release last week, so frankly it doesn't matter too much now
21:31:09 <notmyname> I'm learning some about how they are built, so I think we may be able to have the best of both
21:31:18 <notmyname> speaking of releases...
21:31:36 <notmyname> we just did swiftclient 3.0.0, and that will be the final one (baring any security issues) for mitaka
21:31:51 <notmyname> and for swift itself, we'll have one more release scheduled for mitaka
21:32:01 <notmyname> our final date for that is the week of march 28
21:32:17 <notmyname> but I'll likely try to do it sooner than later
21:32:33 <notmyname> ie next week or the week after
21:33:06 <notmyname> later this week i'll start to organize the starred patches for what we should try to get in
21:33:14 <notmyname> speaking of starred patches...
21:33:33 <notmyname> community dashboard seems to be doing one of it's jobs...
21:33:33 <notmyname> http://not.mn/swift/swift_community_dashboard.html
21:33:41 <notmyname> only 8 unreviewed patches right now!
21:34:26 <notmyname> you all gave me a lot of feedback last week, so I'll be updating this page in the future
21:34:39 <notmyname> (specifically about measuring review vs owner timings
21:34:53 <kota_> maybe a bunch of unreviewd patches got conflicted with current master :/
21:35:17 <notmyname> and finally, (at least in the US) the daylight saving time starts this weekend, but we still will meet next week at 2100UTC
21:35:45 <notmyname> so for some of you, the meeting time will change. pay attention :-=)
21:36:04 <notmyname> and that's all I have for this week
21:36:14 <notmyname> anything else (please keep patch begging in -swift)
21:36:24 <tdasilva> lol
21:36:54 <notmyname> 3...
21:37:04 <notmyname> 2...
21:37:22 <notmyname> 1...
21:37:30 <tdasilva> wait
21:37:32 <tdasilva> jk
21:37:34 <notmyname> tdasilva: go
21:37:37 <notmyname> ok, fine
21:37:39 <notmyname> done
21:37:50 <notmyname> thanks for coming. thank you for your work on swift
21:37:53 <notmyname> #endmeeting