21:00:12 <notmyname> #startmeeting swift
21:00:13 <openstack> Meeting started Wed Jun 21 21:00:12 2017 UTC and is due to finish in 60 minutes.  The chair is notmyname. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:00:15 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:00:17 <openstack> The meeting name has been set to 'swift'
21:00:21 <notmyname> who's here for the swift team meeting?
21:00:36 <kota_> hi
21:00:42 <mattoliverau> o/ (kinda)
21:00:54 <timburke> o/
21:01:17 <notmyname> big crowd ;-)
21:01:28 <acoles> hello
21:01:38 <jungleboyj> o/
21:02:13 <notmyname> looks like a short agenda today, so let's get started so people can get breakfast/to bed
21:02:20 <notmyname> #link https://wiki.openstack.org/wiki/Meetings/Swift
21:02:37 <notmyname> #topic boilerplate announcements
21:02:53 <notmyname> note the PTG and summit links on the agenda
21:03:12 <notmyname> please encourage any ops you know to go to the PTG
21:03:27 <notmyname> #topic deprecate known-bad EC config
21:03:32 <notmyname> patch 468105
21:03:33 <patchbot> https://review.openstack.org/#/c/468105/ - swift - Require that known-bad EC schemes be deprecated
21:03:40 <notmyname> we've got 2 +2 on it
21:03:45 <notmyname> we've emailed different ML
21:03:50 <notmyname> we've talked about it for a long time
21:04:05 <notmyname> is there any reason to not +A it right now?
21:04:24 <mattoliverau> Nope, do it!
21:04:52 <notmyname> kota_: acoles: ok with you for me to +A it right now?
21:05:02 <kota_> ok
21:05:03 <acoles> yes
21:05:11 <notmyname> done
21:05:12 <notmyname> thanks
21:05:20 <notmyname> timburke: thanks for pushing this one
21:05:32 <notmyname> kota_: thanks for setting this whole thing off by finding the bug :-)
21:05:47 <kota_> :-)
21:05:58 <notmyname> #topic previous meeting follow-ups
21:06:03 <notmyname> not much here
21:06:06 <timburke> all i did was write a patch :-) sorry for a bit of a fire drill at work clayg/acoles
21:06:08 <notmyname> romain isn't here this week
21:06:22 <notmyname> although I saw the email ovh sent about LOSF
21:06:27 <notmyname> if you didn't read it, please do
21:06:28 <acoles> timburke: lol
21:06:48 <notmyname> no progress on py3 work afaik
21:06:49 <acoles> timburke: thanks for updating the patch with doc
21:07:08 <notmyname> #topic priority patches
21:07:15 <notmyname> #link https://wiki.openstack.org/wiki/Swift/PriorityReviews
21:07:20 <notmyname> I updated that wiki page
21:07:40 <notmyname> there's 10 patches there, but I have a question about one or two of them
21:07:50 <notmyname> https://review.openstack.org/#/c/455488/
21:07:50 <patchbot> patch 455488 - python-swiftclient - Tolerate RFC-compliant ETags
21:08:03 <notmyname> has +2 from joeljwright1 and timburke, but timburke is also the author
21:08:04 <joeljwright1> reviewed that today
21:08:09 <notmyname> so what needs to happen there?
21:08:12 <notmyname> timburke: ?
21:08:26 <joeljwright1> it's a simple patch, but more eyes couldn't hurt…
21:08:40 <timburke> merge it! there was some hesitance since i'm also the author
21:08:58 <joeljwright1> (I'm the hesitant one)
21:09:02 <joeljwright1> I'll go merge it
21:09:09 <notmyname> joeljwright1: ack
21:09:33 <clayg> i clicked +A on the etag one - I had comments from last week - my bad
21:09:40 <notmyname> ah! thanks clayg
21:09:47 <joeljwright1> dammit clayg beat me to it!
21:09:50 <clayg> sort of - would have been better to do that last week!
21:09:51 <joeljwright1> :D
21:10:48 <notmyname> ok, any other patches to call out or that need discussion during this meeting?
21:11:30 <mattoliverau> The tpool one looks good now, with a fixed SAIO, you can easily test and use servers_per_port
21:11:39 <notmyname> https://review.openstack.org/#/c/289664/
21:11:40 <patchbot> patch 289664 - swift - Make eventlet.tpool's thread count configurable in...
21:11:43 <notmyname> that one?
21:11:55 <mattoliverau> Ta
21:11:57 <mattoliverau> Yup
21:12:29 <timburke> https://review.openstack.org/#/c/307937/ seems like a good idea... haven't actually demonstrated a socket leak, but fairly certain there is one
21:12:30 <patchbot> patch 307937 - swift - Close all versioned_writes subrequests' app_iters
21:12:33 <mattoliverau> So someone go play. Also merged in clays thoughts and using auto config
21:12:58 <notmyname> ah right. thanks timburke. I wanted to track the socket leak ones. didn't you have another?
21:13:28 <timburke> Romain's https://review.openstack.org/#/c/474767/ merged already
21:13:29 <patchbot> patch 474767 - swift - Fix a socket leak in copy middleware (MERGED)
21:13:32 <acoles> rledisez fixed one last week
21:13:40 <acoles> timburke: beta me
21:13:48 <acoles> beat*
21:14:00 <notmyname> tim as a beta version of al?
21:14:02 <notmyname> ;-)
21:14:04 <timburke> oh, https://review.openstack.org/#/c/471993/ ? i forget whether that actually full fixes the problem
21:14:05 <patchbot> patch 471993 - swift - Send Connection:close headers when we haven't read...
21:14:14 <mattoliverau> Lol
21:14:46 <joeljwright1> It also seems like more people need to weigh in on whether we want to add & expose pre/postambles to the SLO API… https://review.openstack.org/#/c/365371
21:14:46 <patchbot> patch 365371 - swift - Add Preamble and Postamble to SLO and SegmentedIte...
21:15:07 <notmyname> joeljwright1: ah, right. you and i talked in irc about it
21:15:23 <timburke> i haven't played around with a bad client on it yet -- see if eventlet closes the connection even if the client tries to hold it open
21:15:24 <clayg> @timburke I've been growing increasingly convinced sending connection: close to eventlet won't do the trick (because of the EC connection: close backend header leaking bug)
21:15:33 <timburke> :-(
21:15:52 <clayg> lp bug #1680731
21:15:53 <openstack> Launchpad bug 1680731 in OpenStack Object Storage (swift) "GET EC policy object includes connection:close header which causes external LB returns 502 bad gateway to client" [Medium,Fix released] https://launchpad.net/bugs/1680731
21:16:27 <clayg> also when I looked at the code in patch 471993 I remember being totally confused - but never tested if the pipelined request with 413 acctually broke
21:16:27 <patchbot> https://review.openstack.org/#/c/471993/ - swift - Send Connection:close headers when we haven't read...
21:18:40 <notmyname> so... more testing is warranted
21:19:28 <notmyname> joeljwright1: when did we talk about the TLO stuff? I want to find the link. what day was it?
21:19:33 <notmyname> yesterday? or monday?
21:19:40 <joeljwright1> Monday
21:19:57 <notmyname> #link http://eavesdrop.openstack.org/irclogs/%23openstack-swift/%23openstack-swift.2017-06-19.log.html#t2017-06-19T15:48:47
21:20:05 <timburke> https://review.openstack.org/#/c/459023/ would be nice (particularly, i feel, if tdasilva is still working on that release)
21:20:05 <patchbot> patch 459023 - liberasurecode - Consistently use zlib for crc32
21:20:08 <notmyname> ah yes. easy to search logs when not much has been said in them ;-)
21:21:05 <notmyname> joel's pre/post amble addition to SLO is for enabling tarfiles. and potentially other archive formats
21:21:22 <notmyname> and while I've heard some bit of discussion about it, there's not a lot of record of that in gerrit
21:21:48 <notmyname> see the above IRC log link for monday's conversation we had. gives a little more detail on the goals for after this patch
21:22:06 <notmyname> but it's a change that, if we add, we'll have to support, so now's the time to speak up about it
21:22:32 <notmyname> #topic open discussion
21:22:43 <notmyname> anything else from anyone to bring up in today's meeting?
21:23:57 <timburke> https://review.openstack.org/#/c/465776/ and https://review.openstack.org/#/c/449394/ are interesting things that kinda popped out of reviewing sharding and symlinks... wouldn't mind getting opinions on them
21:23:58 <patchbot> patch 465776 - swift - Increase name-length limits for internal accounts
21:23:59 <patchbot> patch 449394 - swift - WIP: Move listing formatting out to proxy middleware
21:25:02 <mattoliverau> I'm kinda busy (yes even tho unemployed) until Friday, but will try and have a look at em then
21:25:35 <mattoliverau> The first one would be needed and awesome for Sharding name edgecases
21:26:46 <timburke> mattoliverau: i think the second one would be useful too -- we're splicing listing results across shards in the proxy, right?
21:27:01 <timburke> only having one format to deal with could be handy
21:27:06 <mattoliverau> Oh yeah, I like that idea too :)
21:27:24 <mattoliverau> Yup
21:27:30 <notmyname> timburke: thanks for mentioning those
21:27:41 <mattoliverau> And so we always ask for json
21:28:05 <notmyname> no need to sit around in here for another 30 minutes if theres not a pressing topic... last call?
21:28:31 <notmyname> this is easier in person when i can look at your faces :-)
21:28:58 <notmyname> thanks for the work you do on swift, and thanks for coming to the meeting
21:28:58 <timburke> someone (zaitcev?) should go land https://review.openstack.org/#/c/337297/
21:28:59 <patchbot> patch 337297 - swift - Add support to increase object ring partition power
21:29:16 <tdasilva> hello! sorry i'm late
21:29:35 <timburke> notmyname: see, this is why we can't end early :P
21:29:38 <notmyname> well, exciting things happen in what was going to be the last 5 seconds of the meeting...
21:29:55 <notmyname> seriously, i had #endmeeting already about to hit return
21:29:58 <mattoliverau> Lol
21:30:07 <notmyname> tdasilva: anything to bring up this week?
21:30:15 <notmyname> tdasilva: otherwise, I'm gonna call it :-)
21:30:35 <tdasilva> no, sorry, i'm going to read up on what was said and am happy to follow up on the channel
21:30:47 <notmyname> cool
21:30:54 <notmyname> #endmeeting