21:00:07 #startmeeting swift 21:00:08 Meeting started Wed Oct 2 21:00:07 2019 UTC and is due to finish in 60 minutes. The chair is timburke. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:00:09 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:00:12 The meeting name has been set to 'swift' 21:00:20 who's here for the swift meeting? 21:00:24 o/ 21:01:07 o/ 21:01:11 o/ 21:01:13 o/ 21:02:00 hi 21:02:04 as usual, agenda's at https://wiki.openstack.org/wiki/Meetings/Swift 21:02:21 #topic shanghai 21:02:35 i'm going to keep plugging the etherpad ;-) 21:02:38 #link https://etherpad.openstack.org/p/swift-ptg-shanghai 21:02:44 Could we move this meeting an hour in either direction? 21:03:37 clayg, after shanghai, i think it might move on its own ;-) 21:03:46 (at least, for us) 21:04:31 anyway, i added a couple topics -- everyone should! :D 21:04:58 clayg: I'm not seeing tsync on the etherpad ;) 21:05:24 it's like he's been distracted by something 21:05:48 #topic train release 21:06:31 https://review.opendev.org/#/c/684041/ is working its way through the gate 21:06:32 patch 684041 - swift - Fix some request-smuggling vectors on py3 - 3 patch sets 21:07:00 once it lands, i'll rebase and land https://review.opendev.org/#/c/685503/ 21:07:00 patch 685503 - swift - Authors/changelog for 2.23.0 - 2 patch sets 21:07:17 then request the release and branching 21:07:25 thanks for all the reviews everyone! 21:07:44 as usual, this is going to be the best release yet :-) 21:08:17 Best ever 21:08:25 if anyone wants to spot-check the release notes, now's the time 21:08:40 mattoliverau, no! what about *next* time? 21:08:43 I wish we required eventlet-0.25 before Train. 21:10:00 lol,.. up until this point in time.. ok ok best yet :P 21:10:25 zaitcev, 0.25 only came out in May ... sorry to push you to the bleeding edge so late :-/ 21:11:22 any other questions or comments on the release? 21:12:51 on to updates! 21:13:02 #topic versioning 21:13:12 clayg, tdasilva, how's it going? 21:13:51 looks like there's been good progress on p 682138 21:13:51 https://review.opendev.org/#/c/682138/ - swift - Allow internal clients to use null namespace - 6 patch sets 21:14:26 Some. The prefix/delim subdir thing needs work. 21:14:46 Sorry. I’m not really online right now. Trying. I’ll be on later. 21:14:50 on p 682382 i'm continuing to work on tests, both unit and functional tests 21:14:50 https://review.opendev.org/#/c/682382/ - swift - WIP: New Object Versioning mode - 9 patch sets 21:15:12 I tried to start documenting the API here: https://etherpad.openstack.org/p/swift-object-versioning 21:15:34 👍 good call 21:16:09 i've got more tests to write and there are some corner cases that still need to be addressed 21:16:59 in documenting some of the new api, were there any places where you realized you wanted feedback from people about how it *should* work? 21:18:27 mmm..to be honest, right now in terms of how it should work i've been kinda of looking up to s3 behavior, there's going to be very little deviation from that if any 21:19:24 one of the difficult corner cases is handling when users suspend/enable versioning back and forth 21:20:04 so trying to understand how s3 handles it and also come up with a good definition of our own is important and challenging at the same time 21:20:30 makes sense. is there anything you guys are blocked on? anything that other people can/should be doing? 21:21:46 i'd really welcome if others can start looking at the code and testing it out. I just noticed that mattoliverau wrote down some questions on the etherpad and that's really great, thank you 21:22:15 nps 21:23:18 all right 21:23:18 I'm only gone through the etherpad. next is looking at code :) 21:23:25 *i've 21:23:32 #topic lots of small files 21:23:46 alecuyer, i saw a fix for the auditor go by 21:23:48 https://review.opendev.org/#/c/685699/ 21:23:48 patch 685699 - swift (feature/losf) - LOSF - make invalidate/consolidate_hashes() noop (MERGED) - 1 patch set 21:24:00 and of course so far i've been putting off the integration with null namespace, but should start thinkign about rebasing on top of that soon 21:24:30 and a new patchset for https://review.opendev.org/#/c/666378/ to get tests passing 21:24:31 patch 666378 - swift (feature/losf) - Add tests for vfile.py - 10 patch sets 21:24:34 yes, so that the auditor does not have to be patched to work with LOSF 21:24:52 and for the second one as you know I hit an issue with mocks in the tests, related to 21:24:56 https://www.python.org/dev/peps/pep-3151/ 21:25:57 So I should fix the tests (more details in the patch comments) 21:26:47 the IOError/OSError change did touch more than i was expecting... i suppose you're right, we should probably propose similar changes to master 21:28:03 or not :) I'll let people look at the pep, but I feel like the "deprecation" mentionned there would not happen anytime soon (enough issues with py2/py3 right?) but if anyone wants to leave a comment about that in the patch or on irc thats great 21:28:03 ... i guess? i'm trying to think if there are any places that we might *rely* on there being a difference between the two... and what that would mean for py3 support... 21:30:26 alecuyer, is there much else going on with losf? i'll be sure to get you a merge from master once i've for the new tag 21:30:57 timburke: no, I'll have a few more patches to push but otherwise no 21:31:19 anything you're blocked on or would like input on? 21:31:41 no, nothing blocking, thanks for your comments on that latest patch! 21:31:52 👍 21:32:05 #topic sharding 21:32:46 Things have landed! I've started looking at the latch patch. 21:32:55 https://review.opendev.org/#/c/675820/ and https://review.opendev.org/#/c/681970/ both landed in time for the release! 21:32:56 patch 675820 - swift - sharder: Keep cleaving on empty shard ranges (MERGED) - 5 patch sets 21:32:57 patch 681970 - swift - Sharding: Clean up old CleaveConext's during audit (MERGED) - 2 patch sets 21:33:04 \o/ 21:33:48 hopefully they'll make life easier for everyone with sharding enabled not just NM 21:34:08 i'm pretty sure :-) 21:34:26 So now there's just timburke's latch patch then I will loop back around to playing with auto-sharding. 21:34:46 Who is NM? 21:35:06 zaitcev: the person who had the header issue 21:35:38 mattoliverau, anything you need help on? 21:36:13 not at the moment. I'll review you patch and if anyone feels like they want to comment on the auto sharding code/approach great.. 21:36:21 but we all have alot on our plate atm. 21:36:33 So all good. 21:36:45 gotta reload autosharding into my head and test test teat. 21:36:49 *test 21:37:05 sounds good 21:37:11 #topic open discussion 21:37:37 ...and i feel like clayg had a good one. what do we think about trying to shift the meeting time? 21:38:51 Like you said, it's going to shift on its own soon anyway. 21:39:44 no action is easiest action ;-) 21:40:03 yeah, it'll shift for me in like a week or so. 21:40:16 it's really hard with all the timezones we have 21:40:35 oh, right! tdasilva, you're in a new one now, too, aren't you? 21:40:45 nope, still same as alecuyer 21:40:46 lol, or all the timezones tdasilva is in :P 21:40:54 lol 21:41:48 even taking me out of the pictures, still really hard to find good time between SF, monstreal, brest, tokyo, sydney 21:42:25 clayg, is this time slot bad all around, or is it this time on *this day*? 21:43:36 clayg mentioned he's not really here. So we may need to take this async 21:43:38 https://www.timeanddate.com/worldclock/meetingtime.html?iso=20191109&p1=111&p2=195&p3=224&p4=248&p5=240&p6=43&p7=70 21:44:03 looking at november because of the time changes happening soon 21:46:40 well, something to keep thinking about. timezones are hard :-( 21:47:02 yeah it'll be 8am for me soon, which is much more civilised. Although any pushing forward is great for me (and kota_), no so much for alecuyer. 21:48:47 i'd be ok with pushing it to some weird time for me... but i'd feel bad about doing that to clayg, zaitcev, and tdasilva (when he gets home) 21:49:07 Naah, I'm fine 21:49:10 Maybe we need to get more info from clayg, always happy to make life easier for him. 21:49:38 we'll keep talking about it async 21:49:47 mattoliverau, thinking about the sharding stuff: any opposition to just landing https://review.opendev.org/#/c/670673/ once the release stuff is done? 21:49:47 patch 670673 - swift - ring: Track more properties of the ring - 3 patch sets 21:51:14 not from me, it seems to have my +2 :P 21:51:46 hehe -- was mainly checking to see that you didn't have a strong opinion on wanting more eyes 21:52:01 anything else we should discuss? 21:52:04 I'll give it a re-review just in case :) 21:53:55 all right then 21:54:10 thank you all for coming, and thank you for working on swift! 21:54:14 #endmeeting