21:00:59 #startmeeting swift 21:01:00 Meeting started Wed Nov 28 21:00:59 2018 UTC and is due to finish in 60 minutes. The chair is mattoliverau. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:01:01 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:01:03 The meeting name has been set to 'swift' 21:01:14 who's here for the swift team meeting? 21:01:21 o/ 21:01:27 hello 21:01:37 \o/ hey guys 21:01:40 hi o/ 21:02:36 this will be another status update meeting. notmyname is travelling so I'll have to try and fill his enourmous shoes. 21:02:45 agenda is at 21:02:51 #link https://wiki.openstack.org/wiki/Meetings/Swift 21:03:23 Just a few topics and mainly priority reviews and status updates on task queue and losf 21:04:07 this might be a quick meeting. 21:04:09 mattoliverau: o/ 21:04:13 clayg: o/ 21:04:24 cool, let's get started 21:04:32 #topic priority reviews 21:04:41 #link https://wiki.openstack.org/wiki/Swift/PriorityReviews 21:04:54 that list looks good - mattoliverau did you curate that? 21:05:15 nah, I'm too lazy to do that :P 21:05:17 I noticed kazuhiro updated some general task queue patches 21:06:07 So last week we talked about adding my multi-kms patch to the prioirty reviews, which I forgot, but doesn't matter because thanks to timburke it's been +Aed! 21:06:29 tdasilva: yeah, wish we was here so we could ask about it's state 21:06:34 *he 21:06:39 just needs to get through the gate ;-) 21:07:09 yeah, easier said then done. I was hoping to say it's landed.. but alas not :P 21:07:49 how are the pending s3api patches going? timburke? 21:08:23 no real progress -- i at least did some rebasing, though! 21:08:26 any more to add? need more eyes for reviews? 21:08:36 ahh rebasing is good. 21:08:54 also, it has been Thanksgiving so a slower week then normal 21:08:57 eyes are good. https://review.openstack.org/#/dashboard/?S3+API=status:open+file:%255Eswift/common/middleware/s3api/.*+project:openstack/swift is a handy dash to look at just s3 changes 21:09:19 also to all you Americans, happy thanksgiving and I hope your fridges are full of leftovers :) 21:09:47 #link https://review.openstack.org/#/dashboard/?S3+API=status:open+file:%255Eswift/common/middleware/s3api/.*+project:openstack/swift 21:10:10 oh yeah! https://review.openstack.org/#/c/609860/ is a reasonably cool patch that would open up some possibilities like https://review.openstack.org/#/c/610747/ 21:10:55 ahh cool 21:10:58 but i need to come up with a better name for the callback than "slo_validator", since it can make modifications, too... 21:11:23 slo-matic :P 21:11:43 slo-mematron 21:11:56 too early for me :P 21:12:09 timburke: the name is always the hardest part 21:12:09 i was thinking slo_manifest_hook but i'm certainly open to suggestions 21:12:24 hooks are good! 21:12:38 clayg: real handy for catching fish! 21:12:42 we (old rackspace team) used to use a project naming engine 21:13:18 it's where we got turbohipster (for some nova 3rd party api runner we wrote) 21:13:34 anyway, now I'm distracted and off topic :P 21:14:05 Anything else re:priority reviews? anything we need to add? 21:14:28 let's move on 21:15:34 #topic task queue 21:15:51 kazuhiro isn't here unfortuantly 21:16:03 So not much to really say about this. 21:16:29 rledisez: any chance you had to time to look at the follow up doc patch we want to merge in? kazuhiro wanted you to take a look. 21:17:12 mattoliverau: no, sorry for that, i've been busy fighting eventlet (but I won, i think…) 21:17:13 As tdasilva mentioned, kazuhiro pushed up a new change set yesterday (my time). I haven't looked closely but think it might be a rebase 21:17:34 rledisez: 1 eventlet: 0 21:18:18 rledisez: you were right, that was a good bug :-) 21:18:23 rledisez: nps, when you get a change. it's just documentation regarding the implementation in the overview_expiring_objects section 21:18:31 *chance 21:18:36 man, I can't type today 21:19:17 let's move on so rledisez can tell us about eventlet battles and losfs :) 21:19:22 #topic LOSF 21:19:54 rledisez: any updates here, we saw the eventlet bug and patch you talked about last week 21:20:06 clayg looked at it and I think it's even landed already 21:20:23 so, 3 bugs this week, not all related to losf, but all related to eventlet (one fix merged), the second was in losf (finally was not always called in try/except/finally). last one was in a patch I proposed related to ssync and concurrency 21:20:28 we won the 3 battles :D 21:20:55 #FTW 21:21:42 on pure losf side, it' quite calm, alexandre is fixing some bugs we discover from time to time. nothing really serious (aka, it's not eating data), some memory leaks & co 21:21:43 \o/ nice 21:22:19 no that's good, if your running in prod and it's nice and quite, that sounds like it's getting pretty stable, which is awesome news 21:23:11 Finally block _not_ getting called sounds.. well.. rather broken, wow. 21:23:27 anything the rest of us need to do to help losf's out? 21:23:40 depends on how hard things fail -- i could *totally* see it not necessarily getting called 21:23:58 SIGKILL = no finally (for example) 21:24:34 "oh, i'm sorry were you saying something?" -kernel 21:24:38 the process was not crashing. alexandre could give you the details, but it was just not calling finally, and was continuing to run the function 21:24:45 well yeah I guess, but when I'm coding I tend, maybe wrongly, that my finally block would be run :) 21:24:56 lol 21:25:08 OOM 21:25:28 mattoliverau: yeah i agree sounds wild 21:25:56 alex may indeed have to give us the details - without some careful comments it's likely the issue would get re-introduced if we don't have a communal understanding 21:26:09 no to mention raising concerns of other places the same issue might exists 😬 21:26:10 +1 21:26:28 but I'm sure it'll all be sorted in py3 :P 21:26:38 or go 21:27:32 rledisez: anything else? anywhere we can help? 21:28:16 mattoliverau: not for now I think. i'll try to get alex to join the meeting next time so he can gives more updates and ask help if he needs some 21:28:26 sweet 21:28:33 let's move on then. 21:28:40 #topic open discussion 21:28:57 The floor is open 21:29:28 anything else to bring up at the meeting, any interesting ideas, problems, patches to highlight 21:29:45 any funning holiday stories. 21:30:02 😁 21:30:20 *funny *sigh* 21:30:39 I really should just give up, go back to bed and start again :P 21:31:09 *chirp* *chirp* 21:31:26 Cool, a shorter meeting, we all get an early mark 21:31:32 meeting ending in... 21:31:36 3.. 21:31:42 2.. 21:31:48 1.. 21:31:53 #endmeeting