21:00:01 <timburke> #startmeeting swift 21:00:02 <openstack> Meeting started Wed Jul 17 21:00:01 2019 UTC and is due to finish in 60 minutes. The chair is timburke. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:00:03 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:00:05 <openstack> The meeting name has been set to 'swift' 21:00:08 <timburke> who's here for the swift meeting? 21:00:17 <alecuyer> hello! 21:00:22 <mattoliverau> o/ 21:00:41 <rledisez> hi! 21:00:41 <kota_> o/ 21:00:46 <tdasilva> hi 21:01:36 <timburke> clayg? 21:02:01 <clayg> oh yeah 21:02:29 <timburke> agenda: 21:02:31 <timburke> #link https://wiki.openstack.org/wiki/Meetings/Swift 21:02:52 <timburke> #topic shanghai 21:03:03 <timburke> i made an etherpad! 21:03:05 <timburke> #link https://etherpad.openstack.org/p/swift-ptg-shanghai 21:03:32 <kota_> +1 21:03:35 <mattoliverau> \o/ 21:03:41 <clayg> oh snap is it etherpad time already! 😓 21:03:47 <timburke> if you're coming (or even just hopeful that you'll be coming), please add yourself! 21:04:05 <timburke> clayg, mainly because of logistics -- foundation wants a rough headcount 21:04:51 <timburke> i also started seeding some potential onboarding topics; please add anything else that might be good "intro" material 21:05:08 <mattoliverau> I really need to poke Suse to make a decision, ecp because visa's will be involved. 21:05:15 <mattoliverau> *esp 21:05:30 <timburke> i feel like this is an awesome opportunity to reach out and meet potential new contributors 21:05:33 <alecuyer> Same with OVH I think (romain?) 21:05:47 <rledisez> alecuyer: you can copy/paste my line ;) 21:05:54 <alecuyer> right :-) 21:06:17 <tdasilva> anyone has the link for the visa form? can you add to etherpad? can't seem to find it 21:06:38 <rledisez> #link https://openstackfoundation.formstack.com/forms/travelsupportshanghai 21:06:41 <rledisez> this one? ^ 21:06:52 <timburke> i'll be sure to bring up shanghai again when i see the discount codes go out 21:06:54 <rledisez> oups, no :) 21:07:13 <tdasilva> rledisez: yep, thanks! 21:07:28 <tdasilva> just noticed the dealine too, hadn't seen that yet 21:07:39 <timburke> #link http://www.china-embassy.org/eng/visas/ 21:07:44 <timburke> #link https://openstackfoundation.formstack.com/forms/visa_form_shanghai_summit 21:09:21 <timburke> etherpad was the main thing i wanted to bring up regarding the summit; does anyone have other questions about it? 21:09:25 <clayg> is this one of those places you can't bring a laptop/phone? 21:09:55 <alecuyer> I was wondering about that 21:09:59 <timburke> ...maybe. i'll probably have a freshly-formatted computer, fwiw 21:10:22 <tdasilva> or keep it with you at all times? 21:10:31 <rledisez> alecuyer: i'll ask our SOC team what they think about that 21:10:32 <tdasilva> that's my hope/plan 21:11:31 <timburke> moving on 21:11:34 <kota_> not sure it must be fresh, but fwiw we cannot reach out any cloud service including google thre. 21:11:35 <timburke> #topic releases 21:11:46 <kota_> w/o roaming 21:11:51 <timburke> kota_, good to know 21:12:02 <timburke> there's a new swiftclient release! 21:12:43 <timburke> it'd been a while, and there was an interesting/nasty bug related to SLO that bit a customer of ours 21:12:57 <timburke> hopefully *just* us ;-) 21:13:03 <clayg> 3.8.0? https://pypi.org/project/python-swiftclient/ 21:13:12 <clayg> looks like it! https://pypi.org/project/python-swiftclient/#history 21:13:14 <clayg> WTG! 21:13:27 <timburke> that's the one! 21:13:56 <timburke> and i merged the authors/changelog patch for swift 2.22.0 21:14:08 <clayg> release machine!!! 21:14:20 <timburke> and requested a release: https://review.opendev.org/#/c/671350/ 21:14:21 <patchbot> patch 671350 - releases - swift 2.22.0 release - 1 patch set 21:14:23 <kota_> great 21:14:26 <clayg> 🙌 21:14:53 <mattoliverau> ncie 21:14:56 <mattoliverau> nice even 21:15:41 <timburke> a lot of great stuff is going into these releases and i want to thank you all for your hard work 21:15:58 <timburke> which leads nicely into updates... 21:16:02 <timburke> #topic py3 21:16:20 <timburke> we've got py2 func tests against py3 services in the gate! 21:17:12 <timburke> i've proposed https://review.opendev.org/#/c/670353/ to get swift on py3 in devstack 21:17:13 <patchbot> patch 670353 - devstack - Remove Swift from default DISABLED_PYTHON3_PACKAGES - 1 patch set 21:17:35 <mattoliverau> awesome work 21:17:43 <timburke> and already fixed up vagrant-swift-all-in-one to let you run services on py3 21:18:24 <clayg> so py3 21:18:43 <timburke> once i had that, i figured we'd get the most bang-for-the-buck with probe tests 21:19:05 <timburke> so now we've got https://review.opendev.org/#/c/671333/ 21:19:06 <patchbot> patch 671333 - swift - py3: (mostly) port probe tests - 2 patch sets 21:19:29 <timburke> which already found bugs! that kicked out... 21:19:34 <timburke> https://review.opendev.org/#/c/671167/ 21:19:35 <patchbot> patch 671167 - swift - py3: fix up listings on sharded containers - 1 patch set 21:19:38 <timburke> https://review.opendev.org/#/c/670932/ 21:19:39 <patchbot> patch 670932 - swift - py3: fix up swift-orphans - 1 patch set 21:19:43 <timburke> https://review.opendev.org/#/c/670933/ 21:19:43 <patchbot> patch 670933 - swift - py3: fix object-replicator rsync output parsing - 2 patch sets 21:19:49 <timburke> and https://review.opendev.org/#/c/671168/ 21:19:50 <patchbot> patch 671168 - swift - Fix up errno checking - 1 patch set 21:20:27 <timburke> so... release with initial support is great! but there's also more work ahead ;-) 21:20:44 <kota_> ic 21:20:49 <mattoliverau> good to find these now. 21:20:54 <alecuyer> nice finds:) 21:21:19 <timburke> func tests still need to get ported, of course. and we've got a lot of progress there, too 21:21:46 <timburke> next up 21:21:53 <timburke> #topic lots of small files 21:22:51 <timburke> we've got a working gate again! and merges from master are happening again (thanks kota_!) 21:23:12 <timburke> alecuyer, rledisez, kota_ how's it going? 21:23:29 <alecuyer> thanks kota_ for the merge. 21:23:46 <kota_> yup 21:24:07 <alecuyer> haven't been able to work on LOSF this week, but I have read the comments you made on the test patch tim, thanks. I will work on that , 21:24:38 <timburke> i *almost* got tests passing :P 21:24:42 <alecuyer> and I will get back to the HTTP patch 21:25:00 <alecuyer> timburke on SAIO you mean? :) 21:25:23 <timburke> 👍 21:25:49 <timburke> on py3. there's just this one tricky guy where OSError and IOError get confused... 21:26:15 <alecuyer> ouch oh yes , we can discuss that after the meeting or on the review, I have to check on that 21:26:25 <timburke> k 21:26:47 <timburke> is there anything else the rest of us could/should be doing? 21:27:13 <alecuyer> I saw your comments about the headers (volume header, object header) 21:27:38 <alecuyer> kota_ also had comments about it a while back (I think kota you wished for a fixed size) 21:28:03 <alecuyer> so if anyone wants to comment on that and make the headers better, please comment (swift/obj/header.py) 21:28:25 <timburke> #link https://review.opendev.org/#/c/666378/ 21:28:26 <patchbot> patch 666378 - swift (feature/losf) - Add tests for vfile.py - 7 patch sets 21:28:47 <kota_> hopefully. but it's general comments so if we should make it flexible, it's not strong opinion. I'll check the comment. 21:28:59 <alecuyer> yes, thanks 21:29:31 <timburke> all right 21:29:48 <timburke> #topic auto-sharding 21:30:25 <timburke> mattoliverau, how's it going? 21:30:25 <mattoliverau> I haven't had much of a chance to progress this much, been a busy week. But I did see your patches timburke 21:31:07 <mattoliverau> looked at your ring change which some extra data. Looking good. Haven't tested it much though 21:31:31 <mattoliverau> the ring service is interesting. 21:32:04 <mattoliverau> I really need to write a probe test re:shrinking with the new autosharding to make sure it behaves. 21:32:48 <mattoliverau> With timburke's rewored ring change, I might remove the patch chain around it 21:32:55 <mattoliverau> *reworked 21:33:13 <mattoliverau> wow, I can't type pre breakfast :P 21:33:28 <timburke> oh, right! i'd already started paging this stuff out. the ring-as-a-service thing was just an idea i had since we always seem hesitant to add rings to a daemon that didn't have them before... 21:34:16 <mattoliverau> let's try that again.. with timburke's reworked ring change I might rebase and rework the patch chain around it. To keep patches more generic 21:34:52 <timburke> sounds good 21:35:15 <mattoliverau> that's all I got as an update. 21:35:32 <timburke> #topic symlink-backed versioning 21:35:52 <timburke> i mainly added this because clayg said he might have an update for the meeting earlier :-) 21:36:05 <clayg> oh good! 21:36:21 <clayg> please feel free to look at hardlinks -> p 633094 21:36:21 <patchbot> https://review.opendev.org/#/c/633094/ - swift - Allow "harder" symlinks - 14 patch sets 21:36:40 <clayg> hardlinks are really cool - they're symlinks that put a lot of useful info into your container listing for you 21:37:19 <clayg> versioned writes is very close to moving to using these hardlinks to make things operate a lot smoother on the backend -> p 633857 21:37:20 <patchbot> https://review.opendev.org/#/c/633857/ - swift - WIP: symlink-backed versioned_writes - 3 patch sets 21:38:01 <clayg> instead of COPYs to move things in and out of the stack - we just need to update some links in the version container 21:38:28 <timburke> clayg, you really ought to add yourself as co-author on those ;-) 21:38:59 <clayg> ok 21:39:21 <timburke> i definitely like the reduced data movement aspect; i'm super-excited 21:39:31 <mattoliverau> +1 21:40:06 <clayg> yeah it's super great 21:40:49 <timburke> anything else we can do for you other than go review it? 21:41:13 <clayg> right, yeah play with hardlinks - I was toying around with adding something to python-swiftclient `swift link ...` 21:41:42 <clayg> but i think i set that aside realizing that it was a pretty significant set of work 21:42:10 <timburke> yeah -- i remember you trying to bend it to creating a symlink before and not having much luck... we surely *should* add a `link` subcommand... 21:43:34 <timburke> maybe that's a thing i could take on... i realized as part of the probe tests getting ported that i need to do some updates to swiftclient to get around https://bugs.python.org/issue37093... 21:43:41 <clayg> but aside from that yeah just a review on the hardlinks - it's a foundational piece of the versioned_writes update - which is pretty significant 21:43:45 <timburke> as long as i'm looking at the client anyway... 21:44:00 <clayg> but at this point (EoD) it's going to be significantly well tested - so I'm happy with how that's been going 21:44:23 <clayg> ok, cool! 21:44:25 <clayg> thanks timburke 21:44:35 <timburke> np 21:44:44 <timburke> #topic open discussion 21:44:52 <timburke> anything else we'd like to bring up today? 21:46:06 <mattoliverau> nope 21:47:37 <timburke> all right 21:47:49 <timburke> thank you all for coming, and thank you for working on swift! 21:47:57 <timburke> #endmeeting