| klamath | I noticed that new OS's format XFS with CRC=1 and reflink=1, is there any best practices on this front? im assuming if we want the disk mountable past 2038 we should include these options for bigtime but is there any performance implications noticed on these two flags yet? | 16:42 |
|---|---|---|
| opendevreview | Alistair Coles proposed openstack/swift master: sq: call _calculate_slo_attributes directly when needed https://review.opendev.org/c/openstack/swift/+/896696 | 18:15 |
| opendevreview | Clay Gerrard proposed openstack/swift master: Refactor SLO size/etag sysmeta tests https://review.opendev.org/c/openstack/swift/+/896466 | 19:09 |
| opendevreview | Clay Gerrard proposed openstack/swift master: slo: refactor GET/HEAD response handling https://review.opendev.org/c/openstack/swift/+/893578 | 19:09 |
| opendevreview | Clay Gerrard proposed openstack/swift master: slo: part-number=N query paramater support https://review.opendev.org/c/openstack/swift/+/894570 | 19:09 |
| opendevreview | Clay Gerrard proposed openstack/swift master: SLO drain manifests https://review.opendev.org/c/openstack/swift/+/896721 | 19:09 |
| timburke | klamath, i don't know that anyone's looked into it specifically from swift's perspective. i'd expect the crc to be reasonably cheap, but i'm not at all sure what to expect out of reflink. we'd be interested in hearing about any performance differences you see, though! | 19:40 |
| opendevreview | Merged openstack/swift master: diskfile: Pass extension when creating DiskFileWriters https://review.opendev.org/c/openstack/swift/+/894820 | 20:11 |
| opendevreview | Merged openstack/swift master: proxy-server: fix node error limiting https://review.opendev.org/c/openstack/swift/+/890527 | 20:17 |
| opendevreview | Merged openstack/swift master: proxy-server: add replicated GET path tests https://review.opendev.org/c/openstack/swift/+/895802 | 20:17 |
| opendevreview | Clay Gerrard proposed openstack/swift master: slo: refactor GET/HEAD response handling https://review.opendev.org/c/openstack/swift/+/893578 | 20:44 |
| opendevreview | Clay Gerrard proposed openstack/swift master: slo: part-number=N query paramater support https://review.opendev.org/c/openstack/swift/+/894570 | 20:44 |
| timburke | #startmeeting swift | 21:00 |
| opendevmeet | Meeting started Wed Sep 27 21:00:49 2023 UTC and is due to finish in 60 minutes. The chair is timburke. Information about MeetBot at http://wiki.debian.org/MeetBot. | 21:00 |
| opendevmeet | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 21:00 |
| opendevmeet | The meeting name has been set to 'swift' | 21:00 |
| timburke | who's here for the swift team meeting? | 21:00 |
| acoles | o/ | 21:01 |
| zaitcev | o/ | 21:01 |
| timburke | as usual, the agenda's at | 21:02 |
| timburke | #link https://wiki.openstack.org/wiki/Meetings/Swift | 21:02 |
| timburke | (though i forgot to update the date) | 21:02 |
| timburke | first up | 21:03 |
| timburke | #topic vPTG | 21:03 |
| timburke | only a month away now! | 21:03 |
| timburke | please vote on meeting times; i'll plan on selecting and scheduling them this coming week | 21:03 |
| timburke | #link https://framadate.org/liKzqQw5lZQW8Pvf | 21:03 |
| timburke | and please add topics to the etherpad -- i haven't yet, either, but i know we have interesting topics to discuss! | 21:04 |
| timburke | #link https://etherpad.opendev.org/p/swift-ptg-caracal | 21:04 |
| timburke | any questions/comments/ideas around the ptg? | 21:05 |
| zaitcev | Well, it's a vPTG. How about that. | 21:06 |
| timburke | all right, next up | 21:08 |
| timburke | #topic py312 | 21:08 |
| timburke | python keeps moving forward! and it keeps removing things :P | 21:09 |
| zaitcev | Le sigh. | 21:10 |
| timburke | 3.12.0 will be coming out next week, and i wanted to try to stay ahead of things a little, so i started testing out the latest rc | 21:10 |
| timburke | dependencies look like they'll be a little interesting... | 21:11 |
| timburke | greenlet seems on top of it -- they've got their own rc out that seems to work on py312 | 21:11 |
| timburke | eventlet still doesn't technically support py310... though it seems to work well enough on py310 and py311 for our purposes | 21:12 |
| timburke | (and least, for a development environment -- there may well be some issues lurking that you'd notice in prod) | 21:13 |
| timburke | master eventlet on py312, though, doesn't work -- there are some ssl helpers that got removed, some changes to _thread that need to get mirrored in eventlet.green.thread... | 21:14 |
| timburke | and keystonemiddleware (at least) is using pkg_resources, which has been removed | 21:15 |
| timburke | coming back to code in swift: distutils has also been removed | 21:17 |
| zaitcev | Next step, port Swift from eventlet to Gunicorn. | 21:17 |
| zaitcev | Ugh | 21:17 |
| timburke | which is the real story behind the chain ending at | 21:17 |
| timburke | #link https://review.opendev.org/c/openstack/swift/+/896473 | 21:17 |
| timburke | there's a handful of patches all removing our use of distutils; most of them are only in tests, only that last one touches code under swift/ | 21:18 |
| acoles | keystonemiddleware is out of tree right? | 21:19 |
| timburke | zaitcev, you say that -- i wonder if we might be able to really truly drop py2, then switch it all out to async stuff... | 21:19 |
| timburke | acoles, yeah -- there are some unit tests that cause us to import it, though | 21:20 |
| acoles | ah | 21:20 |
| zaitcev | acoles: if you're willing to do that, there's a far wider selection that just gunicorn. | 21:20 |
| timburke | https://github.com/openstack/swift/blob/master/test/unit/common/middleware/s3api/test_s3api.py#L38 in particular | 21:21 |
| timburke | zaitcev, and as a happy side-effect, own the wsgi server/http parsing part | 21:23 |
| timburke | we've been doing that more and more anyway; may as well rip off the bandaid, right? | 21:23 |
| acoles | 🩹 | 21:24 |
| timburke | two bits of good news on the py312 front: | 21:25 |
| timburke | first, we already got rid of our use of pkg_resources under swift/ -- though a quick grep says we still need to fix up test/unit/obj/test_auditor.py | 21:26 |
| timburke | second, there didn't seem to be any show-stoppers like the segfault we discovered when we started testing under py311 | 21:27 |
| zaitcev | How unfortunate, I bet it's what loads watchers. | 21:29 |
| zaitcev | I can look at that. | 21:29 |
| zaitcev | Meanwhile, acoles would do great to look at https://review.opendev.org/c/openstack/swift/+/787656, which I updated recently. | 21:29 |
| timburke | i think that's all i've got to say about py312 -- if you've got cycles to spare helping out, i could use some reviews on that chain (they should all be pretty quick!) and we ought to stamp out that last `import pkg_resources` that isn't in some `try/except ImportError` handling | 21:30 |
| acoles | ack | 21:30 |
| timburke | zaitcev, yes and no -- see https://github.com/openstack/swift/blob/master/swift/common/utils/__init__.py#L5847-L5852 -- on newer python, we're fine using importlib | 21:31 |
| timburke | all right, last topic i had | 21:33 |
| timburke | #topic stable gates | 21:33 |
| timburke | recently, i wanted to backport a patch to wallaby to fix an issue noticed by klamath | 21:33 |
| timburke | then was sad to see that most (all?) of the py2 jobs failed | 21:34 |
| timburke | this got me looking at the broader state of stable gates in general | 21:34 |
| timburke | 2023.1 and zed were fine; yoga and xena were also broken | 21:35 |
| timburke | yoga is now fixed | 21:35 |
| timburke | #link https://review.opendev.org/c/openstack/swift/+/896282 | 21:35 |
| timburke | but i want to fix xena and wallaby as well -- so i'll probably be working on that over the next week | 21:36 |
| timburke | once those are working again (and i get that patch backported), i'll start checking even older branches | 21:36 |
| timburke | all right, that's all i've got | 21:37 |
| timburke | #topic open discussion | 21:37 |
| timburke | anything else we should bring up this week? | 21:37 |
| zaitcev | \()/ | 21:38 |
| timburke | all right, i'll let you guys go then | 21:41 |
| timburke | thank you all for coming, and thank you for working on swift! | 21:42 |
| timburke | #endmeeting | 21:42 |
| opendevmeet | Meeting ended Wed Sep 27 21:42:10 2023 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 21:42 |
| opendevmeet | Minutes: https://meetings.opendev.org/meetings/swift/2023/swift.2023-09-27-21.00.html | 21:42 |
| opendevmeet | Minutes (text): https://meetings.opendev.org/meetings/swift/2023/swift.2023-09-27-21.00.txt | 21:42 |
| opendevmeet | Log: https://meetings.opendev.org/meetings/swift/2023/swift.2023-09-27-21.00.log.html | 21:42 |
| timburke | thanks for all the reviews, zaitcev! | 22:35 |
| opendevreview | Jianjian Huo proposed openstack/swift master: Proxy: Use shard-format for auto cont GETs and updating https://review.opendev.org/c/openstack/swift/+/895602 | 22:38 |
| opendevreview | Tim Burke proposed openstack/swift stable/xena: [stable-only] Fix gate https://review.opendev.org/c/openstack/swift/+/896322 | 22:41 |
| opendevreview | ASHWIN A NAIR proposed openstack/swift master: slo: refactor GET/HEAD response handling https://review.opendev.org/c/openstack/swift/+/893578 | 23:40 |
| opendevreview | Merged openstack/swift master: tests: swiftclient supports insecure https://review.opendev.org/c/openstack/swift/+/896470 | 23:41 |
| opendevreview | Merged openstack/swift master: tests: boto is always <3.0 https://review.opendev.org/c/openstack/swift/+/896471 | 23:41 |
| opendevreview | Merged openstack/swift master: tests: Stop using distutils.dir_util.mkpath https://review.opendev.org/c/openstack/swift/+/896472 | 23:41 |
| opendevreview | ASHWIN A NAIR proposed openstack/swift master: slo: refactor GET/HEAD response handling https://review.opendev.org/c/openstack/swift/+/893578 | 23:45 |
| opendevreview | Merged openstack/swift master: Swap find_executable for which https://review.opendev.org/c/openstack/swift/+/896473 | 23:50 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!