| *** diablo_rojo has quit IRC | 00:01 | |
| *** ociuhandu has joined #openstack-meeting | 00:04 | |
| *** jmasud has joined #openstack-meeting | 00:04 | |
| *** deardooley has quit IRC | 00:09 | |
| *** ociuhandu has quit IRC | 00:12 | |
| *** mlavalle has quit IRC | 00:23 | |
| *** rfolco has quit IRC | 00:24 | |
| *** martial has quit IRC | 00:30 | |
| *** diablo_rojo has joined #openstack-meeting | 00:32 | |
| *** armax has quit IRC | 00:49 | |
| *** Liang__ has joined #openstack-meeting | 00:59 | |
| *** jmasud has quit IRC | 01:09 | |
| *** yamamoto has joined #openstack-meeting | 01:49 | |
| *** ociuhandu has joined #openstack-meeting | 01:50 | |
| *** ociuhandu has quit IRC | 02:03 | |
| *** ircuser-1 has joined #openstack-meeting | 02:05 | |
| *** apetrich has quit IRC | 02:09 | |
| *** rh-jelabarre has quit IRC | 02:29 | |
| *** rh-jelabarre has joined #openstack-meeting | 02:30 | |
| *** jmasud has joined #openstack-meeting | 02:40 | |
| *** yamamoto has quit IRC | 02:48 | |
| *** jmasud has quit IRC | 02:49 | |
| *** hyunsikyang has joined #openstack-meeting | 02:50 | |
| *** gyee has quit IRC | 02:52 | |
| *** hyunsikyang__ has quit IRC | 02:54 | |
| *** ykatabam has joined #openstack-meeting | 02:58 | |
| *** jmasud has joined #openstack-meeting | 03:01 | |
| *** ianychoi_ has joined #openstack-meeting | 03:08 | |
| *** yamamoto has joined #openstack-meeting | 03:08 | |
| *** ianychoi has quit IRC | 03:10 | |
| *** psachin has joined #openstack-meeting | 03:21 | |
| *** ociuhandu has joined #openstack-meeting | 03:39 | |
| *** ircuser-1 has quit IRC | 03:40 | |
| *** ociuhandu has quit IRC | 03:53 | |
| *** jmasud has quit IRC | 03:56 | |
| *** masahito has quit IRC | 03:57 | |
| *** rh-jelabarre has quit IRC | 03:59 | |
| *** yamamoto has quit IRC | 04:10 | |
| *** yamamoto has joined #openstack-meeting | 04:25 | |
| *** hyunsikyang__ has joined #openstack-meeting | 04:27 | |
| *** hyunsikyang has quit IRC | 04:31 | |
| *** evrardjp has quit IRC | 04:35 | |
| *** evrardjp has joined #openstack-meeting | 04:35 | |
| *** ociuhandu has joined #openstack-meeting | 04:39 | |
| *** masahito has joined #openstack-meeting | 04:43 | |
| *** masahito has quit IRC | 04:47 | |
| *** jmasud has joined #openstack-meeting | 04:47 | |
| *** ociuhandu has quit IRC | 04:48 | |
| *** diablo_rojo has quit IRC | 04:51 | |
| *** dklyle has quit IRC | 05:02 | |
| *** yamamoto has quit IRC | 05:15 | |
| *** jmasud has quit IRC | 05:15 | |
| *** bnemec has quit IRC | 05:15 | |
| *** slaweq has joined #openstack-meeting | 05:19 | |
| *** yamamoto has joined #openstack-meeting | 05:21 | |
| *** tetsuro has joined #openstack-meeting | 05:28 | |
| *** hyunsikyang has joined #openstack-meeting | 05:29 | |
| *** jmasud has joined #openstack-meeting | 05:32 | |
| *** tetsuro has quit IRC | 05:32 | |
| *** hyunsikyang__ has quit IRC | 05:32 | |
| *** links has joined #openstack-meeting | 05:40 | |
| *** hugokuo has joined #openstack-meeting | 05:43 | |
| *** timburke has joined #openstack-meeting | 05:44 | |
| *** ysandeep|away is now known as ysandeep | 05:45 | |
| *** rbudden has quit IRC | 05:54 | |
| *** ysandeep is now known as ysandeep|brb | 05:55 | |
| *** ociuhandu has joined #openstack-meeting | 06:01 | |
| *** slaweq has quit IRC | 06:06 | |
| *** ociuhandu has quit IRC | 06:15 | |
| *** hyunsikyang__ has joined #openstack-meeting | 06:16 | |
| *** jmasud has quit IRC | 06:17 | |
| *** jmasud has joined #openstack-meeting | 06:19 | |
| *** johanssone has joined #openstack-meeting | 06:19 | |
| *** jmasud has quit IRC | 06:23 | |
| *** jmasud has joined #openstack-meeting | 06:26 | |
| *** maciejjozefczyk has joined #openstack-meeting | 06:28 | |
| *** rpittau|afk is now known as rpittau | 06:32 | |
| *** ociuhandu has joined #openstack-meeting | 06:36 | |
| *** slaweq has joined #openstack-meeting | 06:40 | |
| *** ysandeep|brb is now known as ysandeep | 06:51 | |
| *** ttsiouts has joined #openstack-meeting | 06:51 | |
| *** brinzhang has quit IRC | 06:53 | |
| *** dustinc has quit IRC | 06:53 | |
| *** brinzhang has joined #openstack-meeting | 06:53 | |
| *** brinzhang has quit IRC | 06:56 | |
| *** brinzhang has joined #openstack-meeting | 06:56 | |
| *** belmoreira has joined #openstack-meeting | 06:57 | |
| *** ociuhandu has quit IRC | 06:59 | |
| *** bbowen has quit IRC | 07:00 | |
| *** bbowen has joined #openstack-meeting | 07:00 | |
| *** larainema has joined #openstack-meeting | 07:02 | |
| *** links has quit IRC | 07:03 | |
| *** jmasud has quit IRC | 07:03 | |
| *** links has joined #openstack-meeting | 07:04 | |
| *** yamamoto has quit IRC | 07:07 | |
| *** yamamoto has joined #openstack-meeting | 07:09 | |
| *** ircuser-1 has joined #openstack-meeting | 07:10 | |
| *** brinzhang has quit IRC | 07:13 | |
| *** witek has joined #openstack-meeting | 07:13 | |
| *** brinzhang has joined #openstack-meeting | 07:13 | |
| *** takamatsu has joined #openstack-meeting | 07:16 | |
| *** apetrich has joined #openstack-meeting | 07:24 | |
| *** rcernin has quit IRC | 07:27 | |
| *** ociuhandu has joined #openstack-meeting | 07:29 | |
| *** ysandeep is now known as ysandeep|lunch | 07:42 | |
| *** ralonsoh has joined #openstack-meeting | 07:48 | |
| *** masahito has joined #openstack-meeting | 07:50 | |
| *** masahito has quit IRC | 07:55 | |
| *** ysandeep|lunch is now known as ysandeep | 08:26 | |
| *** e0ne has joined #openstack-meeting | 08:38 | |
| *** jmasud has joined #openstack-meeting | 08:45 | |
| *** ociuhandu has quit IRC | 08:50 | |
| *** ttsiouts has quit IRC | 08:50 | |
| *** ociuhandu has joined #openstack-meeting | 08:50 | |
| *** ociuhandu has quit IRC | 08:50 | |
| *** ttsiouts has joined #openstack-meeting | 08:51 | |
| *** ociuhandu has joined #openstack-meeting | 08:53 | |
| *** yamamoto has quit IRC | 08:55 | |
| *** yamamoto has joined #openstack-meeting | 09:09 | |
| *** yamamoto has quit IRC | 09:10 | |
| *** yamamoto has joined #openstack-meeting | 09:10 | |
| *** ykatabam has quit IRC | 09:23 | |
| *** ociuhandu has quit IRC | 09:47 | |
| *** ociuhandu has joined #openstack-meeting | 09:48 | |
| *** ociuhandu has quit IRC | 09:53 | |
| *** ykatabam has joined #openstack-meeting | 09:59 | |
| *** armax has joined #openstack-meeting | 10:01 | |
| *** brinzhang_ has quit IRC | 10:11 | |
| *** ociuhandu has joined #openstack-meeting | 10:12 | |
| *** ykatabam has quit IRC | 10:16 | |
| *** Liang__ has quit IRC | 10:25 | |
| *** brinzhang has quit IRC | 10:29 | |
| *** rpittau is now known as rpittau|bbl | 10:55 | |
| *** Yuli has joined #openstack-meeting | 11:03 | |
| *** Yuli has joined #openstack-meeting | 11:05 | |
| *** ociuhandu has quit IRC | 11:06 | |
| *** ociuhandu has joined #openstack-meeting | 11:07 | |
| *** Yuli has left #openstack-meeting | 11:08 | |
| *** ykatabam has joined #openstack-meeting | 11:09 | |
| *** yamamoto has quit IRC | 11:11 | |
| *** ociuhandu has quit IRC | 11:13 | |
| *** tetsuro has joined #openstack-meeting | 11:16 | |
| *** yamamoto has joined #openstack-meeting | 11:22 | |
| *** jmasud has quit IRC | 11:23 | |
| *** ttsiouts has quit IRC | 11:25 | |
| *** ykatabam has quit IRC | 11:26 | |
| *** ttsiouts has joined #openstack-meeting | 11:27 | |
| *** yamamoto has quit IRC | 11:29 | |
| *** ysandeep is now known as ysandeep|coffee | 11:29 | |
| *** raildo has joined #openstack-meeting | 11:39 | |
| *** ttsiouts has quit IRC | 11:43 | |
| *** rh-jelabarre has joined #openstack-meeting | 11:45 | |
| *** raildo has quit IRC | 11:49 | |
| *** raildo has joined #openstack-meeting | 11:50 | |
| *** tetsuro has quit IRC | 11:52 | |
| *** tetsuro has joined #openstack-meeting | 11:53 | |
| *** tetsuro has quit IRC | 11:54 | |
| *** edagawa_kc has quit IRC | 11:55 | |
| *** yamamoto has joined #openstack-meeting | 12:04 | |
| *** ociuhandu has joined #openstack-meeting | 12:08 | |
| *** takahashi-tsc has quit IRC | 12:10 | |
| *** tsukasasa has quit IRC | 12:10 | |
| *** rfolco has joined #openstack-meeting | 12:14 | |
| *** ttsiouts has joined #openstack-meeting | 12:16 | |
| *** ttsiouts has quit IRC | 12:21 | |
| *** rpittau|bbl is now known as rpittau | 12:31 | |
| *** yamamoto has quit IRC | 12:36 | |
| *** yamamoto has joined #openstack-meeting | 12:42 | |
| *** yamamoto has quit IRC | 12:43 | |
| *** ociuhandu has quit IRC | 12:44 | |
| *** ociuhandu has joined #openstack-meeting | 12:45 | |
| *** ociuhandu has quit IRC | 12:49 | |
| *** ttsiouts has joined #openstack-meeting | 12:55 | |
| *** links has quit IRC | 12:59 | |
| *** ociuhandu has joined #openstack-meeting | 12:59 | |
| *** rbudden has joined #openstack-meeting | 13:13 | |
| *** yamamoto has joined #openstack-meeting | 13:21 | |
| *** ttsiouts has quit IRC | 13:29 | |
| *** yamamoto has quit IRC | 13:29 | |
| *** yamamoto has joined #openstack-meeting | 13:30 | |
| *** moguimar has joined #openstack-meeting | 13:30 | |
| *** psachin has quit IRC | 13:36 | |
| *** TrevorV has joined #openstack-meeting | 13:52 | |
| *** Liang__ has joined #openstack-meeting | 13:53 | |
| *** bnemec has joined #openstack-meeting | 13:56 | |
| *** dklyle has joined #openstack-meeting | 13:57 | |
| *** Liang__ is now known as LiangFang | 13:57 | |
| *** andrebeltrami has joined #openstack-meeting | 13:59 | |
| *** ttsiouts has joined #openstack-meeting | 14:00 | |
| *** yamamoto has quit IRC | 14:06 | |
| *** ttsiouts has quit IRC | 14:12 | |
| *** yamamoto has joined #openstack-meeting | 14:23 | |
| *** yamamoto has quit IRC | 14:23 | |
| *** priteau has joined #openstack-meeting | 14:23 | |
| *** yamamoto has joined #openstack-meeting | 14:24 | |
| *** yamamoto has quit IRC | 14:28 | |
| *** yamamoto has joined #openstack-meeting | 14:41 | |
| *** yamamoto has quit IRC | 14:41 | |
| *** mlavalle has joined #openstack-meeting | 14:41 | |
| *** yamamoto has joined #openstack-meeting | 14:42 | |
| *** zbitter is now known as zaneb | 14:44 | |
| *** yamamoto has quit IRC | 14:46 | |
| *** ociuhandu has quit IRC | 14:52 | |
| *** trident has quit IRC | 15:04 | |
| *** trident has joined #openstack-meeting | 15:05 | |
| *** LiangFang has quit IRC | 15:05 | |
| *** ociuhandu has joined #openstack-meeting | 15:15 | |
| *** yamamoto has joined #openstack-meeting | 15:18 | |
| *** jmasud has joined #openstack-meeting | 15:49 | |
| *** gyee has joined #openstack-meeting | 15:53 | |
| *** TrevorV has quit IRC | 15:59 | |
| *** ociuhandu has quit IRC | 16:07 | |
| *** ociuhandu has joined #openstack-meeting | 16:08 | |
| *** ysandeep|coffee is now known as ysandeep|away | 16:12 | |
| *** ociuhandu has quit IRC | 16:13 | |
| *** witek has quit IRC | 16:13 | |
| *** jmasud has quit IRC | 16:18 | |
| *** jmasud has joined #openstack-meeting | 16:30 | |
| *** jamesden_ has joined #openstack-meeting | 16:30 | |
| *** jamesdenton has quit IRC | 16:31 | |
| *** evrardjp has quit IRC | 16:35 | |
| *** evrardjp has joined #openstack-meeting | 16:35 | |
| *** rpittau is now known as rpittau|afk | 16:37 | |
| *** ociuhandu has joined #openstack-meeting | 17:12 | |
| *** ociuhandu has quit IRC | 17:17 | |
| *** maciejjozefczyk has quit IRC | 17:22 | |
| *** apetrich has quit IRC | 17:22 | |
| *** priteau has quit IRC | 17:32 | |
| *** diablo_rojo has joined #openstack-meeting | 17:33 | |
| *** ralonsoh has quit IRC | 17:41 | |
| *** jmasud has quit IRC | 17:58 | |
| *** ociuhandu has joined #openstack-meeting | 18:00 | |
| *** e0ne has quit IRC | 18:07 | |
| *** dmacpher has quit IRC | 18:08 | |
| *** armstrong has joined #openstack-meeting | 18:10 | |
| *** dmacpher has joined #openstack-meeting | 18:12 | |
| *** dmacpher has quit IRC | 18:19 | |
| *** dmacpher has joined #openstack-meeting | 18:22 | |
| *** dmacpher has quit IRC | 18:27 | |
| *** dmacpher has joined #openstack-meeting | 18:30 | |
| *** ociuhandu has quit IRC | 18:31 | |
| *** ociuhandu has joined #openstack-meeting | 18:33 | |
| *** Lucas_Gray has joined #openstack-meeting | 18:33 | |
| *** ociuhandu has quit IRC | 18:37 | |
| *** jmasud has joined #openstack-meeting | 18:41 | |
| *** ociuhandu has joined #openstack-meeting | 19:12 | |
| *** ociuhandu has quit IRC | 19:25 | |
| *** ociuhandu has joined #openstack-meeting | 19:26 | |
| *** ociuhandu has quit IRC | 19:32 | |
| *** priteau has joined #openstack-meeting | 19:44 | |
| *** jmasud has quit IRC | 19:48 | |
| *** jmasud has joined #openstack-meeting | 19:51 | |
| *** jmasud has quit IRC | 19:51 | |
| *** priteau has quit IRC | 19:51 | |
| *** jmasud has joined #openstack-meeting | 19:54 | |
| *** belmoreira has quit IRC | 19:58 | |
| *** armstrong has quit IRC | 20:02 | |
| *** maciejjozefczyk has joined #openstack-meeting | 20:27 | |
| *** ykatabam has joined #openstack-meeting | 20:46 | |
| *** patchbot has joined #openstack-meeting | 20:58 | |
| timburke | #startmeeting swift | 21:00 |
|---|---|---|
| openstack | Meeting started Wed Apr 29 21:00:13 2020 UTC and is due to finish in 60 minutes. The chair is timburke. Information about MeetBot at http://wiki.debian.org/MeetBot. | 21:00 |
| openstack | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 21:00 |
| *** openstack changes topic to " (Meeting topic: swift)" | 21:00 | |
| openstack | The meeting name has been set to 'swift' | 21:00 |
| timburke | who's here for the swift meeting? | 21:00 |
| seongsoocho | o/ | 21:00 |
| kota_ | hi | 21:00 |
| rledisez | hi o/ | 21:00 |
| tdasilva | o/ | 21:00 |
| clayg | o/ | 21:00 |
| *** ociuhandu has joined #openstack-meeting | 21:01 | |
| *** slaweq has quit IRC | 21:01 | |
| timburke | agenda's at https://wiki.openstack.org/wiki/Meetings/Swift | 21:01 |
| timburke | first up | 21:01 |
| timburke | #topic swift-get-nodes | 21:01 |
| *** openstack changes topic to "swift-get-nodes (Meeting topic: swift)" | 21:01 | |
| *** alecuyer has joined #openstack-meeting | 21:02 | |
| clayg | so i haven't really looked at this - i don't see how it ever worked with paths with spaces? Do you just like wrap it all in "quotes or something" | 21:02 |
| timburke | so i realized earlier this week -- we never made it so you could find %00versions containers with swift-get-nodes! | 21:02 |
| alecuyer | o/ sorry im late | 21:02 |
| mattoliverau | o/ | 21:03 |
| timburke | clayg, yeah, with spaces you'd do something like `swift-get-nodes ring.gz 'acct/cont/obj with spaces'` | 21:03 |
| clayg | oic, but that strategy won't work with the null character | 21:03 |
| timburke | but you can't really do that with null bytes -- pretty sure things break down because of how args get sent to C programs | 21:04 |
| timburke | which is a *bit much* to try to hack around ;-) | 21:04 |
| timburke | #link https://launchpad.net/bugs/1875734 | 21:05 |
| openstack | Launchpad bug 1875734 in OpenStack Object Storage (swift) "swift-get-nodes cannot be used with %00versions containers" [High,In progress] | 21:05 |
| clayg | yeah, ok so un-url-encoded sort of kind of worked fro some names on accident | 21:05 |
| timburke | yup | 21:05 |
| timburke | so i wrote a patch to have it accept percent-encoded paths, like acct/cont/obj%20with%20spaces | 21:06 |
| timburke | #link https://review.opendev.org/#/c/724141/ | 21:06 |
| patchbot | patch 724141 - swift - swift-get-nodes: Allow users to specify either quo... - 1 patch set | 21:06 |
| clayg | I think it should try and take quoted names by default - with a `--for-some-reason-i-did-not-quote-this` flag for when you want an object named `%beef` and can't be bothered to type `%25beef` | 21:06 |
| *** ociuhandu has quit IRC | 21:06 | |
| clayg | I feel like for most of the cases where it was working it will still work - if you "unquote" "object with spaces" it turns into "object with spaces" doesn't it? what's the big deal? | 21:07 |
| timburke | out of an abundance of caution, i kept the old behavior and have you opt-in to quoted paths with --quoted (or just -Q) | 21:07 |
| *** jmasud has quit IRC | 21:07 | |
| clayg | then we don't need to "deprecate" anything - we just get better (and leave an escape hatch if somehow someone was scripting this and figured out a way to make it work) | 21:07 |
| clayg | timburke: I'd advocate for the opposite instead, that's why I wanted to get other opinions | 21:08 |
| timburke | clayg, yeah, i was mainly worried about the %beef sort of case | 21:08 |
| timburke | i'd be perfectly happy to just work the one way :D | 21:08 |
| *** rbudden has quit IRC | 21:08 | |
| timburke | what's everybody else think? | 21:08 |
| clayg | perhaps I'm being overly optimistic about most of the time unquote of unquoted names is the identity function - obviously in SOME cases that's not true | 21:09 |
| *** slaweq has joined #openstack-meeting | 21:09 | |
| rledisez | It seems reasonable to move fast as it's not an API and I doubt everybody did some scripting around that tool. Just displaying a proper warning after the default changed for sometime should be enough | 21:09 |
| clayg | 100%beef is the canonical example of the case where the output would unexpectedly change | 21:10 |
| timburke | fwiw, i only emit the warning when there is a difference | 21:10 |
| rledisez | but the warning shoud be obvious. I can already tell I wouldn't see it because I'm never reading the first lines of the output... | 21:10 |
| clayg | hahah | 21:11 |
| timburke | heh, fair enough -- i probably ought to move it to the bottom we continue down this route | 21:11 |
| rledisez | yeah, I think it gets more chance to be see by operators | 21:11 |
| mattoliverau | we need like a cli equiv of <blink> tag :P | 21:12 |
| clayg | @mattoliverau hahaha | 21:13 |
| timburke | k -- i'll respin with no --quoted option (because it'll be the default behavior) and the warning moved to the bottom | 21:13 |
| *** jmasud has joined #openstack-meeting | 21:13 | |
| *** slaweq has quit IRC | 21:14 | |
| timburke | and just for mattoliverau, i'll probably spend more time than i should digging through ANSI escape sequence docs ;-) | 21:14 |
| clayg | @timburke but what will the warning say exactly? If they're doing the right thing and quoting paths that need it .... won't they get a WARNING: I did exactly what you expected | 21:14 |
| mattoliverau | lol | 21:14 |
| clayg | oh no mattoliverau | 21:14 |
| timburke | hrm. good point :-/ | 21:14 |
| clayg | i think just put it in the release notes | 21:15 |
| timburke | oh, it's for-sure going in the release notes. and i know rledisez would see it that way; do we think that's true of most operators? | 21:15 |
| mattoliverau | snowman says "warning .." :P | 21:16 |
| rledisez | I hope so… :) | 21:16 |
| timburke | also, i can tell that i'm going to hate poking at old swift clusters after this ;-) | 21:17 |
| rledisez | after that, it's just about reminding the change. so many years of passing crazy args to the tool, can't be erased in a blink ;) | 21:17 |
| clayg | i feel like I mostly get these paths from loglines - there they're probably "double quoted" 🙄 | 21:18 |
| *** zaitcev has joined #openstack-meeting | 21:18 | |
| *** raildo has quit IRC | 21:18 | |
| timburke | the only consistency is inconsistency! | 21:19 |
| clayg | eventaully | 21:19 |
| kota_ | lol | 21:19 |
| mattoliverau | lol | 21:20 |
| timburke | well... i'll try something. at least the change is publicized! i *do* want something like this available in the near term, though, 'cause it's not great that there are containers and objects getting created that we can't easily find | 21:20 |
| timburke | #topic S3 MPU deletes | 21:20 |
| *** openstack changes topic to "S3 MPU deletes (Meeting topic: swift)" | 21:20 | |
| timburke | earlier this week, we had an availability issue because of request amplifications coming from deleting S3 MPUs | 21:21 |
| mattoliverau | oh, opps | 21:22 |
| timburke | there were a bunch of confounding factors (old-style versioning being enabled on the segments container, and everything being in an EC policy) | 21:22 |
| timburke | but even setting those aside, you can have a single client DELETE kick off like 1000 subrequests to clean up the segments :-( | 21:23 |
| kota_ | :( | 21:23 |
| timburke | if the client has a short-ish timeout (after all, AWS responds to deletes real quick), they'll likely retry and make the problem worse | 21:24 |
| timburke | i don't really have any action items for this -- mostly just highlighting the problem so people know about it | 21:25 |
| rledisez | i guess it can happen too with bulk delete? | 21:25 |
| timburke | yeah, though bulk delete has the advantage of being able to dribble out bytes to keep the connection alive | 21:26 |
| rledisez | right, so it cannot happen by accident (but still it can be intentional). ratelimit middleware could be a protection for that maybe. I don't remember exactly how it ratelimit (sleeping or returning a 4xx response?) | 21:28 |
| tdasilva | async_delete ? | 21:28 |
| timburke | rledisez, yeah, i was just thinking about ratelimiting... | 21:28 |
| clayg | yeah, it's gunna have to be something like that | 21:28 |
| timburke | tdasilva, the async delete stuff is definitely on my mind :-) i think long-term, we want a | 21:28 |
| timburke | new large object type that behaves much more like MPUs | 21:29 |
| timburke | and async-deletes will likely be part of that solution | 21:29 |
| timburke | #topic PTG | 21:30 |
| *** openstack changes topic to "PTG (Meeting topic: swift)" | 21:30 | |
| timburke | one more reminder about the call for topics | 21:30 |
| timburke | #link https://etherpad.openstack.org/p/swift-ptg-victoria | 21:30 |
| timburke | i think i'm also supposed to sign up for some timeslots or something -- i really need to make sure i figure out the plan for this "virtual PTG" thing this coming week | 21:31 |
| timburke | we're only like a month away! | 21:31 |
| clayg | wow that's nuts | 21:31 |
| timburke | ok, on to ongoing-work! | 21:32 |
| timburke | #topic lots of small files | 21:32 |
| *** openstack changes topic to "lots of small files (Meeting topic: swift)" | 21:32 | |
| timburke | alecuyer, i saw a patch for an updated key format | 21:32 |
| *** rcernin has joined #openstack-meeting | 21:32 | |
| alecuyer | Yes, I pushed a WIP patch for that here : https://review.opendev.org/#/c/723609/ | 21:33 |
| patchbot | patch 723609 - swift (feature/losf) - New key format for objects in the index-server - 1 patch set | 21:33 |
| alecuyer | So, still need to work on it, but I'm happy that I'm removing more code than adding new code. It makes listdir() functions simpler | 21:33 |
| timburke | \o/ | 21:33 |
| alecuyer | and I guess that's about it unless you have questions :) | 21:33 |
| *** rcernin has quit IRC | 21:33 | |
| timburke | i'll try to take a look at it soon :) | 21:33 |
| timburke | i should also revisit my attempt to fix the losf gate job... never enough time lately | 21:34 |
| *** rcernin has joined #openstack-meeting | 21:34 | |
| timburke | #topic swiftclient socket leak | 21:35 |
| *** openstack changes topic to "swiftclient socket leak (Meeting topic: swift)" | 21:35 | |
| timburke | so a week or two ago i saw a message on the ML | 21:35 |
| timburke | #link http://lists.openstack.org/pipermail/openstack-discuss/2020-April/014221.html | 21:35 |
| timburke | about swiftclient 3.9.0 on py2 leaking sockets | 21:36 |
| timburke | #link https://bugs.launchpad.net/python-swiftclient/+bug/1873435 | 21:36 |
| openstack | Launchpad bug 1873435 in python-swiftclient "Established connection is never closed on Python 2.7" [High,In progress] - Assigned to Tim Burke (1-tim-z) | 21:36 |
| clayg | i was looking at that bug today - the patch was pretty invasive | 21:36 |
| timburke | well, part of it was that much of p 674320 probably should have been done like that from the get-go | 21:37 |
| patchbot | https://review.opendev.org/#/c/674320/ - python-swiftclient - Cleanup session on delete (MERGED) - 2 patch sets | 21:38 |
| timburke | clayg, would you prefer that p 721051 was more limited, just adding the `if not six.PY2:`? | 21:38 |
| patchbot | https://review.opendev.org/#/c/721051/ - python-swiftclient - Only add __del__ to HTTPConnection shim on py3 - 2 patch sets | 21:38 |
| *** andrebeltrami has quit IRC | 21:39 | |
| clayg | no, i mean not if that won't fix the bug | 21:39 |
| timburke | i think that'd probably be enough? not entirely sure, honestly. but it'd get us back to the prior behavior on py2, at least | 21:40 |
| timburke | i could have the explcit closing as a follow-up that we separately debate -- that'd be fine | 21:41 |
| zaitcev | Did anyone manage to reproduce the leak outside of that weird test environment? | 21:41 |
| zaitcev | I mean, the original leak that Schulz patched with __del__. | 21:42 |
| timburke | zaitcev, i remember being able to get that ResourceWarning just fine when i was reviewing it | 21:43 |
| zaitcev | I see. | 21:43 |
| timburke | might only get emitted on some versions of py3... i forget when they added that | 21:44 |
| zaitcev | You know | 21:44 |
| zaitcev | I was reading keystoneclient a bit in the past week. And it has client.sess=sess and sess.client=client | 21:45 |
| zaitcev | Not always, but sometimes, depending what options are supplied. | 21:45 |
| zaitcev | Using __del__ in that situation is pure madness. I only reviewed our own stuff, so I saw no loops and no excuse to -1 on that patch. | 21:46 |
| timburke | hmm... i thought the gc was pretty good about cycle detection these days, though? i'd just assumed there was some sort of situation that caused it to not get run *shrug* | 21:47 |
| timburke | anyway, something to be aware of -- i think we ought to decide on a fix and backport it to ussuri | 21:49 |
| timburke | #topic s3api + versioning | 21:49 |
| *** openstack changes topic to "s3api + versioning (Meeting topic: swift)" | 21:49 | |
| timburke | i've seen client trying to request specific versions of object on buckets that have never had versioning enabled | 21:50 |
| kota_ | interecting | 21:51 |
| kota_ | interesting | 21:51 |
| timburke | i *think* it came down to an AWS version id tagging along during a migration? not entirely sure | 21:51 |
| timburke | but s3api didn't handle it well -- object_versioning responds with a 400, then s3api responds 500 | 21:51 |
| timburke | but https://review.opendev.org/#/c/722552/ should fix it! | 21:51 |
| patchbot | patch 722552 - swift - s3api: Check whether versioning is enabled more - 3 patch sets | 21:51 |
| timburke | #topic versioning + container acls | 21:52 |
| *** openstack changes topic to "versioning + container acls (Meeting topic: swift)" | 21:52 | |
| timburke | i've also got a customer that turned on versioning for a container, then was sad that container acls basically stopped working | 21:52 |
| timburke | existing data was fine, but once they wrote new data (new name or an overwrite), reads that previously would have been successful all 403 | 21:53 |
| timburke | clayg mentioned it a bit in -swift before the meeting, but i think the approach i settled on in https://review.opendev.org/#/c/724393/ is reasonable? | 21:54 |
| patchbot | patch 724393 - swift - versioning: Have versioning symlinks make pre-auth... - 1 patch set | 21:54 |
| timburke | basically, since the user needed to be authorized to read the symlink object, we should grant the same authorization to following the link to the versions container | 21:55 |
| clayg | timburke: is there a significant difference between the wsgi utils preauth request helper and installing a noop lambda auth callback? | 21:55 |
| zaitcev | I can't find it now, but I seem to recall that there's some semantic thing that prevents interpreter to invoke __del__ on looped object. Or, rather, it will even try to free anything if one of the objects in the loop has __del__. But I don't remember and I can't find it. | 21:55 |
| timburke | zaitcev, oh, interesting. so the __del__ may not be a good idea on py3, either... | 21:55 |
| clayg | zaitcev: that sounds familiar to me! I always treated __del__ as dark magic! | 21:55 |
| zaitcev | I think I read about it in Bezeley's book. | 21:56 |
| timburke | clayg, i forget -- i'll look into it. i think there might be some other header-scrubbing that goes on, something like that | 21:56 |
| zaitcev | Anyway, sorry. We're on ACLs now. | 21:57 |
| timburke | sorry, running low on time :-) | 21:57 |
| *** ykatabam has quit IRC | 21:57 | |
| timburke | does that seem reasonable? it's only in the get path, and this still lets us define another way of saying "hey, allow these users to list and read old versions" | 21:59 |
| clayg | @timburke what were you saying about "popping" off that stack? | 21:59 |
| clayg | @timburke i'm sure it's reasonable but I need to look at it more - it's on my list for tomorrow | 22:00 |
| timburke | well, i kinda push some state (the allow-reserve-names header and the always-authed callback), then immediately after the app call i restore the previous state (so if the version was a symlink to some *other* container, the new request still needs to be authorized) | 22:01 |
| timburke | i started out just inserting the no-op callback where we used to add the header, but realized that'd be open to abuse | 22:02 |
| timburke | anyway, thanks for thinking about it! | 22:02 |
| timburke | thank you all for coming, and thank you for working on swift! | 22:02 |
| timburke | #endmeeting | 22:02 |
| *** openstack changes topic to "OpenStack Meetings || https://wiki.openstack.org/wiki/Meetings/" | 22:02 | |
| openstack | Meeting ended Wed Apr 29 22:02:54 2020 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 22:02 |
| openstack | Minutes: http://eavesdrop.openstack.org/meetings/swift/2020/swift.2020-04-29-21.00.html | 22:02 |
| openstack | Minutes (text): http://eavesdrop.openstack.org/meetings/swift/2020/swift.2020-04-29-21.00.txt | 22:02 |
| openstack | Log: http://eavesdrop.openstack.org/meetings/swift/2020/swift.2020-04-29-21.00.log.html | 22:03 |
| *** zaitcev has left #openstack-meeting | 22:03 | |
| *** patchbot has left #openstack-meeting | 22:03 | |
| *** jmasud has quit IRC | 22:04 | |
| *** mlavalle has quit IRC | 22:09 | |
| *** mlavalle has joined #openstack-meeting | 22:12 | |
| *** mlavalle has quit IRC | 22:20 | |
| *** jmasud has joined #openstack-meeting | 22:22 | |
| *** Lucas_Gray has quit IRC | 22:23 | |
| *** mlavalle has joined #openstack-meeting | 22:26 | |
| *** Adri2000 has quit IRC | 22:26 | |
| *** ociuhandu has joined #openstack-meeting | 22:36 | |
| *** jmasud has quit IRC | 22:38 | |
| *** ykatabam has joined #openstack-meeting | 22:40 | |
| *** ykatabam has quit IRC | 22:41 | |
| *** ykatabam has joined #openstack-meeting | 22:42 | |
| *** jmasud has joined #openstack-meeting | 22:43 | |
| *** ociuhandu has quit IRC | 22:58 | |
| *** Adri2000 has joined #openstack-meeting | 23:00 | |
| *** moguimar has quit IRC | 23:00 | |
| *** jmasud has quit IRC | 23:02 | |
| *** ociuhandu has joined #openstack-meeting | 23:33 | |
| *** armax has quit IRC | 23:43 | |
| *** armax has joined #openstack-meeting | 23:43 | |
| *** ociuhandu has quit IRC | 23:53 | |
| *** armax has quit IRC | 23:53 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!