21:00:43 <timburke> #startmeeting swift
21:00:44 <openstack> Meeting started Wed Jan 15 21:00:43 2020 UTC and is due to finish in 60 minutes.  The chair is timburke. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:00:45 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:00:47 <openstack> The meeting name has been set to 'swift'
21:00:52 <timburke> who's here for the swift meeting?
21:00:57 <alecuyer> o/
21:01:12 <tdasilva> eu
21:01:14 <kota_> hi
21:01:36 <seongsoocho_> o/
21:01:56 <timburke> as usual, the agenda's at
21:01:59 <timburke> #link https://wiki.openstack.org/wiki/Meetings/Swift
21:02:23 <timburke> i've been out of town for much of the last week, so i don't really have much more than updates to bring up
21:02:30 <timburke> #topic versioning
21:03:01 <timburke> i've been taking a look, and tdasilva has been busy addressing the last few issues we've found
21:03:35 <timburke> if you have any concerns about the new API, now's the time to check it out!
21:03:38 <timburke> #link https://review.opendev.org/#/c/682382/
21:03:38 <patchbot> patch 682382 - swift - New Object Versioning mode - 72 patch sets
21:03:46 <timburke> #link https://review.opendev.org/#/c/673682/
21:03:46 <patchbot> patch 673682 - swift - s3api: Implement object versioning API - 44 patch sets
21:04:15 <timburke> i hope to have both of those landed within the next week or so, with any lingering issues being addressed as follow-up patches
21:04:36 <timburke> tdasilva, anything i forgot to mention about that?
21:04:53 <timburke> and everyone else, anything else you'd like to know about it?
21:05:43 <tdasilva> nope, hoping to review a bit more of the swift client too, it would be nice to get that merged as closed together with the swift patches as possible
21:05:59 <timburke> agreed
21:06:09 <kota_> sounds reasonable
21:06:54 <timburke> i'll also go back and abandon some of the alternate approaches we looked at on our way to this solution
21:07:19 <timburke> oh yeah, and the client patch is
21:07:21 <timburke> #link https://review.opendev.org/#/c/691877/
21:07:22 <patchbot> patch 691877 - python-swiftclient - object versioning features - 9 patch sets
21:07:41 <timburke> #topic bulk regression
21:07:58 <timburke> #link https://launchpad.net/bugs/1857546
21:07:58 <openstack> Launchpad bug 1857546 in OpenStack Object Storage (swift) "Auto-extract looses X-Delete-At and X-Delete-After headers" [High,Confirmed]
21:08:11 <timburke> i saw rledisez put up an alternate fix
21:08:21 <timburke> #link https://review.opendev.org/#/c/701736/
21:08:21 <patchbot> patch 701736 - swift - Allow bulk to fwd some headers at tar extraction - 2 patch sets
21:08:57 <timburke> has anyone had a chance to look at either of the proposed fixes?
21:10:10 <kota_> that option 2 looks good to me.
21:10:15 <kota_> just looking now
21:10:39 <timburke> i was just about to ask if anyone had a preference for which way we fix it ;-)
21:11:18 <zaitcev> If we're talking about the versioning prompted by S3, I haven't reviewed it, sorry.
21:11:19 <timburke> i guess my only concern is all the other allowed_headers at https://github.com/openstack/swift/blob/2.23.0/etc/object-server.conf-sample#L135
21:12:27 <tdasilva> I haven't looked at either yet, but I'll try to take a look today
21:13:18 <kota_> ah, it's in object-server config...
21:13:48 <timburke> zaitcev, this is the bulk-upload fix -- we've now got two solutions proposed: p 700652 (which mostly backs out the subrequest change that caused the regression but still fixes logging) and p 701736 (which starts copying over a subset of headers to the cleaned-up subrequest env)
21:13:48 <patchbot> https://review.opendev.org/#/c/700652/ - swift - Revert "bulk: Use make_subrequest to make subreque... - 1 patch set
21:13:50 <patchbot> https://review.opendev.org/#/c/701736/ - swift - Allow bulk to fwd some headers at tar extraction - 2 patch sets
21:16:15 <kota_> hmm...
21:16:41 <tdasilva> is allowed headers available in /info?
21:16:57 <timburke> nope; wrong layer :-(
21:17:02 <timburke> proxy vs object
21:17:03 <tdasilva> i guess it would not be if it's a object server config
21:17:07 <tdasilva> right
21:20:02 <timburke> i'm not sure i like the hard-coded list -- seems like it'd be better as a configurable list... but then also allowing whole prefixes? hmm... trying to remember if we have some prior art for that...
21:20:58 <kota_> good point
21:21:47 <timburke> or have object meta hardcoded, then configurable list for extra headers *shrug*
21:22:03 <timburke> anyway, i suppose we can argue approaches more on the reviews
21:22:19 <timburke> one extra topic to inject:
21:22:26 <timburke> #topic next swift release
21:22:46 <timburke> it's been a while since we did a release, i think it's about time for a new one!
21:23:14 <zaitcev> I lean towards Romain's approach, BTW
21:23:28 <timburke> i'll look to see what else we could merge for it, but i'm looking at three things, mainly:
21:23:41 <timburke> versioning (ideally, with s3 support)
21:23:47 <timburke> bulk fix
21:23:55 <timburke> and py3 formpost fix
21:24:13 <timburke> #link https://launchpad.net/bugs/1858259
21:24:13 <openstack> Launchpad bug 1858259 in OpenStack Object Storage (swift) "Upload failed if filename is unicode string in formpost (python3)" [High,Confirmed]
21:24:35 <timburke> seongsoocho_ already looked at https://review.opendev.org/#/c/701497/ (thanks!)
21:24:35 <patchbot> patch 701497 - swift - py3: Fix formpost unicode filename issues - 2 patch sets
21:24:47 <seongsoocho_> timburke: yey !!! o/
21:24:55 <timburke> if anyone's got a moment to give it a spin, that'd be great
21:25:02 <tdasilva> I think p 697739  might also be able to make it to that list
21:25:02 <patchbot> https://review.opendev.org/#/c/697739/ - swift - Have slo tell the object-server that it wants whol... - 6 patch sets
21:25:17 <timburke> sounds good
21:25:44 <timburke> *maybe* even https://review.opendev.org/700056
21:25:44 <patchbot> patch 700056 - swift - Middleware that allows a user to have quoted Etags - 5 patch sets
21:26:30 <timburke> i'll get a fresh version up soon that actually has the right names everywhere ;-)
21:26:45 <timburke> that's all i've got for today
21:26:49 <timburke> #topic open discussion
21:26:59 <zaitcev> What about client 3.8.2 or something?
21:27:04 <timburke> anything else people want to bring up?
21:27:13 <timburke> zaitcev, yeah, that sounds good too
21:27:56 <timburke> if we can get the versioning patch for the client, all the better :-)
21:28:41 <zaitcev> On the dark data patch, or, rather auditor plugin patch https://review.opendev.org/212824, I had a change of heart and decided to agitate against the process separation.
21:28:42 <patchbot> patch 212824 - swift - Let developers/operators add watchers to object audit - 20 patch sets
21:28:47 <tdasilva> there's a also a client link patch that would be great to have, not sure its state tho
21:29:34 <tdasilva> p 694211
21:29:35 <patchbot> https://review.opendev.org/#/c/694211/ - python-swiftclient - Support uploading Swift symlinks without content. - 7 patch sets
21:32:14 <zaitcev> This is going to be a week of catching up on reviews.
21:32:28 <tdasilva> zaitcev: was just thinking the same :)
21:32:30 <timburke> zaitcev, i'd love that :D
21:32:58 <timburke> all right, that's all i've got
21:33:01 <zaitcev> Anyway, Sam and Alistair are not with us anymore.
21:33:16 <zaitcev> So, I guess I own the auditor API if nobody objects.
21:33:24 <zaitcev> Romain had some input...
21:33:32 <alecuyer> Ill have to read up.. missed that patch, but that's a really nice idea
21:33:37 <timburke> right after this i'll update the priority reviews wiki to reflect the release plan
21:33:38 <alecuyer> (I see there's some history there :) )
21:33:46 <timburke> alecuyer, just a little ;-)
21:35:26 <timburke> well, let's let seongsoocho_, kota_, and tdasilva grab breakfast if they haven't yet
21:35:37 <timburke> thank you all for coming, and thank you for working on swift!
21:35:37 <kota_> thx
21:35:42 <timburke> #endmeeting