21:00:38 <timburke> #startmeeting swift
21:00:39 <openstack> Meeting started Wed May  5 21:00:38 2021 UTC and is due to finish in 60 minutes.  The chair is timburke. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:00:40 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:00:42 <openstack> The meeting name has been set to 'swift'
21:00:44 <timburke> who's here for hte swift meeting?
21:01:12 <seongsoocho> o/
21:01:50 <mattoliverau> o/
21:03:25 <timburke> agenda's at https://wiki.openstack.org/wiki/Meetings/Swift -- just following up on  in-progress work
21:03:37 <timburke> first up
21:03:42 <timburke> #topic sharding/shrinking
21:04:02 <timburke> mattoliverau, i saw you broke up the new state patches; how's that looking?
21:05:26 <mattoliverau> yeah, just trying to keep it managable
21:05:50 <mattoliverau> I wanted to play with the row_cleaved tracking idea in the Cleave Contexts.
21:06:11 <mattoliverau> It has tidied the ACTIVE states patch alot (still working on it though)
21:06:51 <mattoliverau> and I "think" we can do something similar to stop the listing gaps caused by a similar bug with CLEAVED, so have something around that as a follow up.
21:06:54 <kota_> hello, sorry for late
21:07:02 <timburke> nice
21:07:06 <timburke> kota_, no worries!
21:07:12 <mattoliverau> so 3 patches deep so far, but still just a POC WIP.
21:08:07 <timburke> :-)
21:08:43 <mattoliverau> So far so good. Haven't had a chance to play with your other idea yet. just wantd to see what I already had looked like. But if it solves another issue with the "same" approach then that's a win.
21:09:36 <timburke> i know acoles is out, but his tombstone counting in https://review.opendev.org/c/openstack/swift/+/782832 is looking pretty good -- i hope to have a review up for it later today
21:10:18 <mattoliverau> oh yeah, I need to look at that again too. Hopefully we'll have the last peices of shrinking in place soon!
21:11:06 <timburke> iirc those were the two main efforts on sharding
21:11:12 <timburke> next up
21:11:17 <timburke> #topic relinker
21:11:36 <timburke> mattoliverau, i'm digging the new recon data at https://review.opendev.org/c/openstack/swift/+/779655/
21:12:18 <timburke> want me to squash in my followup at https://review.opendev.org/c/openstack/swift/+/789643/ (once i fix the flake8 error)?
21:12:43 <mattoliverau> thanks :) and thanks for the follow up! Yes! it's a good clean up
21:14:30 <timburke> i still wouldn't mind getting https://review.opendev.org/c/openstack/swift/+/783731 in -- any chance you'd be able to take a look?
21:14:43 <timburke> (Rehash the parts actually touched when relinking)
21:16:24 <mattoliverau> sure thing, I did take a quick look yesterday, just haven't tested it yet
21:16:28 <mattoliverau> in my SAIO
21:16:35 <timburke> thanks!
21:18:10 <timburke> aside from those, we've got the signal-handling and recon-aggregation patches -- i started taking a look at the first; i'll plan to rebase it when i push up the re-spin of the recon patch
21:18:47 <timburke> next up
21:18:53 <timburke> #topic stale EC frags
21:19:25 <timburke> looks like the current approach is at https://review.opendev.org/c/openstack/swift/+/788833
21:19:40 <timburke> and you've started to take a look, is that right mattoliverau?
21:20:05 <mattoliverau> yeah, it's pretty great
21:20:48 <mattoliverau> I have a bunch of questions in it. But mostly my -1 is around some testing. But I've got a follow up that addresses those https://review.opendev.org/c/openstack/swift/+/789129
21:21:37 <mattoliverau> I'll chat to acoles when he's back and hopefully get it squared up and then I can +2 it :)
21:21:51 <timburke> great! i look forward to deleting some lurking frags that have now been scattered across thousands of disks :-)
21:22:39 <timburke> (it *does* make me wonder what our replicated policy looks like... i suspect the dark data watcher would find some replicated data that should be deleted...)
21:23:09 <timburke> ...which brings us to our last topic
21:23:18 <timburke> #topic dark data watcher
21:24:39 <timburke> i still haven't reviewed either of zaitcev or dsariel's patches, unfortunately. and i've still not done anything with my own patch, either
21:25:03 <mattoliverau> me either.. so I don't have much to say
21:25:32 <timburke> i think that's most of what's been going on this week
21:25:38 <timburke> #topic open discussion
21:25:46 <timburke> anything else we ought to bring up?
21:26:46 <mattoliverau> not off the top of my head. I've been meaning to go through the action items from PTG but haven't yet.
21:27:16 <timburke> oh yeah... i promised to look at getting container creation time into account listings...
21:27:35 <mattoliverau> But maybe something we can think about in some future meeting (no rush though). Just to make sure we (ahem I), don't forget
21:28:14 <mattoliverau> yeah, and I was going to look at cluster health with current tools and at least write up some doc or something.
21:29:30 <mattoliverau> but anyway might be something we add to the meeting agena every month or so to make sure we don't forget :shrug: just thinking out loud :)
21:29:41 <timburke> pretty sure it'll require a new column. shouldn't be too bad, but i was debating for a bit about whether to stuff it in with the last modified, similar to what we do with container listings iirc. but the upgrade path makes me nervous, if an old account server gets a new db with packed columns rsynced to it...
21:30:14 <timburke> good idea! i'll put it on for next time (and dig through notes to remember what action items we came up with :-)
21:30:48 <timburke> if there isn't anything else, i'll let you guys get on with your mornings
21:31:20 <timburke> thank you all for coming, and thank you for working on swift!
21:31:24 <timburke> #endmeeting