| *** rcernin has quit IRC | 00:05 | |
| openstackgerrit | Merged openstack/swift stable/train: Stop syncing empty suffixes list https://review.opendev.org/743614 | 00:06 |
|---|---|---|
| openstackgerrit | Merged openstack/python-swiftclient master: Add max_backoff and starting_backoff for get_conn in swift service https://review.opendev.org/743551 | 00:12 |
| *** rcernin has joined #openstack-swift | 00:31 | |
| *** samueldmq has quit IRC | 00:57 | |
| *** samueldmq has joined #openstack-swift | 01:01 | |
| *** gyee has quit IRC | 01:02 | |
| *** tkajinam has quit IRC | 01:06 | |
| *** tkajinam has joined #openstack-swift | 01:06 | |
| *** tkajinam has quit IRC | 01:33 | |
| *** tkajinam has joined #openstack-swift | 01:34 | |
| *** dsariel has quit IRC | 02:08 | |
| *** psachin has joined #openstack-swift | 03:38 | |
| *** evrardjp has quit IRC | 04:33 | |
| *** evrardjp has joined #openstack-swift | 04:33 | |
| *** dsariel has joined #openstack-swift | 05:44 | |
| *** m75abrams has joined #openstack-swift | 06:01 | |
| *** dsariel has quit IRC | 06:47 | |
| seongsoocho | kota_: Hi, kota. My co-worker found a unicode problem. Does japanese have the same problem? Or have you ever heard of this? | 07:05 |
| seongsoocho | https://bugs.launchpad.net/swift/+bug/1889363 | 07:05 |
| openstack | Launchpad bug 1889363 in OpenStack Object Storage (swift) "Prefix filtering issue due to Unicode normalization" [Undecided,New] | 07:05 |
| *** evrardjp has quit IRC | 08:06 | |
| *** evrardjp has joined #openstack-swift | 08:08 | |
| *** rcernin has quit IRC | 09:12 | |
| *** dsariel has joined #openstack-swift | 09:21 | |
| *** dsariel has quit IRC | 09:42 | |
| *** tkajinam has quit IRC | 10:01 | |
| *** dsariel has joined #openstack-swift | 10:31 | |
| *** hoonetorg has joined #openstack-swift | 13:29 | |
| *** rcernin has joined #openstack-swift | 14:41 | |
| *** rcernin_ has joined #openstack-swift | 14:51 | |
| *** irclogbot_1 has quit IRC | 14:52 | |
| *** early` has joined #openstack-swift | 14:52 | |
| *** rcernin_ has quit IRC | 14:56 | |
| *** rcernin has quit IRC | 14:58 | |
| *** early has quit IRC | 14:58 | |
| *** irclogbot_0 has joined #openstack-swift | 14:59 | |
| *** hoonetorg has quit IRC | 15:09 | |
| *** m75abrams has quit IRC | 15:20 | |
| *** hoonetorg has joined #openstack-swift | 15:21 | |
| *** zaitcev has joined #openstack-swift | 15:59 | |
| *** ChanServ sets mode: +v zaitcev | 15:59 | |
| openstackgerrit | Pete Zaitcev proposed openstack/swift master: WIP: Band-aid and test the crash of the account server https://review.opendev.org/743797 | 16:26 |
| *** gyee has joined #openstack-swift | 16:28 | |
| *** psachin has quit IRC | 16:53 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: Allow direct and internal clients to use the replication network https://review.opendev.org/735751 | 17:49 |
| timburke | fleshed out the upgrade impact on ^^^ some more | 17:50 |
| openstackgerrit | Tim Burke proposed openstack/swift master: Add py3 probe tests on CentOS 8 https://review.opendev.org/690717 | 18:01 |
| *** jv_ has quit IRC | 18:29 | |
| clayg | do we have "pairwise" in Swift already somewhere by a different name? https://docs.python.org/3/library/itertools.html#itertools-recipes | 19:13 |
| timburke | clayg, not that i'm aware of, though i know i've written things like `zip(x, x[1:])` a fair few times | 19:20 |
| clayg | oh yeah, that's fine - the things I have support getitem anyway! 🙏 | 19:22 |
| *** dsariel has quit IRC | 19:27 | |
| *** jv_ has joined #openstack-swift | 19:50 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: sharding: Add some more shard-related output to swift-container-info https://review.opendev.org/721377 | 20:21 |
| kota_ | morning | 20:58 |
| seongsoocho | morning | 20:59 |
| mattoliverau | Morning | 20:59 |
| kota_ | seongsoocho: about https://bugs.launchpad.net/swift/+bug/1889363, i didn't hit the problem so far. Usually my customer uses English file name instead Japanese. | 21:03 |
| openstack | Launchpad bug 1889363 in OpenStack Object Storage (swift) "Prefix filtering issue due to Unicode normalization" [Undecided,New] | 21:03 |
| timburke | sorry i'm a little late ;-) | 21:03 |
| timburke | meeting time! | 21:03 |
| kota_ | it would be potentially occurred but it looks client problem, doesn't it? | 21:05 |
| seongsoocho | kota_: oh I see.. I also think It looks client problem, but the user doesn't care about there unicode encoding. So when they try to filtering the result, they can't get a proper result. | 21:07 |
| timburke | mattoliverau, clayg: sorry, i'd also wanted to talk through the various shrinking/overlapping shard ranges patches, but i ate up too much time with the py3 bug | 22:03 |
| clayg | oh right shrinking! oh well there's always the next meeting | 22:03 |
| clayg | are you unblocked on the py3 bug? | 22:04 |
| clayg | I'm making progress on shrinking w/ overlaps | 22:04 |
| clayg | I'm stuck on waterfall ec tho - @timburke how do you link to a line number in gerrit? I thought it was #L2952 | 22:04 |
| timburke | @2952 iirc | 22:05 |
| clayg | mattoliverau: i know you had some intrest in concurrent gets for ec, I wonder what you think about https://review.opendev.org/#/c/737096/4/swift/proxy/controllers/obj.py@2952 (it's a little hard to see with all the per-policy stuff in there) | 22:06 |
| patchbot | patch 737096 - swift - Add concurrent_ec_extra_requests - 4 patch sets | 22:06 |
| timburke | i think i'm unblocked? i'll squash the two patches together, change the default to v2, and make sure there's an UpgradeImpact calling out that (1) once all proxies are upgraded, you should switch to v3 and (2) if you're on py2 and have encryption enabled, you need to upgrade beyond this patch before switching to py3 | 22:06 |
| clayg | is that new concurrent_ec_extra_requests obvious enough to squash into p 711342 and then the other one will JUST be making the "existing" concurrent_gets options per policy? | 22:07 |
| patchbot | https://review.opendev.org/#/c/711342/ - swift - Add concurrent_gets to EC GET requests - 14 patch sets | 22:07 |
| clayg | timburke: that sounds 100% sufficient for now! | 22:07 |
| timburke | clayg, i think that'd be fine. should we increase the pool size, too, though? | 22:08 |
| clayg | pool size... 🤔 | 22:13 |
| clayg | like make it bigger than ndata by default? | 22:13 |
| timburke | i was thinking, make it ec_request_count | 22:14 |
| timburke | otherwise, we don't kick off that ndata+1 request until we get one of the other guys to respond | 22:14 |
| clayg | oh oh, i don't think i quite grokked your comment yet - i'll look at your review more closely before i respin after I finish this shrink stuff | 22:15 |
| clayg | are you on board with squashing the new option? or should the per-replica stuff stay separate until we nail down what we got | 22:16 |
| clayg | nm, i'll just make them 3 different patches 😁 | 22:16 |
| timburke | i'd be fine with squashing the new option -- or separate commits, w/e | 22:28 |
| mattoliverau | clayg: I saw your new config patch for it. Started looking yesterday. At first glance seems similar. And been testing your shrink command. Left a review on the latter | 22:29 |
| mattoliverau | *simplified | 22:30 |
| mattoliverau | IE the N made sense but explaining that to ops | 22:30 |
| *** rcernin_ has joined #openstack-swift | 22:36 | |
| *** rcernin_ has quit IRC | 22:48 | |
| *** rcernin has joined #openstack-swift | 22:48 | |
| *** tkajinam has joined #openstack-swift | 22:53 | |
| timburke | :-/ as i'm writing the upgrade notes, i'm really starting to wonder whether i should just leave py3 writing the funky meta and make sure both py2 and py3 can deal with it... and avoid the v3 bump entirely. then some day in the nebulous future, maybe we remember to switch py3 to writing the format we actually want | 23:41 |
| timburke | it just bugs me that we always have to keep checking for whether things might be in the wrong/unexpected format :-( | 23:42 |
| timburke | this is exactly the kind of thing that made me hesitant to officially declare py3 support | 23:44 |
| timburke | as an operator, there's *no good reason* to ever tell your cluster to start writing v3 -- it's just cognitive load we've pushed onto them | 23:45 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!