| *** mahatic has quit IRC | 00:37 | |
| *** tomha has joined #openstack-swift | 00:39 | |
| *** tomha has quit IRC | 00:43 | |
| *** BjoernT has joined #openstack-swift | 00:57 | |
| *** BjoernT has quit IRC | 00:58 | |
| *** gyee has quit IRC | 01:00 | |
| *** BjoernT has joined #openstack-swift | 01:17 | |
| *** BjoernT has quit IRC | 01:24 | |
| *** mahatic has joined #openstack-swift | 01:35 | |
| *** ChanServ sets mode: +v mahatic | 01:35 | |
| *** tkajinam has quit IRC | 02:21 | |
| *** tkajinam has joined #openstack-swift | 02:21 | |
| *** psachin has joined #openstack-swift | 03:37 | |
| *** pcaruana has joined #openstack-swift | 04:37 | |
| *** new_student1411 has joined #openstack-swift | 04:55 | |
| *** tkajinam has quit IRC | 05:01 | |
| *** mrjk_ has joined #openstack-swift | 05:46 | |
| *** jistr has quit IRC | 05:47 | |
| *** jistr has joined #openstack-swift | 05:48 | |
| *** rdejoux has quit IRC | 05:49 | |
| *** mrjk has quit IRC | 05:49 | |
| *** tkajinam has joined #openstack-swift | 05:55 | |
| *** rcernin has quit IRC | 06:02 | |
| *** hoonetorg has quit IRC | 06:39 | |
| *** hoonetorg has joined #openstack-swift | 06:56 | |
| *** rdejoux has joined #openstack-swift | 07:10 | |
| *** baojg has quit IRC | 07:35 | |
| *** baojg has joined #openstack-swift | 07:36 | |
| *** baojg has quit IRC | 07:39 | |
| *** baojg has joined #openstack-swift | 07:41 | |
| *** baojg has quit IRC | 07:43 | |
| *** baojg has joined #openstack-swift | 07:44 | |
| *** ccamacho has joined #openstack-swift | 08:04 | |
| *** zaitcev_ has joined #openstack-swift | 08:54 | |
| *** ChanServ sets mode: +v zaitcev_ | 08:54 | |
| *** zaitcev has quit IRC | 08:57 | |
| *** tkajinam has quit IRC | 09:05 | |
| *** zaitcev__ has joined #openstack-swift | 09:08 | |
| *** ChanServ sets mode: +v zaitcev__ | 09:08 | |
| *** zaitcev_ has quit IRC | 09:12 | |
| *** e0ne has joined #openstack-swift | 09:48 | |
| *** e0ne has quit IRC | 10:06 | |
| *** dsariel has joined #openstack-swift | 11:14 | |
| *** dsariel has quit IRC | 11:15 | |
| *** tdasilva has joined #openstack-swift | 11:32 | |
| *** ChanServ sets mode: +v tdasilva | 11:32 | |
| *** tesseract has joined #openstack-swift | 11:33 | |
| *** tdasilva_ has joined #openstack-swift | 12:29 | |
| *** ChanServ sets mode: +v tdasilva_ | 12:29 | |
| *** tdasilva has quit IRC | 12:32 | |
| *** zaitcev__ has quit IRC | 12:51 | |
| *** zaitcev__ has joined #openstack-swift | 13:03 | |
| *** ChanServ sets mode: +v zaitcev__ | 13:03 | |
| *** dsariel has joined #openstack-swift | 13:10 | |
| *** dsariel has quit IRC | 13:25 | |
| *** zaitcev_ has joined #openstack-swift | 13:34 | |
| *** ChanServ sets mode: +v zaitcev_ | 13:34 | |
| *** zaitcev__ has quit IRC | 13:37 | |
| *** tdasilva_ is now known as tdasilva | 13:43 | |
| *** e0ne has joined #openstack-swift | 14:12 | |
| *** baojg has quit IRC | 14:27 | |
| *** zaitcev_ has quit IRC | 14:42 | |
| *** zaitcev_ has joined #openstack-swift | 14:58 | |
| *** ChanServ sets mode: +v zaitcev_ | 14:58 | |
| *** baojg has joined #openstack-swift | 15:16 | |
| *** e0ne has quit IRC | 15:29 | |
| openstackgerrit | Clay Gerrard proposed openstack/swift master: symlink-backed versioned_writes https://review.opendev.org/633857 | 15:49 |
|---|---|---|
| *** pcaruana has quit IRC | 15:49 | |
| clayg | tdasilva: timburke: ok, i'm going to try and rework the hardlinks patch so that hardlinks to SLOs use the manifest etag and hope that simplifies somethings? | 15:50 |
| tdasilva | clayg: ack...currently going through p 633857 | 15:51 |
| patchbot | https://review.opendev.org/#/c/633857/ - swift - symlink-backed versioned_writes - 6 patch sets | 15:51 |
| *** new_student1411 has quit IRC | 15:56 | |
| *** gyee has joined #openstack-swift | 15:58 | |
| *** baojg has quit IRC | 16:15 | |
| *** tdasilva has quit IRC | 16:56 | |
| *** tdasilva has joined #openstack-swift | 16:56 | |
| *** ChanServ sets mode: +v tdasilva | 16:56 | |
| *** tdasilva_ has joined #openstack-swift | 17:13 | |
| *** ChanServ sets mode: +v tdasilva_ | 17:13 | |
| *** tdasilva has quit IRC | 17:15 | |
| *** tesseract has quit IRC | 17:24 | |
| *** tdasilva_ has quit IRC | 17:37 | |
| *** tdasilva_ has joined #openstack-swift | 17:37 | |
| *** ChanServ sets mode: +v tdasilva_ | 17:37 | |
| tdasilva_ | clayg: the question on versioning and object expiration is an interesting one. I'm not sure we have added tests for this before...one idea (similar behavior to s3) would be to disallow expiration headers to objects in versioned containers?? | 17:40 |
| openstackgerrit | Merged openstack/swift master: py3: Bring functional/test_object.py under test; add func-ec-py37 job https://review.opendev.org/645895 | 17:44 |
| timburke | clayg, tdasilva_ i wonder how well it'd work to ensure that expiration headers only make it to the link and not the real data... it'd be a break from pre-link behavior, but i think in a good way? | 17:57 |
| timburke | wow... do we really *never* check the status when we refetch the manifest at https://github.com/openstack/swift/blob/2.21.0/swift/common/middleware/slo.py#L817-L821 ? | 18:00 |
| timburke | the non-atomic bits of swift suck... | 18:00 |
| *** thiago__ has joined #openstack-swift | 18:00 | |
| *** ChanServ sets mode: +v thiago__ | 18:00 | |
| *** thiago__ is now known as tdasilva | 18:01 | |
| timburke | otoh, i don't really want to plumb *more* slo logic into the object server, so... | 18:01 |
| *** tdasilva_ has quit IRC | 18:03 | |
| clayg | @tdasilva yes, i'm not clear on expiration behavior when versions are involved - I assumed the version would get uploaded and set to expire (potentially causing the link to 404 after expiration, or still be valid if it's already pointing to a newer version) | 18:05 |
| clayg | I think it gets easier if we're able to move toward bucket policies | 18:05 |
| tdasilva | clayg: how so? | 18:06 |
| clayg | @timburke just expiring the link seems reasonable from a data protection stand point - and also seems like it'd potentially still be useful once we can apply a bucket policy to cleanup the old versions... | 18:07 |
| clayg | @tdasilva well just cause we can stop trying to set which object expires when when they're uploaded and instead you just set the policy on the bucket! | 18:07 |
| timburke | clayg, it *does* mean that versioned_writes should be in the expirer pipeline, though -- i'm not sure that we currently recommend that? but it'd certainly be useful to have that delete marker in there in case there was a re-upload | 18:13 |
| clayg | i'm pretty sure we recommend that you do NOT put versioned_writes in your expirer pipeline - there was a bug on this not that long ago IIRC | 18:14 |
| tdasilva | so in the case of supporting versioning with s3api, would we block users from using expiration to conform with aws s3? | 18:15 |
| clayg | @timburke I think it'd be pretty wild if expirer triggered stack based action | 18:16 |
| tdasilva | assuming we also have bucket policy in the future | 18:16 |
| clayg | tdasilva: I think it's a safe assumption we'll have some solution to bucket policy compatibility in the future | 18:16 |
| timburke | clayg, on the policies thing... and you could have separate policies for the primary vs the main... i think i like that idea! well... *if* we get some sort of exemption for "currently-pointed-at version" (or should it be "most-recent"? 🤔) | 18:17 |
| clayg | tdasilva: I think we'll continue to try and make s3api emulation/compatibility as near-as-possible translation of aws s3 when swift has the requisite underlying compatible capabilities (even to the extent of disallowing combining features that work together in swift because they don't in s3; sure) | 18:18 |
| timburke | swift-ring-builder --help | 18:19 |
| timburke | Ring Builder file does not exist: --help 😞 | 18:19 |
| clayg | timburke: tdasilva: we need to narrow the actionable design - we're talking about "how versioning works today" "how it should work after symlinks" "how it should work with bucket policies" "how it should work when with s3api + bucket policies" that's alot to figure out at the same time | 18:20 |
| timburke | meanwhile, the actual help spans ~350 lines, which is a fair bit of reading... | 18:20 |
| clayg | we could *just* figure out how s3api + bucket policies should work (because that's "as near as possible to how they work in aws s3) - then try and work backwards from there? | 18:21 |
| clayg | timburke: lp bug #1553030 maybe? | 18:22 |
| openstack | Launchpad bug 1553030 in OpenStack Object Storage (swift) "Change optparse to argparse, optparse is Deprecated since python version 2.7" [Undecided,In progress] https://launchpad.net/bugs/1553030 - Assigned to dharmendra (dharmendra-kushwaha) | 18:22 |
| timburke | 👍, at least in part | 18:23 |
| timburke | maybe entirely? the subcommand handling in argparse is actually pretty nice... | 18:24 |
| tdasilva | clayg: I think it might be worth looking at the end goal so that we don't end up having to bend the code later to conform with a new requirement | 18:24 |
| tdasilva | clayg: i guess i'm just thinking that it would be weird if swift allows one thing and s3 does sort of the opposite | 18:30 |
| timburke | ^^^ definitely true. to design the swift thing, we need to figure out the superset of the s3 thing that still makes sense in swift | 18:31 |
| *** psachin has quit IRC | 19:20 | |
| *** tdasilva has quit IRC | 20:06 | |
| *** tdasilva has joined #openstack-swift | 20:06 | |
| *** ChanServ sets mode: +v tdasilva | 20:06 | |
| openstackgerrit | Clay Gerrard proposed openstack/swift master: Allow "harder" symlinks https://review.opendev.org/633094 | 20:44 |
| clayg | hrmm... ubuntu-bionic | ERROR: unknown environment 'func-ec-py3' | 20:45 |
| *** e0ne has joined #openstack-swift | 20:51 | |
| *** e0ne has quit IRC | 20:57 | |
| *** zaitcev_ is now known as zaitcev | 22:02 | |
| zaitcev | timburke: did you try to run functests against py3? | 22:02 |
| zaitcev | your posted reviews implied that you did | 22:03 |
| zaitcev | sorry, brb | 22:07 |
| *** baojg has joined #openstack-swift | 22:18 | |
| *** baojg has quit IRC | 22:22 | |
| zaitcev | So, anyway. According to tcpdump, account server does not send a reply, but just closes the connection. But it logs a success. So..... maybe a flush is missing somewhere, but it worked fine on py2. | 22:27 |
| *** gyee has quit IRC | 22:34 | |
| *** gyee has joined #openstack-swift | 22:57 | |
| *** takamatsu has quit IRC | 23:31 | |
| timburke | zaitcev, what version of eventlet are you using? | 23:52 |
| timburke | that'd be my first guess, anyway | 23:52 |
| zaitcev | python3-eventlet-0.24.1-3.fc30.noarch | 23:58 |
| zaitcev | should be fine | 23:58 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!