| *** gyee has quit IRC | 00:06 | |
| *** ianychoi has quit IRC | 00:12 | |
| *** ianychoi has joined #openstack-swift | 00:15 | |
| openstackgerrit | Merged openstack/swift master: Authors/changelog for 2.23.0 https://review.opendev.org/685503 | 00:22 |
|---|---|---|
| *** ianychoi has quit IRC | 00:32 | |
| *** ianychoi has joined #openstack-swift | 00:48 | |
| *** BjoernT has joined #openstack-swift | 02:15 | |
| *** mkrai_ has joined #openstack-swift | 02:37 | |
| *** BjoernT has quit IRC | 04:09 | |
| *** rcernin has quit IRC | 04:29 | |
| *** rcernin has joined #openstack-swift | 04:33 | |
| *** mkrai_ has quit IRC | 05:01 | |
| *** mkrai_ has joined #openstack-swift | 05:02 | |
| *** spsurya has joined #openstack-swift | 05:05 | |
| *** mkrai_ has quit IRC | 05:56 | |
| *** mkrai__ has joined #openstack-swift | 05:57 | |
| *** ianychoi has quit IRC | 06:20 | |
| *** ianychoi has joined #openstack-swift | 06:23 | |
| *** ianychoi has quit IRC | 06:29 | |
| *** ianychoi has joined #openstack-swift | 06:31 | |
| *** new_student1411 has joined #openstack-swift | 06:32 | |
| *** mkrai_ has joined #openstack-swift | 06:52 | |
| *** mkrai__ has quit IRC | 06:56 | |
| *** mkrai_ has quit IRC | 06:59 | |
| *** mkrai_ has joined #openstack-swift | 06:59 | |
| *** rcernin has quit IRC | 07:06 | |
| *** tesseract has joined #openstack-swift | 07:12 | |
| *** rdejoux has joined #openstack-swift | 07:20 | |
| *** pcaruana has joined #openstack-swift | 07:28 | |
| *** ianychoi has quit IRC | 07:45 | |
| *** e0ne has joined #openstack-swift | 07:52 | |
| *** e0ne has quit IRC | 07:53 | |
| *** mikecmpbll has joined #openstack-swift | 07:58 | |
| *** tkajinam has quit IRC | 08:00 | |
| *** mkrai_ has quit IRC | 08:24 | |
| *** mkrai_ has joined #openstack-swift | 08:25 | |
| *** ccamacho has joined #openstack-swift | 08:31 | |
| *** e0ne has joined #openstack-swift | 08:49 | |
| *** NM has joined #openstack-swift | 09:01 | |
| *** NM has quit IRC | 09:07 | |
| *** diablo_rojo has joined #openstack-swift | 09:17 | |
| *** diablo_rojo has quit IRC | 09:24 | |
| new_student1411 | timburke , thanks for the reply. It was not the exact issue the way I explained it. Let me do some more research on ec2 credentials authentication. | 09:33 |
| *** diablo_rojo has joined #openstack-swift | 10:16 | |
| *** diablo_rojo has quit IRC | 10:40 | |
| *** rcernin has joined #openstack-swift | 10:44 | |
| *** rcernin has quit IRC | 11:34 | |
| *** dosaboy has quit IRC | 11:45 | |
| *** baojg has joined #openstack-swift | 11:47 | |
| *** dosaboy has joined #openstack-swift | 11:54 | |
| *** dosaboy has quit IRC | 12:03 | |
| *** dosaboy has joined #openstack-swift | 12:03 | |
| *** baojg has quit IRC | 12:04 | |
| *** mkrai_ has quit IRC | 12:13 | |
| *** mkrai_ has joined #openstack-swift | 12:18 | |
| *** mkrai_ has quit IRC | 13:01 | |
| *** mkrai_ has joined #openstack-swift | 13:02 | |
| *** BjoernT has joined #openstack-swift | 13:18 | |
| openstackgerrit | Alexandre Lécuyer proposed openstack/swift feature/losf: Add tests for vfile.py https://review.opendev.org/666378 | 13:28 |
| *** mkrai_ has quit IRC | 13:29 | |
| *** BjoernT has quit IRC | 13:29 | |
| *** BjoernT has joined #openstack-swift | 13:32 | |
| *** mkrai_ has joined #openstack-swift | 14:22 | |
| *** mkrai_ has quit IRC | 14:28 | |
| openstackgerrit | Thiago da Silva proposed openstack/swift master: Start building py3 docker images https://review.opendev.org/686419 | 15:06 |
| *** BjoernT has quit IRC | 15:18 | |
| *** BjoernT has joined #openstack-swift | 15:20 | |
| *** mkrai has joined #openstack-swift | 15:21 | |
| *** gyee has joined #openstack-swift | 15:24 | |
| tdasilva | tburke, clayg, mattoliverau: as I'm making progress on tests, i'm now hitting the use-cases of overwrites and deletes when versioning is suspended. Currently the difference we have over s3 is that on deletes we will keep adding delete markers with a version_id, while on overwrite we don't delete any delete_marker. This is different from from s3 in that aws will add a delete_marker only one and with a 'null' | 15:27 |
| tdasilva | version_id. On a overwrite, it will also remove that delete_marker. I don't think our implementation is bad, but i am concerned about the increase version counts as it is counting delete marker (versioning is supposed to be suspended) and the fact that our s3api implementation will probably have to follow swift's behavior and have to deviate a bit from aws s3 | 15:27 |
| *** tesseract has quit IRC | 15:36 | |
| *** mkrai has quit IRC | 15:39 | |
| *** mkrai__ has joined #openstack-swift | 15:39 | |
| clayg | yeah I think when we're suspended we shouldn't create any new delete markers? Can't we just make sure the static link is a tombstone and call it day? | 15:43 |
| clayg | The only one I worry about is an overwrite while suspended - it's fine to overwrite the static link (since the data is still versioned) but when we re-enable versioning we have to make sure to copy any "not-static-links" out of the public container before we overwrite them with a static link to the new data!? | 15:45 |
| clayg | ... but isn't that basically what we have to do when we're enabling versioning for the first time anyway? | 15:45 |
| clayg | Maybe I'm misunderstanding the delete_marker situation on suspended... | 15:46 |
| tdasilva | clayg: regarding the overwrite after re-enable. Yes, we copy non-static-links out of the public container so we are good there... | 16:01 |
| tdasilva | clayg: re: delete markers. aws will do something a bit weird that when in suspended mode, if you issue a delete, it will write down a delete-marker with version-id='null' | 16:02 |
| tdasilva | clayg: for reference: http://paste.openstack.org/show/779176/ | 16:04 |
| *** rdejoux has quit IRC | 16:05 | |
| clayg | yeah #11 is *bogus* - i want to break api compatibility there - i think it's just an accident of how they've implemented | 16:06 |
| clayg | @timburke and I had at one point also talked about in #6 we'd go ahead and assign a version-id (based on the timestamp of the GET) to the object before we add it to the versioned listing as well... | 16:07 |
| clayg | the only time I'd like to see 'null' versions is when we're adding in an for-realzy-object-with-data from the public container - everything in the null-versions container would have a version-id (it's part of the path name!) | 16:08 |
| clayg | do you think our tests would be able to make "useful" assertions about behavior of s3 & swift without having to be much of a stickler on some minor/strange/unimportant differences like the version-id of an object written while suspended and overwritten while enabled? 😬 | 16:10 |
| tdasilva | clayg: sure, i'm not arguing for strictness, but just wanted to make sure we are all on the same page about that. | 16:12 |
| tdasilva | "the only time I'd like to see 'null' versions is when we're adding in an for-realzy-object-with-data from the public container - everything in the null-versions container would have a version-id (it's part of the path name!)" << this is the current implementation | 16:12 |
| clayg | no no no - it's good! we have to do at least two things: | 16:12 |
| clayg | 1) know *exactly* how s3api behaves | 16:12 |
| clayg | 2) know *exactly* how we want swift-object-versioning to behave! | 16:13 |
| clayg | I 100% agree, *if* we decide they should be different we need to be on the same page about that!!! | 16:13 |
| tdasilva | I'd add a 3rd: know (and document) *exactly* where s3api deviates from aws s3 so that our future selves don't have to wonder about the decisions that were made in case people ask about those differences | 16:15 |
| tdasilva | 'cause I know I won't remember :) | 16:15 |
| clayg | same here | 16:15 |
| clayg | so but it sounds like 1) the behavior I'm describing makes sense to you for swift-object-versioning and 2) that's what you've implemented! | 16:16 |
| clayg | so the only think left is 3) how are we going to document how this is different from s3api | 16:16 |
| tdasilva | yeah | 16:16 |
| clayg | I think writing it down is a good first step - but eventually I'd like to codify it somehow... like via a test of some-kinda? | 16:17 |
| tdasilva | me wonders if ceph tests are super strict to pick up on those differences | 16:17 |
| clayg | i haven't managed to get ceph's s3 object-versoining tests running again out versoining support yet... | 16:17 |
| tdasilva | clayg: yeah, in the case of swift tests, i'm making sure to highlight that behavior | 16:18 |
| clayg | yeah, the swift-object-versioning tests - for sure - and you could even drop comments "this isn't what s3 does" | 16:18 |
| clayg | but later in the s3api tests - especially the ones we want to run again swift and s3 (and ceph?) - there's an oppertunity to say something about the differences there | 16:19 |
| clayg | ... maybe | 16:19 |
| tdasilva | +1 | 16:29 |
| *** ccamacho has quit IRC | 16:31 | |
| *** mikecmpbll has quit IRC | 16:32 | |
| openstackgerrit | OpenStack Release Bot proposed openstack/swift stable/train: Update .gitreview for stable/train https://review.opendev.org/686438 | 16:35 |
| openstackgerrit | OpenStack Release Bot proposed openstack/swift stable/train: Update TOX/UPPER_CONSTRAINTS_FILE for stable/train https://review.opendev.org/686439 | 16:35 |
| openstackgerrit | OpenStack Release Bot proposed openstack/swift master: Update master for stable/train https://review.opendev.org/686440 | 16:35 |
| *** e0ne has quit IRC | 16:36 | |
| *** mkrai__ has quit IRC | 17:06 | |
| *** mkrai_ has joined #openstack-swift | 17:06 | |
| *** mkrai_ has quit IRC | 17:11 | |
| *** openstackgerrit has quit IRC | 17:21 | |
| *** e0ne has joined #openstack-swift | 17:59 | |
| *** e0ne has quit IRC | 18:06 | |
| *** spsurya has quit IRC | 18:07 | |
| *** openstackgerrit has joined #openstack-swift | 18:19 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: Update master for stable/train https://review.opendev.org/686440 | 18:19 |
| timburke | yay! we have a train release! | 18:19 |
| timburke | also, i have my china visa :-) | 18:19 |
| clayg | ROFL | 18:27 |
| clayg | File "/home/vagrant/swift/swift/common/db.py", line 155, in <lambda> | 18:27 |
| clayg | self, *args, **kwargs)) | 18:27 |
| clayg | ValueError: the query contains a null character | 18:27 |
| timburke | whomp whomp | 18:27 |
| timburke | *way* down in there https://github.com/python/cpython/blob/v3.7.4/Modules/_sqlite/statement.c#L67-L70 | 18:30 |
| clayg | s'fine - i forgot to retell myself the fable of ye 'ol "bobby tables" before writing sql | 18:52 |
| clayg | query += " name NOT LIKE ? AND" | 18:52 |
| clayg | query_args.append('%%\x00%%') | 18:52 |
| clayg | ^ is a lot more like what I wanted | 18:52 |
| clayg | and sqli's bind/escape whatever is just fine with it | 18:52 |
| clayg | I mean prefix=null%00 was working just fine, so I knew something should work 🤷♂️ | 18:53 |
| clayg | relevant XKCD -> https://xkcd.com/327/ | 18:53 |
| tdasilva | timburke: I also tagged a v2.23.0 release of the docker container: https://hub.docker.com/r/openstackswift/saio/tags | 18:53 |
| clayg | tag that docker build! | 18:54 |
| openstackgerrit | Thiago da Silva proposed openstack/swift master: Start building py3 docker images https://review.opendev.org/686419 | 18:58 |
| timburke | thanks tdasilva! | 18:59 |
| *** itlinux has quit IRC | 19:02 | |
| *** henriqueof1 has joined #openstack-swift | 19:04 | |
| *** henriqueof has quit IRC | 19:07 | |
| openstackgerrit | Merged openstack/swift stable/train: Update .gitreview for stable/train https://review.opendev.org/686438 | 19:44 |
| *** gmann is now known as gmann_afk | 19:52 | |
| *** pcaruana has quit IRC | 20:18 | |
| *** camelCaser has quit IRC | 21:28 | |
| *** camelCaser has joined #openstack-swift | 21:28 | |
| *** rcernin has joined #openstack-swift | 22:08 | |
| *** BjoernT has quit IRC | 22:31 | |
| *** henriqueof1 has quit IRC | 22:58 | |
| *** tkajinam has joined #openstack-swift | 23:01 | |
| timburke | well that's new: https://storage.bhs1.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_654/652184/3/gate/swift-tox-py37/654bebe/job-output.txt | 23:21 |
| timburke | `AssertionError: Invalid switch into Queue.put: ()` coming up out of eventlet | 23:21 |
| openstackgerrit | Merged openstack/swift master: Update master for stable/train https://review.opendev.org/686440 | 23:51 |
| *** NM has joined #openstack-swift | 23:54 | |
| *** NM has quit IRC | 23:58 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!