| mattoliverau | a json body. | 00:00 |
|---|---|---|
| mattoliverau | maybe a '' is fine, or we can create a shard and dump the json. | 00:00 |
| mattoliverau | I've moving back over the latching sharding patch now. it's been around a while, so want to give it a spin on one of my SAIOs. | 00:01 |
| timburke | \o/ thanks! | 00:04 |
| timburke | makes sense about it needing to deserialize the body -- did you notice the test around https://review.opendev.org/#/c/731653/1/test/unit/container/test_server.py@2408 though? | 00:06 |
| patchbot | patch 731653 - swift - Don't auto-create shard containers based on object... - 1 patch set | 00:06 |
| timburke | i'm actually realizing that the `not any(sr.state in SHARD_LISTING_STATES for ...)` in my paste would still allow the empty list to create the shard container, but i'm now not sure that it *should* :-/ | 00:08 |
| openstackgerrit | Tim Burke proposed openstack/swift master: Don't auto-create shard containers based on object updates https://review.opendev.org/731653 | 00:09 |
| timburke | eh, w/e -- squash 'em all in! | 00:09 |
| openstackgerrit | Matthew Oliver proposed openstack/swift master: Add branch and tag tags to docker image https://review.opendev.org/732480 | 00:51 |
| mattoliverau | ^ just testing | 00:51 |
| *** jv|afk has joined #openstack-swift | 01:27 | |
| *** gyee has quit IRC | 01:40 | |
| *** rcernin has quit IRC | 02:24 | |
| *** rcernin has joined #openstack-swift | 02:31 | |
| *** jv|afk has quit IRC | 02:32 | |
| *** rcernin has quit IRC | 03:23 | |
| *** rcernin has joined #openstack-swift | 03:31 | |
| *** psachin has joined #openstack-swift | 03:33 | |
| *** rcernin has quit IRC | 03:46 | |
| *** manuvakery has joined #openstack-swift | 04:06 | |
| *** rcernin has joined #openstack-swift | 04:23 | |
| *** rcernin has quit IRC | 04:28 | |
| *** rcernin has joined #openstack-swift | 04:29 | |
| *** evrardjp has quit IRC | 04:33 | |
| *** evrardjp has joined #openstack-swift | 04:33 | |
| *** ccamacho has quit IRC | 04:41 | |
| manuvakery | guys the https://docs.openstack.org/swift/latest/s3_compat.html doc says object tagging is supported but when i tested on saio it return NotImplemented .. am i missing something or the doc is wrong | 04:46 |
| openstackgerrit | Tim Burke proposed openstack/swift master: docs: s3api doesn't support tagging https://review.opendev.org/733047 | 04:51 |
| timburke | the docs are lying, sorry | 04:51 |
| timburke | ^^^ will correct them | 04:51 |
| manuvakery | yeah just had a look at the code | 04:51 |
| manuvakery | timburke: what about the Virtual-Hosted-Style requests . i have tried few month back but didnt work for me .. now again came back as the path style is getting depricacted in Sep 2020 in aws | 04:55 |
| *** rcernin has quit IRC | 05:36 | |
| *** rcernin has joined #openstack-swift | 05:37 | |
| openstackgerrit | Andreas Jaeger proposed openstack/swift master: Remove <py3.5 dependencies from requirements.txt https://review.opendev.org/732989 | 05:48 |
| openstackgerrit | Andreas Jaeger proposed openstack/swift master: Switch to newer openstackdocstheme and reno versions https://review.opendev.org/732741 | 05:51 |
| *** rcernin has quit IRC | 05:51 | |
| *** rcernin has joined #openstack-swift | 05:56 | |
| *** rcernin has quit IRC | 06:10 | |
| *** rcernin has joined #openstack-swift | 06:14 | |
| *** rcernin has quit IRC | 06:28 | |
| AJaeger | clayg, timburke, it looks like we're good now; I've left a comment in requirements.txt with our findings that might help packagers ;) | 06:31 |
| openstackgerrit | Andreas Jaeger proposed openstack/swift master: Remove <py3.5 dependencies from requirements.txt https://review.opendev.org/732989 | 06:32 |
| openstackgerrit | Andreas Jaeger proposed openstack/swift master: Switch to newer openstackdocstheme and reno versions https://review.opendev.org/732741 | 06:32 |
| openstackgerrit | Andreas Jaeger proposed openstack/python-swiftclient master: Remove <py3.5 dependencies from requirements.txt https://review.opendev.org/733050 | 06:35 |
| *** dtantsur|afk is now known as dtantsur | 06:43 | |
| *** tkajinam has quit IRC | 06:56 | |
| *** tkajinam has joined #openstack-swift | 06:56 | |
| *** rdejoux has quit IRC | 07:07 | |
| *** rdejoux has joined #openstack-swift | 07:12 | |
| openstackgerrit | Andreas Jaeger proposed openstack/python-swiftclient master: Remove <py3.5 dependencies from requirements.txt https://review.opendev.org/733050 | 07:20 |
| openstackgerrit | Andreas Jaeger proposed openstack/python-swiftclient master: Switch to newer openstackdocstheme and reno versions https://review.opendev.org/732742 | 07:20 |
| AJaeger | same trick done for python-swifthclient ^ | 07:20 |
| * AJaeger waits now for all jobs to run | 07:21 | |
| *** ravsingh has joined #openstack-swift | 07:34 | |
| AJaeger | timburke: you need a change for governance repo to retire swift-specs as well - and an email to openstack-discuss as courtesy is also good practice | 07:39 |
| *** ravsingh has quit IRC | 07:44 | |
| *** ravsingh has joined #openstack-swift | 08:17 | |
| *** m75abrams has joined #openstack-swift | 08:34 | |
| *** ccamacho has joined #openstack-swift | 08:50 | |
| *** tkajinam has quit IRC | 08:53 | |
| *** tkajinam has joined #openstack-swift | 08:54 | |
| *** tkajinam has quit IRC | 09:09 | |
| *** m75abrams has quit IRC | 10:18 | |
| *** TobbeCN has joined #openstack-swift | 11:23 | |
| *** jv|afk has joined #openstack-swift | 11:50 | |
| *** rpittau|afk is now known as rpittau | 11:57 | |
| *** tkajinam has joined #openstack-swift | 12:57 | |
| *** psachin has quit IRC | 13:14 | |
| clayg | i still don't think I fully grok the significance of the allow_create_shard in p 731653 | 14:28 |
| patchbot | https://review.opendev.org/#/c/731653/ - swift - Don't auto-create shard containers based on object... - 2 patch sets | 14:28 |
| clayg | it's frustrating because I'm pretty sure keeping object PUTs in async pendings till the primary dbs rebalance around is helpful | 14:30 |
| clayg | ... but best I can tell that's only indirectly controlled by the container server auto create behavior and not related to shard range PUTs at ALL 😞 | 14:31 |
| *** jv|afk has quit IRC | 15:32 | |
| *** gyee has joined #openstack-swift | 15:42 | |
| timburke | clayg, so the one place i *know* we need the autocreate behavior for shard ranges is https://github.com/openstack/swift/blob/2.25.0/swift/container/sharder.py#L1159-L1161 and i think it was mainly just to avoid the need to make two requests (or think about what to do when one or the other fails) | 15:43 |
| timburke | i wonder if i should be even more defensive and only carve out the exception if all shard ranges have state CREATED... | 15:43 |
| *** jv|afk has joined #openstack-swift | 15:55 | |
| clayg | timburke: where did you find that line in logs about the transitive dependency installs, I don't know why futures is installed in py27 jobs on p 733050 | 15:57 |
| patchbot | https://review.opendev.org/#/c/733050/ - python-swiftclient - Remove <py3.5 dependencies from requirements.txt - 2 patch sets | 15:57 |
| *** jv|afk has quit IRC | 16:01 | |
| timburke | clayg, https://storage.gra.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_dc5/732989/4/check/swift-tox-py27/dc56501/tox/py27-2.log | 16:02 |
| clayg | timburke: ok, thanks - so maybe i can find something similar in the python-swiftclient jobs | 16:03 |
| AJaeger | clayg: or https://zuul.opendev.org/t/openstack/build/4fb5d5fbf2544396b53d88561deaaa02/log/job-output.txt#540 | 16:03 |
| AJaeger | clayg: just search in job-output.txt for it | 16:04 |
| clayg | AJaeger: well, but that line futures, and since it's not clear to me why it's being installed after you removed it from requirements | 16:06 |
| timburke | yeah, i'm pretty fuzzy on that myself | 16:07 |
| timburke | it's definitely getting installed... but idk *why* | 16:07 |
| AJaeger | It's installed ;) | 16:07 |
| AJaeger | some other package pulls it in - but I do not know which one. | 16:08 |
| AJaeger | You would need to install them one by one to see which does it | 16:08 |
| clayg | AJaeger: it's only relevant to the extent "despite py27 tests passing; this change breaks py27" | 16:08 |
| clayg | with the swift change we had a good story for "people packaing swift aren't using requirements.txt and our tests work for a predictable and understand reason" - none of that holds for swiftclient as best we can tell at this point | 16:09 |
| clayg | timburke: so I'm looking at _send_shard_ranges - it'd used all over the place | 16:10 |
| clayg | timburke: we send shard-ranges to create new shard-dbs and also to update the root | 16:10 |
| clayg | timburke: i'm less sure how it's involved in shrinking 😬 | 16:11 |
| AJaeger | clayg: it does break? Where? | 16:11 |
| AJaeger | clayg: that means the testing is not compelete, doesn't it? | 16:11 |
| *** jv|afk has joined #openstack-swift | 16:12 | |
| clayg | AJaeger: yes | 16:12 |
| *** dtantsur is now known as dtantsur|afk | 16:13 | |
| clayg | timburke: since it seems like we want to conceptually separate the system name space from "auto-create" maybe we can start thinking about which specific requests we want to create brokers if they don't exist (instead of doing it implicitly on name) | 16:14 |
| clayg | I could imagine an X-Backend-Create or something | 16:14 |
| clayg | we could imagine that's a reasonable way to go w/o having to do all the work in this change - but I'm starting to realize the container-server shouldn't be trying to infer the intent of the caller based on contours of the request | 16:16 |
| timburke | yeah, that'd be fine too -- i'm pretty sure we could get the logic tight enough around the data from the shard range(s), though -- like, if there's exactly one shard range and it matches this container and it's supposed to have state CREATED -- yeah, go ahead and create the thing! | 16:17 |
| clayg | right - exactly that - that kind of sharder logic living in the behaviors of the container-server is backwards | 16:17 |
| timburke | mm. fair enough | 16:18 |
| timburke | let me give it another spin | 16:18 |
| clayg | like conceptually - for future maintainers - the computers just do whatever we tell them | 16:18 |
| *** rpittau is now known as rpittau|afk | 16:24 | |
| timburke | clayg, i like how this is shaping up -- so far probe tests are happy; i'll clean up unit tests then push another version | 16:36 |
| zaitcev | I totally use requirements.txt for packaging, just not in an automated way. | 16:46 |
| zaitcev | Guys | 16:47 |
| zaitcev | Our regular IRC meeting is exactly at the time when our Zoom room slot is, right? It's 4 p.m. Central or 2 p.m. California. | 16:47 |
| openstackgerrit | Merged openstack/python-swiftclient master: Remove references to swift-specs and blueprints https://review.opendev.org/732998 | 17:03 |
| openstackgerrit | Tim Burke proposed openstack/swift master: Don't auto-create shard containers https://review.opendev.org/731653 | 17:05 |
| timburke | zaitcev, yes -- i'm thinking we'll just continue our PTG discussions, like we usually do during these things | 17:06 |
| zaitcev | timburke: 1. I'm trying to confirm if my calculations are correct and 2. which one we join (I presume the PTG Zoom) | 17:07 |
| timburke | yes, zoom is the way to go | 17:15 |
| *** ravsingh has quit IRC | 17:20 | |
| *** jv|afk has quit IRC | 17:42 | |
| openstackgerrit | Merged openstack/swift master: docs: s3api doesn't support tagging https://review.opendev.org/733047 | 17:43 |
| *** m75abrams has joined #openstack-swift | 17:47 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: relinker: Skip locked data dirs https://review.opendev.org/733296 | 18:06 |
| openstackgerrit | Tim Burke proposed openstack/swift master: Don't auto-create shard containers https://review.opendev.org/731653 | 18:13 |
| openstackgerrit | Merged openstack/swift master: Remove <py3.5 dependencies from requirements.txt https://review.opendev.org/732989 | 18:32 |
| openstackgerrit | Merged openstack/swift master: Switch to newer openstackdocstheme and reno versions https://review.opendev.org/732741 | 18:33 |
| *** rdejoux has quit IRC | 19:02 | |
| openstackgerrit | Clay Gerrard proposed openstack/swift master: Simplify container auto create param plumbing https://review.opendev.org/733343 | 19:04 |
| openstackgerrit | Tim Burke proposed openstack/swift master: Don't auto-create shard containers https://review.opendev.org/731653 | 19:04 |
| timburke | 🤔 should X-Backend-Auto-Create work for DELETEs? | 19:07 |
| *** jv|afk has joined #openstack-swift | 19:09 | |
| clayg | 🤯 off the cuff - YES | 19:10 |
| timburke | i wasn't thinking too hard about it before, so i'd just allowed it for PUTs since that was the only case i really cared about | 19:19 |
| timburke | all right, i'll roll in your patch, too clayg -- it'll take me a bit to get some tests for those DELETEs in place, though | 19:19 |
| timburke | and i think i'm going to go grab some lunch before our session | 19:20 |
| clayg | 💡 | 19:20 |
| *** TobbeCN has quit IRC | 19:24 | |
| *** rdejoux has joined #openstack-swift | 19:25 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: Don't auto-create shard containers https://review.opendev.org/731653 | 19:56 |
| openstackgerrit | Tim Burke proposed openstack/swift master: Don't auto-create shard containers https://review.opendev.org/731653 | 20:26 |
| timburke | clayg, so are we going to try to talk more waterfall EC in this next session? did we come up with a whiteboard-ish thing that'd work for you? | 20:28 |
| clayg | i don't have a whiteboard-ish thing ready - we could try and talk about it | 20:28 |
| clayg | i'm still thinking about it and scratching my head | 20:28 |
| *** AJaeger has left #openstack-swift | 20:28 | |
| *** jv|afk has quit IRC | 20:38 | |
| kota_ | good morning | 20:48 |
| * kota_ brewed coffee | 20:48 | |
| *** mattia has joined #openstack-swift | 20:49 | |
| *** jv|afk has joined #openstack-swift | 20:50 | |
| seongsoocho | morning | 20:50 |
| kota_ | seongsoocho: o/ | 20:51 |
| timburke | kota_, coffee's a good ide :) i hadn't gotten around to that last time | 20:54 |
| kota_ | hehe | 20:54 |
| mattoliverau | morning | 21:00 |
| zaitcev | https://etherpad.opendev.org/p/swift-ptg-victoria just in case | 21:00 |
| tdasilva | meeting link: https://zoom.us/j/94203142320 | 21:00 |
| *** ccamacho has quit IRC | 21:15 | |
| *** manuvakery has quit IRC | 21:45 | |
| zaitcev | mattia: hopefully we didn't bore you to tears | 22:26 |
| mattia | it was actually interesting to know what you are working on :) | 22:27 |
| *** jv|afk has quit IRC | 22:41 | |
| *** m75abrams has quit IRC | 22:43 | |
| *** jv|afk has joined #openstack-swift | 22:58 | |
| *** rcernin has joined #openstack-swift | 23:14 | |
| *** sorrison has quit IRC | 23:39 | |
| *** sorrison has joined #openstack-swift | 23:42 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!