| *** gmmaha has joined #swift3 | 00:36 | |
| *** tdasilva has quit IRC | 05:11 | |
| *** tdasilva has joined #swift3 | 05:30 | |
| kota_ | meeting agenda updaetd | 07:00 |
|---|---|---|
| kota_ | s/updaetd/updated/ | 07:18 |
| *** cschwede has quit IRC | 07:43 | |
| *** cschwede has joined #swift3 | 07:45 | |
| *** cschwede has quit IRC | 07:51 | |
| *** cschwede has joined #swift3 | 07:52 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift3: WIP: Fix last modified on copy result https://review.openstack.org/242578 | 11:59 |
| *** openstackgerrit has quit IRC | 12:31 | |
| *** openstackgerrit has joined #swift3 | 12:32 | |
| *** openstackgerrit has quit IRC | 16:02 | |
| *** openstackgerrit has joined #swift3 | 16:02 | |
| *** kota_ has quit IRC | 16:17 | |
| *** kota_ has joined #swift3 | 16:17 | |
| *** chsc has joined #swift3 | 16:29 | |
| *** chsc has quit IRC | 16:35 | |
| *** chsc has joined #swift3 | 16:36 | |
| *** chsc has quit IRC | 22:37 | |
| *** pgbridge has quit IRC | 22:44 | |
| kota_ | meeting time | 23:00 |
| kota_ | who one else here? | 23:01 |
| timburke | hello! | 23:02 |
| kota_ | timburke: hello! | 23:03 |
| kota_ | just time I'm wondering it would be good to skip :P | 23:03 |
| timburke | i wonder if other people weren't thinking about the meeting time as UTC... | 23:03 |
| kota_ | mihgt be time changed? | 23:04 |
| *** bill_az has joined #swift3 | 23:04 | |
| kota_ | PDT -> PST? i mean. | 23:04 |
| kota_ | i'm not sure Japan doesn't have such time clock change culture. | 23:04 |
| kota_ | always we use just JST w/o day light saving. | 23:05 |
| timburke | yeah, it's surely better that way :/ | 23:05 |
| kota_ | I can be up in one hour later. It seems better? | 23:06 |
| gmmaha | kota_: timburke: Do you run meeting in #openstack-meeting channel? | 23:06 |
| kota_ | gmmaha: no, that happens in this channel. | 23:07 |
| gmmaha | aah ok. thanks | 23:07 |
| *** chsc has joined #swift3 | 23:07 | |
| bill_az | Hi everyone - sorry I'm a little late | 23:08 |
| bill_az | Hopefully it was a great summit kota! | 23:08 |
| chsc | Hi | 23:08 |
| chsc | Sorry, i am late, too. | 23:09 |
| kota_ | bill_az: chsc: no worries | 23:09 |
| chsc | I would have a question about bucket creation in swift3, but we can discuss other items first. | 23:10 |
| kota_ | chsc: ok | 23:10 |
| kota_ | it's good time to start now? | 23:11 |
| kota_ | let get started. | 23:11 |
| kota_ | today's meeting agenda is here, https://wiki.openstack.org/wiki/Meetings/swift3 | 23:11 |
| kota_ | 2 items there, and 1 item from chsc just now. | 23:12 |
| kota_ | moving on 1st one. | 23:12 |
| kota_ | summit recap | 23:12 |
| kota_ | we had a great productive week the last week in the last october. | 23:13 |
| kota_ | a bunch of talk about Swift and other ecosystem (definitely also swift3) happend there. | 23:14 |
| kota_ | about Swift, I saw notmyname sent an E-mail to summurize the discussion conclusion yesterday, please check it too. | 23:15 |
| kota_ | the E-mail was sent to mailing list. | 23:16 |
| kota_ | about swift3 side, alpha_ori, dean from IBM, and I had some talks for next step in Mitaka cycle. | 23:17 |
| kota_ | is alpha_ori here? | 23:17 |
| kota_ | seems not... | 23:18 |
| timburke | i know he's not behind me... | 23:18 |
| kota_ | ok | 23:18 |
| kota_ | in tokyo summit, we thought we can forcus on the current way to improve some swift3 stuff, | 23:19 |
| kota_ | i.e. feature docs and versioning | 23:19 |
| kota_ | i saw symlink discussion which is needed by versioning looked to go well | 23:20 |
| kota_ | so i think we don't have significant blocker for the improvement right now. | 23:21 |
| kota_ | that's all from me for summit recap. | 23:23 |
| kota_ | ah, one thing | 23:23 |
| kota_ | i think we want a scheduled room for swiftt3 discussion next Austin but currently i have no idea how we can do so... | 23:24 |
| kota_ | anyways, it was a good summit. | 23:25 |
| kota_ | if you all have questions about the summit topics, please bring it up to this channel. | 23:26 |
| gmmaha | kota_: are there documents i can read into to see what improvements are being planned for this cycle? | 23:27 |
| kota_ | gmmaha: good question | 23:27 |
| kota_ | currently we are planning to put an auto-generation (semi-auto?) documentation tool from swiftstack which shows "which feature is supported by swift3" | 23:28 |
| kota_ | it will make us easy to make sure the avilable stuff on swift3 | 23:30 |
| gmmaha | ohh thats great. Guessing it will update the docs within the project ? or we will/have a wiki page to feed the data into? | 23:30 |
| *** lyrrad has joined #swift3 | 23:30 | |
| timburke | speak of the devil. lyrrad has been working on the compatibility testing tools | 23:32 |
| kota_ | timburke: I meant that one! | 23:33 |
| lyrrad | A few weeks ago I added custom categorization, and I went through and categorized all existing tests. It should then generate a custom report that could be used in MediaWiki format. Has anyone tried the updated tool? | 23:33 |
| kota_ | not yet but sounds cool. | 23:34 |
| kota_ | so next step for that is making sure/using it in upstream :) | 23:35 |
| kota_ | anything else for summit recap? | 23:36 |
| kota_ | moving on next one | 23:37 |
| lyrrad | We can also look into adding additional tests and integrating them into the generated reports, separate from ceph-tests if we want to test issues/features that aren’t already covered. | 23:37 |
| kota_ | lyrrad: good | 23:37 |
| kota_ | the next one is about requet timestamp | 23:38 |
| kota_ | that came from a patch landed recently | 23:39 |
| kota_ | in my mind | 23:39 |
| kota_ | the patch is https://review.openstack.org/#/c/223880/ | 23:39 |
| kota_ | patch 223880 | 23:39 |
| patchbot | kota_: https://review.openstack.org/#/c/223880/ - Ensure object copy response timestamp tags consist... | 23:39 |
| kota_ | the patch seems pretty reasonable lyrrad did a good work for s3 compatibility with making sure the actuall behavior | 23:40 |
| kota_ | that patch changed the response of COPY request | 23:41 |
| kota_ | COPY request have a docs in the response body explain the status (e.g. last-modified) | 23:41 |
| kota_ | and the last-modified should be the timestamp when the COPY is requested. | 23:42 |
| kota_ | i think that is good improvement *BUT* still wondering it might break something. | 23:43 |
| kota_ | with that change, current swift3 uses "Response.environ['HTTP_X_TIMESTAMP']" as the last-modified timestamp. | 23:44 |
| kota_ | However, X-Timestamp header is added to *request* by proxy-server app basically. | 23:46 |
| kota_ | i know we should init the response with a request instance which adds request environ (i.e. also HTTP_* headers) into request. | 23:47 |
| kota_ | but depending on such a coding constraint, we can break easily if we missed to make a code like just 'raise HTTPOk()' | 23:48 |
| kota_ | that will work well in Swift but it will break current swift3 | 23:49 |
| kota_ | and it seems hard to maintain thirdparty middleware also follow the correct semantics (or not9 | 23:50 |
| kota_ | (or not) | 23:50 |
| kota_ | that's why i bring this up to this meeting. | 23:50 |
| timburke | kota_: how about use the X-Timestamp if we have it, or time.time() (or some equivalent) if we don't? | 23:51 |
| kota_ | to solve this issue, i am working with my poc code on 242578 | 23:51 |
| kota_ | timburke: i think it's what i'm working on patch 242578 | 23:52 |
| patchbot | kota_: https://review.openstack.org/#/c/242578/ - WIP: Fix last modified on copy result | 23:52 |
| lyrrad | Will this timestamp match the timestamp that’s stored with the newly created object? For example, the timestamp in a bucket listing? | 23:52 |
| kota_ | lyrrad: when we add 'X-Timestamp' for PUT/POST (COPY, too) request, swift will handles as it is requested at the timestamp. | 23:54 |
| lyrrad | The reason I tried to use the timestamp from the request was so that it would match the timestamp used in other contexts. One might imagine a client doing a naive check to see if an object they put was still there by comparing the timestamp returned when the object was created to the current timestamp. | 23:54 |
| kota_ | the X-Timestamp should be matched with actual created timestamp. | 23:57 |
| kota_ | I don't guess patch 242578 changes the semantics following your intention that we want to use requested put timestamp as last-modified. | 23:58 |
| patchbot | kota_: https://review.openstack.org/#/c/242578/ - WIP: Fix last modified on copy result | 23:58 |
| kota_ | just moving the creation from swift into swift3 to be controllable in swift3. | 23:59 |
| lyrrad | Ah, okay. If the times match then that should work well. | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!