| kota_ | and current swift has a way to set it as 'X-Timestamp' in request header. | 00:00 |
|---|---|---|
| timburke | kota_: and there shouldn't be any way for an S3 client to specify a custom timestamp, right? seems reasonable | 00:00 |
| kota_ | timburke: not sure, need to figure out. | 00:00 |
| kota_ | ok, my task is making sure whether or not S3 has a way for custom timestamp. | 00:02 |
| timburke | i suppose either way, it's unlikely that the way to specify it in S3 would be via an X-Timestamp header, so we're certainly no worse off than we were before | 00:02 |
| kota_ | timburke: i'm with :) | 00:04 |
| lyrrad | I’m pretty sure S3 doesn’t allow you to set the timestamp of an object. It’s assigned by the server. | 00:04 |
| kota_ | lyrrad: nice!! I always ask you "how about swift3?" :) | 00:05 |
| kota_ | ok | 00:06 |
| kota_ | thanks for those comments. they are worthfull. | 00:07 |
| kota_ | ok, moving on to next. | 00:07 |
| lyrrad | I’m not sure what Swift/Swift3’s behaviour is for X-TimeStamp | 00:07 |
| kota_ | lyrrad: sorry I meant "how about actual S3" | 00:08 |
| kota_ | my poorly typing :/ | 00:08 |
| kota_ | lyrrad: you know S3 behavior and it benefits to get improved the swift3 implementation for developpers (including me) | 00:10 |
| kota_ | lyrrad: thanks for working ;) | 00:10 |
| kota_ | so | 00:10 |
| lyrrad | The main thing to know about S3 is that the documentation is sometimes wrong, so it’s necessary to test against live servers (sometimes in different regions). You’re welcome. | 00:10 |
| kota_ | the next topic is for chsc | 00:11 |
| kota_ | sorry too late to moving to this :/ | 00:11 |
| chsc | No problem. Maybe it is a short topic. | 00:11 |
| chsc | With native swift, only users with an admin role can create containers. | 00:11 |
| kota_ | exactly | 00:12 |
| chsc | With swift3, any user with access to an account can create buckets/containers. | 00:12 |
| chsc | I assume this matches s3, where anybody can sign up and create buckets. | 00:12 |
| chsc | Depending on the environment, an administrator might not want everybody to suddenly start creating buckets/containers. | 00:13 |
| chsc | Is there a way to restrict bucket creation in swift3? What do others think of adding such a feature? | 00:13 |
| kota_ | tbh, i'm not sure but... | 00:14 |
| kota_ | s3 seems not to have such an api... | 00:15 |
| kota_ | instead of that | 00:16 |
| kota_ | s3 has a constraint an user can make only 100 buckets | 00:17 |
| kota_ | 100? 1000? not sure. | 00:17 |
| kota_ | but it has an upper limit. | 00:17 |
| lyrrad | 100 per account, though you can request more. | 00:17 |
| chsc | Interesting. Would it be useful to add a swift3 setting, so that when this is enabled only operator_roles can create swift3 buckets? | 00:17 |
| chsc | We could also follow-up on this topic during the next meeting. | 00:20 |
| kota_ | i think the behaivor depands on swif3 setting | 00:21 |
| kota_ | currently we have a couple of choice for acl setting | 00:22 |
| kota_ | i mean s3acl = True or False | 00:22 |
| kota_ | when setting it as True, all users (include just user not admin) also can create a bucket | 00:23 |
| chsc | So with s3acl = False, only admins can create buckets? | 00:23 |
| kota_ | however, when setting it as False, the acl and authentication depends on native swift so it will make user disabled to create a bucket. | 00:24 |
| chsc | ok | 00:24 |
| kota_ | but | 00:24 |
| kota_ | it drops object ACL instead. | 00:24 |
| chsc | yes | 00:24 |
| chsc | So, what about a setting that enables object and bucket acls like s3acl = True, but restricts bucket creation to admins? | 00:25 |
| kota_ | i think, we can do that to consider just implementation side but... | 00:26 |
| kota_ | not sure how we make sure the behavior which is not supported in S3... | 00:27 |
| kota_ | if we do so, i think we should do so carefully, imho. | 00:27 |
| chsc | agreed | 00:28 |
| chsc | Sorry, i have to step away. Thank you for the discussion. | 00:29 |
| kota_ | chsc: no worries, thanks for bringing up this :) | 00:29 |
| kota_ | all of agendas were addressed. | 00:30 |
| kota_ | let's get finished up today's meeting. | 00:30 |
| kota_ | thaks for working swift3 :) | 00:30 |
| gmmaha | thanks for the meeting. was very informative | 00:31 |
| kota_ | timburke: oh, i forgot to say, "congrats timburke for swiftclient-core" | 01:03 |
| timburke | thanks kota_! now i just need more patches to review :) | 01:03 |
| kota_ | timburke: no time to sleep :\ | 01:04 |
| kota_ | kidding | 01:05 |
| timburke | eh, who needs sleep? :P | 01:06 |
| timburke | (also kidding. my daughter was sick the last couple days, and i was reminded of just how much it sucked being repeatedly woken in the middle of the night) | 01:06 |
| *** bill_az has quit IRC | 01:08 | |
| kota_ | please take care of you and your faimily. | 01:08 |
| timburke | thanks kota_ -- she's doing much better today :) | 01:11 |
| kota_ | sounds great :) | 01:11 |
| *** chsc has quit IRC | 01:15 | |
| *** lyrrad has quit IRC | 01:36 | |
| *** bill_az has joined #swift3 | 02:12 | |
| *** bill_az has quit IRC | 02:21 | |
| *** openstackgerrit has quit IRC | 09:46 | |
| *** openstackgerrit has joined #swift3 | 09:47 | |
| *** openstackgerrit has quit IRC | 10:16 | |
| *** openstackgerrit has joined #swift3 | 10:17 | |
| *** bill_az has joined #swift3 | 14:08 | |
| *** bill_az has quit IRC | 14:17 | |
| *** chsc has joined #swift3 | 16:33 | |
| *** chsc has quit IRC | 23:44 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!