| *** diogogmt has joined #openstack-swift | 00:00 | |
| *** jamielennox|away is now known as jamielennox | 00:00 | |
| notmyname | does swauth support a v2 auth api? | 00:06 |
|---|---|---|
| timburke | notmyname: grepping for v2 in https://github.com/openstack/swauth/blob/master/swauth/middleware.py i'm gonna say "no" -- certainly not in the sense you mean | 00:13 |
| notmyname | timburke: ok, thanks | 00:13 |
| timburke | notmyname: fwiw, i don't think it would be very hard to implement a bare-bones v2 (or even v3) api for tempauth that imposes a heap of assumptions and limitations | 00:15 |
| timburke | like "there's only one domain" | 00:15 |
| notmyname | nah, I was thinking in the context "can we reasonably expect that clusters with swuath could be using v2+ auth APIs". seems the answer is "no" | 00:16 |
| notmyname | in the context of the "right" answer for... | 00:17 |
| openstackgerrit | John Dickinson proposed openstack/python-swiftclient: Update help text around container ACL syntax https://review.openstack.org/324739 | 00:17 |
| notmyname | hogepodge: ^ | 00:17 |
| *** patchbot has quit IRC | 00:19 | |
| *** tqtran has quit IRC | 00:19 | |
| notmyname | oops. didn't mean to ctl-c that window | 00:19 |
| *** patchbot has joined #openstack-swift | 00:19 | |
| notmyname | ok, I gotta run. /later | 00:20 |
| *** nadeem has quit IRC | 00:25 | |
| *** jlvillal has quit IRC | 00:41 | |
| *** jlvillal has joined #openstack-swift | 00:46 | |
| *** jlvillal has quit IRC | 00:47 | |
| *** jlvillal has joined #openstack-swift | 00:47 | |
| *** jlvillal has quit IRC | 00:47 | |
| hogepodge | notmyname: thanks | 00:47 |
| *** jlvillal has joined #openstack-swift | 00:47 | |
| *** manous has joined #openstack-swift | 00:57 | |
| *** m_kazuhiro has joined #openstack-swift | 00:57 | |
| *** klrmn has quit IRC | 01:13 | |
| *** neha__ has quit IRC | 01:16 | |
| *** cdelatte has quit IRC | 01:25 | |
| *** neha__ has joined #openstack-swift | 01:30 | |
| *** neha__ has quit IRC | 01:36 | |
| openstackgerrit | Matthew Oliver proposed openstack/swift: Use more specific asserts in test/unit/common/middleware https://review.openstack.org/342770 | 01:55 |
| mattoliverau | ^^ just a flyby rebase. | 01:56 |
| *** manous has quit IRC | 01:59 | |
| openstackgerrit | Tim Burke proposed openstack/python-swiftclient: Add v1password keystoneauth plugin https://review.openstack.org/300697 | 02:09 |
| *** diogogmt has quit IRC | 02:09 | |
| *** manous has joined #openstack-swift | 02:11 | |
| *** bkopilov has quit IRC | 02:16 | |
| *** manous has quit IRC | 02:24 | |
| openstackgerrit | Merged openstack/python-swiftclient: Make tempurl subcommand insist on whole number seconds https://review.openstack.org/367915 | 02:34 |
| *** manous has joined #openstack-swift | 02:37 | |
| *** manous has quit IRC | 02:42 | |
| *** klrmn has joined #openstack-swift | 02:42 | |
| *** manous has joined #openstack-swift | 02:43 | |
| *** manous has quit IRC | 02:48 | |
| *** manous has joined #openstack-swift | 02:49 | |
| *** baojg has joined #openstack-swift | 03:27 | |
| jamielennox | timburke: hey, nice ^ | 03:28 |
| jamielennox | timburke: question if you're here, is there the need to inherit BaseIdentityPlugin or do you just want BasePlugin? | 03:29 |
| *** bkopilov has joined #openstack-swift | 03:33 | |
| *** chsc has joined #openstack-swift | 03:41 | |
| *** chsc has quit IRC | 03:48 | |
| *** tsg has joined #openstack-swift | 03:51 | |
| *** mingyu has joined #openstack-swift | 04:05 | |
| *** dmorita has quit IRC | 04:05 | |
| *** dmorita has joined #openstack-swift | 04:05 | |
| *** dmorita has quit IRC | 04:10 | |
| *** zaitcev has quit IRC | 04:16 | |
| *** psachin has joined #openstack-swift | 04:21 | |
| mahatic | good morning | 04:34 |
| *** ppai has joined #openstack-swift | 04:35 | |
| *** tsg has quit IRC | 04:43 | |
| *** tsg has joined #openstack-swift | 04:43 | |
| *** dmorita has joined #openstack-swift | 04:46 | |
| *** aswadr_ has joined #openstack-swift | 04:49 | |
| kota_ | mahatic: good morning | 04:49 |
| *** dmorita has quit IRC | 04:51 | |
| *** CaioBrentano has quit IRC | 04:58 | |
| *** CaioBrentano has joined #openstack-swift | 04:59 | |
| mahatic | kota_: o/ | 05:00 |
| *** Jeffrey4l__ has joined #openstack-swift | 05:00 | |
| *** mingyu has quit IRC | 05:03 | |
| *** Jeffrey4l_ has quit IRC | 05:04 | |
| *** tqtran has joined #openstack-swift | 05:17 | |
| *** rcernin has joined #openstack-swift | 05:17 | |
| *** tqtran has quit IRC | 05:22 | |
| *** ChubYann has quit IRC | 05:38 | |
| *** McMurlock1 has joined #openstack-swift | 05:45 | |
| *** tsg has quit IRC | 05:49 | |
| *** pcaruana has joined #openstack-swift | 06:04 | |
| *** pcaruana is now known as pcaruana|afk| | 06:13 | |
| openstackgerrit | Ha Van Tu proposed openstack/swift: Update link reference https://review.openstack.org/370554 | 06:28 |
| *** asettle has joined #openstack-swift | 06:29 | |
| *** manous has quit IRC | 06:33 | |
| *** asettle has quit IRC | 06:34 | |
| *** pcaruana|afk| is now known as pcaruana | 06:40 | |
| *** jamielennox is now known as jamielennox|away | 06:45 | |
| *** manous has joined #openstack-swift | 06:47 | |
| *** klrmn has quit IRC | 06:59 | |
| *** manous_ has joined #openstack-swift | 07:04 | |
| *** manous has quit IRC | 07:07 | |
| *** tqtran has joined #openstack-swift | 07:08 | |
| *** zul has joined #openstack-swift | 07:10 | |
| *** rledisez has joined #openstack-swift | 07:14 | |
| *** mingyu has joined #openstack-swift | 07:14 | |
| *** mingyu has quit IRC | 07:16 | |
| *** mingyu has joined #openstack-swift | 07:17 | |
| *** sams-gleb has joined #openstack-swift | 07:30 | |
| *** asettle has joined #openstack-swift | 07:37 | |
| *** asettle has quit IRC | 07:39 | |
| *** zul has quit IRC | 07:57 | |
| *** rledisez has quit IRC | 07:57 | |
| *** jamielennox|away is now known as jamielennox | 07:57 | |
| *** rledisez has joined #openstack-swift | 07:59 | |
| *** cbartz has joined #openstack-swift | 08:14 | |
| *** hseipp has joined #openstack-swift | 08:22 | |
| *** asettle has joined #openstack-swift | 08:52 | |
| *** asettle has quit IRC | 08:57 | |
| *** asettle has joined #openstack-swift | 08:58 | |
| *** jordanP has joined #openstack-swift | 09:11 | |
| *** mingyu has quit IRC | 09:17 | |
| *** mingyu has joined #openstack-swift | 09:17 | |
| *** mvk has quit IRC | 09:20 | |
| *** gabor_antal has joined #openstack-swift | 09:21 | |
| *** tqtran has quit IRC | 09:26 | |
| *** mingyu has quit IRC | 09:30 | |
| *** mingyu has joined #openstack-swift | 09:31 | |
| *** cbartz has quit IRC | 09:49 | |
| *** mvk has joined #openstack-swift | 09:50 | |
| *** mingyu has quit IRC | 09:54 | |
| *** mingyu has joined #openstack-swift | 09:57 | |
| *** kei_yama has quit IRC | 10:03 | |
| *** mingyu has quit IRC | 10:08 | |
| *** mingyu has joined #openstack-swift | 10:11 | |
| *** mingyu has quit IRC | 10:12 | |
| *** mingyu has joined #openstack-swift | 10:14 | |
| *** acoles_ is now known as acoles | 10:18 | |
| *** manous_ has quit IRC | 10:18 | |
| *** mingyu has quit IRC | 10:21 | |
| *** mingyu has joined #openstack-swift | 10:25 | |
| *** sams-gle_ has joined #openstack-swift | 10:49 | |
| *** sams-gleb has quit IRC | 10:49 | |
| *** baojg has quit IRC | 11:01 | |
| *** bkopilov has quit IRC | 11:08 | |
| *** baojg has joined #openstack-swift | 11:18 | |
| *** mingyu_ has joined #openstack-swift | 11:20 | |
| *** 7ITAAJJC3 has joined #openstack-swift | 11:20 | |
| *** 7JTABRCL5 has joined #openstack-swift | 11:20 | |
| *** mingyu has quit IRC | 11:23 | |
| *** ppai has quit IRC | 11:30 | |
| *** asettle has quit IRC | 11:30 | |
| *** ppai has joined #openstack-swift | 11:31 | |
| *** mingyu_ has quit IRC | 11:56 | |
| *** mingyu has joined #openstack-swift | 11:58 | |
| *** CaioBrentano has quit IRC | 11:59 | |
| *** CaioBrentano has joined #openstack-swift | 12:00 | |
| *** CaioBrentano has quit IRC | 12:08 | |
| *** CaioBrentano has joined #openstack-swift | 12:10 | |
| *** donagh has quit IRC | 12:11 | |
| *** cbartz has joined #openstack-swift | 12:18 | |
| *** CaioBrentano has quit IRC | 12:18 | |
| *** CaioBrentano has joined #openstack-swift | 12:19 | |
| *** geaaru has quit IRC | 12:21 | |
| *** CaioBrentano has quit IRC | 12:26 | |
| *** CaioBrentano has joined #openstack-swift | 12:27 | |
| *** asettle has joined #openstack-swift | 12:30 | |
| *** mingyu has quit IRC | 12:30 | |
| *** bkopilov has joined #openstack-swift | 12:32 | |
| *** CaioBrentano has quit IRC | 12:34 | |
| *** CaioBrentano has joined #openstack-swift | 12:35 | |
| *** geaaru has joined #openstack-swift | 12:36 | |
| *** asettle has quit IRC | 12:36 | |
| *** CaioBrentano has quit IRC | 12:42 | |
| *** CaioBrentano has joined #openstack-swift | 12:43 | |
| *** m_kazuhiro has quit IRC | 12:54 | |
| *** CaioBrentano has quit IRC | 12:59 | |
| *** CaioBrentano has joined #openstack-swift | 13:00 | |
| *** StraubTW has joined #openstack-swift | 13:04 | |
| *** nikivi has joined #openstack-swift | 13:06 | |
| *** dmorita has joined #openstack-swift | 13:06 | |
| *** CaioBrentano has quit IRC | 13:08 | |
| *** vinsh has quit IRC | 13:08 | |
| *** CaioBrentano has joined #openstack-swift | 13:09 | |
| *** dmorita has quit IRC | 13:11 | |
| openstackgerrit | ChangBo Guo(gcb) proposed openstack/swift: Add unit test content https://review.openstack.org/353444 | 13:17 |
| *** CaioBrentano has quit IRC | 13:22 | |
| *** CaioBrentano has joined #openstack-swift | 13:22 | |
| *** zaitcev has joined #openstack-swift | 13:25 | |
| *** ChanServ sets mode: +v zaitcev | 13:25 | |
| *** CaioBrentano has quit IRC | 13:29 | |
| *** CaioBrentano has joined #openstack-swift | 13:30 | |
| *** asettle has joined #openstack-swift | 13:33 | |
| *** tsg has joined #openstack-swift | 13:36 | |
| *** caiobrentano__ has joined #openstack-swift | 13:37 | |
| *** asettle has quit IRC | 13:37 | |
| *** CaioBrentano has quit IRC | 13:37 | |
| *** caiobrentano__ has quit IRC | 13:38 | |
| *** caiobrentano__ has joined #openstack-swift | 13:39 | |
| *** nikivi has quit IRC | 13:43 | |
| *** vint_bra1 has joined #openstack-swift | 13:43 | |
| *** dstepanenko_ has quit IRC | 13:45 | |
| *** vint_bra has quit IRC | 13:46 | |
| *** dstepanenko_ has joined #openstack-swift | 13:46 | |
| *** nikivi has joined #openstack-swift | 13:49 | |
| *** 7JTABRCL5 has quit IRC | 13:57 | |
| *** 7ITAAJJC3 has quit IRC | 13:57 | |
| *** chsc has joined #openstack-swift | 14:02 | |
| *** sgundur_ has joined #openstack-swift | 14:08 | |
| *** asettle has joined #openstack-swift | 14:09 | |
| *** nikivi has quit IRC | 14:12 | |
| *** sgundur_ has quit IRC | 14:12 | |
| *** asettle has quit IRC | 14:15 | |
| *** sgundur_ has joined #openstack-swift | 14:16 | |
| *** neha__ has joined #openstack-swift | 14:23 | |
| *** vinsh has joined #openstack-swift | 14:29 | |
| *** caiobrentano__ has quit IRC | 14:34 | |
| *** ppai has quit IRC | 14:35 | |
| *** joeljwright has joined #openstack-swift | 14:36 | |
| *** ChanServ sets mode: +v joeljwright | 14:36 | |
| *** CaioBrentano has joined #openstack-swift | 14:38 | |
| *** sgundur_ has quit IRC | 14:40 | |
| *** klamath has joined #openstack-swift | 14:40 | |
| *** tsg has quit IRC | 14:41 | |
| *** sgundur_ has joined #openstack-swift | 14:41 | |
| klamath | question on container sync, If i delete a sync container on both sides and setup the exact same container names again, container sync refuses to run, seeing a 404 when trying to transfer over database files | 14:41 |
| *** cbartz has quit IRC | 14:47 | |
| *** chsc has quit IRC | 14:52 | |
| *** diogogmt has joined #openstack-swift | 15:03 | |
| *** sgundur_ has quit IRC | 15:05 | |
| *** sgundur_ has joined #openstack-swift | 15:07 | |
| *** klrmn has joined #openstack-swift | 15:10 | |
| *** asettle has joined #openstack-swift | 15:11 | |
| *** zul has joined #openstack-swift | 15:12 | |
| *** sams-gle_ has quit IRC | 15:14 | |
| *** asettle has quit IRC | 15:16 | |
| *** pcaruana has quit IRC | 15:19 | |
| *** neha__ has quit IRC | 15:19 | |
| *** klamath has quit IRC | 15:19 | |
| *** diogogmt has quit IRC | 15:20 | |
| *** zul has quit IRC | 15:22 | |
| *** klamath has joined #openstack-swift | 15:22 | |
| *** tsg has joined #openstack-swift | 15:25 | |
| *** nadeem has joined #openstack-swift | 15:42 | |
| *** tsg has quit IRC | 15:54 | |
| *** thumpba has joined #openstack-swift | 16:00 | |
| *** tsg has joined #openstack-swift | 16:00 | |
| *** sgundur_ has quit IRC | 16:02 | |
| *** dmorita has joined #openstack-swift | 16:03 | |
| *** dmorita has quit IRC | 16:07 | |
| notmyname | good morning | 16:08 |
| *** asettle has joined #openstack-swift | 16:12 | |
| *** McMurlock1 has quit IRC | 16:16 | |
| *** asettle has quit IRC | 16:17 | |
| *** hseipp has quit IRC | 16:23 | |
| *** rcernin has quit IRC | 16:26 | |
| *** chsc has joined #openstack-swift | 16:28 | |
| *** bkopilov has quit IRC | 16:30 | |
| *** asettle has joined #openstack-swift | 16:31 | |
| clayg | notmyname: and what makes you so sure it's good? | 16:32 |
| *** dmorita has joined #openstack-swift | 16:33 | |
| clayg | notmyname: swauth has a catalog/endpoints auth api tho - but it was an extension of v1.1 like auth instead of a whole new thing | 16:33 |
| *** chsc has quit IRC | 16:34 | |
| *** dmorita has quit IRC | 16:35 | |
| *** dmorita has joined #openstack-swift | 16:35 | |
| notmyname | TIL this exists. that's great! http://docs.openstack.org/user-guide/cli-swift-manage-access-swift.html | 16:35 |
| *** asettle has quit IRC | 16:36 | |
| clayg | notmyname: doesn't look like that support v3 style domains | 16:37 |
| clayg | notmyname: aiui, acoles and donagh worked real hard to make acl's still be backwards compatible and also secure - but it requires some sort of legacy/migration syntax where you specifiy a default domain (default is DEFAULT) | 16:38 |
| clayg | but going forward I assumed the recommended syntax was to define users and projects based on their domain (because id's can conflict) | 16:39 |
| clayg | or something? | 16:39 |
| *** sgundur_ has joined #openstack-swift | 16:39 | |
| *** jordanP has quit IRC | 16:40 | |
| clayg | on, nm, that last part says project_uuid:user_uuid format is fine - i guess... maybe we like regex match to decide if it's a uuid? | 16:41 |
| acoles | clayg: we made it so that names in acls would still work but only in the legacy (default) domain, going forwards ids should be used, and AIUI ids are globally unique | 16:42 |
| acoles | IIRC we considered regex matching etc but seemed impossible to distinguish a name from an id | 16:43 |
| clayg | acoles: phew | 16:43 |
| clayg | acoles: so clusters with a legacy/default domain can never upgrade to use multiple auth domains? | 16:43 |
| acoles | clayg: BUT... the default domain used to have id=default but now it doesn't (I learnt recently) .. so when I have time i need to re-check the logic we used for the backwards compatibility piece :/ | 16:44 |
| clayg | rofl | 16:44 |
| notmyname | clayg: see? it's a great morning. you're laughing already! :-) | 16:44 |
| acoles | clayg: yes they can upgrade, but you can't use names in acls in "new" non-default domains (well, you can use them but they will not be validated) | 16:45 |
| acoles | clayg: I do NOT rofl when I learnt that, more pcbabhod | 16:45 |
| clayg | acoles: but how do you know if a ACL is for the default domain or the new non-default domains? | 16:45 |
| notmyname | push chair back and bash head on desk? | 16:46 |
| * acoles pushed chair back and beat head on desk | 16:46 | |
| acoles | yeah, that! | 16:46 |
| notmyname | I win! ;-) | 16:46 |
| clayg | wow, you guys broke goolge! "No results found for pcbabhod | 16:46 |
| patchbot | Error: No closing quotation | 16:46 |
| clayg | STFU PATCHBOT | 16:46 |
| acoles | clayg: we learn which domain an account is in so we know when to allow names or not | 16:47 |
| notmyname | I just went through the patchbot config yesterday to turn off everything but my plugin. didn't seem to work | 16:47 |
| clayg | acoles: keystone does not have accounts - you mean projects/tenants? | 16:47 |
| acoles | clayg: account sysmeta has domain id, if its missing we assume the account was created under v2 and is therefore in default domain | 16:48 |
| clayg | acoles: i'm so glad you remember/understand this | 16:48 |
| *** thumpba has quit IRC | 16:48 | |
| acoles | clayg: yep, we learn which domain a swift account is in by looking for keystone project_domain_id in token info | 16:48 |
| clayg | acoles: do projects have domains or only users? | 16:48 |
| *** thumpba has joined #openstack-swift | 16:49 | |
| acoles | since account <-> project | 16:49 |
| acoles | both project and users are in domains | 16:49 |
| clayg | acoles: ok, it's starting to make sense (really - you're doing a great job) | 16:49 |
| clayg | thank you and sorry that I don't seem willing to just read/test it myself | 16:50 |
| clayg | but someday... i'm going to be devstack and keystone wizard! | 16:50 |
| acoles | BTW, I think the change of default domain default id from being default to something else is not harmful. (I love typing that sentence! it makes me grin) | 16:50 |
| acoles | clayg: np, in return I reserve the right to never learn the ring and always pester you :P | 16:51 |
| clayg | acoles: you wanna talk about optimistic frag serving? | 16:51 |
| *** nadeem has quit IRC | 16:52 | |
| acoles | notmyname: clayg: while working on the api-ref changes donagh and I identified the need for a single, comprehensive doc on ACLs. As a bonus, it would also be correct. | 16:52 |
| *** nadeem has joined #openstack-swift | 16:53 | |
| acoles | clay: can do. I'm just working on kota_ 's comments, haven't got to yours but go ahead | 16:53 |
| *** thumpba has quit IRC | 16:54 | |
| notmyname | "as a bonus, it would also be correct" | 16:54 |
| acoles | clayg: did you find a bug wrt reconstructor or just notice that this patch does not make the recnstructor optimistic (but should make it no worse) | 16:54 |
| *** asettle has joined #openstack-swift | 16:55 | |
| clayg | http://blog.equalrightsinstitute.com/wp-content/uploads/2015/06/IM.png | 16:56 |
| timburke | jamielennox: i wrote it long enough ago that i don't remember :-/ i'll take a look at using BaseAuthPlugin instead, thanks for the tip! | 16:56 |
| clayg | oh, no you did a pretty good job - the second one | 16:56 |
| *** rledisez has quit IRC | 16:57 | |
| *** McMurlock1 has joined #openstack-swift | 16:59 | |
| *** asettle has quit IRC | 17:01 | |
| acoles | clayg: so i thought we could make the reconstructor optimistic as a follow on (in fact not could but should) - like your gist or a refactor so that reconstructor can re-use the proxy obj controller code maybe | 17:02 |
| *** tsg_ has joined #openstack-swift | 17:03 | |
| *** itlinux has joined #openstack-swift | 17:03 | |
| acoles | clayg: kota raised a concern about the chance of different etags being found at same timestamp - I have added a check but I can't think how we would ever expect that to happen other than two concurrent puts use exact same timestamp, but then the consistency engine would never fix up the object so we're hosed anyway | 17:04 |
| clayg | acoles: yeah we mostly decided that it can't use a proxy like thing because it's trying to output just the rebuilt frag - not the whole object (which would then be encoded and most of the frags thrown away) | 17:04 |
| clayg | acoles: but if the proxy could use a lib that was "connect to all the backend and provide a frag iterator" - then the proxy could feed that to decode and the reconstrutor could feed thtat to rebuild? | 17:05 |
| clayg | acoles: agree - not a concern | 17:06 |
| acoles | clayg: oh yeah, so we did, ...but as you say there must be some kind of refactoring/interface that could be shared...haven't had chance to give it much thought yet tbh | 17:06 |
| clayg | but the reconstructor bug is pretty bad - i haven't filed it yet - but i'm sorry worried I should be honest and set it to critical despite the fact that it's pre-existed - it's really not great | 17:06 |
| *** tsg has quit IRC | 17:07 | |
| clayg | then while thinking of that I'm thinking that I don't really know when someone asking a node for a piece of something wouldn't ever want the object server to 404 if it has *something* | 17:07 |
| *** ChubYann has joined #openstack-swift | 17:07 | |
| clayg | basically everywhere we collect frags we already have to have all this code to group by frag's and timestamps - so the 404 is just always not helpful | 17:08 |
| acoles | clayg: thanks for test_safe_json_loads | 17:08 |
| acoles | right | 17:08 |
| clayg | if open just started kicking up non-durable frags without any additional params - who would break? ssync sender probably right? but not suffix hashing - suffix hashing doesn't open diskfiles? | 17:13 |
| *** McMurlock1 has quit IRC | 17:16 | |
| *** thumpba has joined #openstack-swift | 17:17 | |
| clayg | gah, i need a diskfile map :'( | 17:23 |
| *** sgundur_ has quit IRC | 17:23 | |
| acoles | clayg: if you mean we make object server return non-durable without needing the prefs header at all, I think non-upgraded proxy GET might break, or start not getting objects when a partial put has left some non-durables have landed on top of the durable frags | 17:24 |
| *** sgundur_ has joined #openstack-swift | 17:24 | |
| acoles | clayg: at least IIRC thats why I made optimism in obj server depend on new proxy server sending the header | 17:24 |
| clayg | well that would be a pretty good reason to do it... but I thought proxy already had some bucketizing? | 17:25 |
| acoles | suffix hashing shouldn't need to open diskfiles | 17:25 |
| clayg | acoles: it does not - but it reuses all the cleanup/get ondisk_files methods | 17:25 |
| acoles | clayg: it does have buckets per etag but will only ask each node once, so if it gets a non-durable from a partial put then that would prevent it ever getting the durable, whereas on master it only gets the durables | 17:26 |
| timburke | acoles: if we've been just *terrible* about writing down .durables, and find ourselves in a situation where we've got enough frags for an object at both t1 and t2 but only *one* .durable for either, and on separate nodes... the object that we reconstruct depends on the order of the responses we collect, despite everyone knowing that there's a t2 out there? do we just chalk that up to eventual consistency? | 17:27 |
| clayg | it looks like the current proxy code will dedupe etags (and pick the first one to > ndata) - but it doesn't dedupe frag indexes - so we can 503 anyway | 17:29 |
| acoles | timburke: first set of frags to reach a valid set wins, so yes there is a dependency on response order...but with a replicated policy we may get back obj at t1 or t2, depends on which node the proxy chooses. The algorithm could go on to try to gather all the frags at t2, its just more complex | 17:29 |
| clayg | timburke: is that possible? only if you have ndata + nparity > 2 x ndata right? | 17:29 |
| clayg | acoles: boooo complexity! | 17:29 |
| acoles | clayg: this is true "it doesn't dedupe frag indexes - so we can 503 anyway" | 17:30 |
| clayg | acoles: EC has a lot of crappy bugs | 17:31 |
| *** mvk has quit IRC | 17:32 | |
| acoles | timburke: not saying it could not be a follow up enhancement though. if the proxy did continue hunting down frags at t2 there is no guarantee of course that all those frags hinted at for t2 *will* be successfully fetched, if failures occur etc, so could result in wasted extra GETs. I guess I opted for "once I can reconstruct an object then return it" | 17:33 |
| acoles | clayg: pretty sure I add a test for dupe frag indexes | 17:33 |
| clayg | acoles: I think doing anything after you have an object which can be succesfully reconstructed is a mis-feature | 17:34 |
| acoles | clayg: https://github.com/openstack/swift/blob/5dc5f6d3d59a293f2048d5c446aa6996e61f765c/test/unit/proxy/controllers/test_obj.py#L2665-L2665 | 17:34 |
| clayg | acoles: yeah that should be easy to unittest - in the real world i'm not sure how often it comes up outside of the handoff becomes a primary in a rebalance situation | 17:35 |
| clayg | that test looks familiar-ish | 17:35 |
| acoles | handoff fails to clean up after sync'ing to primary? | 17:36 |
| acoles | I probably test a bunch of crazy scenarios that are hard to happen IRL | 17:36 |
| clayg | acoles: how did you do that - is that on master? | 17:37 |
| acoles | no its in the patch - I right click on the method in pycharm and select "Show on github", the elves do stuff and it appears in my browser | 17:38 |
| clayg | holy crap pycharm!? | 17:39 |
| *** chsc has joined #openstack-swift | 17:39 | |
| *** ppai has joined #openstack-swift | 17:39 | |
| clayg | i didn't even know the gerrit review shas were mirror'd into github!? | 17:40 |
| acoles | come to think of it it surprises me that it works when i am not on master | 17:40 |
| clayg | don't stare directly at the elves | 17:40 |
| acoles | lol. but it is in the patch not master: https://review.openstack.org/#/c/215276/40/test/unit/proxy/controllers/test_obj.py@2665 | 17:41 |
| patchbot | patch 215276 - swift - Enable object server to return non-durable data | 17:41 |
| clayg | yeah no worries | 17:41 |
| *** itlinux has quit IRC | 17:43 | |
| *** Suyi_ has joined #openstack-swift | 17:43 | |
| clayg | oh yuk - we call cleanup_ondisk_files once when we're yeilding hashes - then again when we open_diskfile_from_hash | 17:49 |
| clayg | ssync you make me crazy! | 17:50 |
| *** itlinux has joined #openstack-swift | 17:53 | |
| clayg | mahatic: did you make any progress on splitting out/up get_ondisk_files and cleanup_ondisk_files as it's used by ssync and suffix hashing to minimize extraneous calls to listdir? | 17:55 |
| clayg | that was you right? working on the cleanup old tombstones | 17:55 |
| clayg | notmyname: this one should be on priority review -> https://review.openstack.org/#/c/346865/ | 18:00 |
| patchbot | patch 346865 - swift - Delete old tombstones | 18:00 |
| clayg | we still have like 48 hours right? we can *totally* merge all this stuff | 18:00 |
| clayg | notmyname: I mean just look at how *old* lp bug #1301728 is! | 18:01 |
| openstack | Launchpad bug 1301728 in OpenStack Object Storage (swift) "tombstone (.ts) object will never be deleted if the hash_suffix exists in the hashes.pkl " [High,In progress] https://launchpad.net/bugs/1301728 - Assigned to Mahati Chamarthy (mahati-chamarthy) | 18:01 |
| notmyname | clayg: ok! let's get it landed! :-) | 18:03 |
| *** tqtran has joined #openstack-swift | 18:10 | |
| *** tsg_ has quit IRC | 18:10 | |
| *** mvk has joined #openstack-swift | 18:12 | |
| *** asettle has joined #openstack-swift | 18:13 | |
| *** tsg_ has joined #openstack-swift | 18:13 | |
| *** caiobrentano__ has joined #openstack-swift | 18:14 | |
| *** CaioBrentano has quit IRC | 18:16 | |
| *** sgundur_ has quit IRC | 18:17 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: Enable object server to return non-durable data https://review.openstack.org/215276 | 18:18 |
| acoles | clayg: kota_ timburke : ^^ I *think* that is up to date with comments addressed | 18:18 |
| *** sgundur_ has joined #openstack-swift | 18:19 | |
| acoles | good night | 18:20 |
| *** psachin has quit IRC | 18:20 | |
| *** acoles is now known as acoles_ | 18:20 | |
| *** pcaruana has joined #openstack-swift | 18:22 | |
| *** lcurtis has joined #openstack-swift | 18:23 | |
| *** _JZ_ has joined #openstack-swift | 18:24 | |
| *** _JZ_ has quit IRC | 18:28 | |
| *** _JZ_ has joined #openstack-swift | 18:28 | |
| *** _JZ_ has quit IRC | 18:28 | |
| *** _JZ_ has joined #openstack-swift | 18:29 | |
| *** sgundur_ has quit IRC | 18:31 | |
| *** sgundur_ has joined #openstack-swift | 18:32 | |
| *** asettle has quit IRC | 18:34 | |
| *** _JZ_ has quit IRC | 18:36 | |
| *** tongli has joined #openstack-swift | 18:39 | |
| *** ppai has quit IRC | 18:43 | |
| *** tgtanya has joined #openstack-swift | 18:44 | |
| *** _JZ_ has joined #openstack-swift | 18:45 | |
| *** vint_bra1 has quit IRC | 19:02 | |
| *** vint_bra has joined #openstack-swift | 19:02 | |
| *** sgundur_ has quit IRC | 19:04 | |
| *** tgtanya_ has joined #openstack-swift | 19:04 | |
| *** baojg has quit IRC | 19:05 | |
| *** baojg has joined #openstack-swift | 19:05 | |
| *** tgtanya has quit IRC | 19:06 | |
| *** sgundur_ has joined #openstack-swift | 19:15 | |
| *** tsg_ has quit IRC | 19:19 | |
| *** ametts has joined #openstack-swift | 19:26 | |
| *** aswadr_ has quit IRC | 19:32 | |
| *** pcaruana has quit IRC | 19:32 | |
| *** tsg_ has joined #openstack-swift | 19:33 | |
| *** joeljwright has quit IRC | 19:33 | |
| *** asettle has joined #openstack-swift | 19:34 | |
| *** tongli has quit IRC | 19:38 | |
| *** itlinux has quit IRC | 19:39 | |
| *** asettle has quit IRC | 19:40 | |
| *** david-lyle has quit IRC | 19:40 | |
| *** david-lyle has joined #openstack-swift | 19:40 | |
| *** manous_ has joined #openstack-swift | 19:47 | |
| *** itlinux has joined #openstack-swift | 19:48 | |
| *** rcernin has joined #openstack-swift | 19:53 | |
| *** asettle has joined #openstack-swift | 19:58 | |
| *** ametts has quit IRC | 20:01 | |
| *** dmorita has quit IRC | 20:01 | |
| *** dmorita has joined #openstack-swift | 20:01 | |
| *** itlinux has quit IRC | 20:13 | |
| *** ametts has joined #openstack-swift | 20:13 | |
| *** asettle has quit IRC | 20:18 | |
| *** manous_ has quit IRC | 20:19 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 20:27 |
| *** joeljwright has joined #openstack-swift | 20:30 | |
| *** ChanServ sets mode: +v joeljwright | 20:30 | |
| *** sgundur_ has quit IRC | 20:32 | |
| *** manous_ has joined #openstack-swift | 20:33 | |
| *** ametts has quit IRC | 20:42 | |
| openstackgerrit | Thiago da Silva proposed openstack/swift: Return 404 on a GET if tombstone is newer https://review.openstack.org/371150 | 20:49 |
| *** tgtanya_ has quit IRC | 20:52 | |
| *** oshritf_ has joined #openstack-swift | 20:52 | |
| *** oshritf__ has joined #openstack-swift | 20:52 | |
| *** manous_ has quit IRC | 20:56 | |
| *** tgtanya has joined #openstack-swift | 20:57 | |
| *** thumpba has quit IRC | 20:59 | |
| *** thumpba has joined #openstack-swift | 20:59 | |
| *** tgtanya has quit IRC | 21:02 | |
| *** joeljwright has quit IRC | 21:06 | |
| *** thumpba has quit IRC | 21:10 | |
| openstackgerrit | Merged openstack/swift: Update link reference https://review.openstack.org/370554 | 21:12 |
| *** catintheroof has joined #openstack-swift | 21:15 | |
| *** asettle has joined #openstack-swift | 21:19 | |
| *** asettle has quit IRC | 21:24 | |
| notmyname | https://review.openstack.org/#/c/371166/ | 21:25 |
| patchbot | patch 371166 - election - Adding John Dickinson candidacy for Swift | 21:25 |
| *** oshritf_ has quit IRC | 21:25 | |
| *** oshritf__ has quit IRC | 21:25 | |
| torgomatic_ | when the only tool you have is Gerrit, everything looks like a repository | 21:26 |
| *** gyee has joined #openstack-swift | 21:27 | |
| *** dfg_ has joined #openstack-swift | 21:27 | |
| notmyname | torgomatic_: how do you do anything in openstack? `git review` | 21:28 |
| timburke | notmyname: don't forget the interminable waiting! | 21:30 |
| timburke | although, come to think of it, my git-review patches *did* land. now i'm just waiting on meetbot, os-loganalyze, and lodgeit | 21:31 |
| *** tsg__ has joined #openstack-swift | 21:32 | |
| *** rvasilets___ has joined #openstack-swift | 21:33 | |
| notmyname | https://i.imgflip.com/1aocg9.jpg | 21:33 |
| torgomatic_ | I have a hilarious reply to that which I will post in six weeks | 21:34 |
| *** tsg_ has quit IRC | 21:34 | |
| *** joeljwright has joined #openstack-swift | 21:35 | |
| *** ChanServ sets mode: +v joeljwright | 21:35 | |
| jrichli | notmyname: lol. no really, i actually laughed out loud. | 21:39 |
| jrichli | good write up for candidacy, btw (previous comment was about the posted image!) | 21:43 |
| *** manous has joined #openstack-swift | 21:45 | |
| *** AndyWojo has quit IRC | 21:45 | |
| *** ogelbukh has quit IRC | 21:45 | |
| *** Guest59079 has quit IRC | 21:45 | |
| *** CrackerJackMack has quit IRC | 21:45 | |
| *** wasmum has quit IRC | 21:46 | |
| *** geaaru has quit IRC | 21:46 | |
| *** cargonza has quit IRC | 21:46 | |
| *** madorn has quit IRC | 21:46 | |
| *** jamielennox has quit IRC | 21:46 | |
| *** amit213 has quit IRC | 21:47 | |
| *** mattoliverau has quit IRC | 21:47 | |
| *** matt6434 has joined #openstack-swift | 21:47 | |
| *** manous_ has joined #openstack-swift | 21:47 | |
| *** matt6434 is now known as mattoliverau | 21:49 | |
| *** manous has quit IRC | 21:49 | |
| *** ChanServ sets mode: +v mattoliverau | 21:50 | |
| *** ogelbukh has joined #openstack-swift | 21:50 | |
| *** amit213 has joined #openstack-swift | 21:51 | |
| *** AndyWojo has joined #openstack-swift | 21:51 | |
| *** cargonza has joined #openstack-swift | 21:52 | |
| *** CrackerJackMack has joined #openstack-swift | 21:52 | |
| *** manous_ has quit IRC | 21:52 | |
| *** wasmum has joined #openstack-swift | 21:54 | |
| *** madorn has joined #openstack-swift | 21:59 | |
| *** jamielennox has joined #openstack-swift | 22:01 | |
| *** joeljwright has quit IRC | 22:10 | |
| *** asettle has joined #openstack-swift | 22:21 | |
| *** asettle has quit IRC | 22:26 | |
| clayg | tdasilva: how'd you bump into lp bug #1560574? | 22:31 |
| openstack | Launchpad bug 1560574 in OpenStack Object Storage (swift) "GET/HEAD should stop continuing to search more nodes if a tombstone is reached" [Medium,Confirmed] https://launchpad.net/bugs/1560574 | 22:31 |
| mattoliverau | Morning | 22:33 |
| *** nadeem has quit IRC | 22:37 | |
| *** tsg__ has quit IRC | 22:39 | |
| *** chsc has quit IRC | 22:41 | |
| *** thumpba has joined #openstack-swift | 22:41 | |
| *** mmotiani_ has joined #openstack-swift | 22:43 | |
| *** baojg_ has joined #openstack-swift | 22:47 | |
| notmyname | just picked up the tshirts from the shop final_barca_shirts.jpg | 22:48 |
| notmyname | http://d.not.mn/final_barca_shirts.jpg | 22:48 |
| notmyname | that's better | 22:48 |
| zaitcev | ooooookay | 22:49 |
| *** baojg has quit IRC | 22:49 | |
| zaitcev | Never mind, I see what's going on. I thought it was a big swift attacking a furry at first. | 22:50 |
| notmyname | lol | 22:50 |
| klrmn | nice! | 22:52 |
| zaitcev | So if this is a toreador, then who is the bull? Ceph? | 22:52 |
| notmyname | nah. there's no hate. no competition (from this side) | 22:54 |
| notmyname | no symbolism (other than "spain, where the summit is"). just looks good | 22:55 |
| *** StraubTW has quit IRC | 22:57 | |
| jrichli | They are great! love the blue :-) | 23:02 |
| *** lcurtis has quit IRC | 23:04 | |
| mattoliverau | notmyname: so awesome! | 23:13 |
| *** neha__ has joined #openstack-swift | 23:15 | |
| *** dfg_ has quit IRC | 23:19 | |
| *** catintheroof has quit IRC | 23:21 | |
| *** asettle has joined #openstack-swift | 23:22 | |
| *** manous has joined #openstack-swift | 23:26 | |
| *** asettle has quit IRC | 23:27 | |
| *** vinsh has quit IRC | 23:28 | |
| *** thumpba has quit IRC | 23:30 | |
| *** neha__ has quit IRC | 23:37 | |
| *** hoonetorg has quit IRC | 23:37 | |
| *** kei_yama has joined #openstack-swift | 23:37 | |
| *** rvasilets___ has quit IRC | 23:37 | |
| *** Jeffrey4l__ is now known as Jeffrey4l | 23:40 | |
| *** neha__ has joined #openstack-swift | 23:42 | |
| *** tqtran has quit IRC | 23:52 | |
| *** hoonetorg has joined #openstack-swift | 23:55 | |
| *** tqtran has joined #openstack-swift | 23:55 | |
| *** thumpba has joined #openstack-swift | 23:57 | |
| clayg | timburke: boom! three tries! swift info --json | jq .swift.policies[].name | 23:58 |
| timburke | clayg: getting so close now! | 23:58 |
| clayg | rofl :D | 23:58 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
