| *** lespaul has quit IRC | 00:14 | |
| si1ver | Can swift3 use two auth backends? | 00:20 |
|---|---|---|
| si1ver | I know aws4 requires keystone, and aws2 worked fine with tempauth, so now I am hoping I can use both. | 00:20 |
| timburke | si1ver: yes, but pipeline ordering requires a bit of thought. as i recall, you want something like ... swift3 s3token authtoken tempauth keystoneauth ... with delay_auth_decision=true for both s3token and authtoken | 00:22 |
| timburke | it might be ... swift3 s3token authtoken keystoneauth tempauth ... though -- you'll probably want to experiment a little. maybe both are fine? i forget | 00:23 |
| si1ver | that's pretty close to what I have now. Maybe all I need to do is include both reseller prefixes in the s3token middleware, hopefully. | 00:24 |
| timburke | s3token's reseller prefix should match keystoneauth's. i'm pretty sure you can only give it one? | 00:25 |
| si1ver | so s3token is only for keystone then. | 00:28 |
| timburke | yes. its sole reason for existence is to talk to keystone to validate the signature provided | 00:29 |
| si1ver | can swift3 use s3token and tempauth at the same time. | 00:34 |
| timburke | yes, provided that you've got a s3token that includes https://review.openstack.org/#/c/326765/ and you've set delay_auth_decision=true | 00:36 |
| patchbot | patch 326765 - swift3 - Add delay_auth_decision config option (MERGED) | 00:36 |
| timburke | ...and we may not have had an official release since that landed... hrmm... :-/ | 00:36 |
| *** sams-gleb has joined #openstack-swift | 00:37 | |
| timburke | yep, only on master: https://github.com/openstack/swift3/commit/ca9d9d0 | 00:37 |
| timburke | i should probably poke kota_ about that, but i think we may want to get https://review.openstack.org/#/c/406424/ in before a release? | 00:38 |
| patchbot | patch 406424 - swift3 - Only make one request to Keystone per S3 request | 00:38 |
| *** sams-gleb has quit IRC | 00:41 | |
| *** JimCheung has quit IRC | 00:49 | |
| si1ver | delay auth decision is required to be able to use multiple backends right? | 00:51 |
| timburke | yes. otherwise, authtoken (& s3token) either reject the request before any other auth middlewares have a chance to respond or ignore the other middlewares' authing | 00:54 |
| timburke | you *might* be able to get something like ... swift3 tempauth s3token authtoken keystoneauth ... to work for both through the S3 api, but i'm fairly certain the swift API would only work for keystone users | 00:55 |
| timburke | i really really ought to document what we actually expect to work :-/ | 00:56 |
| si1ver | I need swift3 to use tempauth for aws2 requests and keystone for aws4 requests. | 00:57 |
| si1ver | tempauth would actually be fine for both, but since aws4 requires keystone I am working with that constraint. | 00:57 |
| si1ver | I am using swift3 from master, so maybe I have a shot? | 00:58 |
| si1ver | (Probably should pin that....) | 00:58 |
| timburke | yeah, we should be able to make that work | 00:58 |
| timburke | fwiw, i promise to really really try hard not to break it :-) | 00:59 |
| si1ver | Ha, sure I'll hold you to that. :p | 00:59 |
| kota_ | good morning | 00:59 |
| si1ver | spinning up a test env now. | 00:59 |
| timburke | hi kota_! | 00:59 |
| kota_ | hi timburke | 00:59 |
| kota_ | it looks like swift3 conversation in openstack-swift channel? | 01:00 |
| timburke | yep. it was too quiet, anyway :-) | 01:00 |
| kota_ | looking at the log | 01:01 |
| kota_ | done, so we need to bump the version release | 01:02 |
| kota_ | nice to have https://review.openstack.org/#/c/406424/ | 01:02 |
| patchbot | patch 406424 - swift3 - Only make one request to Keystone per S3 request | 01:02 |
| timburke | well, we could, anyway. sounds like that shouldn't really affect si1ver, who's following master | 01:03 |
| kota_ | and i got an answer from IBM that we can bump the minimum requirements of swift for swift3 to be newton stable | 01:03 |
| kota_ | timburke: ^^ | 01:03 |
| kota_ | timburke: so i'll push the patch soon | 01:04 |
| kota_ | timburke: i'd like to include it too | 01:04 |
| kota_ | for the release | 01:04 |
| timburke | sounds good! | 01:04 |
| *** _JZ_ has quit IRC | 01:04 | |
| *** JimCheung has joined #openstack-swift | 01:08 | |
| *** klamath has joined #openstack-swift | 01:09 | |
| *** jerrygb has quit IRC | 01:11 | |
| *** jerrygb has joined #openstack-swift | 01:14 | |
| *** klamath has quit IRC | 01:18 | |
| si1ver | ok cool, getting both tempauth and keystone working with aws2 only required that delay auth decision directive in the swift3 block. | 01:20 |
| si1ver | now if aws4 still works, I'm all good. | 01:20 |
| timburke | si1ver: awesome! if v2 works, i totally expect v4 to work as well (at least for keystone) | 01:21 |
| *** stevemar has quit IRC | 01:27 | |
| *** Raymii has quit IRC | 01:27 | |
| *** stevemar has joined #openstack-swift | 01:29 | |
| *** Raymii has joined #openstack-swift | 01:29 | |
| si1ver | timburke: thanks, adding delay_auth_decision to [filter:swift3] and [filter:s3token] did the trick. | 01:36 |
| timburke | glad to hear it | 01:36 |
| *** JimCheung has quit IRC | 01:47 | |
| *** clu_ has quit IRC | 02:00 | |
| *** mingyu has joined #openstack-swift | 02:11 | |
| *** sams-gleb has joined #openstack-swift | 02:39 | |
| *** sams-gleb has quit IRC | 02:44 | |
| *** mingyu has quit IRC | 02:44 | |
| *** JimCheung has joined #openstack-swift | 02:46 | |
| *** mingyu has joined #openstack-swift | 02:50 | |
| *** JimCheung has quit IRC | 02:51 | |
| *** NM has joined #openstack-swift | 02:54 | |
| notmyname | hello, world | 02:55 |
| *** mingyu has quit IRC | 02:57 | |
| *** mingyu has joined #openstack-swift | 02:58 | |
| *** mingyu has quit IRC | 03:00 | |
| *** mingyu has joined #openstack-swift | 03:00 | |
| *** mingyu has quit IRC | 03:01 | |
| *** NM has quit IRC | 03:13 | |
| *** tqtran has joined #openstack-swift | 03:16 | |
| *** tqtran has quit IRC | 03:18 | |
| klrmn | notmyname: oh, stop being so melodramatic =) | 03:22 |
| *** jlwhite has quit IRC | 03:22 | |
| notmyname | klrmn: oh am i? I just hadn't actually opened my laptop today | 03:22 |
| notmyname | just taking some time to catch up | 03:22 |
| notmyname | seeing what new drama is happening on the ML ;-) | 03:22 |
| klrmn | the 'hello world' all by itself there was looking very … drat, i hate it when my vocabulary runs and hides | 03:23 |
| notmyname | normally I'd start with "good morning", but I kinda missed that window | 03:24 |
| klrmn | hugokuo: usually says good morning about this time… | 03:26 |
| hugokuo | good...... morning ! | 03:27 |
| jrichli | notmyname: o/ is there a meeting tomorrow? | 03:30 |
| notmyname | yes | 03:30 |
| jrichli | okie dokie | 03:30 |
| jrichli | you having fun down under? | 03:31 |
| notmyname | yes, definitely. some great serendipitous conversations, good talks, beautiful venue | 03:31 |
| jrichli | nice. I saw a tweet about a talk from cschwede - retweeted actually, but haven't gotten to watch yet. | 03:33 |
| jrichli | I didn't realize he would be there. figured it was only you and mattoliverau | 03:33 |
| notmyname | yeah, I think I saw that too. that was an EU conference. not here | 03:34 |
| jrichli | OH. oops :p | 03:34 |
| *** sheel has joined #openstack-swift | 03:34 | |
| notmyname | and mattoliverau is probably the only other person you know here. although, there's quite a few IBM (Oz Labs) people here | 03:34 |
| jrichli | like the people working on EC lib? | 03:35 |
| jrichli | I think there was one that I have talked to from there who wanted to make some enhancements | 03:36 |
| notmyname | actually, there is someone here I talked to who is working on that. but he hasn't pushed any code yet | 03:36 |
| notmyname | but yeah, getting libec to run on power8 | 03:36 |
| jrichli | first name starts with a D ... will have to look at my mail | 03:36 |
| jrichli | dja is his handle. Daniel? | 03:37 |
| notmyname | I don't know his IRC handle. I was just looking for it | 03:37 |
| jrichli | good you got to chat with him | 03:38 |
| notmyname | but yeah DANIEL AXTENS | 03:38 |
| notmyname | yes, I did. well, briefly | 03:38 |
| jrichli | :-) | 03:38 |
| notmyname | csmart is now working at rackspace and might be able to start contributing to swift. he's been lurking in here a while | 03:39 |
| jrichli | great to hear it, csmart :-) | 03:42 |
| *** links has joined #openstack-swift | 04:02 | |
| *** dmorita_ has quit IRC | 04:03 | |
| *** psachin has joined #openstack-swift | 04:14 | |
| *** jlwhite has joined #openstack-swift | 04:20 | |
| *** dmorita has joined #openstack-swift | 04:29 | |
| *** dmorita has quit IRC | 04:33 | |
| *** sams-gleb has joined #openstack-swift | 04:41 | |
| *** sams-gleb has quit IRC | 04:46 | |
| *** eckesicle has quit IRC | 04:55 | |
| *** manij has joined #openstack-swift | 04:58 | |
| *** ppai has joined #openstack-swift | 05:03 | |
| *** SkyRocknRoll has joined #openstack-swift | 05:04 | |
| openstackgerrit | Janie Richling proposed openstack/swift: WIP: Symlink support in container-sync https://review.openstack.org/421669 | 05:08 |
| *** tanee_ has joined #openstack-swift | 05:10 | |
| openstackgerrit | Janie Richling proposed openstack/swift: WIP: Symlink support in container-sync https://review.openstack.org/421669 | 05:13 |
| jrichli | btw, speaking of cschwede, I saw that patch 337297 got a +2 : yay! | 05:21 |
| patchbot | https://review.openstack.org/#/c/337297/ - swift - Add support to increase object ring partition power | 05:21 |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift: Optimize ec duplication and its md5 hashing https://review.openstack.org/421673 | 05:38 |
| kota_ | clayg: you may like patch 421673 | 05:39 |
| patchbot | https://review.openstack.org/#/c/421673/ - swift - Optimize ec duplication and its md5 hashing | 05:39 |
| *** psachin has quit IRC | 05:44 | |
| *** jerrygb has quit IRC | 05:45 | |
| *** jerrygb has joined #openstack-swift | 05:46 | |
| *** psachin has joined #openstack-swift | 05:49 | |
| *** jerrygb has quit IRC | 05:50 | |
| *** mingyu has joined #openstack-swift | 05:57 | |
| *** mingyu has quit IRC | 06:05 | |
| *** tqtran has joined #openstack-swift | 06:05 | |
| *** mingyu has joined #openstack-swift | 06:11 | |
| *** mingyu has quit IRC | 06:23 | |
| *** mingyu has joined #openstack-swift | 06:24 | |
| *** sams-gleb has joined #openstack-swift | 06:43 | |
| *** sams-gleb has quit IRC | 06:44 | |
| *** sams-gleb has joined #openstack-swift | 06:45 | |
| *** mingyu has quit IRC | 06:46 | |
| *** mingyu has joined #openstack-swift | 06:48 | |
| *** klrmn has quit IRC | 07:00 | |
| sanchitmalhotra | Hi, I am looking for some information on the following scenario for replication. | 07:08 |
| sanchitmalhotra | Is it possible in SWIFT that i replicate the existing data to some other remote location? | 07:08 |
| sanchitmalhotra | Or i have to do this at the time of initial system setup only while definiting the no. of replicas of each object and their device/location? | 07:08 |
| *** bapalm has quit IRC | 07:16 | |
| *** bapalm has joined #openstack-swift | 07:18 | |
| *** Manish_ has joined #openstack-swift | 07:19 | |
| *** ChubYann has quit IRC | 07:22 | |
| *** mingyu has quit IRC | 07:22 | |
| *** tesseract has joined #openstack-swift | 07:25 | |
| *** tesseract has quit IRC | 07:25 | |
| *** tesseract has joined #openstack-swift | 07:26 | |
| *** dfflanders has quit IRC | 07:29 | |
| *** mingyu has joined #openstack-swift | 07:30 | |
| *** Manish_ has quit IRC | 07:30 | |
| *** mingyu has quit IRC | 07:32 | |
| *** Manish_ has joined #openstack-swift | 07:43 | |
| *** hseipp has joined #openstack-swift | 07:45 | |
| *** tqtran has quit IRC | 07:47 | |
| *** tanee has quit IRC | 07:51 | |
| *** onboard has quit IRC | 07:52 | |
| *** tanee_ is now known as tanee | 07:52 | |
| *** oshritf_ has joined #openstack-swift | 07:56 | |
| *** rledisez has joined #openstack-swift | 08:10 | |
| sanchitmalhotra | Hi... Can anyone help on the above query related to replication? | 08:15 |
| *** mingyu has joined #openstack-swift | 08:23 | |
| *** sams-gleb has quit IRC | 08:26 | |
| *** sams-gleb has joined #openstack-swift | 08:26 | |
| *** sams-gleb has quit IRC | 08:30 | |
| *** mingyu has quit IRC | 08:34 | |
| *** SkyRocknRoll has quit IRC | 08:38 | |
| *** silor has joined #openstack-swift | 08:40 | |
| *** mingyu has joined #openstack-swift | 08:41 | |
| *** silor has quit IRC | 08:44 | |
| *** silor1 has joined #openstack-swift | 08:44 | |
| charz | sanchitmalhotra: maybe you can setup another remote region and add nodes for the region, replicator will replicate data to that. | 08:45 |
| charz | sanchitmalhotra: if you're asking to rpelicate date to remove server that isn't a part of swift cluster. I think the answer is no, as I know. | 08:46 |
| charz | you need 3rd party tools to do that. | 08:47 |
| *** mingyu has quit IRC | 08:47 | |
| *** silor1 is now known as silor | 08:47 | |
| sanchitmalhotra | charz: Thank you very much for the response | 08:47 |
| sanchitmalhotra | charz: Can I replicate data to a node which is not a part of swift cluster? | 08:48 |
| charz | for backup purpose? | 08:48 |
| sanchitmalhotra | yes. as well as a replica site which can be used just in case | 08:51 |
| charz | object-replicator follow the ring file to check the replicas and move replica to correct places, so if your node isn't a part of cluster (devices don't add to object ring), object-replicator won't do anything for drives. | 08:51 |
| sanchitmalhotra | charz: Ok. This is key information. Thank you very much. | 08:52 |
| sanchitmalhotra | It means storage node and replication nodes cannot be entirely different | 08:52 |
| charz | maybe you can consider to setup a global swift cluster. | 08:53 |
| sanchitmalhotra | charz: Okay.... I'll look into it. | 08:55 |
| sanchitmalhotra | charz: To get started, I'll search more information about global swift cluster. | 08:57 |
| sanchitmalhotra | charz: I hope I'll be able to get some more information from you if I get any further query on this topic. | 08:57 |
| charz | sanchitmalhotra: I'm not sure how you defince the "storage" and "replication" node. So maybe you can check swift docs to know each servers. | 08:58 |
| charz | sanchitmalhotra: you can ask question here, I believe developers will happy to answer your quesitons. | 09:00 |
| sanchitmalhotra | charz: Thank you very much :) | 09:00 |
| *** anurgms has joined #openstack-swift | 09:01 | |
| charz | sanchitmalhotra: np | 09:01 |
| *** mingyu has joined #openstack-swift | 09:02 | |
| *** mingyu_ has joined #openstack-swift | 09:03 | |
| *** mingyu has quit IRC | 09:03 | |
| *** SkyRocknRoll has joined #openstack-swift | 09:03 | |
| *** mingyu_ has quit IRC | 09:05 | |
| *** Shashikant86 has joined #openstack-swift | 09:16 | |
| openstackgerrit | Anh Tran proposed openstack/swift: API: Correcting status code https://review.openstack.org/421759 | 09:16 |
| *** dfflanders has joined #openstack-swift | 09:26 | |
| *** mattoliverau has quit IRC | 09:27 | |
| *** matt6434 has joined #openstack-swift | 09:28 | |
| *** Shashikant86 has quit IRC | 09:35 | |
| mathiasb | acoles_: it looks like my problem was that I had both tempauth and keystoneauth in the pipeline, which led to the 401 issue for performing a GET on the container, even though everything else seemed to work | 09:39 |
| *** Shashikant86 has joined #openstack-swift | 09:40 | |
| *** pcaruana has joined #openstack-swift | 09:41 | |
| *** mvk has quit IRC | 09:45 | |
| *** andymccr has joined #openstack-swift | 09:58 | |
| *** asettle has joined #openstack-swift | 09:58 | |
| *** sams-gleb has joined #openstack-swift | 10:15 | |
| *** SkyRocknRoll has quit IRC | 10:16 | |
| *** SkyRocknRoll has joined #openstack-swift | 10:18 | |
| *** mvk has joined #openstack-swift | 10:21 | |
| *** vint_bra has joined #openstack-swift | 10:24 | |
| *** pcaruana has quit IRC | 10:38 | |
| *** vint_bra has quit IRC | 10:38 | |
| *** acoles_ is now known as acoles | 10:39 | |
| acoles | mathiasb: glad you found problem, curious though, it should be ok to have both auths as long as they have distinct reseller prefixes. | 10:40 |
| *** JimCheung has joined #openstack-swift | 10:40 | |
| *** kei_yama has quit IRC | 10:42 | |
| acoles | sanchitmalhotra: here's a good place to start reading about global clusters http://docs.openstack.org/developer/swift/admin_guide.html#geographically-distributed-swift-considerations | 10:46 |
| *** tqtran has joined #openstack-swift | 10:47 | |
| *** tqtran has quit IRC | 10:50 | |
| *** _JZ_ has joined #openstack-swift | 10:54 | |
| *** Manish_ has quit IRC | 10:57 | |
| *** NM has joined #openstack-swift | 11:00 | |
| *** Shashikant86 has quit IRC | 11:00 | |
| *** Shashikant86 has joined #openstack-swift | 11:04 | |
| *** dmorita has joined #openstack-swift | 11:04 | |
| *** dfflanders has quit IRC | 11:05 | |
| *** dmorita has quit IRC | 11:09 | |
| *** oshritf has joined #openstack-swift | 11:28 | |
| *** oshritf_ has quit IRC | 11:29 | |
| *** oshritf has quit IRC | 11:35 | |
| *** oshritf has joined #openstack-swift | 11:35 | |
| *** ppai has quit IRC | 11:36 | |
| *** oshritf has quit IRC | 11:43 | |
| *** ganders has joined #openstack-swift | 11:54 | |
| *** pcaruana has joined #openstack-swift | 11:56 | |
| *** geaaru has joined #openstack-swift | 11:56 | |
| csmart | notmyname: thanks :-) | 12:14 |
| csmart | jrichli: \o cheers | 12:14 |
| *** cdelatte has joined #openstack-swift | 12:15 | |
| csmart | jrichli: notmyname daniel axtens (dja) is an ex-colleague of mine | 12:15 |
| * csmart looks forward to contributing to swift ;-) | 12:17 | |
| *** peterlisak has quit IRC | 12:18 | |
| *** onovy has quit IRC | 12:18 | |
| *** anurgms has quit IRC | 12:22 | |
| *** peterlisak has joined #openstack-swift | 12:26 | |
| *** onovy has joined #openstack-swift | 12:27 | |
| *** Shashikant86 has quit IRC | 12:28 | |
| *** manij has quit IRC | 12:29 | |
| *** NM has quit IRC | 12:38 | |
| *** links has quit IRC | 12:46 | |
| *** sams-gle_ has joined #openstack-swift | 12:50 | |
| *** sams-gleb has quit IRC | 12:50 | |
| *** klamath has joined #openstack-swift | 12:57 | |
| *** klamath has quit IRC | 12:57 | |
| *** klamath has joined #openstack-swift | 12:57 | |
| *** Shashikant86 has joined #openstack-swift | 12:58 | |
| *** jamielennox is now known as jamielennox|away | 12:59 | |
| *** tongli has joined #openstack-swift | 13:01 | |
| *** vint_bra has joined #openstack-swift | 13:14 | |
| *** JimCheung has joined #openstack-swift | 13:17 | |
| *** JimCheung has quit IRC | 13:17 | |
| *** jamielennox|away is now known as jamielennox | 13:19 | |
| *** NM has joined #openstack-swift | 13:28 | |
| *** vint_bra1 has joined #openstack-swift | 13:32 | |
| *** vint_bra has quit IRC | 13:32 | |
| *** vint_bra1 has joined #openstack-swift | 13:32 | |
| *** JimCheung has joined #openstack-swift | 13:38 | |
| *** ganders has quit IRC | 13:41 | |
| *** JimCheung has quit IRC | 13:42 | |
| *** ganders has joined #openstack-swift | 13:43 | |
| *** catintheroof has joined #openstack-swift | 13:45 | |
| *** jerrygb has joined #openstack-swift | 13:56 | |
| *** oshritf has joined #openstack-swift | 13:58 | |
| *** jerrygb has quit IRC | 13:59 | |
| *** jerrygb has joined #openstack-swift | 14:03 | |
| *** jerrygb has quit IRC | 14:03 | |
| *** jerrygb has joined #openstack-swift | 14:04 | |
| *** sams-gle_ has quit IRC | 14:16 | |
| *** sams-gleb has joined #openstack-swift | 14:17 | |
| *** sams-gleb has quit IRC | 14:21 | |
| *** jordanP has joined #openstack-swift | 14:25 | |
| *** NM has quit IRC | 14:27 | |
| *** NM has joined #openstack-swift | 14:27 | |
| *** sams-gleb has joined #openstack-swift | 14:34 | |
| *** tqtran has joined #openstack-swift | 14:40 | |
| *** Shashikant86 has quit IRC | 14:53 | |
| *** links has joined #openstack-swift | 14:56 | |
| *** links has quit IRC | 14:56 | |
| *** Shashikant86 has joined #openstack-swift | 15:00 | |
| *** tongli has quit IRC | 15:00 | |
| *** Shashikant86 has quit IRC | 15:02 | |
| *** NM has quit IRC | 15:02 | |
| *** NM has joined #openstack-swift | 15:03 | |
| *** mingyu has joined #openstack-swift | 15:03 | |
| *** jistr is now known as jistr|mtg | 15:05 | |
| *** openstack has joined #openstack-swift | 15:19 | |
| *** bobby2 has joined #openstack-swift | 15:19 | |
| *** asettle has joined #openstack-swift | 15:19 | |
| *** swifterdarrell has joined #openstack-swift | 15:20 | |
| *** ChanServ sets mode: +v swifterdarrell | 15:20 | |
| *** jlvillal_ is now known as jlvillal | 15:20 | |
| *** jlvillal is now known as Guest88977 | 15:20 | |
| *** fungi has joined #openstack-swift | 15:20 | |
| *** BlackDex_ has joined #openstack-swift | 15:21 | |
| *** zacksh has joined #openstack-swift | 15:22 | |
| *** mgagne has joined #openstack-swift | 15:22 | |
| *** ejat has joined #openstack-swift | 15:22 | |
| *** mgagne has quit IRC | 15:23 | |
| *** mgagne has joined #openstack-swift | 15:23 | |
| *** mgagne is now known as Guest58531 | 15:23 | |
| BlackDex_ | Hello there, i wonder if it is possible to link openstack using an external swift-storage/proxy to an openstack deployment, and if this maybe is possible with a juju charm? | 15:23 |
| *** BlackDex_ is now known as BlackDex | 15:23 | |
| *** DuncanT has joined #openstack-swift | 15:26 | |
| *** briancurtin has joined #openstack-swift | 15:29 | |
| *** klrmn has joined #openstack-swift | 15:29 | |
| *** psachin has quit IRC | 15:33 | |
| *** Shashikant86 has joined #openstack-swift | 15:33 | |
| *** sheel has quit IRC | 15:37 | |
| *** kencjohnston_ has quit IRC | 15:42 | |
| *** kencjohnston has joined #openstack-swift | 15:43 | |
| *** david-lyle has joined #openstack-swift | 15:47 | |
| *** mvk has quit IRC | 15:48 | |
| openstackgerrit | Christopher Bartz proposed openstack/swift: Exclude containers for account quota https://review.openstack.org/414232 | 15:49 |
| *** jistr|mtg is now known as jistr | 15:51 | |
| caiobrentano_ | Hi all | 15:51 |
| caiobrentano_ | I'm writing a middleware for proxy-server. How much overload "swift.proxy.controllers.base.get_container_info" does to check a specific custom header? | 15:52 |
| cbartz | caiobrentano_: In principle it does a container HEAD request. But the information gets cached afterwards. | 15:57 |
| caiobrentano_ | cbartz: this HEAD would happen anyway along the request flow? | 15:59 |
| *** Jeffrey4l has quit IRC | 16:01 | |
| *** Jeffrey4l_ has joined #openstack-swift | 16:01 | |
| *** Jeffrey4l_ is now known as Jeffrey4l | 16:02 | |
| *** geaaru has quit IRC | 16:02 | |
| caiobrentano_ | cbartz: I mean, even if I don't call swift.proxy.controllers.base.get_container_info, somewhere in the proxy-server pipeline, a HEAD would be necessary to complete the request? | 16:02 |
| cbartz | caiobrentano_: I think this depends on what your middleware does. | 16:03 |
| *** mingyu has quit IRC | 16:04 | |
| caiobrentano_ | cbartz: I mean in a "default" proxy-server pipeline. For example, if I make a PUT request to create an object, does the proxy-server needs to HEAD the container to complete the request? | 16:07 |
| cbartz | caiobrentano_: I think so, yes. | 16:08 |
| tdasilva | caiobrentano_: unless it is cached, i'm pretty sure yes: https://github.com/openstack/swift/blob/master/swift/proxy/controllers/obj.py#L634 | 16:08 |
| *** geaaru has joined #openstack-swift | 16:09 | |
| caiobrentano_ | cbartz: If it always do a HEAD in order to complete a PUT request, I can assume this "HEAD" operation is always necessary. So, the overload would be minimal | 16:09 |
| caiobrentano_ | cbartz: thanks! :) | 16:10 |
| *** ganders has quit IRC | 16:10 | |
| cbartz | caiobrentano_: np | 16:10 |
| cbartz | tdasilva: Hello Thiago. Do you might have time for reviewing patch 408596 ? Its the swiftclient side for the prefix-based tempurls support. Its needing only one c reviewer more. Would be cool to have support on the cli side, too. | 16:13 |
| patchbot | https://review.openstack.org/#/c/408596/ - python-swiftclient - prefix-based tempurls support | 16:13 |
| *** SkyRocknRoll has quit IRC | 16:15 | |
| tdasilva | cbartz: hi! cool, i will try to review it. | 16:15 |
| *** zul has quit IRC | 16:15 | |
| cbartz | tdasilva: Thx | 16:15 |
| *** zul has joined #openstack-swift | 16:15 | |
| tdasilva | cbartz: did you see the idea notmyname added to the ideas page regarding tempurl? | 16:15 |
| tdasilva | looked interesting too | 16:16 |
| tdasilva | cbartz: https://blueprints.launchpad.net/swift/+spec/human-readable-tempurl-timestamp | 16:16 |
| cbartz | tdasilva: Yes, sounds nice. Somebody working on it? | 16:20 |
| tdasilva | cbartz: honestly, i don't know, notmyname or somebody from swiftstack might know better....I don't remember seeing any patches...but I could have easily missed... | 16:20 |
| *** ganders has joined #openstack-swift | 16:21 | |
| *** Guest88977 is now known as jlvillal | 16:25 | |
| *** ganders has quit IRC | 16:27 | |
| *** vint_bra has joined #openstack-swift | 16:35 | |
| *** chsc has joined #openstack-swift | 16:35 | |
| *** chsc has joined #openstack-swift | 16:35 | |
| *** vint_bra1 has quit IRC | 16:37 | |
| cbartz | Hey all. If somebody here is using git-lfs and needs a swift binding: On my github account I have published a git-lfs server implementation for swift and a transfer agent for the custom transfer type swift. This allows to upload objects larger > 5 GiB to your swift cluster with git-lfs. More information here: https://github.com/cbartz/git-lfs-swift-server and https://github.com/cbartz/git-lfs-swift-transfer-agent . | 16:38 |
| *** vint_bra1 has joined #openstack-swift | 16:42 | |
| *** vint_bra has quit IRC | 16:45 | |
| *** sheel has joined #openstack-swift | 16:48 | |
| *** ChubYann has joined #openstack-swift | 16:57 | |
| *** Shashikant86 has quit IRC | 17:00 | |
| *** arch-nemesis has joined #openstack-swift | 17:01 | |
| *** ganders has joined #openstack-swift | 17:02 | |
| *** klrmn has quit IRC | 17:14 | |
| *** jordanP has quit IRC | 17:15 | |
| *** tesseract has quit IRC | 17:21 | |
| *** dmorita has joined #openstack-swift | 17:34 | |
| *** dmorita has quit IRC | 17:37 | |
| *** david-lyle is now known as bailing-wire | 17:37 | |
| *** mvk has joined #openstack-swift | 17:38 | |
| *** catintheroof has quit IRC | 17:44 | |
| *** catintheroof has joined #openstack-swift | 17:44 | |
| *** catintheroof has quit IRC | 17:44 | |
| *** catintheroof has joined #openstack-swift | 17:45 | |
| *** bailing-wire has quit IRC | 17:46 | |
| *** rledisez has quit IRC | 17:46 | |
| *** catintheroof has quit IRC | 17:50 | |
| *** JimCheung has joined #openstack-swift | 17:52 | |
| *** jordanP has joined #openstack-swift | 17:58 | |
| *** jordanP has quit IRC | 17:58 | |
| *** chsc has quit IRC | 18:01 | |
| *** klrmn has joined #openstack-swift | 18:02 | |
| *** chsc has joined #openstack-swift | 18:12 | |
| *** chsc has joined #openstack-swift | 18:12 | |
| *** jerrygb has quit IRC | 18:23 | |
| timburke | good morning | 18:26 |
| timburke | acoles: mathiasb: you can have both auth systems set up, but you need to make sure (1) authtoken has delay_auth_decision=true and (2) the middlewares are in the correct order. i *think* it goes ... authtoken keystoneauth tempauth ... but i'd trust experimental results more than my memory | 18:28 |
| *** NM has quit IRC | 18:33 | |
| *** ganders has quit IRC | 18:34 | |
| acoles | timburke: morning. thanks, yes now you say it I also recall there an ordering requirement but like you I'm not sure which way works. | 18:35 |
| *** ganders has joined #openstack-swift | 18:36 | |
| timburke | the big thing i always forget is the delay_auth_decision. really, it should *always* be set to true in a proxy server, but things will mostly still work if you omit it (and thus use the default of false) | 18:37 |
| timburke | ...until you add a second auth system, at which point it starts assuming that the (valid) tempauth or swauth tokens should get 401s immediately | 18:37 |
| openstackgerrit | Alistair Coles proposed openstack/swift: Add assertion to test_consolidate_hashes_raises_exception https://review.openstack.org/422159 | 18:39 |
| *** NM has joined #openstack-swift | 18:43 | |
| *** philipw_ has quit IRC | 18:44 | |
| *** acoles is now known as acoles_ | 18:45 | |
| *** philipw has joined #openstack-swift | 18:45 | |
| *** jerrygb has joined #openstack-swift | 19:24 | |
| *** caiobrentano_ has quit IRC | 19:24 | |
| *** jerrygb has quit IRC | 19:29 | |
| *** caiobrentano_ has joined #openstack-swift | 19:30 | |
| *** caiobrentano_ has quit IRC | 19:45 | |
| *** ganders has quit IRC | 19:45 | |
| *** caiobrentano_ has joined #openstack-swift | 19:45 | |
| *** jidar has left #openstack-swift | 19:46 | |
| *** caiobrentano_ has quit IRC | 19:50 | |
| *** lespaul has joined #openstack-swift | 19:52 | |
| lespaul | Hello. I can upload an object to a container via Dashboard however when trying to download the object from the proxy using swift command I get a permission denied error. Any ideas? | 19:56 |
| *** silor has quit IRC | 19:57 | |
| jrichli | lespaul: when running the swift command, are all the necessary OS env vars set? | 19:57 |
| lespaul | jrichli: yes, i can even upload an object using swift command. Running swift stat shows the right env | 19:59 |
| *** bailing-wire has joined #openstack-swift | 20:00 | |
| jrichli | lespaul: I assume you are using the same credentials and account for both ... sounds like you'd have to look at logs | 20:00 |
| lespaul | ok. also, swift stat still shows the old container name which i have already renamed. I've tried reloading the proxy-server but it persists | 20:01 |
| jrichli | ah, so the inconsistent behavior is happening with a newly re-named container? how much time has passed since the rename? | 20:02 |
| *** bailing-wire is now known as david-lyle | 20:02 | |
| lespaul | 10-15 mins. i was thinking it will update eventually.. | 20:03 |
| jrichli | are all the services running? | 20:03 |
| jrichli | updaters and such | 20:04 |
| lespaul | how long before it updates i don't know. i ran swift-init all restart | 20:04 |
| jrichli | i'd think 10-15 mins would be fine, unless there is some large error on the backend. no major errors from logs? | 20:05 |
| *** jerrygb has joined #openstack-swift | 20:05 | |
| jrichli | can you download from the dashboard? does the dashboard reflect the new name? | 20:08 |
| timburke | what do we mean by "renamed"? like, "created new container with the correct name, sent COPY requests to copy all data in the old container to the new one, then sent DELETE requests to all of the old object, and finally sent a DELETE request for the old container"? | 20:09 |
| timburke | was all of this hidden behind an innocent-looking "rename" button in Horizon? | 20:09 |
| *** klamath has quit IRC | 20:10 | |
| *** jerrygb has quit IRC | 20:13 | |
| lespaul | timburke: i renamed the added storage policy in swift.conf, deleted the object-2 builder and ring files on the proxy and storage nodes. i then created a new ring and distributed them. | 20:27 |
| *** caiobrentano_ has joined #openstack-swift | 20:31 | |
| timburke | i... so the *container* wasn't renamed, but the *storage policy* was? was there only the one container with that policy? did the container previously have any objects in it? | 20:31 |
| timburke | how certain are we that the new storage policy has the same config (replicated vs ec; if ec, the ec_type, ec_num_*_fragments) as the old one, and *only* the name has changed? | 20:34 |
| *** zaitcev has quit IRC | 20:40 | |
| *** dmorita has joined #openstack-swift | 20:44 | |
| *** JimCheung has quit IRC | 20:46 | |
| *** dmorita has quit IRC | 20:48 | |
| lespaul | I think the problem is on the EC storage policy since the default replication policy works fine. I have a storage policy of an EC. If I want to change to a new EC type, do I just change the type in swift.conf then distribute the conf file to the nodes? | 20:48 |
| *** NM has quit IRC | 20:49 | |
| notmyname | good morning | 20:52 |
| *** jerrygb has joined #openstack-swift | 20:53 | |
| notmyname | tdasilva: wrt the human-readable tempurl timestamps, AFAIK there's not currently anyone working on it (but I should check back in the office) | 20:53 |
| tdasilva | cbartz: ^^^ | 20:54 |
| notmyname | cbartz: if you're interested in working on it, that would be great :-) | 20:54 |
| timburke | lespaul: once you have data stored in a storage policy, you never want to change the parameters of that policy. the name can change (although we recommend including the old name as an alias so as not to break existing clients that may expect the old name), but the ec_type, number of data/parity fragments, and fragment size have to stay the same or else swift won't know how to read the objects | 20:55 |
| notmyname | onovy: I see you've proposed a few backports. I'll look into those asap | 20:55 |
| onovy | notmyname: thanks! | 20:56 |
| notmyname | onovy: but FWIW, I'm at a conference this week, so my time in front of a computer is somewhat reduced :-) | 20:56 |
| onovy | notmyname: np | 20:56 |
| onovy | i already backported that for Debian package :) | 20:57 |
| kota_ | morning | 20:57 |
| notmyname | timburke: you're ready to talk about the api thing in the meeting? or were you needing clayg (IIRC he's sick today?) | 20:59 |
| *** m_kazuhiro has joined #openstack-swift | 20:59 | |
| notmyname | meeting time in #opestack-meeting | 21:00 |
| timburke | notmyname: sure. should make it easier to get something landed :P | 21:00 |
| notmyname | kk | 21:00 |
| lespaul | timburke: if i understand correctly, if i want to use a new ec type i will need to create a new storage policy in swift conf, create a new builder/ring then distribute to the nodes. is that correct? | 21:01 |
| timburke | lespaul: yes. storage policies are an append-only sort of a deal | 21:03 |
| lespaul | thanks for the clarification | 21:04 |
| *** acoles_ is now known as acoles | 21:06 | |
| *** caiobrentano_ has quit IRC | 21:07 | |
| *** dmorita has joined #openstack-swift | 21:17 | |
| *** zaitcev has joined #openstack-swift | 21:22 | |
| *** ChanServ sets mode: +v zaitcev | 21:22 | |
| *** zul has quit IRC | 21:25 | |
| *** JimCheung has joined #openstack-swift | 21:26 | |
| *** matt6434 is now known as mattoliverau | 21:26 | |
| *** jamielennox is now known as jamielennox|away | 21:26 | |
| *** ChanServ sets mode: +v mattoliverau | 21:27 | |
| mattoliverau | morning | 21:27 |
| *** zul has joined #openstack-swift | 21:28 | |
| *** jerrygb_ has joined #openstack-swift | 21:30 | |
| zaitcev | tdasilva: thanks for the mail. Imagine that, I updated Fedora, rebooted, and didn't start VPN at once... only got the mail now! | 21:32 |
| tdasilva | zaitcev: :) | 21:32 |
| *** jerrygb has quit IRC | 21:33 | |
| *** Jeffrey4l has quit IRC | 21:35 | |
| * tdasilva wonders if this page would be helpful at all: http://breakingapis.org/ | 21:35 | |
| *** Jeffrey4l has joined #openstack-swift | 21:35 | |
| *** dmorita has quit IRC | 21:38 | |
| *** dmorita has joined #openstack-swift | 21:38 | |
| notmyname | tdasilva: interesting | 21:40 |
| *** dmorita has quit IRC | 21:41 | |
| *** dmorita has joined #openstack-swift | 21:43 | |
| *** vint_bra1 has quit IRC | 21:49 | |
| *** acoles is now known as acoles_ | 21:50 | |
| *** sams-gleb has quit IRC | 21:50 | |
| *** sams-gleb has joined #openstack-swift | 21:51 | |
| *** sams-gleb has quit IRC | 21:55 | |
| *** m_kazuhiro has quit IRC | 21:57 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift: Optimize ec duplication and its md5 hashing https://review.openstack.org/421673 | 22:04 |
| openstackgerrit | Tim Burke proposed openstack/swift: Make test_get_more_nodes_with_zero_weight_region more robust https://review.openstack.org/418082 | 22:04 |
| timburke | i'm still annoyed at all the 412s we send when we mean 400, but clayg tells me we can't ever fix those, either :-/ | 22:08 |
| *** jamielennox|away is now known as jamielennox | 22:10 | |
| *** geaaru has quit IRC | 22:15 | |
| *** jamielennox is now known as jamielennox|away | 22:24 | |
| *** NM has joined #openstack-swift | 22:31 | |
| *** david-lyle has quit IRC | 22:32 | |
| *** david-lyle has joined #openstack-swift | 22:40 | |
| *** Bashta has joined #openstack-swift | 22:46 | |
| Bashta | Hi there | 22:46 |
| *** _JZ_ has quit IRC | 22:50 | |
| Bashta | hey was wondering if someone has already deplyed a vapor serverver with a remote mongodb db? | 22:57 |
| *** Bashta has quit IRC | 23:04 | |
| *** JimCheung has quit IRC | 23:21 | |
| *** Jeffrey4l has quit IRC | 23:24 | |
| *** sheel has quit IRC | 23:24 | |
| *** fungi has quit IRC | 23:24 | |
| *** etienneme has quit IRC | 23:24 | |
| *** furlongm_ has quit IRC | 23:24 | |
| *** tanee has quit IRC | 23:24 | |
| *** stevemar has quit IRC | 23:24 | |
| *** Raymii has quit IRC | 23:24 | |
| *** Jeffrey4l has joined #openstack-swift | 23:24 | |
| *** sheel has joined #openstack-swift | 23:24 | |
| *** fungi has joined #openstack-swift | 23:24 | |
| *** etienneme has joined #openstack-swift | 23:24 | |
| *** furlongm_ has joined #openstack-swift | 23:24 | |
| *** tanee has joined #openstack-swift | 23:24 | |
| *** stevemar has joined #openstack-swift | 23:24 | |
| *** Raymii has joined #openstack-swift | 23:24 | |
| *** Bashta has joined #openstack-swift | 23:25 | |
| *** bobby2 has quit IRC | 23:27 | |
| *** bobby2 has joined #openstack-swift | 23:27 | |
| *** Jeffrey4l has quit IRC | 23:29 | |
| *** kei_yama has joined #openstack-swift | 23:30 | |
| *** chsc has quit IRC | 23:32 | |
| *** JimCheung has joined #openstack-swift | 23:37 | |
| *** StevenK_ is now known as StevenK | 23:40 | |
| *** Jeffrey4l has joined #openstack-swift | 23:41 | |
| *** JimCheung has quit IRC | 23:42 | |
| *** jerrygb_ has quit IRC | 23:42 | |
| *** Jeffrey4l has quit IRC | 23:42 | |
| *** Jeffrey4l has joined #openstack-swift | 23:42 | |
| *** NM has quit IRC | 23:45 | |
| *** jerrygb has joined #openstack-swift | 23:48 | |
| *** cdelatte has quit IRC | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!