| *** rdaly2 has quit IRC | 00:08 | |
| *** rdaly2 has joined #openstack-swift | 00:12 | |
| *** rdaly2 has quit IRC | 00:14 | |
| *** kyles_ne has joined #openstack-swift | 00:23 | |
| *** kyles_ne has quit IRC | 00:27 | |
| *** dmorita has joined #openstack-swift | 00:29 | |
| *** mkollaro has quit IRC | 00:45 | |
| *** nomad411 has joined #openstack-swift | 00:58 | |
| *** nosnos has joined #openstack-swift | 01:01 | |
| *** X019 has joined #openstack-swift | 01:19 | |
| *** kyles_ne has joined #openstack-swift | 01:23 | |
| *** kyles_ne has quit IRC | 01:28 | |
| *** haomaiwa_ has quit IRC | 01:55 | |
| *** sluo_laptop has joined #openstack-swift | 02:07 | |
| *** sluo_laptop has quit IRC | 02:15 | |
| *** haomaiwa_ has joined #openstack-swift | 02:17 | |
| *** tkay has joined #openstack-swift | 02:28 | |
| *** oomichi has joined #openstack-swift | 02:28 | |
| *** X019 has quit IRC | 02:32 | |
| *** tellesnobrega has joined #openstack-swift | 02:32 | |
| *** tellesnobrega has quit IRC | 02:41 | |
| *** abhirc has quit IRC | 02:44 | |
| *** X019 has joined #openstack-swift | 02:44 | |
| *** fifieldt has quit IRC | 02:45 | |
| *** nomad411 has quit IRC | 02:48 | |
| *** rdaly2 has joined #openstack-swift | 03:14 | |
| *** rdaly2 has quit IRC | 03:19 | |
| *** tkay has quit IRC | 03:45 | |
| *** nosnos has quit IRC | 03:49 | |
| *** sgowda has joined #openstack-swift | 03:55 | |
| *** kyles_ne has joined #openstack-swift | 04:16 | |
| *** nosnos has joined #openstack-swift | 04:46 | |
| *** kyles_ne has quit IRC | 04:57 | |
| *** kyles_ne has joined #openstack-swift | 04:58 | |
| *** sluo_laptop has joined #openstack-swift | 05:07 | |
| *** oomichi has quit IRC | 05:11 | |
| *** kyles_ne has quit IRC | 05:26 | |
| *** bkopilov has quit IRC | 05:27 | |
| *** bkopilov has joined #openstack-swift | 05:50 | |
| *** nosnos_ has joined #openstack-swift | 05:50 | |
| *** nosnos has quit IRC | 05:51 | |
| openstackgerrit | yinyin proposed openstack/swift: remove old tombstone(.ts) https://review.openstack.org/134855 | 05:56 |
|---|---|---|
| *** nosnos_ has quit IRC | 05:57 | |
| *** nosnos has joined #openstack-swift | 05:59 | |
| *** miqui has quit IRC | 06:22 | |
| *** nshaikh has joined #openstack-swift | 06:22 | |
| *** nosnos_ has joined #openstack-swift | 06:23 | |
| *** nosnos has quit IRC | 06:23 | |
| openstackgerrit | Merged openstack/swift: Show the sum of every policy's amount in /recon/async https://review.openstack.org/125179 | 07:10 |
| *** sungju has quit IRC | 07:18 | |
| *** miqui_ has quit IRC | 07:22 | |
| openstackgerrit | yinyin proposed openstack/swift: remove old tombstone(.ts) https://review.openstack.org/134855 | 07:39 |
| *** nshaikh has quit IRC | 08:02 | |
| *** silor has joined #openstack-swift | 08:03 | |
| *** fifieldt has joined #openstack-swift | 08:09 | |
| *** kyles_ne has joined #openstack-swift | 08:27 | |
| *** kyles_ne has quit IRC | 08:31 | |
| *** jordanP has joined #openstack-swift | 08:33 | |
| *** jordanP has quit IRC | 08:33 | |
| *** jordanP has joined #openstack-swift | 08:33 | |
| *** geaaru has joined #openstack-swift | 08:34 | |
| *** joeljwright has joined #openstack-swift | 08:40 | |
| *** mmcardle has joined #openstack-swift | 08:49 | |
| *** silor has quit IRC | 08:53 | |
| *** silor has joined #openstack-swift | 08:53 | |
| *** mmcardle1 has joined #openstack-swift | 08:55 | |
| *** mmcardle has quit IRC | 08:57 | |
| *** silor has quit IRC | 09:10 | |
| *** silor has joined #openstack-swift | 09:10 | |
| *** leopoldj has joined #openstack-swift | 09:22 | |
| openstackgerrit | yinyin proposed openstack/swift: remove old tombstone(.ts) https://review.openstack.org/134855 | 09:47 |
| *** ahonda has joined #openstack-swift | 09:50 | |
| *** ahonda has left #openstack-swift | 09:50 | |
| *** nshaikh has joined #openstack-swift | 09:57 | |
| *** silor has quit IRC | 10:07 | |
| *** silor has joined #openstack-swift | 10:08 | |
| *** haomaiwa_ has quit IRC | 10:34 | |
| *** mkerrin has joined #openstack-swift | 10:34 | |
| openstackgerrit | Merged openstack/python-swiftclient: Add unit tests for _encode_meta_headers https://review.openstack.org/133052 | 10:36 |
| *** exploreshaifali has joined #openstack-swift | 10:52 | |
| *** mkollaro has joined #openstack-swift | 10:57 | |
| openstackgerrit | Merged openstack/python-swiftclient: Fix misnamed variable in SwiftReader. https://review.openstack.org/133224 | 11:12 |
| openstackgerrit | yinyin proposed openstack/swift: remove old tombstone(.ts) https://review.openstack.org/134855 | 11:13 |
| *** dblia has joined #openstack-swift | 11:22 | |
| *** leews has quit IRC | 11:27 | |
| *** mahatic has joined #openstack-swift | 11:35 | |
| *** sgowda has quit IRC | 11:42 | |
| *** dblia has left #openstack-swift | 11:45 | |
| *** openstackgerrit has quit IRC | 11:48 | |
| *** openstackgerrit has joined #openstack-swift | 11:49 | |
| openstackgerrit | Daniel Wakefield proposed openstack/python-swiftclient: Fix misplaced check for None in SwiftUploadObject. https://review.openstack.org/133107 | 11:55 |
| *** nellysmitt has joined #openstack-swift | 11:58 | |
| *** dmorita has quit IRC | 12:29 | |
| *** nosnos_ has quit IRC | 12:39 | |
| openstackgerrit | Daniel Wakefield proposed openstack/python-swiftclient: Fix misnamed dictionary key. https://review.openstack.org/129574 | 12:41 |
| openstackgerrit | Daniel Wakefield proposed openstack/python-swiftclient: Fix misnamed dictionary key. https://review.openstack.org/129574 | 12:43 |
| openstackgerrit | Daniel Wakefield proposed openstack/python-swiftclient: Fix misnamed dictionary key. https://review.openstack.org/129574 | 12:44 |
| openstackgerrit | Daniel Wakefield proposed openstack/python-swiftclient: Fix misplaced check for None in SwiftUploadObject. https://review.openstack.org/133107 | 12:44 |
| *** foexle has joined #openstack-swift | 12:47 | |
| *** mg has joined #openstack-swift | 12:48 | |
| mg | Hi all, Swift Browser now has basic support for deleting containers and for copying objects inside an account: https://github.com/zerovm/swift-browser/releases/tag/0.2.0 | 12:49 |
| mg | Give it a try and please open issues for any problems you run into | 12:49 |
| *** exploreshaifali has quit IRC | 12:55 | |
| *** dmsimard_away is now known as dmsimard | 12:58 | |
| *** Manish_ has joined #openstack-swift | 13:01 | |
| Manish_ | Hi, | 13:01 |
| Manish_ | I am getting below memcache error, while executing parallel 1000 requests in SWIFT. "ERROR:root:Timeout connecting to memcached: 127.0.0.1:61010" | 13:02 |
| Manish_ | and the requests are failing. So should memcache error , cause the failure of requests? | 13:02 |
| *** dmsimard is now known as dmsimard_away | 13:16 | |
| *** dmsimard_away is now known as dmsimard | 13:27 | |
| *** tristanC_ is now known as tristanC | 13:27 | |
| *** k4n0_ has quit IRC | 13:39 | |
| *** sandywalsh has joined #openstack-swift | 13:44 | |
| *** foexle has quit IRC | 13:49 | |
| *** lpabon has joined #openstack-swift | 13:52 | |
| *** mmcardle1 has quit IRC | 13:55 | |
| *** mmcardle has joined #openstack-swift | 13:58 | |
| *** zul has joined #openstack-swift | 14:00 | |
| *** tellesnobrega has joined #openstack-swift | 14:05 | |
| *** tdasilva has joined #openstack-swift | 14:07 | |
| *** miqui has joined #openstack-swift | 14:10 | |
| *** tellesnobrega has quit IRC | 14:15 | |
| *** joe__ has quit IRC | 14:16 | |
| *** Manish_ has quit IRC | 14:20 | |
| *** tellesnobrega has joined #openstack-swift | 14:28 | |
| *** tellesnobrega has quit IRC | 14:32 | |
| *** dANO has joined #openstack-swift | 14:44 | |
| *** sudorandom has quit IRC | 14:49 | |
| dANO | Hi , I have a question about the swift replication . I have two servers with OpenStack Swift . The first has the ip: 99.99.99.99 and the second IP : 100.100.100.100 , my question is that I want the second ( 100.100 ) is the replication of the first . So when I created my ring , I just have to write it in the first? : swift-ring-builder object.builder add r1z1-127.0.0.1:6010R100.100.100.100:6010 / sdb1 1 and the second server I h | 14:52 |
| dANO | ave to create the same ring but with the replication ip address of the first server? | 14:52 |
| *** CJM is now known as CrackerJackMack | 14:56 | |
| ctennis | dANO: no, the ring file is the same for all servers | 14:57 |
| dANO | ok | 14:57 |
| ctennis | dANO: the ring file on server 1 should have all drives for server 1 and server2 | 14:57 |
| ctennis | basically, just describe all drives in your system | 14:57 |
| dANO | So if I want when file is upload on my first server , the file is upload on the second | 14:58 |
| dANO | too | 14:58 |
| ctennis | swift will handle that for you | 14:58 |
| *** sandywalsh has quit IRC | 14:58 | |
| dANO | Ok , so I just let my ring like this : swift-ring-builder object.builder add r1z1-127.0.0.1:6010R100.100.100.100:6010 / sdb1 1 on the two servers ? | 14:59 |
| ctennis | don't use 127.0.0.1 | 14:59 |
| ctennis | use the IP address of that server | 14:59 |
| dANO | Ok | 14:59 |
| ctennis | but yes | 14:59 |
| ctennis | plus any drives you have on the 2nd server as well | 14:59 |
| *** sandywalsh has joined #openstack-swift | 15:00 | |
| dANO | Ok so for the first : swift-ring-builder object.builder add r1z1-99.99.99.99:6010:6010R100.100.100.100:6010 / sdb1 1 and on the second swift-ring-builder object.builder add r1z1-100.100.100.100:6010R100.100.100.100:6010 / sdb1 1 | 15:00 |
| ctennis | for the first the replication IP should be 99.99.99.99 as well | 15:00 |
| *** Poe_ has quit IRC | 15:01 | |
| ctennis | you don't need to even specify the replication IP address unless it will be a different interface, which in your case it won't be, so maybe it is simpler for you not to think about it | 15:02 |
| ctennis | swift-ring-builder object.builder add r1z1-99.99.99.99:6010 / sdb1 | 15:02 |
| dANO | be different interface you mean two different physical servers ? | 15:03 |
| ctennis | no, different network interface on the same server | 15:04 |
| *** sudorandom has joined #openstack-swift | 15:06 | |
| dANO | Ok. Because in my case the swift is install on two different physical server. So if i dont need to modify the ring-builder I dont understand where I refer the ip of the two servers , maybe in the proxy config file ? | 15:07 |
| *** silor has quit IRC | 15:13 | |
| *** nshaikh has quit IRC | 15:13 | |
| *** chrisnelson_ is now known as chrisnelson | 15:14 | |
| *** exploreshaifali has joined #openstack-swift | 15:16 | |
| *** dANO has quit IRC | 15:16 | |
| *** aswadr has joined #openstack-swift | 15:17 | |
| *** dANO has joined #openstack-swift | 15:23 | |
| *** silor has joined #openstack-swift | 15:26 | |
| *** pberis has joined #openstack-swift | 15:28 | |
| *** bsdkurt has quit IRC | 15:32 | |
| tdasilva | dANO: I think what ctennis is trying to explain is that you would create one ring like this: https://gist.github.com/thiagol11/8e758db960f225f273a4 | 15:33 |
| tdasilva | dANO: and then deploy the same ring file on both nodes | 15:33 |
| dANO | Oh ok | 15:33 |
| dANO | thank's for your help | 15:34 |
| tdasilva | dANO: so no need to specify a replication ip address and of course you would specify all the disks you are planning to use on each node (e.g., sdb1, sdc1, etc...) | 15:34 |
| tdasilva | dANO: welcome | 15:34 |
| dANO | ok | 15:34 |
| dANO | I will try soon , thank's | 15:35 |
| *** bsdkurt has joined #openstack-swift | 15:36 | |
| tdasilva | acoles: hi, are you around? | 15:43 |
| notmyname | good morning world | 15:44 |
| tdasilva | good morning :-) | 15:44 |
| notmyname | been a while since I've been in here :-) | 15:45 |
| notmyname | or it feels like it | 15:45 |
| tdasilva | welcome back | 15:45 |
| tdasilva | how was vacation? | 15:46 |
| notmyname | nice. barcelona is an interesting place | 15:47 |
| notmyname | got to see some cool stuff. and I didn't think about openstack at all ;-) | 15:47 |
| *** gyee has joined #openstack-swift | 15:47 | |
| *** zul has quit IRC | 15:49 | |
| *** zul has joined #openstack-swift | 15:50 | |
| notmyname | likely abandoned patches have been marked abandoned | 15:52 |
| peluse | morning | 15:53 |
| peluse | clayg, ping me when you're around. want to see about maybe a phone call at around 2 CA time. I have the skeleton of a patch almost working that covers most if not all of the crap we talked about last Fri in Paris. | 15:55 |
| peluse | will grab a bridge and anyone interested would be welcome to join... | 15:55 |
| tdasilva | notmyname: nice! | 15:57 |
| acoles | notmyname: welcome back, did you get to see the upside down string model? | 16:05 |
| acoles | tdasilva: here now | 16:06 |
| notmyname | anticw: yup. pretty cool | 16:06 |
| notmyname | acoles: ^^ | 16:06 |
| acoles | notmyname: forgotten how to type? :) | 16:06 |
| notmyname | acoles: tab-complete fail | 16:07 |
| tdasilva | acoles: hey...I was checking out the in-process patch and noticed that if I remove a middleware from the pipeline the tests for that middleware stills runs, is that expected? | 16:07 |
| notmyname | today's task: figure out what the rest of you have been up to last week ;-) | 16:07 |
| tdasilva | acoles: for example, I remember dlo from the pipeline in the proxy sample file, but the DLO tests still ran fine | 16:07 |
| acoles | tdasilva: looking... | 16:08 |
| tdasilva | acoles: I was expecting those tests would be skipped, wrong assumption i guess :-) | 16:08 |
| notmyname | tdasilva: doesn't DLO automagically work even if it's not in the pipeline? (because it used to be in the proxy server and there are migration thingies in there) | 16:08 |
| tdasilva | s/I remember dlo/I removed dlo/ | 16:08 |
| tdasilva | notmyname: oh, picked wrong middleware to test then :-) | 16:09 |
| tdasilva | notmyname: just picked that one out of the blue | 16:09 |
| notmyname | tdasilva: http://www.geekmelange.com/wp-content/uploads/2014/01/indiana_jones_grail_knight-you-have-chosen-poorly.jpg | 16:09 |
| *** IRTermite1 has quit IRC | 16:09 | |
| tdasilva | notmyname: lol | 16:09 |
| *** IRTermite has joined #openstack-swift | 16:10 | |
| tdasilva | notmyname, acoles: another related question...I noticed that in the sample file dlo and slo are positioned after tempauth, but in the saio files they are before...does that matter? | 16:10 |
| acoles | tdasilva: they should be after | 16:11 |
| notmyname | tdasilva: yeah it matters. you want them after, IIRC, because that way the token auth "wraps" the entire set of subrequests and you don't get edge cases of the token expiring half way through reading the large object | 16:11 |
| *** david-lyle_afk is now known as david-lyle | 16:11 | |
| tdasilva | notmyname: so it's a bug in the saio proxy-server.conf, right? | 16:12 |
| acoles | tdasilva: what notmyname said - they were before but got moved to after | 16:12 |
| notmyname | tdasilva: ya, sounds like the saio docs are wrong. can you throw up a quick patch for it? | 16:13 |
| tdasilva | yeah...doing right now | 16:13 |
| tdasilva | notmyname, acoles: the reason I'm asking is because the I was wondering where to place the object versioning middleware | 16:14 |
| notmyname | tdasilva: after auth, for the same reasons. /me is happy to hear about work on that | 16:14 |
| tdasilva | notmyname: cool, thanks! | 16:15 |
| notmyname | tdasilva: because words matter, can you call the new middleware "versioned_writes" instead of just "versioning"? that way it's more clear what it does and doesn't do | 16:15 |
| tdasilva | notmyname: sure! will do | 16:16 |
| notmyname | also that leaves the door open for another middleware later that actually does handle full versioning with delete handling | 16:16 |
| tdasilva | "delete handling"? | 16:16 |
| notmyname | tdasilva: versioned writes in swift are like a stack (push and pop). so it's not an audit record of "v1, v2, delete, v3, delete" where you can go get any of those at some point in time | 16:17 |
| tdasilva | notmyname: well, you can if you have access to the "versions" container, right? | 16:20 |
| tdasilva | notmyname: do you mean there is no simple api to list an object and its versions? | 16:20 |
| notmyname | no. that you can't roll back an object to the version that was deleted | 16:21 |
| notmyname | that would require a new kind of tombstone file marker | 16:21 |
| tdasilva | oh, i see now | 16:21 |
| *** leopoldj has quit IRC | 16:21 | |
| tdasilva | notmyname: got it, thanks | 16:21 |
| tdasilva | notmyname: why do we attach that prefix to the versioned objects? couldn't figure that one out... | 16:23 |
| *** lpabon has quit IRC | 16:28 | |
| notmyname | tdasilva: ah. that. clever hack to avoid name collisions if you have 2 containers versioning to the same versioning container (IIRC) | 16:28 |
| acoles | tdasilva: so was it just that you picked dlo (which gets put back in the pipeline on the sly)? | 16:31 |
| acoles | tdasilva: here btw https://github.com/openstack/swift/blob/master/swift/proxy/server.py#L62 | 16:32 |
| tdasilva | acoles: yeah, I will try with a different one | 16:33 |
| acoles | tdasilva: otherwise the in-process tests should skip tests based on cluster info just like 'regular' func tests | 16:33 |
| acoles | tdasilva: tempurl tests should skip if you pull it from pipeline | 16:34 |
| openstackgerrit | Thiago da Silva proposed openstack/swift: move slo, dlo after tempauth in pipeline https://review.openstack.org/135008 | 16:34 |
| tdasilva | acoles: nice, running tests now | 16:35 |
| acoles | tdasilva: is your patch ^^ a straight copy/paste from proxy-server.conf-sample? | 16:36 |
| tdasilva | acoles: no, I noticed that there were some differences that I did not want to change, like container-sync | 16:37 |
| acoles | tdasilva: oh, ok, cool | 16:37 |
| tdasilva | acoles: do you think they should be the equal? is that why you are asking? | 16:38 |
| acoles | tdasilva: no, i was just being lazy ;) | 16:38 |
| tdasilva | acoles: hehe, no problem... | 16:39 |
| tdasilva | acoles: so, I noticed the tests were skipped, which is nice | 16:39 |
| acoles | tdasilva: phew! | 16:40 |
| cebruns | Hi all - noob question here - is there a way to find the total capacity available in a Swift cluster? 'swift stat' and 'swift capabilities' look promising, but don't give total capacity installed. | 16:40 |
| tdasilva | nosetests can be deceiving I guess, because it seems to count skipped tests as "ran" | 16:40 |
| openstackgerrit | Alistair Coles proposed openstack/swift: move slo, dlo after tempauth in pipeline https://review.openstack.org/135008 | 16:41 |
| notmyname | reminder that our weekly meeting is at 1900UTC. which means that with the US time change, it's now at 11am Pacific and 2pm Eastern | 16:41 |
| notmyname | meeting agenda updated at https://wiki.openstack.org/wiki/Meetings/Swift | 16:42 |
| tdasilva | acoles: sort of related to in-process patch, I noticed that if your saio processes are not running, but if there's a test.conf file in /etc/swift, then func tests just don't run. Shouldn't we change it so that it run in-process instead? | 16:42 |
| tdasilva | acoles: could be a separate patch | 16:42 |
| mahatic | yay, I'm a part of the meeting! | 16:44 |
| acoles | tdasilva: hmm, not sure, i might want the headsup that my saio isn't started, rather than the tests running against sample conf instead. | 16:47 |
| acoles | tdasilva: intersting idea though, but would prefer it be another patch | 16:47 |
| tdasilva | acoles: yes, you make a good point | 16:48 |
| *** annegent_ has joined #openstack-swift | 16:51 | |
| *** haomaiwang has joined #openstack-swift | 16:54 | |
| *** tattabbum has joined #openstack-swift | 17:03 | |
| tattabbum | Hi there! How can I put object using Postman with Swift's REST API? | 17:03 |
| tattabbum | It works using cURL: curl -X PUT -T app.war http://IP_SWIFT:8080/v1/AUTH_043c13d5c76747f3befed94215b0b0fd/Repository/app.war -H "X-Auth-Token: $token" -H "Content-Type: application/zip" | 17:04 |
| tattabbum | where token contains the keystone authentication token related to the tenant | 17:05 |
| *** annegent_ has quit IRC | 17:06 | |
| morganfainberg | notmyname, ping - so out of curiosity [i know 32 bytes, you like that token length], how far off the mark would ~150-200byte tokens be? that is to say if it also made it so keystone didn't need to store the tokens on the backend, which makes tokens *way way* better for everyone. | 17:06 |
| morganfainberg | notmyname, as an option. | 17:06 |
| *** kyles_ne has joined #openstack-swift | 17:09 | |
| *** tattabbum has quit IRC | 17:10 | |
| *** tattabbum has joined #openstack-swift | 17:11 | |
| *** tkay has joined #openstack-swift | 17:30 | |
| *** kyles_ne has quit IRC | 17:31 | |
| *** kyles_ne has joined #openstack-swift | 17:32 | |
| *** nexusz99 has joined #openstack-swift | 17:39 | |
| dANO | Hi again, I use Openstack swift for store some files who is present in my rails app. But at the beginning my file was store in Amazon S3 , so I decide to move the file to Amazon S3 in Openstack Swift but after have move somes files now I have an error 503 service unavailable and i cant upload new file in my Openstack Swift Server. Anyone know how i can resolve this problem ? | 17:43 |
| tdasilva | dANO: can you provide more info? if possible, throw some logs and config files on a etherpad or gist. Also, what command is failing (e.g., PUT)? does GET work? the more information you can provide the easier it will be for people to help you out... | 17:49 |
| dANO | tdasilva: I know but in my syslog nothing appear so I dont understand why. | 17:52 |
| *** jordanP has quit IRC | 17:54 | |
| notmyname | morganfainberg: I'm curious about your "keystone doesn't need to store tokens". meaning it can compute validity rather than going to a lookup table? any chance swift could do it with a shared secret? | 18:03 |
| morganfainberg | We would reconstruct the token data on validate. We know that data isn't changing (fairly bounded data set) | 18:04 |
| morganfainberg | The issue is keystone is currently storing the token data, which results in an unbounded data set we need to store. Most of which is ephemeral data. | 18:04 |
| notmyname | right | 18:05 |
| *** annegent_ has joined #openstack-swift | 18:06 | |
| notmyname | morganfainberg: in general, shorter is better. with v3/v3 you've already broken simple interactions with curl (ie it's hard to type in json bodies to post creds), so 100-200 byte tokens is probably fine. | 18:06 |
| morganfainberg | So we're looking at allowing a token that contains bare minimum data. Userid, scope, audit id, expiry. Encoded and signed. Keystone can construct he data body (most of that data will / can be cached). So validate should be fast-ish. | 18:06 |
| notmyname | morganfainberg: certainly better than 8k+ tokens ;-) | 18:06 |
| morganfainberg | Oh yeah. I know we talked about 8k, 1k, 500b. | 18:07 |
| morganfainberg | So wanted to see your feeling on the 150-200b range | 18:07 |
| notmyname | I'd be happy if keystone tokens were on the order of 100-200 bytes. my understanding of large public clusters is that size wouldn't be unconscionable overhead (like multi kilobytes is). | 18:08 |
| morganfainberg | ++. | 18:08 |
| notmyname | morganfainberg: at least until we can support signed URLs in keystone+swift ;-) | 18:08 |
| notmyname | cause that would be really nice | 18:09 |
| morganfainberg | notmyname: it's a battle I've kept bringing up and kept running into issues with. | 18:09 |
| morganfainberg | Longer term, I'd like to support it. | 18:09 |
| morganfainberg | But we have an issue, nova doing things on behalf of user - talking to another service. | 18:10 |
| morganfainberg | Eg glance or cinder. | 18:10 |
| notmyname | yup | 18:10 |
| *** tattabbum has quit IRC | 18:10 | |
| morganfainberg | Once we solve how to do that with signed requests j think we can revisit. | 18:10 |
| notmyname | for now, swift has tempurl for that. you have to store the secret in swift itself, but it does give a degree of flexibility for signed urls that is very nice | 18:11 |
| *** bkopilov has quit IRC | 18:13 | |
| dANO | If I created three accounts on my Openstack Swift , the total size of my storage server is divise by 3 ? Or I can setup specific size for the accounts ? | 18:16 |
| morganfainberg | notmyname: ++. We might be able to use credential backend in keystone to allow that as well. But revisit when I think about that more. | 18:18 |
| *** geaaru has quit IRC | 18:19 | |
| *** tkay has left #openstack-swift | 18:21 | |
| *** abhirc has joined #openstack-swift | 18:24 | |
| *** elambert has joined #openstack-swift | 18:26 | |
| *** nellysmitt has quit IRC | 18:27 | |
| *** cbaesema has joined #openstack-swift | 18:28 | |
| tdasilva | dANO: you should look at the docs for account quota for determining specific size for a given account | 18:33 |
| *** silor has quit IRC | 18:33 | |
| tdasilva | dANO: http://docs.openstack.org/developer/swift/middleware.html#module-swift.common.middleware.account_quotas | 18:33 |
| *** silor has joined #openstack-swift | 18:38 | |
| clayg | peluse: sorry i missed you, i'm around today | 18:39 |
| dANO | tdasilva: thank's with this command all my openstack swift is reconfigured for allow the correct space right ? | 18:43 |
| *** zaitcev has joined #openstack-swift | 18:44 | |
| *** ChanServ sets mode: +v zaitcev | 18:44 | |
| *** gyee has quit IRC | 18:45 | |
| *** gyee has joined #openstack-swift | 18:45 | |
| tdasilva | dANO: not sure I understand your question...but for example, if you don't have account quota set for any account, your storage capacity will be shared among your 3 accounts and any one account will be able to consume any amount of storage up to the cluster capacity | 18:45 |
| dANO | Ok | 18:46 |
| tdasilva | dANO: If a account quota is set for a given account, then that specific account will be able to consume only up to that limit you set | 18:46 |
| tdasilva | dANO: does that make sense? | 18:46 |
| dANO | yes it's ok | 18:46 |
| dANO | but i find my problem with openstack, is just my disk /dev/root is full so i cant upload file in | 18:47 |
| dANO | I go to increase the max size of him | 18:47 |
| *** zul has quit IRC | 18:48 | |
| notmyname | reading the swift+LTFS thread in the -dev mailing list is interesting | 18:48 |
| tdasilva | notmyname, cschwede: looks like IBM is also looking at swift + tape storage: https://www.youtube.com/watch?v=EMLK5XsU0_8&feature=youtu.be | 18:51 |
| tdasilva | unfortunately audio is pretty bad for some reason | 18:52 |
| *** annegent_ has quit IRC | 18:55 | |
| *** NM has joined #openstack-swift | 19:01 | |
| *** annegent_ has joined #openstack-swift | 19:01 | |
| peluse | clayg, OK one more stumbling block I want to get over before I share the ideas/code w/you... stay tuned :) | 19:16 |
| clayg | peluse: np | 19:17 |
| NM | Hey guys! How are you? Just got back from 20 days of vacation. What is the current version? 10.0 :D | 19:18 |
| *** aswadr has quit IRC | 19:20 | |
| clayg | n | 19:21 |
| tdasilva | hey NM, welcome back! :-) | 19:22 |
| clayg | NM: acctually notmyname was on vacation too - so we didn't really get anything done | 19:22 |
| notmyname | lol | 19:22 |
| notmyname | I myself am trying to figure out what's going on | 19:22 |
| clayg | notmyname: we had a meeting last week - http://eavesdrop.openstack.org/irclogs/%23openstack-swift/%23openstack-swift.2014-11-12.log (search for meeting) | 19:24 |
| notmyname | clayg: cool, thanks | 19:24 |
| clayg | oh you *say* that | 19:25 |
| *** rdaly2 has joined #openstack-swift | 19:25 | |
| *** judd7 has joined #openstack-swift | 19:26 | |
| *** infotection has quit IRC | 19:33 | |
| *** infotection has joined #openstack-swift | 19:34 | |
| *** zul has joined #openstack-swift | 19:35 | |
| *** acoles is now known as acoles_away | 19:37 | |
| *** mahatic has quit IRC | 19:39 | |
| *** dANO has quit IRC | 19:39 | |
| *** rdaly2 has quit IRC | 19:42 | |
| *** rdaly2 has joined #openstack-swift | 19:56 | |
| *** mahatic has joined #openstack-swift | 20:00 | |
| openstackgerrit | Clay Gerrard proposed openstack/python-swiftclient: Capture test output better https://review.openstack.org/131238 | 20:10 |
| *** annegent_ has quit IRC | 20:22 | |
| *** annegent_ has joined #openstack-swift | 20:24 | |
| openstackgerrit | Clay Gerrard proposed openstack/swift: Always use FakeMemcache for in-process tests https://review.openstack.org/135069 | 20:31 |
| *** exploreshaifali has quit IRC | 20:45 | |
| *** openstackgerrit has quit IRC | 20:49 | |
| *** openstackgerrit has joined #openstack-swift | 20:49 | |
| clayg | is anyone worried about https://review.openstack.org/#/c/134419/ changing error limiting to be per device instead of per ring/device ? | 20:50 |
| clayg | like if i have a node that is serving account/container/object - and the account is blowing up - does that mean I should probably not be sending it object requets either? | 20:51 |
| *** mkollaro has quit IRC | 20:52 | |
| notmyname | clayg: I can imaging container requests slowing down when object requests are ok | 20:53 |
| notmyname | if the limit is the db lock | 20:53 |
| clayg | yeah error limiting is hard, it's like if you have a strong signal that server/drive is messed up - but the signal is only coming from objects, you might wanna throw a container request at it and see what happens... but if that horks too maybe you don't bother with the next account request that comes along... | 20:54 |
| clayg | oh you know what - sam's error keys include the port - i'm a jackass | 20:57 |
| *** silor has quit IRC | 20:59 | |
| *** mkollaro has joined #openstack-swift | 21:03 | |
| *** NM has quit IRC | 21:13 | |
| *** tdasilva has quit IRC | 21:16 | |
| openstackgerrit | Anne Gentle proposed openstack/swift-specs: Adds Object Storage v1 API specification information https://review.openstack.org/129384 | 21:21 |
| *** fifieldt_ has joined #openstack-swift | 21:22 | |
| *** fifieldt has quit IRC | 21:23 | |
| *** rdaly2 has quit IRC | 21:25 | |
| *** miqui has quit IRC | 21:29 | |
| *** exploreshaifali has joined #openstack-swift | 21:30 | |
| clayg | notmyname: ^ do we really want to doc the v1 api in the specs repo? that seems to stand in the face of specs are not docs mantra? | 21:36 |
| clayg | why don't we just do that do that in /docs ??? | 21:36 |
| annegent_ | clayg: the rest of the projects are moving towards specs so that it's not a true "contract" per se | 21:40 |
| annegent_ | clayg: and if you ever invent v1.1 or v2 you'll want a place to design | 21:40 |
| notmyname | annegent_: I know that there have been the discussions around distributed docs management, which sounds great. so the answer is that the API docs are going to be in the associated -specs repo, then? | 21:41 |
| annegent_ | notmyname: except for http://developer.openstack.org/api-ref.html that is still considered to be "user" doc | 21:44 |
| notmyname | annegent_: what are the others, if not "user"? | 21:44 |
| annegent_ | so the API docs for contrib devs are in specs, API docs for SDK devs are the API ref | 21:44 |
| annegent_ | swift dev = specs. php-opencloud dev = developer.openstack.org | 21:44 |
| clayg | notmyname: we can merge them into /specs but it's going to be a decidely different workflow than all of the other design documents - the stuff that you argue about during design has little to do with presenting the client facing interface in a consumable fashion | 21:46 |
| *** zaitcev has quit IRC | 21:46 | |
| clayg | I think it's an impedence between what other people think they want from specifications and what we realisticly hope to achieve from garnering concensus on a design prior to a nuts and bolt review of the implementation | 21:48 |
| clayg | either way if it's documentation it's not a design document - and I don't really have a cut and dry definition of "specification" | 21:48 |
| notmyname | annegent_: in gerrit I see you got some comments from donagh about what should be where (-specs vs other). is that what you are basing this on? is there any other general rule for how you chose these things to add to -specs? | 21:50 |
| clayg | annegent_: fwiw I violently agree that when we want to hash out a v1.1 api it'll have a iteration through a design review prior to implementation, but ultimately we'll also want it to have documentation that we almost cirtinaly not match 1-to-1 with everything that comes up during design | 21:51 |
| clayg | an api doc will need to say how it works, but it can skip of why it has to work this way for the implementation to be sane | 21:51 |
| *** judd7 has quit IRC | 21:51 | |
| annegent_ | notmyname: I'm working with each team to fit what they'd like | 21:51 |
| annegent_ | clayg: I don't think it's a justification doc. more a way to convey to swift core reviewers what the api was supposed to be coded as | 21:52 |
| *** tellesnobrega has joined #openstack-swift | 21:52 | |
| annegent_ | notmyname: clayg: these specs should work well with what the API working group is drafting as API guidelines | 21:53 |
| annegent_ | notmyname: clayg: not that that's super well-defined yet, again, we're working with teams on what they see as working well. | 21:53 |
| annegent_ | notmyname: as far as I know you're not really drafting a new API so for you all, it's really just a matter of object-api going away. | 21:53 |
| annegent_ | notmyname: so with object-api going away, I'm harvesting bits to where we think the content should go. best guesses | 21:54 |
| notmyname | ah, ok | 21:54 |
| annegent_ | nova doesn't have it all set yet, really only keystone has theirs processed this way already | 21:55 |
| annegent_ | but those two projects are working on new/updated apis | 21:55 |
| notmyname | annegent_: would you be opposed to moving them to the swift repo under the /docs directory? I think that would work a lot better with both how we use our current in-tree docs and how we want to use -specs | 21:56 |
| annegent_ | notmyname: not really, seems like the right reviewers are there | 21:57 |
| notmyname | annegent_: sorry I haven't been on top of this sooner. summit prep+summit+vacation have kept me busy ;-) | 21:57 |
| notmyname | annegent_: thanks! | 21:57 |
| annegent_ | notmyname: heh, I know what you mean | 21:57 |
| annegent_ | notmyname: do you need to discuss at a swift team meeting or shall I just make the move? | 21:58 |
| notmyname | annegent_: so that being said, I'm not really sure about the distinction of what's here vs not. swift doesn't have an admin api, so all of these things are stuff that end-user clients can use | 21:58 |
| annegent_ | ah, it's the actual use cases that don't belong in an API reference. | 21:59 |
| annegent_ | notmyname: such as "how to make a static website" -- most "how to" info belongs in the user guide | 21:59 |
| notmyname | oh ok | 21:59 |
| clayg | annegent_: if you move it it'll give us a call to action in the swift meeting on wednesday (go review XXX, it'll make /docs better!) | 22:00 |
| annegent_ | so donaugh did a pretty decent job of separating out how to | 22:00 |
| notmyname | annegent_: just make the move. we discussed the specs repo at the summit, and long-lived reference docs is _not_ how we are using it or want to use it | 22:00 |
| annegent_ | clayg: notmyname: ok | 22:00 |
| notmyname | and we can point it out for review on wednesday's meeting | 22:00 |
| clayg | nice | 22:00 |
| annegent_ | notmyname: clayg: is it a separate folder and index.rst, or do you want it as a section? | 22:01 |
| * clayg is super pumped about having a decent api reference in /docs | 22:01 | |
| annegent_ | clayg: heh. don't get too excited, it's prety simplistic | 22:01 |
| *** zaitcev has joined #openstack-swift | 22:02 | |
| *** ChanServ sets mode: +v zaitcev | 22:02 | |
| clayg | annegent_: acctually I think the stuff we struggle with most is when there's a total vaccum, if we have a skeleton we'll sorta hang some meat off it over time | 22:02 |
| notmyname | annegent_: I think getting them in to the docs directory is more important than having the content properly in the right places. I'm fine with either a section or a separate folder for now. we can make it better over time in subsequent patches | 22:03 |
| clayg | annegent_: so currently the first link in "overview and concepts" is "Swift's API docs" => http://docs.openstack.org/api/openstack-object-storage/1.0/content/ | 22:03 |
| clayg | annegent_: will these superseed that? | 22:03 |
| annegent_ | clayg: yes, exactly, those are going away. That's sourced from object-api. | 22:04 |
| annegent_ | clayg: so I can either put an /api/ folder in https://github.com/openstack/swift/tree/master/doc/source or doc/api? or something else | 22:05 |
| clayg | i like /doc/source/api? | 22:05 |
| clayg | notmyname is going to grab coffee - he says I can have whatever I want! | 22:06 |
| clayg | yippee! | 22:06 |
| *** BAKfr has quit IRC | 22:06 | |
| annegent_ | clayg: hee! | 22:06 |
| clayg | our specs repo is sorta weird in that there's a number of top level directories that you have to watch for changes when you're doing something with an auto-build script - so my desire to keep everything rooted in /doc/source is probably self serving | 22:07 |
| clayg | but i don't see any *good* reason to to leave everything rooted in /doc/source either and it obviously makes sense to next everything under a subdir if it's a fork lift | 22:08 |
| *** BAKfr has joined #openstack-swift | 22:09 | |
| clayg | s/reason to to/reason not to/ | 22:09 |
| openstackgerrit | Anne Gentle proposed openstack/swift: Adds v1 API documentation to doc/source/api https://review.openstack.org/135102 | 22:19 |
| *** mahatic has quit IRC | 22:20 | |
| *** exploreshaifali has quit IRC | 22:20 | |
| *** mahatic has joined #openstack-swift | 22:21 | |
| annegent_ | clayg: yeah I think what other projects are doing are a folder per api version | 22:22 |
| annegent_ | clayg: but only keystone has landed so far | 22:22 |
| *** annegent_ has quit IRC | 22:23 | |
| clayg | oh that makes sense i guess | 22:23 |
| *** rdaly2 has joined #openstack-swift | 22:24 | |
| clayg | so /doc/source/api/v1/overview.rst or something... i'm guessing when we make it to v1.1 a lot of stuff will still be the same with v1.1 dunno how sphinx is about inheritence... | 22:24 |
| clayg | probably everything in one bucket is fine until we have a v1.1 - leave it up to the implementation of v1.1 to figure out the specifics of the reoganization of the doc tree | 22:25 |
| *** rdaly2 has quit IRC | 22:33 | |
| *** tellesnobrega has quit IRC | 22:37 | |
| *** rdaly2 has joined #openstack-swift | 22:47 | |
| *** rdaly2 has quit IRC | 22:56 | |
| *** judd7 has joined #openstack-swift | 22:57 | |
| *** dmsimard is now known as dmsimard_away | 22:57 | |
| *** judd7 has quit IRC | 23:10 | |
| *** gyee has quit IRC | 23:21 | |
| *** zul has quit IRC | 23:31 | |
| *** tkay has joined #openstack-swift | 23:39 | |
| *** annegent_ has joined #openstack-swift | 23:40 | |
| *** annegent_ has quit IRC | 23:42 | |
| *** rdaly2 has joined #openstack-swift | 23:53 | |
| *** rdaly2 has quit IRC | 23:56 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
