Wednesday, 2014-07-23

notmynamemattoliverau: so I don't think your patch is bad at all. I was just wondering what I was missing when looking at it00:01
mattoliveraunotmyname: Lol, your weren't missing much, I'm just working through old bugs :) Feel free to -2 any that you don't think needs to be solved and we can just close the bug :)00:03
notmynameit does seem interesting that there isn't an explicit test for DiskFileWriter.write() anywhere. actually, the only 3 lines not tested are in that method00:03
notmynameseems it's just indirectly tested00:03
notmynamemattoliverau: so based on this, do you plan on doing anything with that patch? eg you might look at the test_short_body() test00:04
notmynamemattoliverau: and thanks for going through old bugs!00:05
mattoliveraunotmyname: how about I take a look at the short body test, and see if I can add some more DiskFile.write tests (though the latter might need to be in another change).00:06
notmynamemattoliverau: ok, sounds good. and I agree about a different patch set00:06
notmynamemattoliverau: were you looking in to doing the http method calls to keystone in python-swiftclient so that we don't have to use keystone client?00:08
notmynameI don't remember who was looking in to that00:08
mattoliveraunotmyname: that wasn't me, I was around when your were talking about it, was it goodes or someone else. I can't quite remember.. happy to take a look though, but wont step on someone elses toes, so if we can't find out who it was I'll take a look :)00:10
notmynameheh, ok00:11
notmynameI haven't heard anything or seen anything. just curious. IMO it should be done before the juno release00:12
mattoliveraunotmyname: yeah, the keystone requirement in swiftclient caused some issues in the infra sprint. Specifically in regards to a keystone and gnome-keyring issue Ubuntu precise.00:14
mattoliverauor something like that.. keystone related, but was causing log hosting in swift to fail regularly. until the work around of removing a python keyring package was found.00:17
notmynamehmm..interesting. I hadn't heard that00:18
mattoliveraunotmyname: well we only found it out last week when implementing the moving of the logsto swift :)00:20
openstackgerritA change was merged to openstack/swift: lock_file race fixes  https://review.openstack.org/10830900:23
mattoliveraunotmyname: sdague has already submitted a fix to keystone for it (https://review.openstack.org/#/c/107719/) .00:24
*** dmorita has joined #openstack-swift00:25
*** ho has joined #openstack-swift00:45
*** shri has left #openstack-swift00:45
*** mwstorer has quit IRC00:52
thurloatomg logs are just spam01:24
thurloathah, sorry about that. wrong window01:30
thurloatquestion relating to spam logs though, tried cranking the loglevel down for swiftproxy server, and we're still getting thousands of lines of Jul 22 06:48:28 swproxy03 proxy-server: Handoff requested (2) (txn: txXXXXX) (client_ip: 10.100.51.254)01:31
*** AbyssOne has quit IRC01:33
*** AbyssOne has joined #openstack-swift01:36
*** nexusz99 has joined #openstack-swift01:45
*** nosnos has joined #openstack-swift01:50
*** nexusz99 has quit IRC01:53
*** haomaiw__ has quit IRC02:23
*** haomaiwa_ has joined #openstack-swift02:25
*** sungju has joined #openstack-swift02:26
*** haomaiw__ has joined #openstack-swift02:38
*** haomaiwa_ has quit IRC02:41
*** gyee has quit IRC03:02
torgomaticyeah, someone should probably fix that log message at some point03:07
torgomaticever since the proxy started looking at handoffs on 404, that message has been nothing more than spam03:08
*** nosnos has quit IRC03:29
*** nosnos has joined #openstack-swift03:29
*** haomaiw__ has quit IRC03:32
*** haomaiwang has joined #openstack-swift03:33
*** nosnos has quit IRC03:33
*** nosnos has joined #openstack-swift04:14
*** stevemac has joined #openstack-swift04:19
*** stevemac is now known as stevetan_04:20
*** bszabo has joined #openstack-swift04:26
openstackgerritTushar Gohad proposed a change to openstack/swift: EC: Add support for policy types, 'erasure_coding' policy  https://review.openstack.org/10364404:29
*** haomaiw__ has joined #openstack-swift04:29
*** ho has quit IRC04:30
*** bszabo has quit IRC04:30
*** haomaiwang has quit IRC04:32
*** mrmoje has quit IRC04:37
*** sumit_ has joined #openstack-swift04:39
sumit_Hi, how to generate token when using swift with keystone?04:39
*** stevetan_ has quit IRC04:40
*** stevetan has joined #openstack-swift04:42
*** stevetan has quit IRC04:43
*** anand_ts has joined #openstack-swift04:55
zaitcevit's the same token... so just something like04:59
zaitcevcurl -v -d '{"auth": {"passwordCredentials": {"username": "admin", "password": "passw0rd"}, "tenantName": "admin"}}' -H "Content-type: application/json" http://localhost:5000/v2.0/tokens05:00
zaitcevit shold puke back the token with id: and serviceCatalog with endpoints[ {'publicURL': ... ]05:01
zaitcevSame as any other OpenStack service05:01
*** mrsnivvel has quit IRC05:03
*** ppai has joined #openstack-swift05:04
*** mrmoje has joined #openstack-swift05:09
sumit_I'm facing this error: {"error": {"message": "Unable to sign token. (Disable debug mode to suppress these details.)", "code": 500, "title": "Internal Server Error"}}05:15
sumit_logs are at location: http://paste.openstack.org/show/87693/05:15
*** mrsnivvel has joined #openstack-swift05:16
*** chandankumar has joined #openstack-swift05:18
zaitcevjust go back to bearer tokens for now05:18
zaitcevlooks like your openssl is underrev for PKI tokens05:19
zaitcever05:19
zaitcevthey both are bearer tokens, I meant go back to uuid tokens05:19
*** chandankumar_ has joined #openstack-swift05:20
sumit_Thanks, it worked. for uuid token05:22
sumit_but we need pki token05:22
*** chandankumar has quit IRC05:22
sumit_can you tell us how we can rev up the ssl for pki token05:22
zaitcevSorry, I'm not an expert on that. I just update my openssl with yum and that's it.05:25
sumit_ok thanks,05:26
sumit_we will try to solve the prolem, since we have come to know the cause05:27
sumit_:005:27
*** ciypro|afk has quit IRC05:28
*** chandankumar_ is now known as chandankumar05:28
*** zaitcev has quit IRC05:31
*** k4n0 has joined #openstack-swift05:33
*** nosnos has quit IRC05:35
*** nosnos has joined #openstack-swift05:36
*** nosnos has quit IRC05:41
*** nosnos has joined #openstack-swift05:59
*** els13 has joined #openstack-swift06:03
*** sumit_ has quit IRC06:05
*** nexusz99 has joined #openstack-swift06:05
*** psharma has joined #openstack-swift06:06
*** ppai has quit IRC06:11
nexusz99Hi. Can i ask you a question? . In 'swift/common/storage_policy.py' , why 'StoragePolicySingleton' class wraps _POLICIES. I think it is unnecessary.06:13
*** sungju has quit IRC06:25
*** ppai has joined #openstack-swift06:28
*** zslot has joined #openstack-swift07:26
*** mrsnivvel has quit IRC07:35
*** ppai has quit IRC07:39
*** minisot has quit IRC07:44
*** mrsnivvel has joined #openstack-swift07:51
*** ppai has joined #openstack-swift07:53
*** mrmoje has quit IRC07:55
*** bszabo has joined #openstack-swift08:03
*** nacim has joined #openstack-swift08:05
*** bszabo has quit IRC08:07
*** acoles_away is now known as acoles08:12
*** openstackgerrit_ has joined #openstack-swift08:15
*** foexle has joined #openstack-swift08:15
*** mmcardle has joined #openstack-swift08:18
*** bszabo_ has joined #openstack-swift08:29
*** mrmoje has joined #openstack-swift08:30
*** ppai has quit IRC08:35
*** mmcardle has quit IRC08:41
*** ppai has joined #openstack-swift08:49
*** mmcardle has joined #openstack-swift08:50
*** acoles is now known as acoles_away08:55
*** bszabo_ has quit IRC09:00
*** bszabo has joined #openstack-swift09:01
*** bszabo has quit IRC09:05
*** ppai has quit IRC09:07
*** haomaiw__ has quit IRC09:11
*** haomaiwang has joined #openstack-swift09:11
*** anand_ts has quit IRC09:13
*** haomai___ has joined #openstack-swift09:15
*** acoles_away is now known as acoles09:16
*** haomaiwang has quit IRC09:18
*** ppai has joined #openstack-swift09:20
*** mmcardle has quit IRC09:24
*** ppai has quit IRC09:25
*** mmcardle has joined #openstack-swift09:27
*** ppai has joined #openstack-swift09:39
*** mmcardle has quit IRC09:55
*** nexusz99 has quit IRC09:57
*** mmcardle has joined #openstack-swift10:01
*** kopparam has joined #openstack-swift10:02
openstackgerritLin Yang proposed a change to openstack/swift: Remove duplicate conditional statement  https://review.openstack.org/10894010:03
*** nosnos has quit IRC10:09
*** mmcardle has quit IRC10:17
*** acoles is now known as acoles_away10:18
*** acoles_away is now known as acoles10:25
*** mmcardle has joined #openstack-swift10:36
*** nsaje has left #openstack-swift10:41
*** ppai has quit IRC10:43
*** bszabo_ has joined #openstack-swift10:43
*** ppai has joined #openstack-swift10:56
*** foexle has quit IRC11:01
*** psharma has quit IRC11:03
*** ppai has quit IRC11:05
*** psharma has joined #openstack-swift11:17
*** ppai has joined #openstack-swift11:22
*** kopparam has quit IRC11:24
*** kopparam has joined #openstack-swift11:27
*** ppai has quit IRC11:31
*** psharma has quit IRC11:33
*** kopparam has quit IRC11:42
*** ppai has joined #openstack-swift11:43
*** kopparam has joined #openstack-swift11:43
*** mkollaro has joined #openstack-swift11:45
*** chandankumar has quit IRC11:47
*** kopparam has quit IRC11:47
*** chandankumar has joined #openstack-swift11:49
*** psharma has joined #openstack-swift11:50
*** k4n0 has quit IRC12:05
*** mmcardle has quit IRC12:09
*** ppai has quit IRC12:11
*** kopparam has joined #openstack-swift12:13
*** nexusz99 has joined #openstack-swift12:17
*** kopparam has quit IRC12:20
peluse_nexusz99:  StoragePolicySingleton wrapes _POLICIES to enable unit test code that imports POLICIES to patch them via the decorator @patch_policies.  Take a closer look there (or try running the unit tests without the wrapper) and see if that doesn't help with your understanding of why its needed12:34
nexusz99peluse_: Thanks!12:40
*** psharma has quit IRC12:40
*** bvandenh has joined #openstack-swift12:41
*** dmorita has quit IRC12:43
*** dmsimard_away is now known as dmsimard12:44
*** bszabo_ has quit IRC13:01
*** bszabo_ has joined #openstack-swift13:01
*** bvandenh has quit IRC13:02
*** pberis has joined #openstack-swift13:04
*** zslot has quit IRC13:05
*** zslot has joined #openstack-swift13:06
*** nexusz99 has quit IRC13:12
*** mmcardle has joined #openstack-swift13:25
*** kopparam has joined #openstack-swift13:28
*** kopparam has quit IRC13:31
peluse_nexusz99:  no problem, if you see an alternate solution or simplification please feel free to suggest, would be great to hear other ideas!13:34
acolesportante: did you have any success setting up keystone? let me know if you need more help13:37
portantenot yet13:37
*** mkollaro has quit IRC13:48
*** tdasilva has joined #openstack-swift13:49
pconstantinepeluse_: looks like PatchPolicies uses setUp(self) instead of setUpClass(cls), that's why it's patched in the wrong time13:55
*** kopparam has joined #openstack-swift13:56
*** chandankumar_ has joined #openstack-swift14:01
*** chandankumar has quit IRC14:04
*** chandankumar_ has quit IRC14:04
*** chandankumar has joined #openstack-swift14:06
*** zaitcev has joined #openstack-swift14:07
*** ChanServ sets mode: +v zaitcev14:07
*** kopparam has quit IRC14:13
*** kopparam has joined #openstack-swift14:14
*** kopparam has quit IRC14:18
*** bszabo_ has quit IRC14:23
*** bszabo_ has joined #openstack-swift14:24
openstackgerritOpenStack Proposal Bot proposed a change to openstack/swift: Updated from global requirements  https://review.openstack.org/8873614:25
*** bszabo_ has quit IRC14:28
*** chandankumar_ has joined #openstack-swift14:51
*** chandankumar has quit IRC14:55
*** nacim has quit IRC15:05
*** els13 has quit IRC15:17
*** kevinc_ has joined #openstack-swift15:21
*** bszabo_ has joined #openstack-swift15:24
*** bszabo_ has quit IRC15:25
peluse_clayg:  see note from pconstantine above wrt PatchPolicies ^^15:43
openstackgerritAlistair Coles proposed a change to openstack/swift: Enable object system metadata on PUTs  https://review.openstack.org/7999115:44
pconstantinetorgomatic, creiht, portante: friendly ping about https://review.openstack.org/#/c/72157/15:48
*** marcusvrn has quit IRC15:51
*** mwstorer has joined #openstack-swift15:54
notmynamepeluse_: FWIW, clayg is out today and until the end of next week on vacation15:55
notmynamegood morning, world15:55
openstackgerritAlistair Coles proposed a change to openstack/python-swiftclient: Add keystone v3 auth support  https://review.openstack.org/9178815:56
peluse_notmyname:  OK, thanks.  pconstantine, the comment you are referring to (I assume you mean the one about patching at the wrong time) was claygs so he'd be the best one to address.  I'll ping him about it next week if I don't see you ask later on...15:56
notmynamepeluse_: pconstantine: where is that call?15:57
pconstantinenotmyname: which one?15:57
notmynamePatchPolicies15:58
pconstantinenotmyname: test.unit.__init__.pt15:58
pconstantinenotmyname: test.unit.__init__.py15:58
*** stevemac has joined #openstack-swift15:58
*** gyee has joined #openstack-swift15:58
peluse_in the comment block for the class definition is what I had assumed15:59
*** stevemac is now known as stevetan15:59
pconstantinenotmyname: https://github.com/openstack/swift/blob/master/test/unit/__init__.py#L6716:00
peluse_and clayg spent a decent amount of time getting policy patching to work, correctly reload from .conf versus per unit test patching/resotring so there could very well be more here than meets the eye...16:00
notmynamepconstantine: I feel like I missed something. you said "that's why it's patched at the wrong time". did you see an error?16:00
pconstantinenotmyname: read the comments in the PatchPolicy class :)16:01
pconstantinePatchPolicies16:01
notmynamepconstantine: ah ok. you're jsut answering the question in the docstring16:01
pconstantinenotmyname: yep, there was some sort of discussion earlier on why the patch_policies decorator exists16:02
notmynamegotcha. do the tests still pass when you change it to setUpClass?16:02
pconstantinenotmyname: I think they should, but it should be simplified more, not just changing the setUp into stUpClass16:04
peluse_pconstantine:  that code does take some serious pythonese to follow, any simplication that preserves the functionality would be fantastic (pathced policies override swift.conf and individual unit tests patched policies revert to what they were before the test, policies can be reloaded, etc)16:12
*** chandankumar_ has quit IRC16:13
pconstantinepeluse_ notmyname: using something like this: https://gist.github.com/pkit/7e80f0b749468ad869c416:18
pconstantinepeluse_ notmyname: tests pass16:19
notmyname:-)16:19
notmynameI need to step out for a bit. thanks, pconstantine16:19
pconstantinenotmyname: but that's just a simple 2 minute change, probably can be simplified much more16:20
*** zslot has quit IRC16:25
*** jdag has joined #openstack-swift16:27
*** mmcardle has quit IRC16:44
torgomaticpconstantine: do you remember what creiht's issue was with cross account copy and security? I forget, but i remember it seemed important.16:45
peluse_pconstantine:  nice... sounds like the start of a nice patch16:46
*** shri has joined #openstack-swift16:46
pconstantinepconstantine: it was just issue with 'AUTH_foo' being not the same as tenant name16:46
pconstantinetorgomatic: which means that the header should be `X-Copy-From-Account: AUTH_test`16:47
torgomaticso if X-Copy-From-Account matches the actual Swift account name, there's no security issue?16:48
pconstantinetorgomatic: it just denies response if it's not account name16:49
torgomaticThat seems reasonable enough.16:49
*** stevetan has quit IRC16:49
pconstantinetorgomatic: the main security issue, that was fixed in much earlier patch sets is that __call__() removes authorize() function too early16:49
*** stevemac has joined #openstack-swift16:50
*** stevemac is now known as stevetan16:50
pconstantinetorgomatic: then creiht said that he get access denied to anything16:50
pconstantinetorgomatic: and then I found out that he used the tenant name and not account name16:50
torgomaticpconstantine: got it. Okay, I'll give that one a look once I get off this train and into my office.16:51
pconstantinetorgomatic: cool, it's documented in the comments there :)16:51
*** marcusvrn has joined #openstack-swift16:59
*** mmcardle has joined #openstack-swift17:10
*** pberis has quit IRC17:14
*** stevetan has quit IRC17:15
*** pberis has joined #openstack-swift17:15
openstackgerritguang-yee proposed a change to openstack/swift: Use auth_token from keystonemiddleware  https://review.openstack.org/10904717:28
*** mmcardle has quit IRC17:29
*** tongli has joined #openstack-swift17:34
*** acoles is now known as acoles_away17:35
* notmyname is back17:46
notmynamereminder that there is a swift team meeting in a little over an hour in #openstack-meeting17:47
notmynameanyone familiar with the keystone repo changes? https://review.openstack.org/#/c/109047/17:55
notmynamefor others looking for it: https://github.com/openstack/keystonemiddleware17:58
openstackgerritSamuel Merritt proposed a change to openstack/swift: Use auth_token from keystonemiddleware  https://review.openstack.org/10904717:59
notmynamewait, keystonemiddleware still requires python-keystoneclient?17:59
notmynameI guess packaging one repo was too simple. now we call all package two of them18:00
portantepconstantine: ack, will look at it after the V3 keystone review that I have open18:02
zaitcevmadness18:06
zaitcevI wonder if all the naysayers were right about OpenStack's amateurism.18:06
zaitcevKeystone is not even the worst of our projects, it's just insular and must be used by every other project in an install.18:07
zaitcevNot that I am likely to do better in place of Dolph, Adam, et al18:08
*** portante has quit IRC18:10
*** portante has joined #openstack-swift18:13
*** ChanServ sets mode: +v portante18:13
pconstantinezaitcev: what's keystone? never heard of it :)18:16
notmynamepconstantine: bad beer18:16
zaitcevReally? Wow.18:17
pconstantinezaitcev: nah, I'm just always in denial of its existence18:18
*** tdasilva_ has joined #openstack-swift18:18
peluse_hey keystone light is pretty good!  when to school in Colorado and that was the beer of choice for years.. until we could afford Coors Light :)18:18
*** lpabon has quit IRC18:18
zaitcevI mean that the beer connection goes further back than PBR18:18
pconstantinepeluse_: US beers are all bad, dunno why18:19
zaitcevIt's decided. I'm starting a storage orchestration project "Baltika".18:19
notmynamepconstantine: that isn't true. but the good ones aren't normally made by the major companies (but there are some exceptions)18:19
pconstantinepeluse_: there were a lot of Germans somewhere South, why didn't they bring their good beer with them?18:19
peluse_pconstantine:  because you're not drinking them in large enough volume I suspect :)18:19
peluse_pconstantine:  but I did live in Germany for a while and must say.. yeah, no arugment there!18:20
*** portante has quit IRC18:20
*** tdasilva has quit IRC18:21
*** lpabon has joined #openstack-swift18:23
*** portante has joined #openstack-swift18:24
*** ChanServ sets mode: +v portante18:24
*** mrmoje has quit IRC18:34
*** gvernik has joined #openstack-swift18:48
notmynamemeeting time19:00
*** acoles_away is now known as acoles19:01
*** tsg has joined #openstack-swift19:06
*** fausto has joined #openstack-swift19:33
*** NM has joined #openstack-swift19:53
NMHi everyone :)20:00
notmynameI feel the meeting are going more smoothly these days. thanks everyone for coming and participating20:00
NMDoes anyone knows if swift supports keystone v3? I was looking for informations and just found this: https://blueprints.launchpad.net/swift/+spec/keystone-v3-support20:00
notmynameNM: see the bullets under Keystone v3 on https://wiki.openstack.org/wiki/Swift/PriorityReviews20:01
notmynameNM: short answer: we're working on it20:01
mattoliverauk, I'm going to back to bed. Cya all in a few hours.20:01
*** gvernik has quit IRC20:05
NMnotmyname: thank you so much. I'm really enjoying this project and I'm trying to learn how to follow all good work you're doing.20:05
notmynameNM: it's a whole-team effort. I'm glad you're looking in to things. please feel free to ask when you have questions20:06
*** kevinc_ has quit IRC20:14
*** kevinc_ has joined #openstack-swift20:18
*** acoles is now known as acoles_away20:24
*** dmsimard is now known as dmsimard_away20:28
*** kevinc_ has quit IRC20:29
-openstackstatus- NOTICE: nodepool is unable to build test nodes so check and gate tests are delayed20:42
*** ChanServ changes topic to "nodepool is unable to build test nodes so check and gate tests are delayed"20:42
*** kevinc_ has joined #openstack-swift21:09
*** tdasilva_ has quit IRC21:21
*** fausto has quit IRC21:28
*** tongli has quit IRC21:35
*** kevinc_ has quit IRC22:02
*** NM has quit IRC22:02
-openstackstatus- NOTICE: zuul is working through a backlog of jobs due to an earlier problem with nodepool22:07
*** ChanServ changes topic to "zuul is working through a backlog of jobs due to an earlier problem with nodepool"22:07
mattoliverauMorning all22:08
*** kevinc_ has joined #openstack-swift22:08
*** shakamunyi has joined #openstack-swift22:10
*** shakamunyi has quit IRC22:15
peluse_mattoliverau:  didn't you just go to bed?  :)22:26
*** petek has joined #openstack-swift22:28
mattoliveraupeluse_: yes and no, I just went back to bed.. but now am up again.. yay its morning already :)22:31
*** shakamunyi has joined #openstack-swift22:33
mattoliverauhere's a question, in the abandon search filter I'm adding to the dash, do we want WIP changes to appear if they have also been untouched for 4 weeks and have a -1 score. Or should WIP be omitted from the search?23:37
peluse_my take would be WIP should not be included as they're set conciously by the owner23:41
mattoliveraupeluse_: I was leading that way too, but at what point are WIP patches abandoned.. I guess I could just to a periodic autid of WIP patches (as they are already in the dashboard) and if they get too old poke owners.23:43
mattoliverauand by audit, I might automate some report :P23:45
*** shakamunyi has quit IRC23:56
*** flwang has joined #openstack-swift23:56
*** tusharsg has joined #openstack-swift23:59
*** tsg has quit IRC23:59

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!