Wednesday, 2014-07-30

ayoungjamielennox, no, but even removing the session from that call breaks00:04
jamielennoxi was just surprised they done that00:04
*** gabriel-bezerra has quit IRC00:04
jamielennoxi had some WIP stuff that i had held off on cause there was a lot of tests changes00:05
jamielennoxnope - no need for that nonsense00:05
*** gabriel-bezerra has joined #openstack-keystone00:05
ayoungjamielennox, for v3, how do we set domain for the project?00:07
ayoungIts close to working for V3, just the token request does not put a domain in there00:07
ayoungRESP BODY: {"error": {"message": "Expecting to find domain in project. The server could not comply with the request since it is either malformed or otherwise incorrect. The client is assumed to be in error.", "code": 400, "title": "Bad Request"}}00:07
jamielennoxuser domain, or scope to a domain00:09
ayoungyeah, Its all there.  I'm Ok.  Just need to dig through it.00:11
*** ayoung has quit IRC00:16
*** gokrokve has quit IRC00:21
*** gokrokve has joined #openstack-keystone00:21
*** gokrokve has quit IRC00:25
openstackgerritA change was merged to openstack/keystone: Update middleware that was moved to keystonemiddleware  https://review.openstack.org/10647800:37
openstackgerritA change was merged to openstack/keystone: Check url is in the 'self' link in list responses  https://review.openstack.org/10929000:37
*** mitz has quit IRC00:43
*** mitz has joined #openstack-keystone00:48
*** amcrn has quit IRC00:52
*** dhellmann is now known as dhellmann_00:53
openstackgerritAngus Lees proposed a change to openstack/keystone: Issue multiple SQL statements in separate engine.execute() calls  https://review.openstack.org/11051200:53
*** zzzeek has quit IRC00:56
*** diegows has quit IRC00:57
*** mitz has quit IRC01:00
*** mitz has joined #openstack-keystone01:03
*** mitz has quit IRC01:04
*** mitz has joined #openstack-keystone01:06
*** griff is now known as jgriffith01:11
*** mitz has quit IRC01:17
*** rwsu has quit IRC01:18
*** mrmoje has quit IRC01:23
*** marcoemorais has quit IRC01:27
*** mberlin1 has quit IRC01:40
*** mberlin has joined #openstack-keystone01:41
*** xianghui has joined #openstack-keystone01:49
openstackgerritA change was merged to openstack/python-keystoneclient: Add an example of using v3 client with sessions  https://review.openstack.org/10883901:49
*** bknudson has quit IRC01:59
morganfainbergdolphm, ping, want to say this is a bug https://review.openstack.org/#/c/103493/6/specs/juno/del-tokens-when-del-ec2-credential.rst not a spec02:19
*** ayoung has joined #openstack-keystone02:47
*** harlowja is now known as harlowja_away02:57
*** gabriel-bezerra has quit IRC03:00
*** gabriel-bezerra has joined #openstack-keystone03:01
*** jamielennox is now known as jamielennox|away03:05
*** hrybacki has joined #openstack-keystone03:16
*** hrybacki has quit IRC03:16
*** morganfainberg is now known as morganfainberg_Z03:20
*** morganfainberg_Z is now known as morganfainberg03:22
*** gyee has quit IRC03:23
*** topol has joined #openstack-keystone03:32
*** gabriel-bezerra has quit IRC03:41
*** gabriel-bezerra has joined #openstack-keystone03:42
*** nkinder_away has quit IRC04:16
*** nkinder_away has joined #openstack-keystone04:18
*** gabriel-bezerra has quit IRC04:27
*** gabriel-bezerra has joined #openstack-keystone04:28
*** flwang has quit IRC04:29
*** ajayaa has joined #openstack-keystone04:56
*** morganfainberg is now known as morganfainberg_Z04:59
*** topol has quit IRC05:06
*** jaosorior has joined #openstack-keystone05:12
openstackgerritSteve Martinelli proposed a change to openstack/keystone: PoC - Transform a Keystone token to a SAML assertion  https://review.openstack.org/11054205:32
stevemarmarekd|away,  https://review.openstack.org/11054205:32
openstackgerritSteve Martinelli proposed a change to openstack/keystone: PoC - Transform a Keystone token to a SAML assertion  https://review.openstack.org/11054205:36
*** gabriel-bezerra has quit IRC05:38
*** gabriel-bezerra has joined #openstack-keystone05:39
*** chandankumar has joined #openstack-keystone05:44
*** tomoiaga has joined #openstack-keystone05:51
*** jgriffit1 has joined #openstack-keystone06:01
openstackgerritOpenStack Proposal Bot proposed a change to openstack/keystone: Imported Translations from Transifex  https://review.openstack.org/10693906:04
*** jgriffith has quit IRC06:05
*** henrynash has joined #openstack-keystone06:07
*** gabriel-bezerra has quit IRC06:12
*** gabriel-bezerra has joined #openstack-keystone06:12
*** fifieldt_ is now known as fifieldt06:15
*** stevemar has quit IRC06:15
*** henrynash has quit IRC06:16
*** fifieldt has quit IRC06:19
*** YorikSar has quit IRC06:20
*** YorikSar has joined #openstack-keystone06:22
*** jgriffit1 is now known as jgriffith06:24
*** bvandenh has joined #openstack-keystone06:38
*** ukalifon1 has joined #openstack-keystone06:38
*** henrynash has joined #openstack-keystone06:57
*** henrynash has quit IRC07:08
*** marekd|away is now known as marekd07:10
openstackgerritA change was merged to openstack/keystone: Updated from global requirements  https://review.openstack.org/10900207:21
*** fifieldt has joined #openstack-keystone07:23
*** henrynash has joined #openstack-keystone07:30
*** henrynash has quit IRC08:12
*** mrmoje has joined #openstack-keystone08:14
*** henrynash has joined #openstack-keystone08:14
*** henrynash has quit IRC08:19
*** YorikSar has quit IRC08:44
*** YorikSar has joined #openstack-keystone08:46
*** henrynash has joined #openstack-keystone08:53
*** henrynash has quit IRC09:00
*** marzif has joined #openstack-keystone09:01
*** afazekas has joined #openstack-keystone09:03
openstackgerritAjaya Agrawal proposed a change to openstack/keystone: Implemented caching in identity layer.  https://review.openstack.org/11057509:06
*** henrynash has joined #openstack-keystone09:08
*** henrynash has quit IRC09:08
*** openstackgerrit has quit IRC09:16
*** openstackgerrit has joined #openstack-keystone09:17
*** ByteSore_ has joined #openstack-keystone09:43
*** ajayaa has quit IRC09:44
*** asmacdo has quit IRC09:44
*** afaranha has quit IRC09:44
*** raildo has quit IRC09:44
*** vhoward has quit IRC09:44
*** ByteSore has quit IRC09:44
*** asmacdo has joined #openstack-keystone09:44
*** vhoward has joined #openstack-keystone09:44
*** afaranha has joined #openstack-keystone09:45
*** raildo has joined #openstack-keystone09:46
*** ajayaa has joined #openstack-keystone09:59
*** ajayaa has quit IRC10:12
jaosoriorDoes anybody know what the status of this is https://bugs.launchpad.net/keystone/+bug/1211582 ? it's been there for a while O_o10:38
uvirtbotLaunchpad bug 1211582 in keystone "Filter user list by partial attributes" [Wishlist,New]10:38
*** ajayaa has joined #openstack-keystone10:51
*** ajayaa has quit IRC11:00
*** ajayaa has joined #openstack-keystone11:01
*** henrynash has joined #openstack-keystone11:18
*** mitz has joined #openstack-keystone11:27
*** mitz has quit IRC11:28
*** mitz has joined #openstack-keystone11:30
*** mitz has quit IRC11:31
*** mitz has joined #openstack-keystone11:33
bjornarI have a problem with keystone not able to find keystone.contrib.revoke.backends.sql11:51
*** diegows has joined #openstack-keystone11:53
bjornarfixed11:55
*** xianghui has quit IRC12:01
ajayaaayoung, hi. what does the test "keystone.tests.test_backend_ldap_pool.LdapIdentitySqlAssignment.test_utf8_encoded_is_used_in_pool" deal with?12:05
openstackgerritStuart McLaren proposed a change to openstack/keystonemiddleware: Add composite auth support (service token)  https://review.openstack.org/10838412:06
ayoungajayaa, I'd have to look.  THat was from the pooling patch that just went in.  Your best bet is to find the author of the patch and ask him directly12:11
ayoungUse git blame on that file to find the commit hash, and git show to see the author12:11
*** ayoung is now known as ayoung-afk12:11
openstackgerritStuart McLaren proposed a change to openstack/keystonemiddleware: Add composite auth support (service token)  https://review.openstack.org/10838412:12
ajayaaarunkant, git blames you for "keystone.tests.test_backend_ldap_pool.LdapIdentitySqlAssignment.test_utf8_encoded_is_used_in_pool"12:14
*** cjellick has joined #openstack-keystone12:24
openstackgerritAbhishek Kekane proposed a change to openstack/keystone: Keystone service throws error on SIGHUP signal  https://review.openstack.org/10748212:36
*** henrynash has quit IRC12:47
*** hrybacki has joined #openstack-keystone12:50
*** bvandenh has quit IRC13:02
*** gordc has joined #openstack-keystone13:02
*** joesavak has joined #openstack-keystone13:08
*** jasondotstar has joined #openstack-keystone13:09
*** bvandenh has joined #openstack-keystone13:17
chmoueli was wondering if pkiz should be added to devstack13:17
*** lbragstad has joined #openstack-keystone13:19
*** lbragstad has quit IRC13:20
*** lbragstad has joined #openstack-keystone13:20
*** bknudson has joined #openstack-keystone13:32
*** afazekas has quit IRC13:35
*** bknudson has quit IRC13:36
*** jdennis1 has quit IRC13:43
*** stevemar has joined #openstack-keystone13:45
*** afazekas has joined #openstack-keystone13:49
openstackgerritLance Bragstad proposed a change to openstack/keystone: Implement validation on the Catalog V3 resources  https://review.openstack.org/9626613:58
openstackgerritHarry Rybacki proposed a change to openstack/keystonemiddleware: Convert auth_token middleware to use sessions  https://review.openstack.org/10503114:00
*** gabriel-bezerra has quit IRC14:04
*** gabriel-bezerra has joined #openstack-keystone14:04
*** htruta has quit IRC14:06
openstackgerritA change was merged to openstack/keystone: KeyError instead of exception.KeyError  https://review.openstack.org/11039714:08
*** henrynash has joined #openstack-keystone14:15
*** gabriel-bezerra has quit IRC14:19
*** gabriel-bezerra has joined #openstack-keystone14:20
*** afazekas has quit IRC14:20
openstackgerritSteve Martinelli proposed a change to openstack/keystone-specs: Rename contents to template  https://review.openstack.org/11065814:22
openstackgerritAlexey Miroshkin proposed a change to openstack/keystone: Add filters to the collections 'self' link  https://review.openstack.org/11066114:26
*** afazekas has joined #openstack-keystone14:33
openstackgerritSteve Martinelli proposed a change to openstack/keystone-specs: Rename contents to template  https://review.openstack.org/11065814:39
stevemardstanek, ping?14:40
*** bvandenh has quit IRC14:43
*** gabriel-bezerra has quit IRC14:45
*** gabriel-bezerra has joined #openstack-keystone14:45
*** hrybacki has quit IRC14:46
openstackgerritRodrigo Duarte proposed a change to openstack/keystone: Hierarchical Projects  https://review.openstack.org/10884114:48
*** chandankumar_ has joined #openstack-keystone14:50
*** lbragstad has quit IRC14:52
*** thedodd has joined #openstack-keystone14:52
*** lbragstad has joined #openstack-keystone14:52
stevemarso many AFK's today14:52
stevemareven bknudson is out14:53
*** chandankumar has quit IRC14:53
marekdstevemar: i ma here14:53
stevemarmarekd, brant is offline, must be recharging14:54
marekdstevemar:  LOL14:54
marekdstevemar: sorry i didn't have time to see your patch, i am finishing the adfs stuff :(14:54
stevemarmarekd, np14:54
marekdBTW, do you know how to split patches?14:54
stevemari believe ayoung had a post about that14:55
marekdThere is A and B, B depends on A. I need to abandon A.14:55
marekdok i will take a look at it later.14:55
*** lbragstad has quit IRC14:57
*** lbragstad has joined #openstack-keystone14:57
*** rwsu has joined #openstack-keystone14:58
stevemarthen just abandon A and rebase B on master?14:58
*** jdennis has joined #openstack-keystone14:59
marekdhah, that might even work :-)14:59
marekdthanks.14:59
ajayaamarekd: you could use git add -p14:59
marekdajayaa: i am always using add -p15:00
ajayaaokay. Probably I missed the context. :)15:01
*** ajayaa has quit IRC15:04
*** vhoward has left #openstack-keystone15:09
jaosoriorlbragstad, you around?15:13
stevemarhenrynash, ping?15:13
lbragstadjaosorior: yep15:14
henrynashstevemar: hi15:14
stevemarhenrynash, could you take a very quick look at: https://review.openstack.org/#/c/110658/15:14
*** david-lyle has joined #openstack-keystone15:14
henrynashstevemar: will do15:14
openstackgerrithenry-nash proposed a change to openstack/keystone-specs: Enable filtering of credentials by user ID.  https://review.openstack.org/11067415:14
stevemarhenrynash, I did some work to setup specs.openstack.org/openstack/keystone-specs (to actually publish them)15:14
*** tomoiaga has quit IRC15:14
stevemarbut theres no data, i think we just need to push 1 patch through15:15
henrynashstevemar: OK!15:15
jaosoriorcool, well, regarding the bug https://bugs.launchpad.net/keystone/+bug/1350273  , there is already this: https://bugs.launchpad.net/keystone/+bug/1211582 which would help with filtering stuff from the extra attributes in those models, but it's been there for a while and I'm not sure if it15:16
uvirtbotLaunchpad bug 1350273 in keystone "Filtering services by name doesn't work" [Undecided,New]15:16
jaosoriorit15:16
jaosoriorit's actually being worked on15:16
*** zzzeek has joined #openstack-keystone15:16
*** afazekas has quit IRC15:17
openstackgerritA change was merged to openstack/keystone-specs: Rename contents to template  https://review.openstack.org/11065815:17
lbragstadjaosorior: it's been quiet for a while15:18
openstackgerrithenry-nash proposed a change to openstack/keystone-specs: Enable filtering of credentials by user ID.  https://review.openstack.org/11067415:18
henrynasha quickie spec: https://review.openstack.org/#/c/110674/15:20
jaosoriorany way to contact eugeniya? I didn't take it over because it didn't seem polite somehow O_o15:20
navid_marekd: hi15:20
*** gokrokve has joined #openstack-keystone15:20
navid_marekd: do you have time, questions?15:21
jaosoriorlbragstad: But before that, I think it would actually make sense to add that field to the sql model, and make a migration15:21
*** david-lyle has quit IRC15:22
*** henrynash has quit IRC15:22
*** david-lyle has joined #openstack-keystone15:23
marekdnavid_: hey!15:24
marekdnavid_: what's up?15:24
*** vhoward has joined #openstack-keystone15:24
*** henrynash has joined #openstack-keystone15:25
*** henrynash has quit IRC15:25
navid_marekd: I wanted to know from the resources you send me, the service provider that you were helping to setup, is in what stage.15:25
marekdnavid_: that day Carlos set up Identity Provider, not Service Provider.15:25
marekdi have SPs for a long time :-)15:26
navid_I know , I thought you are working on the service provider15:26
marekdnavid_: what do you mean?15:26
marekdi need both15:26
lbragstadjaosorior: morganfainberg_Z might have an opinion on this come to think of it15:26
stevemaryay http://specs.openstack.org/openstack/keystone-specs/doc/build/html/15:26
navid_marekd: Farhan asked me that, give me couple of minutes.15:27
marekdnavid_: ah i think what you are asking about....so I didn't configure that SP that was meant to be configured with IdP@UTSA.15:28
marekdis it what you are asaking about?15:28
*** bknudson has joined #openstack-keystone15:28
*** gyee has joined #openstack-keystone15:28
*** tomoiaga has joined #openstack-keystone15:29
*** tomoiaga has quit IRC15:32
*** Farhan has joined #openstack-keystone15:33
*** ukalifon1 has quit IRC15:36
*** chandankumar_ has quit IRC15:45
*** mrmoje has quit IRC15:47
*** gyee has quit IRC15:51
*** gokrokve_ has joined #openstack-keystone15:51
*** gokrokve has quit IRC15:55
*** gokrokve_ has quit IRC15:55
*** marcoemorais has joined #openstack-keystone16:00
openstackgerritRodrigo Duarte proposed a change to openstack/keystone: Hierarchical Projects  https://review.openstack.org/10884116:02
*** hrybacki has joined #openstack-keystone16:07
dstanekdolphm: ping16:12
*** thedodd has quit IRC16:17
*** thedodd has joined #openstack-keystone16:19
*** jasondotstar has quit IRC16:22
openstackgerritDavid Chadwick proposed a change to openstack/keystone-specs: Specification for IETF ABFAB federation  https://review.openstack.org/10863116:24
openstackgerritStuart McLaren proposed a change to openstack/keystonemiddleware: Add composite auth support (service token)  https://review.openstack.org/10838416:33
*** marcoemorais has quit IRC16:34
*** marcoemorais has joined #openstack-keystone16:35
openstackgerritLance Bragstad proposed a change to openstack/keystone: Make BaseValidationTestCase  https://review.openstack.org/10909816:35
*** marcoemorais has quit IRC16:35
*** marcoemorais has joined #openstack-keystone16:35
*** marcoemorais has quit IRC16:35
*** marcoemorais has joined #openstack-keystone16:36
*** marcoemorais1 has joined #openstack-keystone16:36
*** marcoemorais1 has quit IRC16:37
*** marcoemorais1 has joined #openstack-keystone16:37
*** marcoemorais1 has quit IRC16:37
*** marcoemorais1 has joined #openstack-keystone16:37
*** jasondotstar has joined #openstack-keystone16:40
*** marcoemorais has quit IRC16:40
*** jasondotstar has quit IRC16:40
*** gokrokve has joined #openstack-keystone16:40
*** jasondotstar has joined #openstack-keystone16:41
*** jasondotstar has quit IRC16:41
*** rodrigods has quit IRC16:43
*** jasondotstar has joined #openstack-keystone16:43
*** joesavak has quit IRC16:44
*** rodrigods has joined #openstack-keystone16:47
*** rodrigods has joined #openstack-keystone16:47
*** rodrigods has quit IRC16:47
*** peluse_ has quit IRC16:55
*** rodrigods has joined #openstack-keystone16:56
*** rodrigods has joined #openstack-keystone16:56
*** hrybacki has quit IRC17:00
*** marcoemorais1 has quit IRC17:00
*** marcoemorais has joined #openstack-keystone17:01
*** henrynash has joined #openstack-keystone17:01
*** marcoemorais has quit IRC17:02
*** marcoemorais has joined #openstack-keystone17:03
henrynashdolphm, ayoung: a micro-spec I’d like to get into Juno-3: https://review.openstack.org/#/c/110674/17:05
dolphmhenrynash: looking17:05
*** markwash has quit IRC17:08
*** morganfainberg_Z is now known as morganfainberg17:09
morganfainbergchmouel, pkiz added to devstack? the default right now is to use pkiz in devstack17:10
chmouelmorganfainberg: ah i guess that because of my config, thanks!17:10
morganfainbergchmouel, sure thing17:10
*** hrybacki has joined #openstack-keystone17:12
hrybackiayoung-afk:  https://review.openstack.org/#/c/105031/ is good to go17:13
hrybackijamielennox|away: ^^17:13
*** bvandenh has joined #openstack-keystone17:14
*** jasondotstar is now known as jasondotstar|afk17:26
*** david-lyle has quit IRC17:30
*** david-lyle has joined #openstack-keystone17:30
*** gabriel-bezerra has quit IRC17:30
*** gabriel-bezerra has joined #openstack-keystone17:31
openstackgerritRaildo Mascena de Sousa Filho proposed a change to openstack/keystone-specs: Hierarchical Multitenacy  https://review.openstack.org/10101717:32
openstackgerritHarry Rybacki proposed a change to openstack/keystonemiddleware: Convert auth_token middleware to use sessions  https://review.openstack.org/10503117:33
*** david-lyle has quit IRC17:35
openstackgerritA change was merged to openstack/keystone: Use config fixture from oslo.config  https://review.openstack.org/10325417:35
*** david-lyle has joined #openstack-keystone17:36
raildomorganfainberg: dolphm henrynash  I created a vm with the hierarchical multitenancy code . If you want to test just run the command: ssh stack@ssh.cloud.lsd.ufcg.edu.br -p 1002217:36
raildopassword: stack17:36
raildoand you can see the API here: https://wiki.openstack.org/wiki/HierarchicalMultitenancy_API17:37
morganfainbergraildo, FYI, i wouldn't put passwords in an open IRC channel (at least not for anything that'll last longer than a few minutes)17:37
morganfainbergraildo, this channel is logged as well btw.17:37
raildomorganfainberg: sorry17:37
morganfainbergraildo, i'm just worreid someone will do something malicious (not us here, someone else) with that VM17:38
morganfainbergraildo, no need to apologize17:38
raildoBut as this password is only access to this VM, we have no problem,17:38
*** chandankumar_ has joined #openstack-keystone17:38
raildoI plan to remove it later17:39
raildobut thanks for the tip :)17:39
morganfainbergraildo, right, but someone could use it to spam, or attempt hacking some other site, proxy, etc17:39
morganfainbergraildo, np. as long as you don't leave it running too long, should be fine17:39
raildook17:39
*** harlowja_away is now known as harlowja17:40
openstackgerritMorgan Fainberg proposed a change to openstack/keystone: Set default token provider to UUID  https://review.openstack.org/11048817:41
*** markwash has joined #openstack-keystone17:42
*** ajayaa has joined #openstack-keystone17:43
ajayaamorganfainberg, hi!17:44
*** packet has joined #openstack-keystone17:44
*** ajayaa has quit IRC17:48
*** morganfainberg is now known as morganfainberg_Z17:51
*** gabriel-bezerra has quit IRC18:00
*** gabriel-bezerra has joined #openstack-keystone18:01
*** packet has quit IRC18:01
*** gyee has joined #openstack-keystone18:04
*** gyee has quit IRC18:05
*** gyee has joined #openstack-keystone18:05
*** packet has joined #openstack-keystone18:06
*** ayoung-afk is now known as ayoung18:09
ayoungdolphm, OK, squabbling aside, what is the plan for PKI tokens?  Are we just going to disable them, or are we going to get rid of them?18:09
ayoungthere are two issues we know about right now.  The first is Horizon, and the second is the increase in request size.  Are either of those insurmountable?18:12
*** bvandenh has quit IRC18:13
*** gyee has quit IRC18:18
*** gyee has joined #openstack-keystone18:19
*** joesavak has joined #openstack-keystone18:20
*** jsavak has joined #openstack-keystone18:22
*** joesavak has quit IRC18:26
*** bvandenh has joined #openstack-keystone18:28
stevemarayoung, i don't think they are going away any time soon18:29
*** jsavak has quit IRC18:29
ayoungstevemar, we went from discussing it to a patch submitted to turn off PKI tokens.  Without them enabled, we would never have discovered the problems with them in the first place18:31
ayoungstevemar, the Horizon thing got me thinking.18:32
ayoungThe only way to make it palatable for Horizon to use PKI tokens is memcached18:32
*** jasondotstar|afk is now known as jasondotstar18:33
ayoungand the only thing that made UUID tokens palatable was memcached18:33
ayoungFor some reason it is "optional" in Horizon,  but required in All of the other services that use auth-token middleware18:33
*** chandankumar_ has quit IRC18:33
ayoungwithout memcached, there would be a clear advantage to PKI tokens, but with it, the advantage is sortof on the side of UUID tokens.  However, I still think that PKI tokens were required to solve usse cases that UUID tokens ignored, like revocations18:34
ayoungUUID tokens were cached in memcached, so if they were revoked, the service would never find out18:35
ayoungwe also have the fact that none of the clients ever cached tokens, and always requested new ones, which means that caching UUID tokens was really kindof wasteful for anything but commo with Horizon18:35
ayoungor complex scripting not going through the clients18:36
ayoungstevemar, I'm just not sure it is that clean cut a case for or against18:36
stevemarayoung, for the record, i like PKI, and i like the advantages it provides (one less round trip)18:37
ayoungstevemar, if Horizon were a single page app, and all calls went from the browser to the services, most of the issues would go away18:37
*** marcoemorais has quit IRC18:37
stevemari'm not a fan of the usability (longggg token), but that's more or less hidden from the user18:37
ayoungstevemar, drop the catalog from the token and the size shrinks, but I don't think even a catalog-less token would be much below 1K18:38
*** marcoemorais has joined #openstack-keystone18:38
*** marcoemorais has quit IRC18:38
*** marcoemorais has joined #openstack-keystone18:39
gyeeayoung, you ain't got nothin on the "size" argument18:39
*** ajayaa has joined #openstack-keystone18:39
openstackgerritA change was merged to openstack/keystone: Remove duplicated asserts  https://review.openstack.org/10976018:39
*** marcoemorais has quit IRC18:39
*** marcoemorais has joined #openstack-keystone18:40
gyeeyou want size, or efficiency? :)18:40
*** marcoemorais has quit IRC18:40
ayounggyee, actually, I do18:40
openstackgerritSteve Martinelli proposed a change to openstack/keystone-specs: Remove an unused import from conf.py  https://review.openstack.org/11074918:40
*** marcoemorais has joined #openstack-keystone18:40
ayounggyee, lets talk swift for a moment18:40
ayoungin the case of swift, the real issue is that the size would be paid up front:18:40
ayounghere is what I mean by that:18:40
ayoungswift needs the token in some form, UUID or PKI18:41
ayoungif it is UUID, it needs to go and query the data from Keystone18:41
gyeeSwift is a bad example, Swift should be using signature access, just like AWS18:41
ayounggyee, the difference is the going to Keystone is a "local" call as opposed to the PKI token18:42
ayoungboth require the same bandwidth, but the PKI token has to go across the public internet18:42
ayoungthe UUID call does not18:42
ayoungnow, if there are multiple calls with the full PKI token payload, that is going to add up, with no benefit18:42
ayounggyee, define "signature access"18:43
gyeeayoung, ec2 & s318:43
ayoungI've never used Amazon18:43
ayoungAside to buy things18:43
notmynameayoung: shared secret that is used to HMAC sign a request18:43
notmynameHMAC generate client-side, then validated server side18:43
gyeeright, so sample and effective18:44
gyeeforget tokens man18:44
ayounggyee, that means that all requests need to be signed, either by a symmetric or asymmetric key18:44
gyeesymmetric18:44
gyeeHMAC is based on shared secret18:44
notmynamegyee: yes, but swift (server side) needs to get the shared secret from keystone. ideally in some cacheable form so every request doesn't still require a request to keystone18:45
notmynameayoung: it's basically what swift has today with the tempurl support.18:45
gyeenotmyname, sure, Keystone or Barbican, and cache it18:45
stevemargyee, ayoung can you guys push this patch through: https://review.openstack.org/#/c/110749/18:45
gyeebut that's a one time deal right?18:45
notmynamebut that is based on account metadata, not something living in keystone18:45
stevemarit's to update http://specs.openstack.org/openstack/keystone-specs/18:45
notmynamegyee: time-based, shared secret is stored in account metadata18:46
ayoungnotmyname, gyee a signature can't be much less than 1k In size,18:46
gyeeayoung, HMAC is pretty small18:46
notmynameayoung: generally it is hex of sha25618:46
ayoungthat is roughly comparable to a pki token without catalog18:46
ayoungnotmyname, do you not include the signer-data in that?18:46
notmynameayoung: what is "Signer-data"?18:47
ayoungso a user only has one key that they can use?18:47
gyeeonly "signer data" is the access key ID18:47
notmynameayoung: 2, in the case of swift's tempurls, so they can be rotated18:47
gyeewhich used to identify the secret18:47
ayoungnotmyname, in CMS (usedfor PKI tokens) it is the identifier of who signed the document18:47
ayoungnotmyname, it can't be a sha256.  It has to be a signed sha256, which is much larger.  about 512 bytes18:48
gyeeayoung, only signer is the "account holder"18:48
notmynameayoung: for tempurls, see https://github.com/openstack/swift/blob/master/bin/swift-temp-url18:48
ayoungso, sure, with HMAC< it would be smaller than token, and you would have a better degree of "I signed this" than you do for PKI tokens18:48
notmyname$ swift-temp-url GET 3600 /v1/AUTH_test/mycontainer/someobject/name/to/sign foobarkey18:49
notmyname/v1/AUTH_test/mycontainer/someobject/name/to/sign?temp_url_sig=fa81cf7f1927fe8087cae1ffea7dd2255c1cc5fb&temp_url_expires=140674977118:49
notmynamefor example18:49
gyeenotmyname, you only allow one key per container?18:51
notmynamegyee: 2 per account18:51
notmynamegyee: stored in account metadata. 2 keys so you can rotate it without worrying about prematurely expiring exising URLs out there18:52
gyeeoh, that explains why your example don't need access key ID in temp_url_sig18:52
ayoungok, lets punt on tempurls for the time being.  The issues with PKI tokens stand regardless of swift for talking to any service.  If the token goes across the wire more than once, its paying twice for no benefit18:52
notmynameayoung: right. tempurls are an example :-)18:52
notmynameayoung: if such were supported by keystone, swift still needs to get the shared secret from keystone. if it's not supported, then swift needs to validate the token18:53
notmynameayoung: and either way, it's cheaper to do nothing than something (ie caching is good and therefore tokens can't be one-time-use)18:53
gyeeayoung, yeah, if I am using a mobile app, I'd hate PKI tokens :)18:53
ayoungnotmyname, lets say i t needs to get the shared secret from somewhere.  Unless that method is going to replace all token usage, we need to figure something out18:54
ayounggyee, right.18:54
gyeeI do see mobile apps accessing Swift btw18:54
openstackgerritA change was merged to openstack/keystone-specs: Remove an unused import from conf.py  https://review.openstack.org/11074918:54
ayounggyee, yeah, but using keystone tokens or using tempurls?18:55
gyeeayoung, no idea, I don't have access to the code18:55
stevemarsweet! http://specs.openstack.org/openstack/keystone-specs/18:56
*** gabriel-bezerra has quit IRC18:56
*** gabriel-bezerra has joined #openstack-keystone18:57
gyeeayoung, I do think we can offer choices, based on application behavior18:58
gyeesomething like POST /v3/auth/tokens?format=PKI18:58
stevemari've been pitching that ?format for saml too :)18:59
gyeestevemar, amen brother!18:59
ayounggyee, that was why it was supposed to be /aut/tokens  vs....18:59
ayoungstevemar, you want /auth/saml18:59
ayoungand so /auth/pkiz18:59
ayoungand so /auth/uuid18:59
ayoungetc18:59
gyee++18:59
notmynameI've got a swift meeting now19:00
ayoungBut what is the default?19:00
ayoungnotmyname, bring up this question and let us know if there is fundamentally a problem with PKI tokens for swift, please19:00
stevemari guess it would be uuid, seems like thats what dolphinator wants19:01
openstackgerritLance Bragstad proposed a change to openstack/keystone: Make BaseValidationTestCase  https://review.openstack.org/10909819:01
lbragstadstevemar: addressed your comments19:01
stevemarthx dude19:01
notmynameayoung: I'm actually not sure what the question is19:01
lbragstadstevemar: we need the special encoding on that test module because we're testing unicode strings19:01
lbragstadwith goofy characters19:01
gyeegoofy is a character19:02
ayounggyee, the thing is, UUID tokens will end up requiring persistence in Keystone19:04
ayoungI would never have wasted the time on revocation events if we were going to keep persistence19:04
gyeeayoung, right, that's one of the tradeoffs19:05
gyeethat's really not much you can do on the client-side with an opaque string19:06
ayounggyee, does it make any real difference for an app to do a memcache call and to make an http call to keystone?19:07
ayoungif all of the apps end up caching the tokens in memcache...but then, they were doing that even with UUID tokens.  What is the difference there?19:08
gyeememcache is also client-server thingy, so yeah, there will be network overhead as well19:09
gyeebut if there's an instance running on the same host then it suppose to be fast19:10
gyeebut if you are running a memcache ring, then I suppose you have to deal with replication as well?19:10
ayounggyee,  I suspect that the way people are using memcache here is not in a ring19:11
ayoungso local call vs remote19:11
gyeeayoung, we are running a ring19:11
ayoungplus going to Keystone is 2 calls:  call to keystone, keystone to datastore19:11
gyeewe = HP19:11
ayounggyee, not on nova19:11
ayounggyee, the auth token caching on nova is in a ring?19:11
gyeeayoung, it depends on how many nova instances, for HA purposes19:12
ayounggyee, that sounds like a mistake19:12
gyeenova API instances I mean19:12
ayounggyee, why not just have each instance of the api server cache tokens when required?19:12
*** gabriel-bezerra has quit IRC19:12
gyeeayoung, for HP, LB can route the call to any instance19:13
gyees/HP/HA/19:13
*** gabriel-bezerra has joined #openstack-keystone19:13
gyeeif you have 50 instance, you don't want to hit Keystone 50 times19:14
openstackgerritMarek Denis proposed a change to openstack/python-keystoneclient: List federated projects and domains  https://review.openstack.org/10739319:23
*** markwash_ has joined #openstack-keystone19:23
marekdstevemar: lol, I am afraid I am gonna ask you for the 3rd +2 on this patch: https://review.openstack.org/#/c/107393 :-)19:24
marekdstevemar: i have abandoned underlying one after jamie's suggestions19:24
marekdstevemar: and I confirm the patch in the current shape also lists my projects/domains :-)19:24
raildogyee: I answered your comments there, I had forgotten to send: P19:24
*** markwash__ has joined #openstack-keystone19:26
*** markwash has quit IRC19:26
openstackgerritA change was merged to openstack/keystone-specs: Enable filtering of credentials by user ID.  https://review.openstack.org/11067419:28
*** markwash has joined #openstack-keystone19:29
*** markwash_ has quit IRC19:29
*** markwash__ has quit IRC19:32
*** bvandenh has quit IRC19:32
*** markwash_ has joined #openstack-keystone19:34
gyeeraildo, thanks, will take another look19:35
*** markwash has quit IRC19:37
*** packet has quit IRC19:37
*** markwash_ has quit IRC19:39
openstackgerritMarek Denis proposed a change to openstack/python-keystoneclient: List federated projects and domains  https://review.openstack.org/10739319:42
ayounggyee, no session affinity?19:47
stevemarmarekd, i don't think httpretty is included19:50
marekdstevemar: yeah19:50
marekdstevemar: i just noticed there is something definitely wrong.....19:51
marekdstevemar: even though the real code (not tests) work pretty well.19:51
marekdbut even fresh pull from master fails for me.19:51
marekdstevemar: what's the solution for that?19:53
marekdstevemar: ah, I didn't rebase properly...:/\19:53
*** markwash_ has joined #openstack-keystone19:53
stevemarmarekd, that would be the problem19:54
ayoungmorganfainberg_Z, let me know when you wake up.  I don't want to sit on this PKI token issue for ever, but I want to make sure we all understand it before moving forward19:56
*** ajayaa has quit IRC19:59
marekdstevemar: eh, it was not me *not rebasing* properly but not fixing the tests where Python was not very helpful in showing where is the problem (have you ever expected the big output log with set of classes and no helpful info what is screwed up?)20:04
marekdrunning full set of tests and uploading new ver.20:04
openstackgerritMarek Denis proposed a change to openstack/python-keystoneclient: List federated projects and domains  https://review.openstack.org/10739320:05
stevemarmarekd, ill check it out when i get back, heading out soon20:05
marekdstevemar: sure!20:05
marekdstevemar: have fun.20:05
openstackgerritBrant Knudson proposed a change to openstack/python-keystoneclient: Redact tokens in request headers  https://review.openstack.org/11011720:12
*** stevemar has quit IRC20:12
openstackgerritMarek Denis proposed a change to openstack/python-keystoneclient: Add v3scopedsaml entry to the setup.cfg.  https://review.openstack.org/11077020:13
marekdjamielennox|away: stevemar: appreciate your eyes on this patch: https://review.openstack.org/#/c/106751/20:17
marekdit's been stuck for a while without a  review :(20:18
*** markwash_ is now known as markwash20:18
*** amerine_ is now known as amerine20:25
*** hemna has joined #openstack-keystone20:30
*** andreaf has quit IRC20:34
openstackgerritDiane Fleming proposed a change to openstack/identity-api: Remove ATOM responses for list versions and get version details.  https://review.openstack.org/11077720:35
*** navid_ has quit IRC20:36
*** cjellick_ has joined #openstack-keystone20:38
*** cjellick_ has quit IRC20:41
*** cjellick_ has joined #openstack-keystone20:41
*** cjellick has quit IRC20:42
openstackgerrithenry-nash proposed a change to openstack/identity-api: Enable filtering of credentials by user ID  https://review.openstack.org/11078220:44
*** cjellick_ has quit IRC20:46
openstackgerritMarek Denis proposed a change to openstack/python-keystoneclient: SAML2 wrapper plugin for full federation authN  https://review.openstack.org/10675120:48
*** marekd is now known as marekd|away20:49
*** jasondotstar has quit IRC20:59
openstackgerrithenry-nash proposed a change to openstack/identity-api: Enable filtering of credentials by user ID  https://review.openstack.org/11078221:03
*** LinStatSDR has joined #openstack-keystone21:06
*** lbragstad has quit IRC21:06
*** hrybacki_ has joined #openstack-keystone21:09
*** lbragsta_ has joined #openstack-keystone21:10
*** lbragsta_ is now known as lbragstad_21:11
lbragstad_dstanek: do you know if there is a jsonschema property for validating a property *isn't* None if it is in the request?21:13
*** hrybacki has quit IRC21:13
*** hrybacki_ has quit IRC21:14
*** henrynash has quit IRC21:15
*** henrynash has joined #openstack-keystone21:18
*** henrynash has quit IRC21:20
dstaneklbragstad_: you mean that if you get the key it can't be None, but you don't require it?21:26
lbragstad_dstanek: thinking about the region case...21:26
lbragstad_if we provide an 'id' in the region request, it can't be None, or null21:27
lbragstad_because, then we create an entry in the DB, or whatever backend, that doens't have an id21:27
lbragstad_dstanek: http://paste.openstack.org/show/89251/21:27
openstackgerritBrant Knudson proposed a change to openstack/identity-api: JSON Home support  https://review.openstack.org/10988121:27
lbragstad_dstanek: pushing a new commit here soon,21:28
dstaneklbragstad_: so you want to allow parent to not be specified, but if it is it must match the pattern right?21:28
lbragstad_dstanek: we should be able to fix https://bugs.launchpad.net/keystone/+bug/1322639 with jsonschema21:29
uvirtbotLaunchpad bug 1322639 in keystone "region creation API should not allow empty id" [Medium,In progress]21:29
lbragstad_if 'id' is passed in a region create request, then it should be checked that it is not None21:29
lbragstad_but 'id' isn't technically required in a region create request. If an 'id' isn't provided by the user, keystone will generate a UUID for it21:30
*** lbragstad_ is now known as lbragstad21:30
dstaneklbragstad_: i think this will work http://paste.openstack.org/show/89252/ but you should probably verify in the interpreter21:30
dstaneklbragstad: id is not required and type is string with a pattern21:31
*** rwsu has quit IRC21:31
openstackgerritA change was merged to openstack/keystone: Add filters to the collections 'self' link  https://review.openstack.org/11066121:32
lbragstaddstanek: my mistake... we shouldn't allow an empty ID string21:33
lbragstaddstanek: the "id": {"type": "string"} part should throw an exception when passing id: None21:35
*** rwsu has joined #openstack-keystone21:39
*** gordc has quit IRC21:39
*** henrynash has joined #openstack-keystone21:42
*** henrynash has quit IRC21:43
openstackgerritLance Bragstad proposed a change to openstack/keystone: Implement validation on the Catalog V3 resources  https://review.openstack.org/9626621:43
*** lbragsta_ has joined #openstack-keystone21:48
*** lbragstad has quit IRC21:52
*** lbragsta_ has quit IRC21:53
*** hrybacki has joined #openstack-keystone21:57
*** henrynash has joined #openstack-keystone22:00
*** bknudson has quit IRC22:08
*** hrybacki_ has joined #openstack-keystone22:15
*** alex_xu has quit IRC22:17
*** hrybacki has quit IRC22:19
*** david-lyle has quit IRC22:42
*** david-lyle has joined #openstack-keystone22:42
*** david-lyle has quit IRC22:47
*** thedodd has quit IRC22:51
openstackgerritDolph Mathews proposed a change to openstack/keystone: revise docs on default _member_ role  https://review.openstack.org/11080322:52
henrynashgyee, dstanek: thx for quick approval of the spec chaneg for filtering credentials by user_id22:57
gyeehenrynash, np, that's one of those no-brainers :)23:00
henrynashgyee: nice to have one of those from time to time....23:00
*** thedodd has joined #openstack-keystone23:05
openstackgerritA change was merged to openstack/identity-api: Enable filtering of credentials by user ID  https://review.openstack.org/11078223:06
*** zzzeek has quit IRC23:09
*** bknudson has joined #openstack-keystone23:13
*** bknudson has quit IRC23:18
*** bknudson has joined #openstack-keystone23:19
*** jaosorior has quit IRC23:22
*** henrynash has quit IRC23:29
*** thedodd has quit IRC23:48
*** thedodd has joined #openstack-keystone23:52
*** jamielennox|away is now known as jamielennox23:53
*** gyee has quit IRC23:55
*** morganfainberg_Z is now known as morganfainberg23:55
*** hrybacki_ has quit IRC23:57
*** thedodd has quit IRC23:57
morganfainbergdolphm, ping, back23:57

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