| *** daneyon has quit IRC | 00:15 | |
| *** achampion has joined #openstack-keystone | 00:45 | |
| *** mhu has joined #openstack-keystone | 00:53 | |
| *** otwieracz has joined #openstack-keystone | 00:53 | |
| *** marcoemorais has joined #openstack-keystone | 00:53 | |
| *** mgagne has joined #openstack-keystone | 00:53 | |
| *** mrda-weekend has joined #openstack-keystone | 00:53 | |
| *** comstud has joined #openstack-keystone | 00:53 | |
| *** vishy has joined #openstack-keystone | 00:54 | |
| *** jraim_ has joined #openstack-keystone | 00:54 | |
| *** Ephur has joined #openstack-keystone | 00:54 | |
| *** dolphm has joined #openstack-keystone | 00:54 | |
| *** dstanek is now known as dstanek_zzz | 01:07 | |
| *** uvirtbot has joined #openstack-keystone | 01:08 | |
| *** marcoemorais has quit IRC | 01:11 | |
| *** dstanek_zzz is now known as dstanek | 01:30 | |
| *** richm has left #openstack-keystone | 01:37 | |
| *** zzzeek has joined #openstack-keystone | 01:37 | |
| *** mberlin1 has joined #openstack-keystone | 01:37 | |
| *** mberlin has quit IRC | 01:38 | |
| *** zzzeek has quit IRC | 02:09 | |
| *** nsquare has quit IRC | 02:16 | |
| *** zzzeek has joined #openstack-keystone | 02:49 | |
| *** zzzeek has quit IRC | 03:09 | |
| *** daneyon has joined #openstack-keystone | 03:32 | |
| *** zhiyan_ has quit IRC | 03:41 | |
| *** lbragstad_ has joined #openstack-keystone | 03:45 | |
| *** lbragstad_ is now known as lbragstad | 03:45 | |
| *** daneyon has quit IRC | 03:50 | |
| *** daneyon has joined #openstack-keystone | 03:51 | |
| *** zhiyan_ has joined #openstack-keystone | 04:08 | |
| *** rodrigods has joined #openstack-keystone | 04:15 | |
| *** harlowja is now known as harlowja_away | 04:26 | |
| *** rodrigods has quit IRC | 04:36 | |
| *** praneshp has joined #openstack-keystone | 04:37 | |
| *** praneshp_ has joined #openstack-keystone | 04:40 | |
| *** praneshp has quit IRC | 04:43 | |
| *** praneshp_ is now known as praneshp | 04:43 | |
| *** stevemar has joined #openstack-keystone | 05:10 | |
| *** ukalifon1 has joined #openstack-keystone | 05:25 | |
| *** ukalifon1 has quit IRC | 05:43 | |
| *** praneshp has quit IRC | 05:43 | |
| *** praneshp has joined #openstack-keystone | 05:44 | |
| *** dstanek is now known as dstanek_zzz | 05:49 | |
| *** henrynash has joined #openstack-keystone | 05:58 | |
| *** dstanek_zzz is now known as dstanek | 06:15 | |
| *** dstanek is now known as dstanek_zzz | 06:25 | |
| *** rwsu has quit IRC | 06:25 | |
| *** praneshp has quit IRC | 06:37 | |
| *** ukalifon has joined #openstack-keystone | 06:57 | |
| morganfainberg | henrynash hey there! | 07:14 |
|---|---|---|
| henrynash | morganfainbeg: hi | 07:14 |
| morganfainberg | henrynash, i see you approving changes/reviewing :) | 07:14 |
| henrynash | morganfainberg: yep…have been a bit absent fomr doing that recently…trying to make up | 07:15 |
| morganfainberg | hehe | 07:15 |
| henrynash | btw, did we come to a collective view on whether are or are no going to let roel assignmnets be made without being able to invalidate whether the user/group exists? | 07:16 |
| morganfainberg | henrynash, oh good question | 07:16 |
| morganfainberg | henrynash, uhhhhhhhhh | 07:16 |
| morganfainberg | henrynash, i think for federation we have to assume we can't tell the user/group exists, right? | 07:17 |
| henrynash | well, I guess group roles might be OK…since I assume we would hang the roles to a group that was generated by the federation mapper (separate discussion as to whether the group is defined bythe mapper or is today’s identity managed group) | 07:18 |
| henrynash | the problem is users…..I’m not sure if we even enable teh federation mapper to map to a role on a specific user (although I know we discussed that, and I suspect some people might want that) | 07:20 |
| henrynash | and in that case, no, we can’t tell if it exists (since the mapping would say something like “if the have the follwing attributes, then let them in and there user_id is based on attribute Y, or something like that) | 07:21 |
| *** dstanek_zzz is now known as dstanek | 07:26 | |
| morganfainberg | henrynash, i don't know if we have a consensus on that :( | 07:27 |
| henrynash | i didn’t think so either :-) | 07:27 |
| morganfainberg | henrynash, mind if i do a pass on the SQL for your IDMapping? | 07:31 |
| henrynash | sure! | 07:31 |
| morganfainberg | henrynash, i've been learning a lot about much much more efficient sql in SQLAlchemy lately | 07:31 |
| morganfainberg | example: | 07:31 |
| *** stevemar has quit IRC | 07:31 | |
| morganfainberg | henrynash, https://review.openstack.org/#/c/86025/11/keystone/assignment/backends/sql.py | 07:32 |
| morganfainberg | henrynash, way less overhead because we don't need to do the bookeeping of the object stuff, we aren't using the object, we just want the values | 07:32 |
| henrynash | morganfainberg: yes, I always kind of worried that we mix and match the levels of sqlalchemy that we use all over teh place….and we don’t really seem to haev set any standard way of doing this….maybe a developing.rst section on best practices? | 07:34 |
| morganfainberg | yeah we need to do that | 07:34 |
| morganfainberg | i also want to dig in and make some massive overhauls to make it more efficient | 07:34 |
| *** chandan_kumar has joined #openstack-keystone | 07:35 | |
| *** dstanek is now known as dstanek_zzz | 07:36 | |
| henrynash | I have no doubt that we can make a massive diferrence if we focus on it….to be honest we have hardly doen any REAL performance work | 07:37 |
| *** henrynash has quit IRC | 07:39 | |
| *** henrynash has joined #openstack-keystone | 07:41 | |
| *** praneshp has joined #openstack-keystone | 07:42 | |
| *** ukalifon has quit IRC | 07:44 | |
| *** ukalifon1 has joined #openstack-keystone | 07:47 | |
| morganfainberg | henrynash, ok patch 5 posted | 07:52 |
| morganfainberg | henrynash, very minor tweak | 07:52 |
| morganfainberg | should work :) | 07:52 |
| morganfainberg | should also be much more efficient than building that object we throw out everytime | 07:52 |
| *** praneshp has quit IRC | 09:01 | |
| henrynash | morganfainberg: yep. nice change, thx | 09:07 |
| morganfainberg | henrynash, interesting thought experiment along the same lines: https://review.openstack.org/103304 | 09:11 |
| henrynash | just heading offline - will check a bit later…. | 09:12 |
| morganfainberg | headed to bed myself | 09:12 |
| *** ukalifon1 has quit IRC | 09:50 | |
| *** xianghui has quit IRC | 10:42 | |
| *** xianghui has joined #openstack-keystone | 10:43 | |
| *** xianghuihui has joined #openstack-keystone | 11:01 | |
| *** chandan_kumar has quit IRC | 11:04 | |
| *** xianghuihui has quit IRC | 11:07 | |
| *** ukalifon1 has joined #openstack-keystone | 11:10 | |
| *** xianghuihui has joined #openstack-keystone | 11:49 | |
| *** openstackgerrit has joined #openstack-keystone | 12:02 | |
| *** dstanek_zzz is now known as dstanek | 12:37 | |
| *** dstanek is now known as dstanek_zzz | 12:46 | |
| *** henrynash has quit IRC | 12:50 | |
| *** ukalifon1 has quit IRC | 12:55 | |
| *** dstanek_zzz is now known as dstanek | 13:07 | |
| openstackgerrit | OpenStack Proposal Bot proposed a change to openstack/keystone: Updated from global requirements https://review.openstack.org/101800 | 13:10 |
| *** shakayumi has joined #openstack-keystone | 13:14 | |
| *** shakayumi has quit IRC | 13:14 | |
| *** dstanek is now known as dstanek_zzz | 13:17 | |
| *** chandan_kumar has joined #openstack-keystone | 13:43 | |
| *** diegows has joined #openstack-keystone | 13:51 | |
| *** diegows has quit IRC | 13:52 | |
| *** diegows has joined #openstack-keystone | 13:52 | |
| *** diegows has joined #openstack-keystone | 13:53 | |
| openstackgerrit | Arun Kant proposed a change to openstack/keystone: Adding support for ldap connection pooling. https://review.openstack.org/95300 | 13:54 |
| *** zzzeek has joined #openstack-keystone | 14:02 | |
| *** xianghuihui has quit IRC | 14:06 | |
| *** xianghuihui has joined #openstack-keystone | 14:07 | |
| *** dstanek_zzz is now known as dstanek | 14:08 | |
| *** xianghuihui has quit IRC | 14:12 | |
| *** xianghuihui has joined #openstack-keystone | 14:12 | |
| *** xianghui has quit IRC | 14:14 | |
| *** xianghui has joined #openstack-keystone | 14:14 | |
| *** xianghui has quit IRC | 14:15 | |
| *** dstanek is now known as dstanek_zzz | 14:19 | |
| *** chandan_kumar has quit IRC | 14:23 | |
| *** ukalifon has joined #openstack-keystone | 14:25 | |
| *** rodrigods has joined #openstack-keystone | 14:29 | |
| *** chandan_kumar has joined #openstack-keystone | 14:35 | |
| -openstackstatus- NOTICE: Gerrit will be offline from 1500-1515 UTC for project renames | 14:38 | |
| *** ChanServ changes topic to "Gerrit will be offline from 1500-1515 UTC for project renames" | 14:38 | |
| *** rodrigods has quit IRC | 14:40 | |
| *** ukalifon has quit IRC | 14:50 | |
| *** ukalifon1 has joined #openstack-keystone | 14:51 | |
| *** rodrigods has joined #openstack-keystone | 14:57 | |
| *** rodrigods has quit IRC | 14:57 | |
| *** rodrigods has joined #openstack-keystone | 14:57 | |
| *** rodrigods has quit IRC | 15:08 | |
| openstackgerrit | Nathan Kinder proposed a change to openstack/keystone: Implicitly ignore attributes that are mapped to None in LDAP https://review.openstack.org/103325 | 15:10 |
| *** dstanek_zzz is now known as dstanek | 15:10 | |
| *** rodrigods has joined #openstack-keystone | 15:18 | |
| *** rodrigods has quit IRC | 15:18 | |
| *** rodrigods has joined #openstack-keystone | 15:18 | |
| *** xianghuihui has quit IRC | 15:19 | |
| *** dstanek is now known as dstanek_zzz | 15:21 | |
| *** rodrigods has quit IRC | 15:29 | |
| *** ukalifon1 has quit IRC | 15:37 | |
| *** xianghui has joined #openstack-keystone | 15:39 | |
| *** openstackgerrit has quit IRC | 15:46 | |
| *** openstackgerrit has joined #openstack-keystone | 15:47 | |
| *** diegows has quit IRC | 15:48 | |
| openstackgerrit | Longgeek proposed a change to openstack/keystone-specs: Update pbr version https://review.openstack.org/103339 | 15:48 |
| *** diegows has joined #openstack-keystone | 15:48 | |
| *** ChanServ changes topic to "Please Review and/or Update Proposed Specs | Hackathon July 9-11: http://dolphm.com/openstack-keystone-hackathon-for-juno/" | 15:52 | |
| *** chandan_kumar has quit IRC | 16:00 | |
| *** stevemar has joined #openstack-keystone | 16:04 | |
| stevemar | morganfainberg, is there any specialness needed to get openstack proposal bot to kick in for keystone-specs? | 16:11 |
| *** dstanek_zzz is now known as dstanek | 16:12 | |
| stevemar | morganfainberg, someone proposed a change to keystone specs requirements, i was going to -2 it, but then realized there havent been any requirements proposals to date in that repo | 16:12 |
| *** dstanek is now known as dstanek_zzz | 16:22 | |
| *** daneyon has quit IRC | 16:46 | |
| *** nsquare has joined #openstack-keystone | 16:47 | |
| *** nsquare_ has joined #openstack-keystone | 16:48 | |
| *** zzzeek has quit IRC | 16:48 | |
| *** nsquare has quit IRC | 16:51 | |
| *** Chicago has quit IRC | 17:01 | |
| *** dstanek_zzz is now known as dstanek | 17:12 | |
| *** dstanek is now known as dstanek_zzz | 17:22 | |
| *** nsquare_ has quit IRC | 17:33 | |
| *** henrynash has joined #openstack-keystone | 17:44 | |
| *** chandan_kumar has joined #openstack-keystone | 17:45 | |
| *** praneshp has joined #openstack-keystone | 17:52 | |
| *** praneshp_ has joined #openstack-keystone | 18:13 | |
| *** dstanek_zzz is now known as dstanek | 18:13 | |
| *** praneshp has quit IRC | 18:14 | |
| *** praneshp_ is now known as praneshp | 18:14 | |
| *** dstanek is now known as dstanek_zzz | 18:23 | |
| *** ChanServ sets mode: +o dolphm | 18:53 | |
| openstackgerrit | Morgan Fainberg proposed a change to openstack/keystone: DO NOT MERGE: Performance profiling https://review.openstack.org/103304 | 19:10 |
| morganfainberg | stevemar, what is proposal bot missing for -specs? | 19:11 |
| morganfainberg | stevemar, it's just a config thing though i think | 19:12 |
| boris-42 | morganfainberg ^ you really need profiler=) | 19:12 |
| morganfainberg | boris-42, yes we do. | 19:12 |
| morganfainberg | boris-42, i'm just looking at the basic rally tests right now. | 19:12 |
| boris-42 | morganfainberg so you can help to sped up | 19:12 |
| boris-42 | morganfainberg speed up | 19:12 |
| boris-42 | morganfainberg we are mostly blocked by oslo.messaging | 19:12 |
| morganfainberg | boris-42, hehe | 19:12 |
| morganfainberg | boris-42 well has the req merged? | 19:12 |
| boris-42 | morganfainberg yep lib is in global requiremnts | 19:13 |
| boris-42 | morganfainberg and at monday ceilometer patch will be in | 19:13 |
| morganfainberg | boris-42, ok so, what do i do to add it for keystone? | 19:13 |
| morganfainberg | i'm happy to add it forthis patch | 19:13 |
| morganfainberg | besides "install it" | 19:13 |
| boris-42 | morganfainberg basically need this patch in oslo.messaging https://review.openstack.org/#/c/103270/ | 19:13 |
| boris-42 | morganfainberg or you can make your local devstack installation | 19:14 |
| morganfainberg | ah | 19:14 |
| *** dstanek_zzz is now known as dstanek | 19:14 | |
| boris-42 | morganfainberg with my messaging + couple of patches in keystone | 19:14 |
| morganfainberg | so gate wont work w/o that patch | 19:14 |
| morganfainberg | darn | 19:14 |
| boris-42 | morganfainberg ? | 19:14 |
| boris-42 | morganfainberg to enable in gates profiling we will need to merge couple of patches in different porjects=) | 19:15 |
| morganfainberg | boris-42, i meant, if i could turn it on right now w/o needing that patch for ^ that profiling patchset | 19:15 |
| morganfainberg | boris-42, yeah that was my point | 19:15 |
| boris-42 | morganfainberg unfortunately we need to merge ceilometer & oslo.messaging patches | 19:15 |
| boris-42 | morganfainberg and you'll be able | 19:15 |
| morganfainberg | *nod* | 19:15 |
| boris-42 | morganfainberg so we are quite close | 19:15 |
| boris-42 | morganfainberg as well I have some patch in rally | 19:15 |
| morganfainberg | let me fix a typo in a patch and maybe i'll have you help me put it together for a local tempest run | 19:15 |
| boris-42 | morganfainberg that capture traces for every iteration | 19:16 |
| morganfainberg | morganfainberg, that patch there is the initial pass on reducing object book keeping overhead for SQLA and LDAP queries | 19:16 |
| morganfainberg | s/morganfainberg/boris-42 | 19:16 |
| boris-42 | morganfainberg heh | 19:17 |
| boris-42 | morganfainberg actually I can publish some patches to keystone | 19:19 |
| morganfainberg | boris-42, no rush if we need extra stuff to merge first | 19:19 |
| morganfainberg | boris-42, don't want you to get stuck rebasing a ton (unless you're bored ;) | 19:19 |
| boris-42 | morganfainberg heh osprofiler is super clean=) | 19:20 |
| boris-42 | morganfainberg for 3 months I had to rebase about 3 times couple of lines of code=) | 19:20 |
| morganfainberg | well i def. want it in keystone :) | 19:20 |
| boris-42 | morganfainberg so pls | 19:20 |
| morganfainberg | i mean, i can't speak for the rest of core. but i do want it | 19:20 |
| boris-42 | morganfainberg +1 that oslo.messaging patch | 19:20 |
| boris-42 | morganfainberg and add comment that you would like to get it=) | 19:21 |
| boris-42 | morganfainberg it makes sense=) | 19:21 |
| morganfainberg | nod will do | 19:21 |
| boris-42 | morganfainberg and in keystone python client we can merge even now patch | 19:21 |
| *** dstanek is now known as dstanek_zzz | 19:24 | |
| boris-42 | morganfainberg oh | 19:26 |
| boris-42 | morganfainberg I'll need your help a bit | 19:26 |
| morganfainberg | sure | 19:27 |
| boris-42 | morganfainberg keystone client is a bit different from others | 19:27 |
| morganfainberg | yeah it is | 19:27 |
| boris-42 | morganfainberg I need to add special header in all reuqest | 19:27 |
| boris-42 | morganfainberg where can I do that?) | 19:27 |
| morganfainberg | uhm.. | 19:27 |
| morganfainberg | that might get sticky, because some of the code keystoneclient is using will be re-used by most other clients | 19:28 |
| morganfainberg | if it has to be specific to keystoneclient (header) | 19:28 |
| boris-42 | morganfainberg nope it is not specific | 19:28 |
| morganfainberg | boris-42, might need to get jamielennox|away to help on that | 19:28 |
| morganfainberg | boris-42, ok so then Session object | 19:28 |
| boris-42 | morganfainberg it's not specific, actually I will be happy only=) | 19:29 |
| boris-42 | morganfainberg if everybody will use one base | 19:29 |
| boris-42 | morganfainberg so it will be just one patch=0 | 19:29 |
| morganfainberg | and eventually all the other clients will consume that instead of custom implementation | 19:29 |
| boris-42 | morganfainberg not 100500 | 19:29 |
| morganfainberg | some clients do already use some of the session object | 19:29 |
| boris-42 | morganfainberg ah I found it thanks | 19:29 |
| boris-42 | morganfainberg one sec and you'll get one of required patches | 19:30 |
| morganfainberg | boris-42, cool! | 19:30 |
| boris-42 | morganfainberg I will work during next week to get in rally support of profiling | 19:35 |
| boris-42 | morganfainberg I mean integration | 19:35 |
| morganfainberg | boris-42, great | 19:35 |
| morganfainberg | boris-42, more profiling work will help us all out :) | 19:35 |
| boris-42 | morganfainberg yep it is really hard to get works work slow | 19:38 |
| boris-42 | morganfainberg especially in projects like nova and heat | 19:38 |
| boris-42 | morganfainberg and in solum & murano it will be even harder | 19:38 |
| openstackgerrit | Boris Pavlovic proposed a change to openstack/python-keystoneclient: Add profiling support to Keystone https://review.openstack.org/103367 | 19:39 |
| boris-42 | morganfainberg ^ | 19:39 |
| morganfainberg | wow that is not a big changeset | 19:40 |
| boris-42 | morganfainberg it's client part | 19:40 |
| morganfainberg | still | 19:40 |
| boris-42 | morganfainberg lemme make project one | 19:40 |
| boris-42 | morganfainberg but it depends on that patch in oslo.messaging | 19:40 |
| morganfainberg | not bad | 19:40 |
| boris-42 | morganfainberg (but we can make it lazy) | 19:40 |
| morganfainberg | *nod* | 19:40 |
| boris-42 | morganfainberg btw there are some issues with grenade jobs | 19:41 |
| boris-42 | morganfainberg if I am adding to requirements osprofiler in clients | 19:41 |
| morganfainberg | grenade can be a pita | 19:41 |
| boris-42 | morganfainberg it says something like: 2014-06-28 18:06:58.031 | 'osprofiler' is not a global requirement but it should be,something went wrong | 19:42 |
| boris-42 | morganfainberg probably we should add to icehouse and havana global-requiremnts osprofiler | 19:42 |
| morganfainberg | huh | 19:42 |
| morganfainberg | interesting | 19:42 |
| morganfainberg | nah | 19:42 |
| morganfainberg | just figure out how to make it not error | 19:42 |
| morganfainberg | we don't want to add more requirements to stable | 19:42 |
| morganfainberg | if avoidable | 19:42 |
| boris-42 | morganfainberg this thing with lazy stuff will work well | 19:42 |
| morganfainberg | ++ | 19:42 |
| openstackgerrit | Morgan Fainberg proposed a change to openstack/keystone: Remove `with_lockmode` use from Trust SQL backend. https://review.openstack.org/97059 | 19:43 |
| boris-42 | morganfainberg but we need this requirement in keystone | 19:43 |
| boris-42 | morganfainberg (in requieremnts ) | 19:43 |
| boris-42 | morganfainberg I will try to add small patch and we will see what will happen=) | 19:43 |
| morganfainberg | boris-42, if it's in global requirements, it should be propsoed to keystone | 19:44 |
| boris-42 | morganfainberg it is in global requirments | 19:44 |
| boris-42 | morganfainberg it is not in global requirements of icehouse and havana | 19:44 |
| morganfainberg | https://review.openstack.org/#/c/101800/ | 19:44 |
| morganfainberg | oh right | 19:45 |
| morganfainberg | right | 19:45 |
| morganfainberg | lazy load ftw then | 19:45 |
| boris-42 | morganfainberg I am not able to lazy load in project | 19:45 |
| boris-42 | morganfainberg only in python client | 19:45 |
| boris-42 | morganfainberg but in oslo.messaging it works well | 19:45 |
| morganfainberg | no? | 19:45 |
| boris-42 | morganfainberg yep I will show | 19:45 |
| morganfainberg | ok | 19:45 |
| boris-42 | morganfainberg wait a sec for one more patch=) | 19:45 |
| morganfainberg | sure | 19:46 |
| boris-42 | morganfainberg heh I need your help one more time=) | 19:52 |
| morganfainberg | sure | 19:52 |
| boris-42 | morganfainberg so projects like nova/cinder have such file service.py | 19:52 |
| boris-42 | morganfainberg that is used to start wsgi servicer (and rpc services) | 19:52 |
| boris-42 | morganfainberg where can I find it in keystone? | 19:53 |
| morganfainberg | boris-42, https://github.com/openstack/keystone/blob/master/keystone/service.py ? | 19:53 |
| morganfainberg | that what you're looking for? | 19:53 |
| boris-42 | morganfainberg yep I was looking at it.. | 19:54 |
| boris-42 | morganfainberg not sure yet=) | 19:54 |
| *** zzzeek has joined #openstack-keystone | 19:54 | |
| morganfainberg | there is also https://github.com/openstack/keystone/blob/master/keystone/common/wsgi.py | 19:54 |
| boris-42 | morganfainberg yep ^ that look more fair | 19:54 |
| boris-42 | morganfainberg hmmm | 19:55 |
| boris-42 | morganfainberg lemme explain what I would like to do=) | 19:56 |
| morganfainberg | we may not be running under eventlet | 19:56 |
| morganfainberg | we may be under apache | 19:56 |
| morganfainberg | what you're probably expecting is https://github.com/openstack/keystone/blob/master/keystone/common/environment/eventlet_server.py | 19:56 |
| boris-42 | morganfainberg https://github.com/boris-42/nova/blob/0e5949fa1e1fac5b9cf75685ee78285da30599ec/nova/service.py#L159-L160 | 19:56 |
| morganfainberg | but soon that wont be where we run keystone from (by default) | 19:56 |
| boris-42 | morganfainberg ^ this is nova | 19:56 |
| boris-42 | morganfainberg I have to setup osproifler lib notifier function | 19:57 |
| boris-42 | morganfainberg to say osprofiler to use oslo.messaging Notifier API instance | 19:57 |
| boris-42 | morganfainberg it should be done on start of service | 19:57 |
| boris-42 | morganfainberg so when we are processing wsgi middleware we already have stetted up osprofiler lib | 19:57 |
| morganfainberg | if you were to put it in service.py before loading the backends, we should be good. | 19:58 |
| morganfainberg | erm | 19:58 |
| boris-42 | morganfainberg ? | 19:59 |
| boris-42 | morganfainberg you don't use oslo.messaging?) | 19:59 |
| morganfainberg | we do. | 19:59 |
| morganfainberg | but it isn't consumed by anything in keystone | 19:59 |
| morganfainberg | we emit notifications but never consume | 20:00 |
| boris-42 | morganfainberg and you won't consume them | 20:00 |
| morganfainberg | nope | 20:00 |
| morganfainberg | we will not | 20:00 |
| boris-42 | morganfainberg I mean | 20:00 |
| morganfainberg | https://github.com/openstack/keystone/blob/master/keystone/notifications.py | 20:00 |
| morganfainberg | you are correct | 20:00 |
| boris-42 | morganfainberg profiler will just create new notifications | 20:00 |
| boris-42 | morganfainberg rally or tempest or whatever will consume them | 20:00 |
| boris-42 | morganfainberg to draw traces | 20:00 |
| morganfainberg | ah | 20:00 |
| morganfainberg | all our notification code is in https://github.com/openstack/keystone/blob/master/keystone/notifications.py | 20:00 |
| morganfainberg | so | 20:01 |
| boris-42 | morganfainberg I can put here https://github.com/openstack/keystone/blob/master/keystone/notifications.py#L180 | 20:01 |
| *** nsquare has joined #openstack-keystone | 20:02 | |
| boris-42 | morganfainberg initialization of profiler | 20:02 |
| morganfainberg | hm | 20:02 |
| morganfainberg | that probably wont be called early enough | 20:02 |
| morganfainberg | we can try it | 20:02 |
| boris-42 | morganfainberg hm | 20:02 |
| boris-42 | morganfainberg it's bad place as well | 20:03 |
| boris-42 | morganfainberg cause we won't know what is the service | 20:03 |
| boris-42 | morganfainberg that calls it | 20:03 |
| boris-42 | morganfainberg actually we can add middleware | 20:05 |
| morganfainberg | sure | 20:05 |
| morganfainberg | that is a good route | 20:05 |
| boris-42 | morganfainberg that will do initialization but not sure that it's what we want=) | 20:05 |
| morganfainberg | could the middleware be used everywhere? | 20:06 |
| boris-42 | morganfainberg hm in other projects? | 20:06 |
| morganfainberg | everyservice that is? | 20:06 |
| morganfainberg | (or most) | 20:06 |
| morganfainberg | because if so, then that would be a good approach, easy to disable/remove if you wont want it | 20:06 |
| morganfainberg | boris-42, yeah | 20:06 |
| boris-42 | morganfainberg it's easy to disable | 20:06 |
| boris-42 | morganfainberg whole osprofiler is lazy | 20:06 |
| morganfainberg | a production environment may not want profiling code at all | 20:06 |
| boris-42 | morganfainberg and there is already osprofiler middleware | 20:06 |
| morganfainberg | if it was in middleware, it's a deployer choice to use | 20:06 |
| boris-42 | morganfainberg nope they want=) | 20:06 |
| morganfainberg | right | 20:06 |
| boris-42 | morganfainberg believe me=) | 20:07 |
| boris-42 | morganfainberg I specially create this stuff to be ON always in production | 20:07 |
| morganfainberg | eh. i know some deployers given the choise would say no | 20:07 |
| morganfainberg | but, in either case i dunno where to put this | 20:07 |
| boris-42 | morganfainberg I mean lemme publish what I have | 20:07 |
| morganfainberg | let me think about it post weekend and maybe we can get something figured out | 20:07 |
| openstackgerrit | Boris Pavlovic proposed a change to openstack/keystone: Add osprofiler wsgi middleware https://review.openstack.org/103368 | 20:08 |
| boris-42 | morganfainberg ^ | 20:08 |
| boris-42 | morganfainberg so profiling by default does nothing | 20:09 |
| morganfainberg | nod. | 20:09 |
| boris-42 | morganfainberg even if this middleware is turned ON | 20:09 |
| boris-42 | morganfainberg it won't do anything | 20:09 |
| boris-42 | morganfainberg unit it get special header with trace info signed by proper HMAC key that is specified in paste.ini | 20:09 |
| boris-42 | morganfainberg so we can keep it turned ON even in production, and if we get some nasty stuff, we can easily trace it (without restarting everything) | 20:10 |
| morganfainberg | nod | 20:10 |
| boris-42 | morganfainberg so I hope that I'll be able to push community to keep it on in gates | 20:12 |
| boris-42 | morganfainberg so in rally gates we will be able to use it without any hacks=) | 20:12 |
| morganfainberg | and in gate i would hope it's on | 20:12 |
| morganfainberg | :) | 20:12 |
| morganfainberg | or at least a toggle in ds-g so it's easy to turn on where we want it | 20:12 |
| boris-42 | morganfainberg yep we should make it simple to find issues | 20:12 |
| *** shakamunyi has joined #openstack-keystone | 20:13 | |
| boris-42 | morganfainberg so seems like in keystone we will have 2 middle wares | 20:14 |
| boris-42 | morganfainberg one from osprofiler.lib and one that init osprofiler | 20:14 |
| *** dstanek_zzz is now known as dstanek | 20:15 | |
| morganfainberg | boris-42, sure | 20:15 |
| boris-42 | morganfainberg but probably we can talk with others | 20:15 |
| morganfainberg | boris-42, ok | 20:15 |
| boris-42 | morganfainberg probably there is way without second middleware=) | 20:15 |
| morganfainberg | boris-42, gonna head out for a bit, be back later :) | 20:15 |
| boris-42 | morganfainberg see you | 20:15 |
| morganfainberg | dstanek, added you to the https://review.openstack.org/#/c/103304/ review. looking at what we can do to generally improve keystone performance, and i think this is a start. | 20:16 |
| *** stevemar has quit IRC | 20:16 | |
| *** shakamunyi has quit IRC | 20:18 | |
| *** dstanek is now known as dstanek_zzz | 20:25 | |
| *** zzzeek has quit IRC | 20:26 | |
| openstackgerrit | A change was merged to openstack/keystone: Remove db, db.sqlalchemy from openstack-common.conf https://review.openstack.org/103251 | 20:30 |
| *** daneyon has joined #openstack-keystone | 20:33 | |
| *** henrynash has quit IRC | 20:34 | |
| *** shakamunyi has joined #openstack-keystone | 20:34 | |
| openstackgerrit | A change was merged to openstack/python-keystoneclient: Updated from global requirements https://review.openstack.org/101404 | 20:41 |
| openstackgerrit | A change was merged to openstack/python-keystoneclient: Add issued handlers to auth_ref and fixtures https://review.openstack.org/100775 | 20:41 |
| *** jimbaker has quit IRC | 20:49 | |
| *** jimbaker has joined #openstack-keystone | 20:54 | |
| *** jimbaker has quit IRC | 20:54 | |
| *** jimbaker has joined #openstack-keystone | 20:54 | |
| *** henrynash has joined #openstack-keystone | 21:12 | |
| *** henrynash has quit IRC | 21:15 | |
| *** dstanek_zzz is now known as dstanek | 21:15 | |
| *** dstanek is now known as dstanek_zzz | 21:25 | |
| openstackgerrit | A change was merged to openstack/python-keystoneclient: Correcting using-api-v2.rst https://review.openstack.org/102986 | 21:26 |
| *** hrybacki has joined #openstack-keystone | 21:28 | |
| openstackgerrit | Boris Pavlovic proposed a change to openstack/keystone: Add sample of rally plugin https://review.openstack.org/98836 | 21:38 |
| boris-42 | morganfainberg ^ could you pls one more time +2 plugin sample I address dolphm comments | 21:40 |
| *** htruta has quit IRC | 21:45 | |
| *** rodrigods has joined #openstack-keystone | 21:53 | |
| *** rodrigods has quit IRC | 21:53 | |
| *** rodrigods has joined #openstack-keystone | 21:53 | |
| *** chandan_kumar has quit IRC | 21:58 | |
| *** htruta has joined #openstack-keystone | 22:03 | |
| *** shakamunyi has quit IRC | 22:08 | |
| *** dstanek_zzz is now known as dstanek | 22:16 | |
| *** dstanek is now known as dstanek_zzz | 22:26 | |
| *** rodrigods has quit IRC | 22:29 | |
| *** dstanek_zzz is now known as dstanek | 22:51 | |
| *** dstanek is now known as dstanek_zzz | 23:01 | |
| *** rodrigods has joined #openstack-keystone | 23:01 | |
| *** rodrigods has quit IRC | 23:01 | |
| *** rodrigods has joined #openstack-keystone | 23:01 | |
| *** daneyon has quit IRC | 23:28 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!