| bknudson1 | this looks bad: http://logs.openstack.org/27/171327/2/gate/gate-keystone-python27/6acb43d/console.html#_2015-04-07_22_27_29_961 | 00:06 |
|---|---|---|
| bknudson1 | py27 tests failed. | 00:06 |
| bknudson1 | this is also an error in dogpile.cache... | 00:08 |
| bknudson1 | new release? | 00:08 |
| bknudson1 | I was able to recreate it locally... | 00:10 |
| bknudson1 | pymongo 3.0 was released. | 00:11 |
| morganfainberg | blah | 00:12 |
| morganfainberg | g-r update time. | 00:12 |
| morganfainberg | and likely needs to be in the kilo-rc for g-r | 00:12 |
| morganfainberg | alternatively... | 00:12 |
| bknudson1 | I'll work on the g-r update. | 00:13 |
| morganfainberg | we drop the mongo driver for dogpile. | 00:13 |
| morganfainberg | :P | 00:13 |
| morganfainberg | (not serious yet) | 00:13 |
| bknudson1 | we can't drop mongo... we needs webscale | 00:13 |
| morganfainberg | hahahaha | 00:13 |
| bknudson1 | morganfainberg: https://review.openstack.org/#/c/171438/ <-- g-r change. | 00:15 |
| bknudson1 | I updated the commit message. | 00:18 |
| bknudson1 | bug 1441393 | 00:19 |
| openstack | bug 1441393 in Keystone "keystone unit tests fail with pymongo 3.0" [Undecided,New] https://launchpad.net/bugs/1441393 | 00:19 |
| bknudson1 | did they rename the method or remove it or what? | 00:19 |
| bknudson1 | it's gone! | 00:22 |
| bknudson1 | no deprecation warning or anything. | 00:22 |
| morganfainberg | major version bump though looks like | 00:23 |
| morganfainberg | which, at least they are using sensible versioning | 00:23 |
| morganfainberg | 2.x = has it, 3.x doesn't | 00:23 |
| bknudson1 | I guess mongos_enum didn't have a dostring, so maybe it wasn't public. | 00:23 |
| morganfainberg | yeah might not have been | 00:23 |
| bknudson1 | _MONGOS_MODES is still there. | 00:24 |
| bknudson1 | oh, it was renamed. | 00:25 |
| morganfainberg | oh | 00:26 |
| morganfainberg | fantastic | 00:26 |
| bknudson1 | looks like it's called read_pref_mode_from_name now? | 00:27 |
| morganfainberg | not that this becomes hard to support multiple versions this way or anything >.> | 00:27 |
| morganfainberg | ah well i guess thats why we have min/cap capabilities | 00:27 |
| bknudson1 | y, it looks like it was just renamed... let me give it a shot. | 00:29 |
| *** browne has joined #openstack-keystone | 00:33 | |
| openstackgerrit | Brant Knudson proposed openstack/keystone: Work with pymongo 3.0 https://review.openstack.org/171442 | 00:36 |
| bknudson1 | we'll see if that works. | 00:36 |
| openstackgerrit | Brant Knudson proposed openstack/keystone: Work with pymongo 3.0 https://review.openstack.org/171442 | 00:42 |
| openstackgerrit | Brant Knudson proposed openstack/keystone: Work with pymongo 3.0 https://review.openstack.org/171442 | 00:43 |
| bknudson1 | added a comment. | 00:43 |
| openstackgerrit | Jamie Lennox proposed openstack/keystone: Move endpoint policy into keystone core https://review.openstack.org/171448 | 00:51 |
| *** stevemar has joined #openstack-keystone | 00:57 | |
| *** ChanServ sets mode: +v stevemar | 00:57 | |
| *** zzzeek has quit IRC | 01:00 | |
| *** alexsyip has quit IRC | 01:05 | |
| *** _cjones_ has quit IRC | 01:08 | |
| *** spandhe has quit IRC | 01:08 | |
| openstackgerrit | Merged openstack/keystonemiddleware: Fix typos in keystonemiddleware https://review.openstack.org/170722 | 01:08 |
| lhcheng | stevemar: when you get the chance, can you remove the -2s for https://review.openstack.org/#/c/167939/ and https://review.openstack.org/#/c/167939/ | 01:11 |
| stevemar | lhcheng, for our newest core? Anything!? | 01:13 |
| stevemar | !! | 01:13 |
| openstack | stevemar: Error: "!" is not a valid command. | 01:13 |
| lhcheng | stevemar: lol thank you sir! | 01:14 |
| stevemar | lhcheng, you linked the same review twice :) | 01:14 |
| stevemar | were there 2 different ones? | 01:14 |
| lhcheng | oops | 01:15 |
| lhcheng | https://review.openstack.org/#/c/168427/ | 01:15 |
| lhcheng | stevemar: I am just making sure you get the link :P | 01:15 |
| *** ayoung has joined #openstack-keystone | 01:16 | |
| *** ChanServ sets mode: +v ayoung | 01:16 | |
| *** iamjarvo has joined #openstack-keystone | 01:19 | |
| stevemar | hey, i am the author on that one | 01:20 |
| stevemar | neat | 01:20 |
| lhcheng | stevemar: thanks! | 01:23 |
| *** lhcheng has quit IRC | 01:23 | |
| stevemar | morganfainberg, i already proposed the placeholder patch :) https://review.openstack.org/#/c/171329/ | 01:23 |
| *** alex_xu has joined #openstack-keystone | 01:37 | |
| *** erkules has quit IRC | 01:42 | |
| *** iamjarvo has quit IRC | 01:49 | |
| *** dims_ has quit IRC | 01:51 | |
| *** spandhe has joined #openstack-keystone | 01:54 | |
| *** spandhe_ has joined #openstack-keystone | 01:59 | |
| *** cyeoh has quit IRC | 02:00 | |
| *** d34dh0r53 has quit IRC | 02:00 | |
| *** spandhe has quit IRC | 02:01 | |
| *** spandhe_ is now known as spandhe | 02:01 | |
| *** hockeynut has quit IRC | 02:01 | |
| *** comstud has quit IRC | 02:01 | |
| *** lbragstad has quit IRC | 02:01 | |
| *** Trozz_ has joined #openstack-keystone | 02:01 | |
| *** boris-42 has quit IRC | 02:02 | |
| *** dtroyer has quit IRC | 02:02 | |
| *** mgagne has quit IRC | 02:02 | |
| *** adam_g has quit IRC | 02:02 | |
| *** gus has quit IRC | 02:02 | |
| *** sudorandom has quit IRC | 02:02 | |
| *** Trozz has quit IRC | 02:02 | |
| *** dolphm has quit IRC | 02:02 | |
| *** raildo has quit IRC | 02:02 | |
| *** gus has joined #openstack-keystone | 02:03 | |
| *** ayoung has quit IRC | 02:05 | |
| *** boris-42 has joined #openstack-keystone | 02:05 | |
| *** sudorandom has joined #openstack-keystone | 02:07 | |
| *** d34dh0r53 has joined #openstack-keystone | 02:07 | |
| *** hockeynut has joined #openstack-keystone | 02:07 | |
| *** comstud has joined #openstack-keystone | 02:07 | |
| *** dolphm has joined #openstack-keystone | 02:07 | |
| *** adam_g has joined #openstack-keystone | 02:07 | |
| *** ChanServ sets mode: +o dolphm | 02:07 | |
| *** adam_g has quit IRC | 02:07 | |
| *** adam_g has joined #openstack-keystone | 02:07 | |
| *** mgagne has joined #openstack-keystone | 02:08 | |
| *** dtroyer has joined #openstack-keystone | 02:08 | |
| *** mgagne is now known as Guest39926 | 02:08 | |
| *** lbragstad has joined #openstack-keystone | 02:11 | |
| *** cyeoh has joined #openstack-keystone | 02:11 | |
| *** SpamapS has joined #openstack-keystone | 02:18 | |
| SpamapS | morganfainberg: poke | 02:18 |
| morganfainberg | SpamapS: eek | 02:18 |
| *** patrickeast has quit IRC | 02:21 | |
| SpamapS | morganfainberg: dogpile.cache is pissing me off | 02:21 |
| morganfainberg | SpamapS: sorry :(. What's happening and can I help make it less sucky. | 02:22 |
| SpamapS | morganfainberg: the entire library seems to rely on the region being available at import time. | 02:22 |
| *** harlowja is now known as harlowja_away | 02:22 | |
| morganfainberg | Sort of | 02:22 |
| SpamapS | which is like, zomg not concurrency safe and makes me cry | 02:22 |
| morganfainberg | I have written some tooling around that. | 02:22 |
| SpamapS | I want a cache per object | 02:22 |
| morganfainberg | Ah a region per object? | 02:22 |
| morganfainberg | In dogpile parlance | 02:23 |
| SpamapS | well really I just care that the object is isolated so I don't have to worry about concurrency. | 02:23 |
| morganfainberg | In-memory cache? | 02:23 |
| SpamapS | any backend will do | 02:23 |
| morganfainberg | So dogpile does thread.local some stuff | 02:24 |
| morganfainberg | When it comes to the backends. | 02:24 |
| SpamapS | It doesn't say in the docs regions can be shared across threads. | 02:24 |
| morganfainberg | What are you poking at so I have a bit more context. | 02:24 |
| SpamapS | If it did that, I might feel better. | 02:25 |
| SpamapS | morganfainberg: shade.... | 02:25 |
| morganfainberg | Ok. | 02:25 |
| SpamapS | morganfainberg: https://git.openstack.org/cgit/openstack-infra/shade/tree/shade/__init__.py | 02:25 |
| morganfainberg | So let me finish food, (have to pay and walk home in the rain) | 02:25 |
| SpamapS | morganfainberg: basically I can't use the cache_on_arguments decorator without having region available at import time. | 02:26 |
| SpamapS | morganfainberg: rain should be stopping any minute | 02:26 |
| morganfainberg | Yeah I keep meaning to fix that | 02:26 |
| SpamapS | I could just generate my own keys | 02:27 |
| morganfainberg | The solution I had was I made a partial that decorated and configured the object's region then the decorator was object local | 02:27 |
| SpamapS | morganfainberg: thats kind of what I'm poking at now. | 02:27 |
| morganfainberg | It's icky. | 02:28 |
| SpamapS | morganfainberg: I've got it working, but then I lose things like the .invalidate() method. | 02:28 |
| SpamapS | morganfainberg: https://review.openstack.org/171456 | 02:28 |
| morganfainberg | That is the problem. | 02:28 |
| SpamapS | morganfainberg: that works, but now no .invalidate() method, which was the whole point | 02:28 |
| morganfainberg | You can lift he invalidate from the object local up. | 02:29 |
| morganfainberg | SpamapS: this is a core issue with memorization via decorators. | 02:30 |
| morganfainberg | SpamapS: decorators are compile time. You could make a version that is passed a region object that then the region is post-configured | 02:31 |
| SpamapS | morganfainberg: it could be accomplished by having the decorator just keep track of methods and then close the loop at __init__ time with registration. | 02:31 |
| SpamapS | morganfainberg: but that kind of has to be done inside dogpile.cache | 02:31 |
| morganfainberg | Sure. It's meta programming for sure. | 02:31 |
| SpamapS | I have a feeling users just give up and make a global region. | 02:31 |
| morganfainberg | It's what we did for keystone. Well a per-module region. | 02:32 |
| morganfainberg | Sort of | 02:32 |
| SpamapS | That won't work for shade. | 02:32 |
| SpamapS | It wants per-cloud cache. | 02:32 |
| SpamapS | some clouds we might want to use memcache servers, some local. It's just not a global problem. | 02:33 |
| morganfainberg | You're asking for a abcmeta mechanism with @memoize being the source for indicating memorization should occur | 02:33 |
| SpamapS | I think what I might do is just ditch the decorators. | 02:33 |
| SpamapS | I can make my own keys | 02:33 |
| SpamapS | sucks | 02:33 |
| SpamapS | but I can. | 02:33 |
| morganfainberg | I know how I'd write a metaclass for this. | 02:33 |
| morganfainberg | Actually. | 02:34 |
| *** browne1 has joined #openstack-keystone | 02:34 | |
| SpamapS | Anyway, kids are going bonkers, I need to take a break. | 02:34 |
| morganfainberg | And the decorator could then look local for the region and call the wrap method as needed. | 02:34 |
| SpamapS | Thanks for confirming what I thought: that it's not ideal and needs thought. | 02:34 |
| SpamapS | morganfainberg: thats what I'm poking at now. | 02:34 |
| SpamapS | morganfainberg: ignore the private class, but the test at the bottom is what I'm aiming at http://paste.ubuntu.com/10769144/ | 02:34 |
| morganfainberg | Yeah. Dogpile is very rough around the edges because it meets a lot of needs just not all | 02:34 |
| morganfainberg | And it has been slow moving because it meets a lot of needs | 02:35 |
| morganfainberg | But I think I know how I want to solve this for you. | 02:35 |
| morganfainberg | In dogpile that is. | 02:35 |
| *** browne has quit IRC | 02:37 | |
| openstackgerrit | Lance Bragstad proposed openstack/keystonemiddleware: Pull echo service out of auth_token. https://review.openstack.org/165171 | 02:42 |
| *** iamjarvo has joined #openstack-keystone | 02:44 | |
| jamielennox | really simple one if people could take a look: https://review.openstack.org/#/c/167162/ | 02:46 |
| jamielennox | +3, -1 including test changes | 02:47 |
| *** dims has joined #openstack-keystone | 02:51 | |
| jamielennox | bknudson1: i get what you mean on https://review.openstack.org/#/c/169947/ but i think we need to up the default somewhat | 02:54 |
| *** dims has quit IRC | 02:57 | |
| *** richm has quit IRC | 03:04 | |
| SpamapS | morganfainberg: http://paste.ubuntu.com/10769407/ <-- the ??? is the part that stumps me, unless we just require passing the object in. | 03:04 |
| morganfainberg | so, when you use invalidate you need to pass self in, e.g. https://github.com/openstack/keystone/blob/master/keystone/assignment/core.py#L966 | 03:05 |
| morganfainberg | because the decorator is actually part of the region object | 03:06 |
| morganfainberg | and the local object is part of the key | 03:06 |
| morganfainberg | it gets a bit weird | 03:06 |
| morganfainberg | but it makes sense | 03:06 |
| morganfainberg | in a pythonic way | 03:06 |
| morganfainberg | like i said i think i have a better approach | 03:06 |
| morganfainberg | but it isn't something i can solve in dogpile w/o a release and chatting with mike bayer | 03:07 |
| SpamapS | morganfainberg: ok so I'm not cray cray on that either. OK. | 03:07 |
| morganfainberg | nope | 03:07 |
| SpamapS | morganfainberg: I'm quite happy to do leg work on that if it would help you. | 03:08 |
| morganfainberg | SpamapS, oh sure. let me mock it up though mocking it up likely will be most of the work :P | 03:08 |
| morganfainberg | i try not to pull a mordred when possible | 03:08 |
| SpamapS | morganfainberg: missing API's lead to frustration, frustration leads to anger, and anger leads to NIH | 03:08 |
| stevemar | jamielennox, done | 03:15 |
| jamielennox | stevemar: thank you - just need to prompt to get the little ones in | 03:16 |
| stevemar | np at all | 03:16 |
| jamielennox | stevemar: do you feel comfortable reviewing https://review.openstack.org/#/c/162529/ | 03:17 |
| stevemar | jamielennox, i've been meaning to look at that one | 03:18 |
| *** boris-42 has quit IRC | 03:18 | |
| stevemar | it'll be a good excuse to delay making my slide decks | 03:19 |
| *** alex_xu has quit IRC | 03:20 | |
| *** alex_xu has joined #openstack-keystone | 03:21 | |
| mordred | morganfainberg: you're one of the dogpile authors? | 03:24 |
| morganfainberg | mordred, i am a contributor | 03:24 |
| mordred | morganfainberg: neat! | 03:24 |
| morganfainberg | mordred, mike bayer (SQLAlchemy) is the author | 03:24 |
| mordred | morganfainberg: then, we may want to have a larger conversation about stupid things I'm doing with cache keys to | 03:25 |
| mordred | too | 03:25 |
| morganfainberg | i gave up and just let dogpile to it's own cache key management | 03:26 |
| mordred | well ... I _Would_ ... | 03:26 |
| morganfainberg | in keystone | 03:26 |
| mordred | except that I will have more than one OpenStackCloud object | 03:26 |
| morganfainberg | right | 03:26 |
| mordred | each with its own, let's say for instance "list_flavors()" method | 03:26 |
| mordred | except - those flavors are not the same | 03:26 |
| lifeless | compositekey time | 03:27 |
| mordred | because it's different clouds | 03:27 |
| morganfainberg | oh what your usecase is makes tons of sense to me | 03:27 |
| morganfainberg | you could use the namespace or your own keygen for it | 03:27 |
| mordred | lifeless: you pop up like a magical bunny, you know that | 03:27 |
| lifeless | mordred: its knack | 03:27 |
| lifeless | mordred: having provoked pip 6.1 | 03:27 |
| mordred | morganfainberg: I looked at namepsace, but it seemed heavy-weight per call-site | 03:27 |
| mordred | lifeless: :) | 03:27 |
| mordred | morganfainberg: so I made my own keygen - but a) wow what a strange API And b) wanted to make sure I wasn't bonghitsing | 03:28 |
| morganfainberg | the keygen is very odd | 03:28 |
| morganfainberg | it's why i stopped trying to write my own | 03:28 |
| mordred | morganfainberg: http://git.openstack.org/cgit/openstack-infra/shade/tree/shade/__init__.py#n263 | 03:28 |
| mordred | I _think_ I may also need to add self.region into the key too | 03:29 |
| morganfainberg | you will likely need to | 03:29 |
| mordred | I wish there was an easy way to add a prefex and then defer to the dogpile keygen | 03:29 |
| morganfainberg | you also probably want to sha that resulting string | 03:29 |
| mordred | the make_cache_key in dogpile didn't? | 03:30 |
| morganfainberg | looks like you just join it | 03:30 |
| lifeless | morganfainberg: why sha specifically? [sha is kindof expensive] | 03:30 |
| morganfainberg | lifeless, because it uses the full argstring of the all the arguments | 03:31 |
| morganfainberg | which is why kwargs are banned in the default impl | 03:31 |
| morganfainberg | lifeless, some (read: memcache) backends are stupid and can't hand;e long strings as keys | 03:31 |
| lifeless | morganfainberg: no, I mean - like, md5 ok? or even a weaker hash ? | 03:31 |
| morganfainberg | oh doesn't matter what you use | 03:31 |
| morganfainberg | hash it | 03:31 |
| mordred | ah | 03:31 |
| morganfainberg | if you don't mind the collision issues, md5 | 03:32 |
| morganfainberg | it's all fine | 03:32 |
| lifeless | morganfainberg: md5 has no collision issues for this domain | 03:32 |
| mordred | so "hash it because there are stupid backends that have length issues" not "hash it because there is a part of the API I didn't catch" | 03:32 |
| * mordred groks | 03:32 | |
| lifeless | morganfainberg: sure :) | 03:32 |
| morganfainberg | lifeless, true. i meant at a "cause $security$ people look at this and bitch about any use of sha1 or worse" | 03:33 |
| morganfainberg | mordred, yep | 03:33 |
| mordred | cool | 03:33 |
| morganfainberg | mordred, we hit the key-length issue in keystone a number of times. | 03:33 |
| lifeless | morganfainberg: heh :). Security folk that don't consider context need edumacation. | 03:33 |
| * mordred sics lifeless on the unedumacated security folk | 03:33 | |
| morganfainberg | lifeless, i wanted to use sha1 for an id that doesnt' need to be crypto secure, just to obscure data in a mostly-unique way | 03:33 |
| mordred | morganfainberg: this was helpful! I like asking questions where I get concrete answers aback | 03:33 |
| morganfainberg | got told it was no fly, gave up and let people just do sha256 and suffer | 03:34 |
| morganfainberg | lifeless, where collision makes zero difference... | 03:34 |
| morganfainberg | mordred :) | 03:34 |
| lifeless | morganfainberg: desk . head () | 03:34 |
| morganfainberg | mordred, i had a way to disable hashing of the keys when i did the initial impl in keystone. it was super valuable to debug | 03:35 |
| mordred | morganfainberg: that's a good idea - perhaps I should tie that to self.debug=True | 03:35 |
| morganfainberg | mordred, but once you get your head wrapped around how dogpile works it's not as important | 03:35 |
| lifeless | morganfainberg: oh, I'm reminded - that IdP php impl - does it need reviewing? | 03:35 |
| lifeless | morganfainberg: or did I escape? | 03:35 |
| morganfainberg | lifeless, you escaped | 03:36 |
| morganfainberg | lifeless, it looks fine, it doesn't make stupid assumptions | 03:36 |
| morganfainberg | we should just endeavor to not carry our own impl | 03:36 |
| morganfainberg | long term | 03:36 |
| lifeless | morganfainberg: so we're going to maintain it ? :( | 03:36 |
| morganfainberg | nah, im harping on the "lets not maintain it" train | 03:36 |
| lifeless | I hope you win | 03:36 |
| lifeless | since OpenStack is currently organised as a project creation engine | 03:37 |
| lifeless | I fear | 03:37 |
| morganfainberg | but ripping it out and replacing it now is unlikely w/o jeapordizing other people's plans | 03:37 |
| morganfainberg | for some testing that needs it | 03:37 |
| *** adrian_otto has joined #openstack-keystone | 03:37 | |
| morganfainberg | but folks like cinerama et al working on the infra things agree | 03:37 |
| morganfainberg | we should work on finding an alternative before too long | 03:37 |
| morganfainberg | the issue is, we can't just turn it off. | 03:38 |
| morganfainberg | but alternatives/replacements are on the table, then we pivot over to new thing | 03:38 |
| * morganfainberg hopes he didn't just put words in stephanie's mouth... i think she said she agreed | 03:39 | |
| mordred | yeah - the "keep the train moving" right now is more important than the "kill with knives" part | 03:39 |
| mordred | but I think most everyone agrees | 03:39 |
| morganfainberg | mordred so the way i want dogpile to work is like abcMEta | 03:40 |
| morganfainberg | apply a metaclass | 03:40 |
| morganfainberg | decorate with the cool @memoize decorator | 03:41 |
| lifeless | morganfainberg: so that we can have two problems? | 03:41 |
| morganfainberg | which tracks the important stuff | 03:41 |
| morganfainberg | mordred, then you call <obj>.configure_cache | 03:41 |
| *** tqtran_ has quit IRC | 03:41 | |
| morganfainberg | mordred, and if you nver configure the cache it doesn't explode like dogpile does today | 03:41 |
| morganfainberg | lifeless, we already have caching... we already have 2 problems | 03:41 |
| morganfainberg | lifeless, this is just making the caching less painful to consume | 03:42 |
| lifeless | morganfainberg: its more that I rather suspect metaclasses are massively overused. | 03:42 |
| morganfainberg | lifeless, this is a case, like with abstract base, it is warranted | 03:42 |
| morganfainberg | due to the desire to memoize via decorator (one of the few cases memoization really makes sense) | 03:43 |
| morganfainberg | lifeless, since decorators are import time | 03:43 |
| lifeless | morganfainberg: mmm. I'd need to look rather closer to really make an assessment. | 03:43 |
| morganfainberg | lifeless, currently you need to instantiate the region object at import to use the decorator | 03:43 |
| morganfainberg | lifeless, if you want per-object cache region (aka i have 5 clouds, each has different cache data) | 03:43 |
| morganfainberg | lifeless, and potentially configs | 03:44 |
| morganfainberg | lifeless, you can't pre-make the regions, you're either metaprogramming the objects or the regions. | 03:44 |
| lifeless | morganfainberg: you'll need to point me at the decorator, I can't see in abstract the connection to import time | 03:44 |
| lifeless | [yes, I know decorators etc....] | 03:44 |
| morganfainberg | sec | 03:44 |
| morganfainberg | lifeless, https://bitbucket.org/zzzeek/dogpile.cache/src/229615be466d00c9c135a90d8965679ab3e4edaa/dogpile/cache/region.py?at=master#cl-839 | 03:45 |
| morganfainberg | lifeless, i have other options. | 03:45 |
| morganfainberg | you could context-manager or write your own memoizer that just uses key-value, this does have the ability to use async data creation with windowed continued cache validation (cache is stale at 10 seconds, invalid at 15, so async runner gets data at 10s vs. blocking calls and does the update before 15) | 03:46 |
| morganfainberg | so, depends on what you want. | 03:47 |
| lifeless | this comment https://bitbucket.org/zzzeek/dogpile.cache/src/229615be466d00c9c135a90d8965679ab3e4edaa/dogpile/cache/region.py?at=master#cl-953 is wrong | 03:47 |
| lifeless | you can lookup the context in the call stack at decorator invocation time | 03:47 |
| lifeless | its a little terrifying of course | 03:47 |
| morganfainberg | oh of course | 03:47 |
| morganfainberg | there are improvements that could be made | 03:47 |
| morganfainberg | easily | 03:47 |
| morganfainberg | i also try not to dive into call stack inspection when possible | 03:48 |
| morganfainberg | it seems overkill | 03:48 |
| lifeless | ok, so the problem statement | 03:48 |
| lifeless | I have a class X | 03:48 |
| lifeless | with a method Y | 03:48 |
| lifeless | and I have M instances of X | 03:48 |
| morganfainberg | yes | 03:49 |
| lifeless | and I want a distinct dogpile cache for each i in M on method Y | 03:49 |
| morganfainberg | yep | 03:49 |
| mordred | I _may_ want - just to be clear | 03:49 |
| morganfainberg | mordred, there are other options. | 03:49 |
| mordred | yah | 03:49 |
| lifeless | and the cache needs to be named - so that it coalesces between different processes,including successors to the same process | 03:49 |
| stevemar | morganfainberg, we upset folks in -chef, by removing xml | 03:49 |
| morganfainberg | lifeless, yep. | 03:49 |
| stevemar | i think we probably need to communicate our changes better to the puppet and chef folks | 03:50 |
| morganfainberg | stevemar, that ship has kind of sailed... | 03:50 |
| lifeless | morganfainberg: ok, so how will metaclasses help that? | 03:50 |
| stevemar | we screwed them over a few times now | 03:50 |
| morganfainberg | stevemar, we've been trying ot be better :( | 03:50 |
| morganfainberg | lifeless, so the metaclass like abstractmeta | 03:50 |
| lifeless | morganfainberg: I can tell you want comes to my mind, if you like [its not metaclasses :)] | 03:50 |
| morganfainberg | lifeless, would track what methods are memoized | 03:50 |
| morganfainberg | when you instantiate the class *if* you configure the cache, you get memoization on your resulting methods | 03:51 |
| morganfainberg | since you decorated with @memoize | 03:51 |
| morganfainberg | lifeless, that is one of the options i was thinking. | 03:51 |
| morganfainberg | lifeless, the other is write a non-instance bound decorator that lets you use whatever region you want | 03:52 |
| lifeless | morganfainberg: what about letting the object define the region? | 03:52 |
| morganfainberg | lifeless, the metaclass is mostly that, just a wrapper for it. it's the same basic mechanism with how @abstractmethod registers on the underlying class | 03:53 |
| morganfainberg | so you could look at self.__memoizedmethods__ and know if you needed to wrap/not wrap it w/ region logic | 03:53 |
| lifeless | morganfainberg: I mean the decorator | 03:54 |
| lifeless | @memoize(region='attributename') | 03:54 |
| morganfainberg | and extract it from the self | 03:54 |
| lifeless | metaclasses are very one-shot. Its hard to combine them safely. | 03:55 |
| lifeless | So I look for a real must-use case for them | 03:55 |
| morganfainberg | anyway in py 2 this would actually probably break anyway | 03:55 |
| lifeless | a variant on the above - you could define a protocol | 03:55 |
| lifeless | __dogpile_region__ | 03:56 |
| *** alexsyip has joined #openstack-keystone | 03:56 | |
| lifeless | and do getattr(self, '__dogpile_region__', None) | 03:56 |
| lifeless | [or spelling variations thereon] | 03:56 |
| mordred | stevemar: to be fair - the xml stuff was talked about TO DEATH for FOREVER | 03:56 |
| morganfainberg | lifeless, i think we're mostly in agreement with the end result. just minor where does this implement go is different | 03:56 |
| lifeless | morganfainberg: sure :) | 03:56 |
| lifeless | morganfainberg: just kibbitzing | 03:57 |
| morganfainberg | lifeless, end result = same exact thing :) | 03:57 |
| morganfainberg | lifeless, i really think we are effectively describing the same thing. | 03:57 |
| lifeless | morganfainberg: well, not if there is a metaclass involved :) | 03:57 |
| morganfainberg | lifeless, and we'd end up at the same solution long term | 03:57 |
| morganfainberg | not a metaclass, and something that is vaguely metaprogramming in small bits | 03:57 |
| morganfainberg | vs. strict getattr() | 03:57 |
| morganfainberg | mordred, stevemar, XML has been slowly being deprecated for how many cycles? | 04:01 |
| morganfainberg | and how much did we say "this is going away" "no really" | 04:01 |
| morganfainberg | "seriously guys" | 04:01 |
| morganfainberg | "it's gone. no this is a stub it is going to die" | 04:01 |
| morganfainberg | lifeless, i think the real answer is "don't make your decorators instance bound or you'll have a bad time" | 04:02 |
| mordred | also - what were the chef folks using xml for? | 04:02 |
| morganfainberg | wondering if it's just the paste pipeline removal | 04:03 |
| morganfainberg | and breaking things that way | 04:03 |
| lifeless | morganfainberg: hmm, I think instance bound is fine, you just need to define a clear protocol for looking up context | 04:03 |
| morganfainberg | lifeless, in this case instance bound doesn't really buy you anything except the issue. | 04:04 |
| morganfainberg | lifeless, all the rest of the logic could be from the defined "this is my region object" | 04:04 |
| morganfainberg | lifeless, protocol as you said. | 04:04 |
| morganfainberg | lifeless, this case is like region is meant to *almost* be a singleton for a given backend configuration.... | 04:05 |
| morganfainberg | which is not what a singleton is... but its what it almost feels like is wanted. | 04:05 |
| lifeless | indeed | 04:06 |
| lifeless | you want to namespace stuff dynamically within the one cache, AIUI | 04:06 |
| lifeless | the namespace parameter to the region sourced decorator is insufficient because its not dynamic | 04:06 |
| openstackgerrit | Jamie Lennox proposed openstack/python-keystoneclient: Specify default_domain to generic plugin https://review.openstack.org/171462 | 04:07 |
| jamielennox | stevemar: probably want ^ for OSC | 04:07 |
| *** rwsu has quit IRC | 04:07 | |
| morganfainberg | lifeless, solution do the same thing you do with should_cache_fn or expiration_time | 04:07 |
| morganfainberg | if namespace is_callable() | 04:07 |
| morganfainberg | call it | 04:07 |
| morganfainberg | vs. static string | 04:07 |
| lifeless | morganfainberg: that would work too. I was proposing a protocol as a way to avoid redundant things | 04:08 |
| lifeless | say you have two methods A and B on class T. | 04:08 |
| lifeless | they both need to namespace the same way | 04:08 |
| morganfainberg | lifeless, sure. the other issue i have is that the memoized decorator explodes with "OMG E_NOT_CONFIGURED" if the region hasn't been configured | 04:08 |
| morganfainberg | and reconfiguring a region... god just don't | 04:08 |
| lifeless | well | 04:08 |
| lifeless | that seems trivial | 04:08 |
| lifeless | if not configured don't cache | 04:08 |
| morganfainberg | its a rats nest i haven't had time to propose a PR for | 04:09 |
| morganfainberg | lifeless, exactly | 04:09 |
| morganfainberg | lifeless, it's a few layers of things that could easily be solved. | 04:09 |
| morganfainberg | lifeless, oh for your case of the A and B on T - i just use a generator to make the function for me. | 04:10 |
| morganfainberg | it's a minor metaprogramming thing | 04:10 |
| morganfainberg | depends on how much extra mechanism you want [protocol] in dogpile. | 04:10 |
| lifeless | morganfainberg: sure, OTOH that then becomes a thing folk need to lookup when reading the code using it | 04:10 |
| morganfainberg | lifeless, right. | 04:11 |
| *** erkules has joined #openstack-keystone | 04:11 | |
| openstackgerrit | Jamie Lennox proposed openstack/python-keystoneclient: Allow saving and caching the plugin auth state https://review.openstack.org/149175 | 04:12 |
| morganfainberg | ok this was a fun convo | 04:13 |
| morganfainberg | :) | 04:13 |
| openstackgerrit | Merged openstack/python-keystoneclient: Return None for missing trust_id in fixture https://review.openstack.org/167162 | 04:13 |
| *** _cjones_ has joined #openstack-keystone | 04:19 | |
| openstackgerrit | Jamie Lennox proposed openstack/python-keystoneclient: Make tests run against original client and session https://review.openstack.org/117089 | 04:19 |
| *** _cjones_ has quit IRC | 04:19 | |
| *** _cjones_ has joined #openstack-keystone | 04:20 | |
| *** samueldmq has quit IRC | 04:28 | |
| *** _cjones_ has quit IRC | 04:28 | |
| jamielennox | morganfainberg: do you want to pull rank on https://review.openstack.org/#/c/169947/ | 04:33 |
| morganfainberg | oh that one? | 04:33 |
| morganfainberg | sec | 04:33 |
| morganfainberg | jamielennox, i would like to see it configurable with a more sane default than 120 | 04:34 |
| stevemar | morganfainberg, mordred fair enough, we did talk about removing XML to death | 04:34 |
| morganfainberg | erm 1* | 04:34 |
| jamielennox | it's not perfect but i think it will cut down some edge cases | 04:34 |
| stevemar | and yeah, it was in regards to the paste pipeline | 04:34 |
| morganfainberg | jamielennox, i'd be ok with this as is if you have a followup task/bug/RFE filed to make it configurable | 04:34 |
| jamielennox | morganfainberg: the reason it's a constant is so that it was overridable from a subclass, i just don't really know any time where it's acceptable to have a user decide how long a token they need | 04:35 |
| stevemar | i'll try to see what we can do about making the communication better | 04:35 |
| morganfainberg | jamielennox, ah | 04:35 |
| jamielennox | morganfainberg: also, proposed https://review.openstack.org/#/c/171448/ just to see how hard it was, once it's been done once i think we could port the rest over fairly quickly (ignoring the migrations) | 04:38 |
| jamielennox | i'm thinking that the enabled flag might not be worth it | 04:39 |
| *** iamjarvo has quit IRC | 04:40 | |
| morganfainberg | jamielennox, commented on the 120s review | 04:42 |
| morganfainberg | +2 | 04:42 |
| morganfainberg | we should check with bknudson1 on this. | 04:42 |
| stevemar | jamielennox, i'm not pro or anti the enabled flag really | 04:42 |
| jamielennox | stevemar: i can see it being useful for new features, but like the trust enabled flag is still there and you just shouldn't be running without them at this point | 04:43 |
| jamielennox | i'm not sure how we ever remove these options once they're in, so for features that have been there for a cycle or two just run it - in endpoint filterings case we should have just made it part of core | 04:44 |
| morganfainberg | i'd like to see the enabled flag go away fro anything that is "stable" | 04:45 |
| morganfainberg | does anyone really turn trusts off? | 04:45 |
| morganfainberg | you could do the same thing with policy | 04:45 |
| jamielennox | i'd expect weird breakages | 04:45 |
| jamielennox | well not so weird i guess | 04:45 |
| morganfainberg | which is where i'd like to see that kind of disable for "stable" features happen | 04:45 |
| openstackgerrit | Merged openstack/keystone: Work with pymongo 3.0 https://review.openstack.org/171442 | 04:45 |
| morganfainberg | if that makes sense. | 04:45 |
| stevemar | finally that thing merged | 04:45 |
| morganfainberg | jamielennox, endpoint filtering should just be core. in fact... there is a review up for merging the drivers at least | 04:46 |
| *** lhcheng has joined #openstack-keystone | 04:46 | |
| *** ChanServ sets mode: +v lhcheng | 04:46 | |
| jamielennox | morganfainberg: not so much the point for this review, i just picked one that was fairly small and isolated | 04:47 |
| morganfainberg | right | 04:47 |
| stevemar | what are we going to do about the routes? | 04:47 |
| jamielennox | stevemar: ? | 04:47 |
| stevemar | api paths | 04:47 |
| morganfainberg | stevemar, pecan | 04:47 |
| stevemar | will we drop the OS-ENDPOINT-FILTERING part? | 04:47 |
| stevemar | or support both for 1 release | 04:48 |
| morganfainberg | we can't drop that | 04:48 |
| morganfainberg | unfortunately | 04:48 |
| jamielennox | stevemar: not a part of this review/blueprint | 04:48 |
| morganfainberg | and what jamielennox said | 04:48 |
| morganfainberg | when contrib is dead | 04:48 |
| morganfainberg | we can talk about the OS-<whatever> | 04:48 |
| morganfainberg | stuff | 04:48 |
| jamielennox | stevemar: i'm just trying to move it into the "application" rather than be a middleware piece | 04:48 |
| jamielennox | ideally jsonhome would help here - though it's probably got OS- in it as well | 04:49 |
| morganfainberg | stevemar, the API route doesn't matter for the "this isn't an extension" things | 04:50 |
| stevemar | mordred, i figured, was just a question :) | 04:50 |
| morganfainberg | stevemar, we could call new stuff whatever we want. | 04:50 |
| morganfainberg | mordred, Aha! my time hash come. someone tab-failed to your name not mine ;) | 04:51 |
| stevemar | gah! | 04:52 |
| stevemar | indeed | 04:52 |
| *** adrian_otto has left #openstack-keystone | 05:17 | |
| *** gyee has quit IRC | 05:23 | |
| morganfainberg | jamielennox, i'll plan to talk to bknudson1 tomorrow re: token minimum. | 05:24 |
| *** spandhe has quit IRC | 05:32 | |
| SpamapS | morganfainberg: https://review.openstack.org/171456 <-- success | 05:32 |
| *** rushiagr_away is now known as rushiagr | 05:33 | |
| morganfainberg | SpamapS, awesome! i would probably memoize the wrap itself as well (do it once and store the resultant function to be recalled via a dict or something) vs. wrap every time | 05:36 |
| morganfainberg | depending on how much you invalidate | 05:37 |
| morganfainberg | SpamapS, looks good, glad it works [though the api is a bit weird to work "around" in this case] | 05:37 |
| *** spandhe has joined #openstack-keystone | 05:37 | |
| *** dims has joined #openstack-keystone | 05:41 | |
| *** dims has quit IRC | 05:47 | |
| *** lhcheng has quit IRC | 05:48 | |
| *** spandhe has quit IRC | 05:48 | |
| bigjools | am I on a hiding to nothing if I try to get federated logins working on icehouse? | 05:54 |
| *** stevemar has quit IRC | 05:54 | |
| *** pnavarro has joined #openstack-keystone | 06:07 | |
| *** ParsectiX has joined #openstack-keystone | 06:12 | |
| *** abhishekk has joined #openstack-keystone | 06:23 | |
| abhishekk | Hi, is eventlet is deprecated from keystone? | 06:25 |
| morganfainberg | abhishekk, yes. | 06:25 |
| abhishekk | is it not used from kilo onwards or it is optional for service provider to use or not | 06:25 |
| morganfainberg | it is optional. but will be removed in M-release | 06:26 |
| abhishekk | morganfainberg: thanks for the update, what about stable branches? | 06:26 |
| morganfainberg | abhishekk, stable branches remain as they are | 06:26 |
| abhishekk | morganfainberg: then should we fix the eventlet issue for kilo and stable branches, I am reffering to your comment on https://bugs.launchpad.net/nova/+bug/1361360/comments/79 | 06:27 |
| openstack | Launchpad bug 1361360 in OpenStack Compute (nova) icehouse "Eventlet green threads not released back to the pool leading to choking of new requests" [High,In progress] - Assigned to Abhishek Kekane (abhishek-kekane) | 06:27 |
| morganfainberg | abhishekk, so icehouse, juno will not warn on startup using eventlet, kilo and liberty will warn, M-release will no longer be able to use eventlet | 06:27 |
| morganfainberg | abhishekk, nope | 06:27 |
| morganfainberg | abhishekk, well i'd rather not | 06:28 |
| morganfainberg | abhishekk, i think the answer was just to turn off keepalives | 06:28 |
| *** pnavarro has quit IRC | 06:29 | |
| morganfainberg | abhishekk, the long answer was don't run keystone in eventlet | 06:29 |
| abhishekk | morganfainberg: but if any service provider wants to use eventlet with juno, icehouse or kilo then? | 06:30 |
| morganfainberg | abhishekk, isn't the response to disable keepalives? | 06:30 |
| morganfainberg | and thats it right? | 06:30 |
| abhishekk | morganfainberg: yes | 06:31 |
| morganfainberg | and it was a "configurable" fix | 06:31 |
| morganfainberg | right? | 06:31 |
| abhishekk | morganfainberg: yes | 06:31 |
| abhishekk | it is configurable to keep backward compatibility | 06:32 |
| morganfainberg | we will need to discuss later it's 2300 | 06:32 |
| morganfainberg | i am inclined to say kilo and later we will simply recommend use mod_wsgi in all cases. | 06:33 |
| abhishekk | morganfainberg: sorry, and thanks for the input | 06:33 |
| morganfainberg | eventlet has many many other issues such as not being able to configure ssl versions and ciphers | 06:33 |
| morganfainberg | for ssl | 06:33 |
| abhishekk | morganfainberg: yes, i know | 06:34 |
| morganfainberg | but i can be convinced to allow the fix for kilo RC. | 06:34 |
| morganfainberg | just not right now when i'm almost alseep ;) | 06:34 |
| morganfainberg | and if kilo RC then liberty | 06:34 |
| morganfainberg | but the life of the fix is very limited. | 06:34 |
| abhishekk | morganfainberg: thanks for your time and inputs | 06:34 |
| morganfainberg | np | 06:34 |
| abhishekk | morganfainberg: good night | 06:35 |
| morganfainberg | lets talk again when we open RC2 window | 06:35 |
| abhishekk | morganfainberg: sure | 06:35 |
| *** mflobo has quit IRC | 06:35 | |
| morganfainberg | abhishekk, which time zone are you in? | 06:35 |
| abhishekk | morganifainberg: I am from India | 06:35 |
| morganfainberg | ok | 06:35 |
| morganfainberg | so ~+5hrs from UTC? | 06:36 |
| *** mflobo has joined #openstack-keystone | 06:36 | |
| abhishekk | morganfainberg: yes | 06:36 |
| morganfainberg | almost 12 hours off from me. ok. cool. | 06:36 |
| abhishekk | right | 06:36 |
| morganfainberg | i'll plan to catch you early my time or late my time when we get close to RC2 | 06:36 |
| morganfainberg | if i don't, don't hesitate to bug me and remind me :) | 06:37 |
| abhishekk | sure, I will | 06:37 |
| morganfainberg | cool. | 06:37 |
| abhishekk | :) | 06:37 |
| morganfainberg | thanks! | 06:37 |
| abhishekk | thank you | 06:37 |
| morganfainberg | have a good day! :) | 06:37 |
| * morganfainberg clocks out for the night. | 06:37 | |
| *** krykowski has joined #openstack-keystone | 06:37 | |
| abhishekk | thank you :) | 06:37 |
| *** lhcheng has joined #openstack-keystone | 06:38 | |
| *** ChanServ sets mode: +v lhcheng | 06:38 | |
| *** mflobo has quit IRC | 06:38 | |
| *** mflobo has joined #openstack-keystone | 06:40 | |
| openstackgerrit | Merged openstack/keystone: Update man pages for the Kilo release https://review.openstack.org/171327 | 06:41 |
| *** abhishekk has left #openstack-keystone | 06:47 | |
| openstackgerrit | Merged openstack/python-keystoneclient: Provide a generic auth plugin loader https://review.openstack.org/162529 | 06:47 |
| openstackgerrit | Merged openstack/python-keystoneclient: Provide a generic auth plugin loader https://review.openstack.org/162529 | 06:47 |
| *** tobberydberg has joined #openstack-keystone | 06:54 | |
| *** jaosorior has joined #openstack-keystone | 06:56 | |
| *** erkules has quit IRC | 06:58 | |
| *** erkules has joined #openstack-keystone | 06:58 | |
| *** pnavarro has joined #openstack-keystone | 07:06 | |
| *** alexsyip has quit IRC | 07:11 | |
| *** fhubik_afk has joined #openstack-keystone | 07:19 | |
| *** afazekas has joined #openstack-keystone | 07:23 | |
| *** pnavarro has quit IRC | 07:26 | |
| *** pnavarro has joined #openstack-keystone | 07:27 | |
| *** lhcheng has quit IRC | 07:28 | |
| *** lhcheng has joined #openstack-keystone | 07:30 | |
| *** ChanServ sets mode: +v lhcheng | 07:30 | |
| *** lhcheng has quit IRC | 07:34 | |
| *** pnavarro has quit IRC | 07:37 | |
| *** browne1 has quit IRC | 08:02 | |
| *** jistr has joined #openstack-keystone | 08:02 | |
| *** ajayaa has joined #openstack-keystone | 08:05 | |
| *** davechen has joined #openstack-keystone | 08:05 | |
| *** henrynash has joined #openstack-keystone | 08:06 | |
| *** ChanServ sets mode: +v henrynash | 08:06 | |
| *** pnavarro has joined #openstack-keystone | 08:12 | |
| openstackgerrit | Thomas Goirand proposed openstack/keystonemiddleware: Removes discover from test-reqs https://review.openstack.org/171516 | 08:13 |
| *** browne has joined #openstack-keystone | 08:15 | |
| *** browne has quit IRC | 08:29 | |
| *** ParsectiX_ has joined #openstack-keystone | 09:00 | |
| *** ParsectiX has quit IRC | 09:15 | |
| *** henrynash has quit IRC | 09:15 | |
| *** davidckennedy has joined #openstack-keystone | 09:16 | |
| *** dims has joined #openstack-keystone | 09:19 | |
| *** pnavarro has quit IRC | 09:21 | |
| *** dims has quit IRC | 09:24 | |
| *** pnavarro has joined #openstack-keystone | 09:43 | |
| openstackgerrit | Dave Chen proposed openstack/python-keystoneclient: Removes discover from test-reqs https://review.openstack.org/171540 | 09:45 |
| openstackgerrit | Dave Chen proposed openstack/keystone: Removes discover from test-reqs https://review.openstack.org/171542 | 09:48 |
| *** davechen has quit IRC | 09:54 | |
| *** rm_work has quit IRC | 10:04 | |
| *** ericksonfgds has quit IRC | 10:05 | |
| *** ericksonsantos has quit IRC | 10:05 | |
| *** htruta has quit IRC | 10:05 | |
| *** raildo|away has quit IRC | 10:06 | |
| *** ericksonsantos has joined #openstack-keystone | 10:06 | |
| *** rm_work has joined #openstack-keystone | 10:06 | |
| *** dims has joined #openstack-keystone | 10:08 | |
| *** raildo|away has joined #openstack-keystone | 10:09 | |
| *** htruta has joined #openstack-keystone | 10:09 | |
| *** pnavarro has quit IRC | 10:17 | |
| *** samueldmq has joined #openstack-keystone | 10:32 | |
| openstackgerrit | David Charles Kennedy proposed openstack/keystonemiddleware: Add Endpoint Enforcement to Keystonemiddleware https://review.openstack.org/153296 | 10:43 |
| openstackgerrit | David Charles Kennedy proposed openstack/keystonemiddleware: Add Endpoint Enforcement to Keystonemiddleware https://review.openstack.org/153296 | 10:45 |
| *** f13o has joined #openstack-keystone | 10:52 | |
| openstackgerrit | Dave Chen proposed openstack/keystone: Removes discover from test-reqs https://review.openstack.org/171542 | 11:31 |
| *** aix has joined #openstack-keystone | 11:42 | |
| *** Olena has quit IRC | 12:16 | |
| *** bknudson1 has quit IRC | 12:16 | |
| openstackgerrit | Samuel de Medeiros Queiroz proposed openstack/keystone: Exposes bug on role assignments creation https://review.openstack.org/171596 | 12:21 |
| *** boris-42 has joined #openstack-keystone | 12:23 | |
| openstackgerrit | Samuel de Medeiros Queiroz proposed openstack/keystone: Adds inherited column to RoleAssignment PK https://review.openstack.org/142472 | 12:23 |
| *** browne has joined #openstack-keystone | 12:32 | |
| *** raildo|away is now known as raildo | 12:34 | |
| *** ajayaa has quit IRC | 12:35 | |
| *** gordc has joined #openstack-keystone | 12:36 | |
| *** mattfarina has joined #openstack-keystone | 12:52 | |
| *** bknudson has joined #openstack-keystone | 13:00 | |
| *** ChanServ sets mode: +v bknudson | 13:00 | |
| *** davechen has joined #openstack-keystone | 13:05 | |
| *** edmondsw has joined #openstack-keystone | 13:12 | |
| *** nkinder has quit IRC | 13:12 | |
| *** lhcheng has joined #openstack-keystone | 13:14 | |
| *** ChanServ sets mode: +v lhcheng | 13:14 | |
| *** amakarov_away is now known as amakarov | 13:16 | |
| *** davechen has left #openstack-keystone | 13:16 | |
| *** stevemar has joined #openstack-keystone | 13:19 | |
| *** ChanServ sets mode: +v stevemar | 13:19 | |
| *** richm has joined #openstack-keystone | 13:20 | |
| *** mattfarina has quit IRC | 13:22 | |
| *** stevemar has quit IRC | 13:27 | |
| *** stevemar has joined #openstack-keystone | 13:28 | |
| *** ChanServ sets mode: +v stevemar | 13:28 | |
| *** lhcheng has quit IRC | 13:29 | |
| *** browne has quit IRC | 13:37 | |
| *** browne has joined #openstack-keystone | 13:41 | |
| *** rushiagr is now known as rushiagr_away | 13:46 | |
| openstackgerrit | Raildo Mascena de Sousa Filho proposed openstack/keystone: Add domain_id checking in create_project https://review.openstack.org/159944 | 13:50 |
| *** dims has quit IRC | 13:52 | |
| openstackgerrit | Victor Sergeyev proposed openstack/keystone: POC: Run SQL migration tests on PostgreSQL and MySQL https://review.openstack.org/171115 | 13:53 |
| *** dims has joined #openstack-keystone | 13:53 | |
| *** ajayaa has joined #openstack-keystone | 13:53 | |
| *** jistr has quit IRC | 13:54 | |
| *** jistr has joined #openstack-keystone | 13:54 | |
| *** aix has quit IRC | 13:56 | |
| *** chlong has quit IRC | 13:58 | |
| *** fhubik_afk has quit IRC | 14:01 | |
| *** nkinder has joined #openstack-keystone | 14:02 | |
| *** browne has quit IRC | 14:04 | |
| *** henrynash has joined #openstack-keystone | 14:08 | |
| *** ChanServ sets mode: +v henrynash | 14:08 | |
| *** jistr has quit IRC | 14:10 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 14:11 | |
| *** jistr has joined #openstack-keystone | 14:12 | |
| *** tobberydberg has quit IRC | 14:13 | |
| openstackgerrit | Raildo Mascena de Sousa Filho proposed openstack/keystone: Add is_domain field in Project Table https://review.openstack.org/157427 | 14:21 |
| openstackgerrit | Raildo Mascena de Sousa Filho proposed openstack/keystone: Change domain_id FK in project table https://review.openstack.org/166354 | 14:21 |
| openstackgerrit | Raildo Mascena de Sousa Filho proposed openstack/keystone: Bye Bye Domain Table https://review.openstack.org/161854 | 14:21 |
| openstackgerrit | Raildo Mascena de Sousa Filho proposed openstack/keystone: Honor domain operations in project table https://review.openstack.org/143763 | 14:21 |
| openstackgerrit | Raildo Mascena de Sousa Filho proposed openstack/keystone: Stop calling domain drivers https://review.openstack.org/165936 | 14:21 |
| openstackgerrit | Raildo Mascena de Sousa Filho proposed openstack/keystone: Change project name constraint https://review.openstack.org/158372 | 14:21 |
| openstackgerrit | Raildo Mascena de Sousa Filho proposed openstack/keystone: Creation of subdomains and filtering by parent_id https://review.openstack.org/161378 | 14:22 |
| openstackgerrit | Raildo Mascena de Sousa Filho proposed openstack/keystone: List projects filtering by is_domain flag https://review.openstack.org/158398 | 14:22 |
| *** aix has joined #openstack-keystone | 14:24 | |
| *** zzzeek has joined #openstack-keystone | 14:28 | |
| *** rushiagr_away is now known as rushiagr | 14:32 | |
| *** browne has joined #openstack-keystone | 14:34 | |
| *** davidckennedy has quit IRC | 14:34 | |
| *** stevemar has quit IRC | 14:34 | |
| *** henrynash has quit IRC | 14:39 | |
| *** mattfarina has joined #openstack-keystone | 14:42 | |
| *** jistr has quit IRC | 14:45 | |
| *** topol has joined #openstack-keystone | 14:45 | |
| *** ChanServ sets mode: +v topol | 14:45 | |
| *** topol_ has joined #openstack-keystone | 14:46 | |
| *** ChanServ sets mode: +v topol_ | 14:46 | |
| *** topol has quit IRC | 14:50 | |
| *** topol_ is now known as topol | 14:50 | |
| *** jistr has joined #openstack-keystone | 14:51 | |
| *** ParsectiX has joined #openstack-keystone | 14:58 | |
| *** ayoung has joined #openstack-keystone | 15:08 | |
| *** ChanServ sets mode: +v ayoung | 15:08 | |
| *** stevemar has joined #openstack-keystone | 15:13 | |
| *** ChanServ sets mode: +v stevemar | 15:13 | |
| *** jistr is now known as jistr|mtg | 15:16 | |
| *** ParsectiX has quit IRC | 15:17 | |
| *** _cjones_ has joined #openstack-keystone | 15:20 | |
| *** _cjones_ has quit IRC | 15:24 | |
| *** jistr|mtg is now known as jistr | 15:30 | |
| *** thedodd has joined #openstack-keystone | 15:34 | |
| *** rwsu has joined #openstack-keystone | 15:37 | |
| *** topol has quit IRC | 15:39 | |
| *** topol has joined #openstack-keystone | 15:40 | |
| *** ChanServ sets mode: +v topol | 15:40 | |
| *** afazekas has quit IRC | 15:43 | |
| ayoung | nkinder, I want to add ispsilon to my public Federation Demo. Do I still need all of the repos etc from https://github.com/nkinder/rdo-vm-factory/blob/master/rdo-federation-setup/vm-post-cloud-init-ipa.sh | 15:43 |
| nkinder | ayoung: no, it's been simplified | 15:44 |
| *** ozialien has joined #openstack-keystone | 15:44 | |
| nkinder | ayoung: I *think* I pushed a patch there last night... | 15:44 |
| ayoung | nkinder, so I don't need the coprs? | 15:45 |
| nkinder | ayoung: https://github.com/nkinder/rdo-vm-factory/blob/master/rdo-federation-setup/vm-post-cloud-init-ipa.sh#L22-L26 | 15:45 |
| nkinder | there are 2 copr repos only now | 15:45 |
| ayoung | nkinder I am having issues on the ipa server resolving outside...musthave messed up my resolvers setup | 15:46 |
| *** ozialien has quit IRC | 15:47 | |
| *** Ephur_ is now known as Ephur | 15:48 | |
| ayoung | Error: Package: ipsilon-0.5.0-1.el7.centos.noarch (nkinder-ipsilon) | 15:48 |
| ayoung | Requires: python-cherrypy | 15:48 |
| ayoung | nkinder, that is EPEL, no? | 15:48 |
| nkinder | ayoung: yes, epel is needed too (it's a few lines above) | 15:48 |
| nkinder | ayoung: for python-cherrypy and python-sqlalchemy | 15:49 |
| ayoung | nkinder, I have it, but can't seem to get those two... | 15:49 |
| nkinder | ayoung: you sure it's enabled? | 15:49 |
| ayoung | nkinder, something did this /etc/yum.repos.d/epel.repo.rpmnew | 15:49 |
| ayoung | with no replacement... | 15:50 |
| *** alex_xu has quit IRC | 15:50 | |
| *** alex_xu has joined #openstack-keystone | 15:50 | |
| openstackgerrit | Kamil Rykowski proposed openstack/keystone: Update openstack-common reference in openstack/common/README https://review.openstack.org/171706 | 15:51 |
| *** jistr has quit IRC | 15:56 | |
| *** _cjones_ has joined #openstack-keystone | 15:57 | |
| *** ozialien has joined #openstack-keystone | 15:59 | |
| dolphm | stevemar: poke me when you're around | 15:59 |
| *** ozialien has quit IRC | 16:01 | |
| *** carlosmarin has joined #openstack-keystone | 16:05 | |
| *** ozialien has joined #openstack-keystone | 16:05 | |
| *** ozialien has quit IRC | 16:06 | |
| *** pnavarro has joined #openstack-keystone | 16:08 | |
| *** alexsyip has joined #openstack-keystone | 16:10 | |
| openstackgerrit | Merged openstack/keystone-specs: Update path for listing a project's endpoint groups https://review.openstack.org/168427 | 16:17 |
| *** ozialien has joined #openstack-keystone | 16:19 | |
| *** gyee has joined #openstack-keystone | 16:20 | |
| *** ChanServ sets mode: +v gyee | 16:20 | |
| *** lhcheng has joined #openstack-keystone | 16:21 | |
| *** ChanServ sets mode: +v lhcheng | 16:21 | |
| openstackgerrit | Brant Knudson proposed openstack/python-keystoneclient: Use pkg_resources to get version https://review.openstack.org/171720 | 16:22 |
| *** ozialien has quit IRC | 16:23 | |
| *** krykowski has quit IRC | 16:23 | |
| *** pnavarro has quit IRC | 16:29 | |
| *** krykowski has joined #openstack-keystone | 16:30 | |
| *** tqtran has joined #openstack-keystone | 16:30 | |
| *** ayoung has quit IRC | 16:30 | |
| dolphm | morganfainberg: all stable builds failed this morning. who's around to maintain our mongo code? | 16:32 |
| morganfainberg | Yeah. | 16:34 |
| morganfainberg | dolphm: we have a fix that landed in master last night. From brant. But the mongo code is on the hitlist that if not maintained... | 16:35 |
| dolphm | morganfainberg: are there backports proposed? | 16:35 |
| morganfainberg | dolphm: this was actually pymongo beig sane and doing breaking changes in a major version bump | 16:35 |
| dolphm | morganfainberg: mongo support was a hit and run patch in the first place | 16:36 |
| morganfainberg | dolphm: yep. | 16:36 |
| *** haneef_ has joined #openstack-keystone | 16:36 | |
| morganfainberg | The two possible fixes: cap pymongo or backport. | 16:36 |
| morganfainberg | I was waiting for the change from brant to land to propose the backports (was late) | 16:36 |
| dolphm | morganfainberg: i'm looking at the code in master - if it's just the conditional, that could be backported | 16:37 |
| morganfainberg | Yep. | 16:37 |
| morganfainberg | That's it. | 16:37 |
| dolphm | master patch https://review.openstack.org/#/c/171442/ | 16:37 |
| morganfainberg | https://review.openstack.org/#/c/171442/ | 16:38 |
| morganfainberg | Yep | 16:38 |
| dolphm | morganfainberg: but, given that no one is testing mongo support, i'd make no promises about 3.0 actually working in any branch | 16:38 |
| dolphm | morganfainberg: so, i'd be inclined to pin stable branches instead | 16:38 |
| morganfainberg | There is also a cap in g-r posted by brant. | 16:38 |
| morganfainberg | I'd like to even cap master. | 16:38 |
| dolphm | morganfainberg: to stable requirements? | 16:38 |
| morganfainberg | Not yet. Again was a today when I was awake plan. | 16:39 |
| haneef_ | morganfainberg: comments on : https://bugs.launchpad.net/keystone/+bug/1441733 ( shouldn't this be included in install) | 16:39 |
| openstack | Launchpad bug 1441733 in Keystone "pip install or python setup.py install should include httpd/keystone.py" [Undecided,New] | 16:39 |
| morganfainberg | dolphm: was about to start proposing those. :) | 16:40 |
| morganfainberg | You asked about when I was getting started. | 16:40 |
| morganfainberg | haneef_: where should it be installed? | 16:40 |
| morganfainberg | haneef_: each distro is different | 16:40 |
| openstackgerrit | Raildo Mascena de Sousa Filho proposed openstack/keystone-specs: Crating Liberty repository for specs https://review.openstack.org/171733 | 16:41 |
| morganfainberg | Not sure the right place for that in Python land for consumption in httpd (I am saying yes, it make sense to include it, but I don't know where) | 16:41 |
| gyee | install it into a known location and let each distro do their symlinks | 16:42 |
| haneef_ | I agree with it, but if some one is doing pip install how can they get this file. I would be better move httpd directory under kesytone | 16:42 |
| gyee | morganfainberg, dolphm, I am fine with putting a pin in pymongo | 16:42 |
| morganfainberg | gyee: that was the starting point. The pymongo dogpile driver is not really maintained or used. We may want to consider dropping it. | 16:43 |
| gyee | dropping it in Liberty? | 16:44 |
| morganfainberg | deprecate in L, drop in M | 16:44 |
| morganfainberg | if no one is using it | 16:45 |
| morganfainberg | and i highly doubt someone is | 16:45 |
| gyee | let me double check with the folks here | 16:45 |
| gyee | s/here/HP/ | 16:45 |
| morganfainberg | right | 16:45 |
| morganfainberg | and if they are using it lets make sure they're really looking at it and maintaining it | 16:45 |
| morganfainberg | because drive-by "this is your problem now" is not ok | 16:45 |
| dolphm | i imagine if someone is using it, they're more likely to have forked their maintenance effort, because we haven't seen any! | 16:46 |
| morganfainberg | exactly | 16:46 |
| morganfainberg | i somehow doubt it is "perfect in everyway" | 16:46 |
| openstackgerrit | Raildo Mascena de Sousa Filho proposed openstack/keystone-specs: Creating Liberty repository for specs https://review.openstack.org/171733 | 16:47 |
| dolphm | every patch to keystone/common/cache/backends/mongo.py since the original merge (besides brant's patch yesterday) is entirely high level maintenance on things like hacking standards | 16:48 |
| dolphm | morganfainberg: ^ just checked | 16:48 |
| dolphm | gyee: ^ | 16:48 |
| lbragstad | morganfainberg: maybe throw it out on the ML just to be safe? | 16:48 |
| dolphm | lbragstad: in the mean time, lay down the deprecator! | 16:48 |
| morganfainberg | stable g-r already fixed for juno | 16:49 |
| lbragstad | dolphm: ++ | 16:49 |
| dolphm | morganfainberg: for pymongo? | 16:49 |
| lbragstad | i'm in favor of droppin' the deprecation hammer | 16:49 |
| morganfainberg | same with icehouse | 16:49 |
| morganfainberg | yeah as of today | 16:49 |
| dolphm | morganfainberg: awesome | 16:49 |
| dolphm | my job here is done. | 16:49 |
| dolphm | beer time! | 16:49 |
| morganfainberg | Ihar Hrachyshka got it taken care of | 16:49 |
| openstackgerrit | Raildo Mascena de Sousa Filho proposed openstack/keystone-specs: Move reseller spec for Liberty release https://review.openstack.org/170926 | 16:50 |
| openstackgerrit | Raildo Mascena de Sousa Filho proposed openstack/keystone-specs: Move reseller spec for Liberty release https://review.openstack.org/170926 | 16:52 |
| gyee | dolphm, beer time at AM?! | 16:55 |
| morganfainberg | bknudson, ping when you have a moment re: https://review.openstack.org/#/c/169947/ | 16:57 |
| morganfainberg | gyee, it's 5 somewhere | 16:57 |
| *** spandhe has joined #openstack-keystone | 16:57 | |
| morganfainberg | how do you know dolphm isn't somewhere where it is 5? | 16:57 |
| gyee | heh | 16:57 |
| dolphm | gyee: just kidding lol | 16:58 |
| dolphm | morganfainberg: i just woke up like 2 hours ago | 16:58 |
| morganfainberg | dolphm, i woke up about 1hr ago:P | 16:59 |
| morganfainberg | 1.5 actually | 16:59 |
| *** topol has quit IRC | 17:06 | |
| *** amakarov is now known as amakarov_away | 17:06 | |
| *** _cjones_ has quit IRC | 17:06 | |
| gyee | fire up the deprecator! | 17:06 |
| morganfainberg | lol | 17:06 |
| morganfainberg | next meeting | 17:06 |
| *** _cjones_ has joined #openstack-keystone | 17:07 | |
| *** sdake has joined #openstack-keystone | 17:07 | |
| sdake | hey quick question about bandit gating in keystone | 17:07 |
| sdake | I want to add bandit to magnum | 17:07 |
| sdake | but I don't see any tox logic in keystone to launch bandit | 17:08 |
| sdake | how would someone run bandit localy? | 17:09 |
| morganfainberg | sdake, bknudson put together the work for bandit | 17:09 |
| morganfainberg | sdake, looking to see, but i dont think it was tox based | 17:09 |
| sdake | the gate name has "tox" in it | 17:09 |
| sdake | in project-config | 17:09 |
| *** drjones has joined #openstack-keystone | 17:09 | |
| morganfainberg | sdake, https://github.com/stackforge/bandit#usage | 17:09 |
| *** _cjones_ has quit IRC | 17:10 | |
| sdake | ya I can run manually but prefer it to run in tox :) | 17:10 |
| sdake | i've already run it manually with good results | 17:10 |
| sdake | (no failures) | 17:10 |
| morganfainberg | https://review.openstack.org/#/c/157930/ | 17:10 |
| sdake | thanks ! | 17:11 |
| raildo | breton, ping about your comment here: https://review.openstack.org/#/c/158372/38 I tested on mysql and works fine, I saw this error here: http://stackoverflow.com/questions/2979369/databaseerror-current-transaction-is-aborted-commands-ignored-until-end-of-tra and probably this is a error in dbsync | 17:17 |
| raildo | breton, what do you think? makes sense? | 17:17 |
| raildo | lhcheng, hey new core :D Do you have some time to review this bug? https://review.openstack.org/#/c/158720/ | 17:19 |
| lhcheng | raildo: lol | 17:20 |
| raildo | haha | 17:20 |
| lhcheng | raildo: sure, will take a look at it | 17:20 |
| raildo | lhcheng, thanks | 17:20 |
| *** harlowja_away is now known as harlowja | 17:21 | |
| *** ozialien has joined #openstack-keystone | 17:22 | |
| *** Guest39926 is now known as mgagne | 17:29 | |
| *** mgagne has joined #openstack-keystone | 17:29 | |
| samueldmq | raildo, try session.rollback() after line 587 | 17:29 |
| *** topol has joined #openstack-keystone | 17:30 | |
| *** ChanServ sets mode: +v topol | 17:30 | |
| samueldmq | raildo, I ran in a similar issue these days, for postgresql we need to either rollback the transaction that had error | 17:30 |
| samueldmq | raildo, (line 586 in your case) | 17:30 |
| samueldmq | raildo, re-instantiate the session object also works | 17:30 |
| bknudson | morganfainberg: I have a moment. | 17:30 |
| morganfainberg | Computer just crashed 3 times in a row. | 17:31 |
| morganfainberg | Anyway. | 17:31 |
| morganfainberg | bknudson: I think we need a better mechanism than x second being static, but 1 second is far too aggressive. | 17:31 |
| morganfainberg | bknudson: you open to landing 120s and looking for a better solution long term? | 17:32 |
| bknudson | morganfainberg: too aggressive for what? I don't think there's ever been a bug opened saying that the current behavior isn't working. | 17:32 |
| bknudson | there's no bug associated with the review | 17:32 |
| morganfainberg | The bugs have been written off mostly I think as "token itself timed out" | 17:33 |
| bknudson | there should at least be a bug so that users can find it. | 17:33 |
| morganfainberg | Happy to force a bug associated with it. | 17:33 |
| raildo | samueldmq, thanks for the tip :) | 17:33 |
| morganfainberg | jamielennox: ^ please make sure we have a clear bug on this so it can be tracked. | 17:33 |
| morganfainberg | bknudson: it's one of those issues that would present like another issue we regularly tell people to increase their token ttl to solve | 17:34 |
| samueldmq | raildo, np | 17:34 |
| morganfainberg | bknudson: so I don't think we have a clear user-originated bug. | 17:34 |
| morganfainberg | But yes having a tracking bug would be good. | 17:35 |
| bknudson | y, everybody just ups the token ttl to deal with any issue like this | 17:35 |
| bknudson | and they don't bother telling anyone that they're going to do it (at least around here) | 17:35 |
| morganfainberg | The issue is it'll be transient and ttl won't matter. | 17:35 |
| morganfainberg | But change in ttl will make it less common. | 17:36 |
| *** drjones has quit IRC | 17:36 | |
| morganfainberg | It's fair to ask for a new token if yours is about to expire a litter earlier than 1 second from expiry. | 17:36 |
| *** pnavarro has joined #openstack-keystone | 17:37 | |
| *** _cjones_ has joined #openstack-keystone | 17:39 | |
| *** henrynash has joined #openstack-keystone | 17:41 | |
| *** ChanServ sets mode: +v henrynash | 17:41 | |
| raildo | henrynash, ping, can you see my comments here? https://review.openstack.org/#/c/170926/2/doc/source/index.rst | 17:48 |
| *** jaosorior has quit IRC | 17:52 | |
| *** krykowski has quit IRC | 18:01 | |
| *** henrynash has quit IRC | 18:05 | |
| *** henrynash has joined #openstack-keystone | 18:08 | |
| *** ChanServ sets mode: +v henrynash | 18:08 | |
| *** ajayaa has quit IRC | 18:21 | |
| openstackgerrit | Merged openstack/keystone: Tox env for Bandit https://review.openstack.org/157930 | 18:22 |
| *** aix has quit IRC | 18:24 | |
| *** boris-42 has quit IRC | 18:28 | |
| *** nkinder has quit IRC | 18:34 | |
| breton | raildo: > re-instantiate the session object also works | 18:36 |
| breton | raildo: that's true, we did it yesterday with samueldmq and it worked | 18:36 |
| *** ayoung has joined #openstack-keystone | 18:39 | |
| *** ChanServ sets mode: +v ayoung | 18:39 | |
| openstackgerrit | Brant Knudson proposed openstack/keystone: Move bandit requirement to test-requirements-bandit.txt https://review.openstack.org/171772 | 18:47 |
| *** stevemar has quit IRC | 18:48 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/keystonemiddleware: Updated from global requirements https://review.openstack.org/166438 | 18:58 |
| ayoung | Who knows Federation in the absense of stevemar? | 19:05 |
| breton | ayoung: just ask. I'd like to hear some questions on federation and try answering them. Maybe not now, but in a couple of hours maybe | 19:11 |
| ayoung | lhcheng, rodrigods in the websso documentation we say that remote_id_attribute "be set on a per-protocol basis." | 19:11 |
| ayoung | breton, yeah, just trying to get the right persons attention. | 19:12 |
| ayoung | lhcheng, rodrigods I +2ed this doc too, but I don't think we actually honor that | 19:12 |
| lhcheng | ayoung: we used it somewhere.. | 19:13 |
| * lhcheng looking it up | 19:13 | |
| lhcheng | https://github.com/openstack/keystone/blob/master/keystone/contrib/federation/controllers.py#L271-L272 | 19:13 |
| ayoung | lhcheng, code doesn't seem to be there...I just tried with the kerberos one, though, and that is...not something in the config file | 19:13 |
| ayoung | lhcheng, that is not per protocol | 19:14 |
| lhcheng | ugh, I get what you're saying | 19:14 |
| lhcheng | in the docs, there is a general setting and per protocol | 19:14 |
| lhcheng | yeah, the per protocol is not honored | 19:15 |
| openstackgerrit | Brant Knudson proposed openstack/keystone: Move bandit requirement to test-requirements-bandit.txt https://review.openstack.org/171772 | 19:15 |
| lhcheng | ayoung: should be too hard to fix that | 19:16 |
| lhcheng | ayoung: do you think it is a rc backport potential? | 19:16 |
| lhcheng | *should not | 19:16 |
| *** henrynash has quit IRC | 19:24 | |
| rodrigods | lhcheng, ayoung was afk | 19:25 |
| rodrigods | we don't honor per protocol and also I don't see why we should | 19:25 |
| ayoung | lhcheng, rodrigods we need ti | 19:25 |
| rodrigods | why? | 19:26 |
| ayoung | rodrigods, I am trying to set up a demo where I do SAML next to SSSD. | 19:26 |
| ayoung | they need a differente remote_id_attributes | 19:26 |
| ayoung | and saml next to oidc won't work, either | 19:26 |
| rodrigods | so I might be confusing what remote_id means | 19:27 |
| lhcheng | ayoung: I agree, we can't setup more than one protocol with the current config. | 19:27 |
| lhcheng | ayoung: open a bug and tag it for rc-potential? | 19:27 |
| ayoung | lhcheng, will do. | 19:28 |
| *** ayoung is now known as ayoung-mtg | 19:28 | |
| *** thedodd has quit IRC | 19:37 | |
| *** rushiagr is now known as rushiagr_away | 19:45 | |
| *** samueldmq has quit IRC | 19:47 | |
| *** _cjones_ has quit IRC | 19:50 | |
| *** sdake_ has joined #openstack-keystone | 19:55 | |
| *** mattfarina has quit IRC | 19:58 | |
| *** sdake has quit IRC | 19:58 | |
| *** thedodd has joined #openstack-keystone | 19:59 | |
| *** ozialien has quit IRC | 19:59 | |
| *** mattfarina has joined #openstack-keystone | 20:04 | |
| *** zzzeek has quit IRC | 20:04 | |
| *** zzzeek_ has joined #openstack-keystone | 20:04 | |
| *** chrisshattuck has joined #openstack-keystone | 20:04 | |
| *** _cjones_ has joined #openstack-keystone | 20:04 | |
| *** wpf has quit IRC | 20:05 | |
| *** wpf has joined #openstack-keystone | 20:07 | |
| openstackgerrit | Brant Knudson proposed openstack/python-keystoneclient: Use pkg_resources to get version https://review.openstack.org/171720 | 20:07 |
| *** stevemar has joined #openstack-keystone | 20:08 | |
| *** ChanServ sets mode: +v stevemar | 20:08 | |
| openstackgerrit | Brant Knudson proposed openstack/python-keystoneclient: Remove support for __version__ https://review.openstack.org/171720 | 20:09 |
| *** nkinder has joined #openstack-keystone | 20:09 | |
| openstackgerrit | Raildo Mascena de Sousa Filho proposed openstack/keystone: Prohibit invalid ids in subtree and parents list https://review.openstack.org/158720 | 20:18 |
| *** sdake_ has quit IRC | 20:28 | |
| openstackgerrit | Henrique Truta proposed openstack/keystone: Add domain_id checking in create_project https://review.openstack.org/159944 | 20:32 |
| openstackgerrit | Henrique Truta proposed openstack/keystone: Add is_domain field in Project Table https://review.openstack.org/157427 | 20:32 |
| openstackgerrit | Henrique Truta proposed openstack/keystone: Honor domain operations in project table https://review.openstack.org/143763 | 20:32 |
| openstackgerrit | Henrique Truta proposed openstack/keystone: Remove domain table references https://review.openstack.org/165936 | 20:32 |
| openstackgerrit | Henrique Truta proposed openstack/keystone: Change project name constraint https://review.openstack.org/158372 | 20:32 |
| *** ayoung-mtg has quit IRC | 20:33 | |
| *** stevemar has quit IRC | 20:34 | |
| *** stevemar has joined #openstack-keystone | 20:34 | |
| *** ChanServ sets mode: +v stevemar | 20:34 | |
| *** stevemar has quit IRC | 20:42 | |
| *** stevemar has joined #openstack-keystone | 20:42 | |
| *** ChanServ sets mode: +v stevemar | 20:42 | |
| *** topol has quit IRC | 20:49 | |
| openstackgerrit | Raildo Mascena de Sousa Filho proposed openstack/keystone: Add domain_id checking in create_project https://review.openstack.org/159944 | 20:51 |
| *** raildo is now known as raildo|away | 20:54 | |
| *** edmondsw has quit IRC | 21:15 | |
| *** luminalflux has joined #openstack-keystone | 21:16 | |
| crinkle | it looks like the openstackclient image create command has a --owner flag to specify a project, but no --domain flag, how does it distinguish between images for projects in different domains? | 21:21 |
| stevemar | crinkle always asking the hard questions | 21:23 |
| crinkle | stevemar: i live to be difficult :) | 21:23 |
| stevemar | crinkle, does glance ever distinguish between images for projects vs domains? cc sigmavirus24 | 21:23 |
| zigo | morganfainberg: I got keystone RC1 built without any issue now, thanks a lot for your work. | 21:24 |
| morganfainberg | zigo, np | 21:24 |
| morganfainberg | zigo, glad to have helped | 21:24 |
| morganfainberg | crinkle, hi! *waves* | 21:24 |
| sigmavirus24 | crinkle: so --owner is always the tenant-id | 21:25 |
| sigmavirus24 | or rather, the owner field in glance is always the tenant, not the specific user | 21:25 |
| crinkle | morganfainberg: hi o/ | 21:25 |
| sigmavirus24 | I'm rather certain we don't have any way of storing a domain for an image though, and that's something we /could/ add but would ned to be brought up probably as a blueprint or at the summit (preferably both) | 21:25 |
| morganfainberg | crinkle, got nothing else, just saying hi cause you were chatting here :) | 21:26 |
| crinkle | sigmavirus24: okay, and since it's the ID and not the name, it doesn't need a domain to distinguish it? | 21:26 |
| sigmavirus24 | But as a newish glance dev, I'm not sure if it hasn't already been brought up | 21:26 |
| morganfainberg | lhcheng, yes on the rc-potential bug | 21:27 |
| sigmavirus24 | crinkle: right. Since the entire tenant owns it, I don't think we've had a need to add a domain to distinguish anything | 21:29 |
| crinkle | sigmavirus24: that makes sense, thanks! | 21:29 |
| sigmavirus24 | crinkle: you're welcome! | 21:29 |
| morganfainberg | sigmavirus24, glance images are about the only thing in OpenStack i see really benefitting (ok heat but thats separate) from Domain awareness | 21:30 |
| morganfainberg | sigmavirus24, but in Liberty with domains being ~= projects it likely becomes less of an issue. | 21:30 |
| sigmavirus24 | morganfainberg: well I need to learn more about domains/(projects|tenants) and how they work and are intended to be used | 21:30 |
| crinkle | morganfainberg: does that mean you see this changing at some point soonish? | 21:31 |
| sigmavirus24 | To be fair, I only got involved in glance around ~k-2 and am still acquainting myself with things | 21:31 |
| morganfainberg | crinkle, any change that is made shouldn't break anything you're doing. if it does, we screwed up | 21:31 |
| sigmavirus24 | Also, the focus in Kilo was fixing bugs but more heavily on getting a semblance of Artifacts and Catalog Index Search (new features) | 21:31 |
| morganfainberg | crinkle, we should give you more flexibility is all we'd change | 21:31 |
| sigmavirus24 | I doubt anyone was paying close attention to things in Keystone closely enough to know we could be benefitting from things | 21:32 |
| stevemar | sigmavirus24, you are le right | 21:32 |
| crinkle | morganfainberg: coolio | 21:32 |
| sigmavirus24 | stevemar: about? | 21:32 |
| stevemar | no one was paying attn to keystone | 21:32 |
| * morganfainberg needs to book train ticket for Vancouver | 21:32 | |
| sigmavirus24 | hah yeah | 21:32 |
| morganfainberg | stevemar, does anyone actually pay attention to keystone? | 21:32 |
| sigmavirus24 | I would like to pay more attention | 21:33 |
| crinkle | morganfainberg: party train through pdx? :D | 21:33 |
| morganfainberg | crinkle, yep | 21:33 |
| * sigmavirus24 's time is already very limited though | 21:33 | |
| crinkle | morganfainberg: \o/ | 21:33 |
| sigmavirus24 | Anyone in here (other than stevemar) going to PyCon this weekend? | 21:33 |
| stevemar | sigmavirus24, dstanek will be (or is) there | 21:34 |
| morganfainberg | sigmavirus24, i think some people from -infra are there. | 21:34 |
| sigmavirus24 | stevemar: yeah, dstanek is doing a tutorial this year | 21:35 |
| morganfainberg | sigmavirus24, you into math [the name] or just like Σ or something else? | 21:35 |
| stevemar | sigmavirus24, weren't you? | 21:35 |
| * sigmavirus24 should have excluded dstanek by default too | 21:35 | |
| sigmavirus24 | stevemar: just a plain ol' talk | 21:35 |
| stevemar | sigmavirus24, when is it? | 21:36 |
| stevemar | i'll be there tomorrow at 2:30ish | 21:36 |
| sigmavirus24 | morganfainberg: I have a B.Sc, and M.Sc in Pure Math | 21:36 |
| sigmavirus24 | stevemar: Saturday after lunch | 21:36 |
| morganfainberg | sigmavirus24, awesome! | 21:36 |
| sigmavirus24 | 3rd track I think | 21:36 |
| stevemar | oh nice, i'll show up | 21:36 |
| sigmavirus24 | stevemar: you don't have to =P | 21:36 |
| morganfainberg | sigmavirus24, math is fun. | 21:36 |
| stevemar | i'll be lurking in the back | 21:36 |
| morganfainberg | sigmavirus24, /me didn't get around to getting a degree in it though :P | 21:36 |
| sigmavirus24 | I might not even show up. I'll be volunteering at the RAX booth around the same time =P | 21:36 |
| stevemar | just clone yourself | 21:37 |
| sigmavirus24 | morganfainberg: I got both in 5 years because I was bored as an undergrad and started taking grad classes and had ~50% of the credits done by the time I had my B.Sc | 21:37 |
| morganfainberg | sigmavirus24, lol. | 21:37 |
| sigmavirus24 | Boredom also led to python and OSS though so | 21:37 |
| sigmavirus24 | That is the story of why I'll never use my degrees =P | 21:37 |
| morganfainberg | sigmavirus24, i ended up w/ a theatre and film degree - bordom lead me to being into CS. | 21:37 |
| sigmavirus24 | Funny how that works, eh? | 21:38 |
| morganfainberg | sigmavirus24, yep | 21:38 |
| sigmavirus24 | stevemar: I get in tomorrow around noon I think | 21:38 |
| stevemar | i don't have a funny story with my degree :( | 21:39 |
| stevemar | it's just there | 21:39 |
| morganfainberg | my story is funny cause i paid too much $ for it | 21:39 |
| morganfainberg | or sad. | 21:39 |
| morganfainberg | something like that | 21:40 |
| * morganfainberg is thankfully debt free now though. | 21:40 | |
| sigmavirus24 | So, I entered my uni because it had a program where you'd take a B.Sc over 5 years but have real world experience at the end | 21:42 |
| sigmavirus24 | And I entered as a Math major and the market went boom | 21:43 |
| stevemar | morganfainberg, when are we cutting KSC? | 21:43 |
| sigmavirus24 | So all of the school's math contacts were either bankrupt or not willing to hire those of us looking | 21:43 |
| stevemar | i have a few OSC patches that i want to include but would require a new KSC version | 21:43 |
| morganfainberg | stevemar, when it's ready | 21:44 |
| stevemar | morganfainberg, well thats vague, are you in manager training? | 21:44 |
| morganfainberg | stevemar, so.. uh when i check with jamielennox what is left before we're ready to cut the "OMG Kilo" version | 21:44 |
| stevemar | morganfainberg, okie dokie | 21:44 |
| morganfainberg | stevemar, no not in training... trial by fire steering a project as PTL :P | 21:44 |
| *** pnavarro has quit IRC | 21:44 | |
| * morganfainberg smooths out the pointy-hair. | 21:44 | |
| stevemar | hmm, i wonder if we can include some stuff in OSC that will depend on a new KSC version then... | 21:45 |
| stevemar | and we would release one at the same time | 21:45 |
| morganfainberg | stevemar, you'd need to make sure it was conditional | 21:45 |
| morganfainberg | stevemar, since g-r wont update | 21:45 |
| stevemar | and have that serve as the OMG kilo version | 21:45 |
| stevemar | okay then nvm | 21:45 |
| stevemar | g-r is not updating for kilo right? | 21:46 |
| * stevemar needs to read the ML from today | 21:46 | |
| morganfainberg | should be frozen barring GATE IS HORKED moments | 21:46 |
| morganfainberg | should have been frozen from ~k3 milestone on | 21:46 |
| morganfainberg | you know... feature freeze-ish | 21:46 |
| sigmavirus24 | whatever that is =P | 21:49 |
| morganfainberg | stevemar, it's that magical time of year when a software project .. no wait.. strike that | 21:49 |
| morganfainberg | sigmavirus24, ^ not stevemar | 21:49 |
| morganfainberg | damn and i was doing so well at not tab<fail> in IRC today. | 21:49 |
| sigmavirus24 | heh | 21:50 |
| stevemar | it's only 2 chars | 21:50 |
| morganfainberg | so, thats 1 more character stroke than is typically needed | 21:50 |
| stevemar | i come before the other st<tab> | 21:50 |
| morganfainberg | a<tab> = ayoung, b<tab> = bknudson, ds<tab> dstanek, s<tab> = stevemar2 (that shady guy) | 21:50 |
| stevemar | mo<tab><tab> is morganfainberg | 21:51 |
| morganfainberg | stevemar, my client does some weird guessing on who last spoke in the channel | 21:51 |
| stevemar | you get bknudson with b<tab>? i have to use bk<tab> | 21:51 |
| bknudson | what client is that? | 21:51 |
| morganfainberg | because mord is earlier alphabetically than morg | 21:52 |
| morganfainberg | bknudson, Textual | 21:52 |
| morganfainberg | bknudson, LimeChat fork for OS X | 21:52 |
| morganfainberg | or was it sublime... one of those things | 21:52 |
| bknudson | it had lime in it | 21:53 |
| morganfainberg | bknudson, exactly! you get it.. limes... which reminds me, tequila anyone? | 21:53 |
| * morganfainberg is allowed to be a little less serious now that we have an RC. | 21:53 | |
| morganfainberg | or so i keep telling myself | 21:54 |
| stevemar | morganfainberg, i am with you on that one | 21:54 |
| bknudson | morganfainberg: I assume the mongo workaround is targeted to rc2 | 21:58 |
| *** ayoung-mtg has joined #openstack-keystone | 21:58 | |
| *** bknudson has left #openstack-keystone | 21:58 | |
| *** bknudson has joined #openstack-keystone | 21:58 | |
| *** ChanServ sets mode: +v bknudson | 21:58 | |
| bknudson | oops, closed the window | 21:58 |
| morganfainberg | bknudson, should be considered when we open RC2 window. | 21:58 |
| morganfainberg | bknudson, 100% | 21:59 |
| bknudson | it's proposed as backport to stable so we'd lose it if it's not in rc2 | 21:59 |
| *** chrisshattuck has quit IRC | 21:59 | |
| openstackgerrit | Brant Knudson proposed openstack/keystone: Refactor common function for loading drivers https://review.openstack.org/166628 | 22:00 |
| openstackgerrit | Brant Knudson proposed openstack/keystone: Refactor MemcachedBackend is not a Manager https://review.openstack.org/166615 | 22:00 |
| openstackgerrit | Brant Knudson proposed openstack/keystone: Tests don't override default config with default https://review.openstack.org/166631 | 22:00 |
| openstackgerrit | Brant Knudson proposed openstack/keystone: Use stevedore https://review.openstack.org/166543 | 22:00 |
| openstackgerrit | Brant Knudson proposed openstack/keystone: Use short names for drivers https://review.openstack.org/166622 | 22:00 |
| stevemar | dammit brant | 22:00 |
| *** chlong has joined #openstack-keystone | 22:01 | |
| bknudson | would be nice to get the stevedore change in somewhat early. | 22:04 |
| *** lhcheng_ has joined #openstack-keystone | 22:05 | |
| *** lhcheng has quit IRC | 22:05 | |
| morganfainberg | bknudson, ++ | 22:05 |
| morganfainberg | stevemar, https://review.openstack.org/#/c/171772/ | 22:06 |
| morganfainberg | stevemar, plz press go | 22:06 |
| bknudson | morganfainberg: stevemar: that one's actually waiting on global-requirements change to be unblocked. | 22:06 |
| bknudson | https://review.openstack.org/#/c/167126/ | 22:07 |
| morganfainberg | oh damn it | 22:07 |
| bknudson | bandit isn't in g-r yet. | 22:07 |
| morganfainberg | i need to look more closely at dependson | 22:07 |
| morganfainberg | fair enough | 22:07 |
| morganfainberg | steve can still +2 | 22:07 |
| morganfainberg | yay for DependsOn :) | 22:07 |
| *** topol has joined #openstack-keystone | 22:07 | |
| *** ChanServ sets mode: +v topol | 22:07 | |
| bknudson | I may have to correct it if the g-r review changes. | 22:07 |
| bknudson | Depends-On is awesome. | 22:08 |
| morganfainberg | stevemar, so | 22:09 |
| morganfainberg | stevemar, re: ksc | 22:09 |
| morganfainberg | stevemar, lets plan to release ksc w/ the RC2 window for keystone opening | 22:09 |
| morganfainberg | jamielennox, ^ cc | 22:09 |
| morganfainberg | and KSM around the same time | 22:09 |
| *** jamielennox is now known as jamielennox|away | 22:11 | |
| stevemar | morganfainberg, that'll be the first for L release then? | 22:14 |
| stevemar | or the final one for K? | 22:14 |
| morganfainberg | K | 22:14 |
| *** mancdaz has quit IRC | 22:21 | |
| *** mancdaz has joined #openstack-keystone | 22:22 | |
| *** sigmavirus24 is now known as sigmavirus24_awa | 22:24 | |
| *** stevemar has quit IRC | 22:30 | |
| bknudson | I noticed there was a requirements update in review for client or middleware... | 22:31 |
| bknudson | middleware: https://review.openstack.org/#/c/166438/ | 22:31 |
| bknudson | also, btw, gate is busted. | 22:31 |
| bknudson | for keystoneclient and keystonemiddleware | 22:32 |
| bknudson | I think there's an uncapped dependency out there somewhere. | 22:32 |
| *** lhcheng_ has quit IRC | 22:32 | |
| morganfainberg | tempest-y-think | 22:32 |
| morganfainberg | thing* | 22:32 |
| morganfainberg | http://logs.openstack.org/38/166438/2/check/gate-tempest-dsvm-neutron-src-keystonemiddleware-icehouse/8f6d2f9/console.html#_2015-04-08_19_45_49_025 | 22:33 |
| morganfainberg | oh yeah | 22:33 |
| morganfainberg | uncapped dep somehow | 22:33 |
| bknudson | there's translations for keystone -- want to pick that up in rc2? https://review.openstack.org/#/c/171472/ | 22:33 |
| bknudson | oops, wrong project. | 22:33 |
| *** bknudson has left #openstack-keystone | 22:33 | |
| *** bknudson has joined #openstack-keystone | 22:34 | |
| *** ChanServ sets mode: +v bknudson | 22:34 | |
| morganfainberg | bknudson, uhm. as long as we haven't moved any string changes in. i checked before we ran the RC1 patch... didn't look like translations were outstanding | 22:34 |
| *** topol has quit IRC | 22:34 | |
| bknudson | ok, it's just the keystonemiddleware requirements update I'd wonder about... | 22:35 |
| bknudson | if that's wanted before a release. | 22:35 |
| morganfainberg | i'd like the requirements to meet the same as in keystone | 22:35 |
| morganfainberg | for this release | 22:35 |
| morganfainberg | so probably want that before we release | 22:36 |
| bknudson | y, that's kind of scary | 22:36 |
| morganfainberg | this is why ksm and ksc will get stable branches with backports | 22:36 |
| morganfainberg | that match the keystone one(s) | 22:36 |
| *** lhcheng has joined #openstack-keystone | 22:36 | |
| *** ChanServ sets mode: +v lhcheng | 22:36 | |
| morganfainberg | for gate sanity | 22:37 |
| morganfainberg | among other things | 22:37 |
| bknudson | looks like the requirements proposed for ksm are the same as keystone -- | 22:37 |
| bknudson | https://review.openstack.org/#/c/166438/2/requirements.txt | 22:37 |
| morganfainberg | yeah just checked | 22:37 |
| bknudson | vs http://git.openstack.org/cgit/openstack/keystone/tree/requirements.txt?id=2015.1.0rc1 | 22:37 |
| morganfainberg | bknudson, so we should land that before we ksm release | 22:37 |
| bknudson | good luck. | 22:37 |
| *** thedodd has quit IRC | 22:42 | |
| openstackgerrit | Merged openstack/keystone: Add placeholders for reserved migrations https://review.openstack.org/171329 | 22:42 |
| *** gordc has quit IRC | 22:45 | |
| breton | oh. | 22:45 |
| breton | I thought we wanted to hold off placeholders for some time | 22:47 |
| breton | anyway, np. | 22:49 |
| morganfainberg | we can always pull them out | 22:50 |
| morganfainberg | better to have them and not need them than someone screwing it up ;) | 22:50 |
| openstackgerrit | Brant Knudson proposed openstack/keystone: Tests enforce use of stevedore loading https://review.openstack.org/171854 | 22:51 |
| openstackgerrit | Brant Knudson proposed openstack/keystone: Use short names for drivers https://review.openstack.org/166622 | 22:57 |
| openstackgerrit | Brant Knudson proposed openstack/keystone: Tests enforce use of stevedore loading https://review.openstack.org/171854 | 22:57 |
| *** mancdaz has quit IRC | 22:57 | |
| *** browne has quit IRC | 23:03 | |
| *** carlosmarin has quit IRC | 23:03 | |
| openstackgerrit | Brant Knudson proposed openstack/keystone: Update sample config file https://review.openstack.org/171860 | 23:03 |
| *** browne has joined #openstack-keystone | 23:04 | |
| *** chlong has quit IRC | 23:05 | |
| *** bknudson has quit IRC | 23:07 | |
| openstackgerrit | guang-yee proposed openstack/keystone: make sure we properly initialize the backends before using the drivers https://review.openstack.org/171864 | 23:08 |
| *** jamielennox|away is now known as jamielennox | 23:12 | |
| *** mancdaz has joined #openstack-keystone | 23:14 | |
| *** dims has quit IRC | 23:16 | |
| jamielennox | no bknudson or stevemar? is this the right channel? | 23:21 |
| *** mattfarina has quit IRC | 23:22 | |
| *** Ephur has quit IRC | 23:34 | |
| *** dims has joined #openstack-keystone | 23:34 | |
| openstackgerrit | Jamie Lennox proposed openstack/python-keystoneclient: Increase minimum token life required https://review.openstack.org/169947 | 23:36 |
| *** sdake has joined #openstack-keystone | 23:40 | |
| *** sdake_ has joined #openstack-keystone | 23:42 | |
| *** sdake has quit IRC | 23:45 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!