| *** dims has quit IRC | 00:00 | |
| *** bearhands is now known as comstud | 00:00 | |
| *** danwent has quit IRC | 00:01 | |
| *** hemna_ is now known as hemnafk | 00:02 | |
| *** noorul`` has quit IRC | 00:02 | |
| *** faramir has joined #openstack-dev | 00:03 | |
| *** amotoki_ has quit IRC | 00:06 | |
| *** amotoki_ has joined #openstack-dev | 00:06 | |
| *** jasdeepH has quit IRC | 00:07 | |
| *** armax has left #openstack-dev | 00:07 | |
| *** danwent has joined #openstack-dev | 00:08 | |
| *** alok_ has quit IRC | 00:09 | |
| *** lbragstad has joined #openstack-dev | 00:11 | |
| *** amotoki_ has quit IRC | 00:12 | |
| *** dkranz has joined #openstack-dev | 00:12 | |
| *** shanewang_away is now known as shanewang | 00:13 | |
| *** colinmcnamara has joined #openstack-dev | 00:14 | |
| *** jasdeepH has joined #openstack-dev | 00:16 | |
| *** michchap has joined #openstack-dev | 00:16 | |
| *** dims has joined #openstack-dev | 00:16 | |
| gyee | ayoung, dolphm, morganfainberg, can you guys please take a look at https://review.openstack.org/#/c/40036/ | 00:18 |
|---|---|---|
| gyee | would be nice if it make the m3 cut today | 00:18 |
| morganfainberg | gyee, you'll never guess what i was just looking at | 00:18 |
| gyee | morganfainberg, I am trying to give nachi a lift | 00:19 |
| gyee | want to gether all the inputs | 00:19 |
| *** sushils has quit IRC | 00:19 | |
| gyee | gather all the inputs so do the patch in one shot | 00:19 |
| *** galstrom_zzz is now known as galstrom | 00:20 | |
| *** jhesketh_ has joined #openstack-dev | 00:21 | |
| ayoung | gyee, lookinj | 00:21 |
| ayoung | ah, he fixed the pep8... | 00:22 |
| ayoung | gyee, I need some background on this one | 00:22 |
| ayoung | what is the problem here? | 00:23 |
| *** sthaha has joined #openstack-dev | 00:24 | |
| *** sthaha has quit IRC | 00:24 | |
| *** sthaha has joined #openstack-dev | 00:24 | |
| *** safchain has quit IRC | 00:24 | |
| *** otherwiseguy has joined #openstack-dev | 00:25 | |
| *** rwsu is now known as rwsu-away | 00:26 | |
| *** dstanek has quit IRC | 00:27 | |
| *** nati_uen_ has quit IRC | 00:28 | |
| *** fbo_away has quit IRC | 00:29 | |
| *** stevemar has joined #openstack-dev | 00:29 | |
| ayoung | morganfainberg, do you get what he means by Protocol Agnostic? | 00:29 |
| morganfainberg | ayoung, not really sure what he means by that | 00:29 |
| morganfainberg | ayoung, i am guessing it's HMAC signature, which is protocol (e.g. EC2?) agnostic? | 00:30 |
| morganfainberg | ayoung, but i think i'm missing the use case. (hasn't clicked for me yet) | 00:31 |
| *** shakayumi has quit IRC | 00:31 | |
| morganfainberg | gyee, any insight you could provide? | 00:31 |
| *** galstrom is now known as galstrom_zzz | 00:32 | |
| *** sridevi has joined #openstack-dev | 00:33 | |
| morganfainberg | ayoung, https://blueprints.launchpad.net/keystone/+spec/generic-signature-validation looks like there is a lot of description there | 00:33 |
| *** Ryan_Lane has quit IRC | 00:33 | |
| *** jamielennox|away is now known as jamielennox | 00:34 | |
| *** diogogmt has joined #openstack-dev | 00:35 | |
| *** flaper87 is now known as flaper87|afk | 00:36 | |
| ayoung | morganfainberg, full of sound and fury | 00:36 |
| ayoung | morganfainberg, thing is, I like the cleanup in that patch all by itself | 00:36 |
| morganfainberg | ayoung, aye | 00:36 |
| morganfainberg | ayoung, much nicer setup. | 00:37 |
| *** tjones has quit IRC | 00:37 | |
| ayoung | HMAC is Hash Based Message Authentication Code, which is really what we do in PKI messages: and one weakness there is that it is limited to MD5 for now | 00:37 |
| ayoung | BUt I don't think he means "Algorithm Agnostic" | 00:38 |
| *** rfolco has quit IRC | 00:38 | |
| *** tjones has joined #openstack-dev | 00:38 | |
| ayoung | and he talks about middleware, but that really shold be in the Keystone client | 00:38 |
| *** safchain has joined #openstack-dev | 00:39 | |
| ayoung | morganfainberg, do we need to support additional HMAC based authentication? I want to get away from all this custom authN stuff. | 00:39 |
| morganfainberg | ayoung, it would remove the need for a special S3 and special EC2 mechanism, it looks like | 00:40 |
| ayoung | morganfainberg, but we don't use those to authenticate to Keystone | 00:41 |
| ayoung | oh, the contrib extensions | 00:41 |
| morganfainberg | ayoung, yes. | 00:41 |
| ayoung | I thought that stuff was legacy | 00:41 |
| ayoung | morganfainberg, do we want to continue to extend that? | 00:42 |
| ayoung | I kindof fail to see the burning need for this. | 00:42 |
| morganfainberg | ayoung, i think there is still some interest in it, but i don't see a need to continue to extend. the current ec2 middleware hasn't worked since nova removed the "FLAG" import | 00:42 |
| ayoung | Heh | 00:42 |
| morganfainberg | ayoung, minimal interest != need to support. | 00:42 |
| ayoung | I think I acked that patch | 00:43 |
| ayoung | so...pocket veto? | 00:43 |
| ayoung | Or do we ack it to get the nicer code layout? | 00:43 |
| morganfainberg | ayoung, thing is, i kinda want the cleanup :P | 00:43 |
| gyee | morganfainberg, ayoung, sorry I got pulled away for a meeting, back now | 00:43 |
| ayoung | OK, help me build a rationalization | 00:43 |
| ayoung | gyee, what is the burning use case for it | 00:43 |
| ayoung | gyee, we like the code cleanup, but can't see justification for the patch itself...help us understand | 00:44 |
| gyee | ayoung, signature-based authentication | 00:44 |
| ayoung | gyee, do we need that? | 00:44 |
| ayoung | TO keystone? | 00:44 |
| gyee | we need a generic HMAC base signature authentication | 00:44 |
| ayoung | NO we don't | 00:44 |
| morganfainberg | gyee, for example, the middleware for ec2 (in keystone) hasn't worked … at least since the patch nova removed the FLAGS import. | 00:44 |
| ayoung | gyee, maybe *you* do, though? | 00:45 |
| morganfainberg | which, iirc was the primary consumer of that. | 00:45 |
| gyee | ayoung, not just ec2 | 00:45 |
| gyee | any HMAC-based signatures | 00:45 |
| ayoung | and, is EC2 HMAC, or is it PKI? I thought EC2 was based on asymmetric keys? | 00:45 |
| gyee | AWS, S3, formpost, tempurl, etc | 00:46 |
| gyee | HMAC is symmetric key | 00:46 |
| ayoung | I know that. | 00:46 |
| ayoung | I thought EC2 was asym | 00:46 |
| *** jasondotstar has joined #openstack-dev | 00:46 | |
| gyee | EC2 is HMAC-based | 00:46 |
| *** dstanek has joined #openstack-dev | 00:46 | |
| morganfainberg | gyee, right | 00:47 |
| ayoung | So EC2 Auth is not based on EC2 Keyspairs? What a clever naming scheme | 00:48 |
| gyee | EC2 keypairs is a bad name | 00:48 |
| gyee | access key ID is not a key | 00:48 |
| gyee | access key ID is more like a user ID | 00:49 |
| *** Tross has left #openstack-dev | 00:49 | |
| *** hugokuo has left #openstack-dev | 00:49 | |
| ayoung | gyee, OK, so we get a V3 version of EC2 and Swift this way. | 00:49 |
| morganfainberg | gyee, makes it easier to implement .. what ayoung said | 00:50 |
| *** rfolco has joined #openstack-dev | 00:50 | |
| gyee | ayoung, right, with this, we can essentially implement any kind of HMAC-based signature validation at middleware | 00:50 |
| ayoung | gyee, this is burning? | 00:50 |
| gyee | ayoung, hell yeah | 00:50 |
| ayoung | gyee, please expound why? | 00:51 |
| gyee | so we can start working on HMAC-based middlewares | 00:51 |
| morganfainberg | gyee, you mean an update to say http://git.openstack.org/cgit/openstack/keystone/tree/keystone/middleware/ec2_token.py ? | 00:52 |
| ayoung | gyee, middlewares? THose should be in keystone client, not in Keystone server | 00:53 |
| *** vladimir3p has quit IRC | 00:53 | |
| *** sridevi has quit IRC | 00:53 | |
| gyee | moranfainberg, ayoung, but we need this working before we can work on keystoneclient side | 00:53 |
| *** nos_ has joined #openstack-dev | 00:54 | |
| *** nos_ has quit IRC | 00:54 | |
| *** hugokuo has joined #openstack-dev | 00:55 | |
| *** nosnos has joined #openstack-dev | 00:55 | |
| *** changbl has quit IRC | 00:55 | |
| ayoung | gyee, who needs this? Is new HMAC based middlewares a high demand Item? | 00:56 |
| morganfainberg | gyee, so there is a demand to maintain the EC2/S3 or new HMAC middlewares? | 00:56 |
| gyee | morganfainberg, correct, https://github.com/openstack/keystone/blob/master/keystone/middleware/ec2_token.py will just be calling the auth plugin for v3 | 00:56 |
| morganfainberg | ayoung, damn it, you beat me to the question again. | 00:57 |
| gyee | ayoung, I know HP needs it :) | 00:57 |
| *** jasondotstar has quit IRC | 00:57 | |
| stevemar | gyee: anyone else? :) | 00:58 |
| gyee | stevemar, I don't have access to the customers list | 00:59 |
| *** Ruetobas has joined #openstack-dev | 00:59 | |
| ayoung | gyee, more info please...Why does HP need it? | 00:59 |
| gyee | ayoung, uh, because we use openstack? | 01:00 |
| *** yjiang5 is now known as yjiang5_away | 01:00 | |
| ayoung | gyee, thing is, I don;t think this code supports EC2 as is | 01:00 |
| *** freedomhui has joined #openstack-dev | 01:00 | |
| stevemar | gyee: i meant more as... any other folks mentioned they would want this (like rax, nebula, etc) | 01:00 |
| *** changbl has joined #openstack-dev | 01:00 | |
| morganfainberg | gyee, ec2 code is moribund at this point (the middleware) that is. | 01:01 |
| *** giroro_ has joined #openstack-dev | 01:01 | |
| gyee | morganfainberg, we're going to fix it | 01:01 |
| ayoung | gyee, if this came in with a working swift and ec2 implementation, complete with unit tests, I might be more favorable | 01:02 |
| *** mars has quit IRC | 01:02 | |
| gyee | stevemar, I can ask but their lawyers may say no :) | 01:02 |
| gyee | ayoung, can't work on middleware till this one is there | 01:03 |
| gyee | unless you want it as multiple patches | 01:03 |
| *** Ruetobas has quit IRC | 01:04 | |
| gyee | morganfainberg, nova doesn't support euca2ools anymore? | 01:04 |
| ayoung | gyee, you are asking me to buy a pig in a poke. I don't know that this code as is does anything useful, and the only justification you give me is something HP confidential. How would you react if the roles were reversed? | 01:04 |
| *** giroro_ has quit IRC | 01:04 | |
| morganfainberg | gyee, not sure. | 01:04 |
| gyee | ayoung, how do we support HMAC-based sig authentication today? | 01:06 |
| morganfainberg | gyee, iirc we don't. | 01:06 |
| gyee | pretty much all extensions | 01:06 |
| morganfainberg | well, contrib extensions | 01:06 |
| ayoung | gyee, but that would not change with this patch | 01:07 |
| gyee | all I am saying is, with this we don't need extensions | 01:07 |
| ayoung | maybe it would if it went in in H1 and then the Swift etc patches followed | 01:07 |
| *** colinmcnamara has quit IRC | 01:07 | |
| ayoung | but putting it in the night of the deadline with no functioning mechanisms seems fooloish | 01:07 |
| morganfainberg | gyee, would it be viable to push this for Icehouse-1? | 01:07 |
| ayoung | gyee, I think you are wrong | 01:08 |
| gyee | I1 is fine | 01:08 |
| ayoung | I think we would still need additional code at least for EC2 | 01:08 |
| ayoung | from what I've seen | 01:08 |
| gyee | ayoung, nope | 01:08 |
| morganfainberg | it seems like the likelyhood of hitting H-final with anything useful out of it (besides the code-cleanup) is slim | 01:08 |
| ayoung | gyee, I'm more than happy to have it in I0 | 01:08 |
| ayoung | but with working implementations for Swift and EC2, can be dependent patches | 01:09 |
| gyee | ayoung, that's fine, the earlier it gets in, the earlier we can make middleware changes | 01:09 |
| ayoung | deal | 01:09 |
| morganfainberg | ayoung ++ | 01:09 |
| *** safchain has quit IRC | 01:09 | |
| gyee | ayoung, morganfainberg, sure, I suppose we can make middleware changes in parallel | 01:10 |
| gyee | that would give you guys more context on why the patch is needed | 01:10 |
| morganfainberg | gyee, yes, i am on board with that | 01:11 |
| *** Ruetobas has joined #openstack-dev | 01:11 | |
| gyee | sorry I can disclosed any more without the corporate lawyers all over my ass | 01:11 |
| gyee | can't | 01:11 |
| ayoung | gyee, statements like that make me want to whip out the -2 gun and start shooting at your reviews | 01:12 |
| *** Mandell has quit IRC | 01:12 | |
| morganfainberg | gyee, i can understand corporate lawyer issues, but it makes it hard to review/see what the benefit (or need) is for a given featureset | 01:12 |
| morganfainberg | gyee, especially an AuthN or AuthZ addition at the 11th hour | 01:13 |
| ayoung | now, that being said, anyone want to look at some PKI code cleanups? | 01:13 |
| gyee | morganfainberg, ayoung, yeah, lets do this on etherpad and whiteboard | 01:13 |
| *** erkules_ has joined #openstack-dev | 01:14 | |
| *** SumitNaiksatam has quit IRC | 01:14 | |
| morganfainberg | gyee, that makes sense. | 01:14 |
| morganfainberg | gyee, good plan | 01:14 |
| ayoung | https://review.openstack.org/#/c/43768/ gyee morganfainberg | 01:14 |
| gyee | wow, what's a commit message | 01:15 |
| ayoung | and https://review.openstack.org/#/c/43767/4 | 01:15 |
| ayoung | gyee, I know, he's a little terse | 01:15 |
| morganfainberg | ayoung, wow. | 01:16 |
| gyee | https://review.openstack.org/#/c/43767 is a no brainer | 01:16 |
| *** jhesketh has quit IRC | 01:16 | |
| morganfainberg | gyee, agreed | 01:16 |
| *** erkules has quit IRC | 01:16 | |
| *** jhesketh has joined #openstack-dev | 01:16 | |
| *** angdraug has quit IRC | 01:17 | |
| morganfainberg | gyee, the other one looks sane, it's mostly code htat should land in oslo soon | 01:17 |
| morganfainberg | gyee, ayoung, i want to once over it of course. | 01:18 |
| ayoung | jdennis1, is a pretty good python coder. With he wouldn't be so stingy with the wordcount on his commit messages, though. | 01:18 |
| ayoung | please do | 01:19 |
| ayoung | I'd almost classify it as a bug fix, but of course, it goes beyond that | 01:19 |
| morganfainberg | yes. | 01:19 |
| *** amotoki_ has joined #openstack-dev | 01:20 | |
| ayoung | gyee, you should know this code. I think you and I both have worked on it over time | 01:21 |
| *** tjones has quit IRC | 01:22 | |
| gyee | ayoung, I only claim the credit when something is working :) | 01:22 |
| ayoung | he made the mistake of reusing a bug id...but an earlier commit claimed it...must have been a partial fix | 01:22 |
| *** tjones has joined #openstack-dev | 01:23 | |
| gyee | ayoung, the stuff in keystone.common.utils will be part of oslo? | 01:23 |
| *** briancurtin has joined #openstack-dev | 01:24 | |
| *** amotoki_ has quit IRC | 01:24 | |
| ayoung | gyee, my bad, I was thinking of the Cert setup stuff you did in examples. | 01:24 |
| ayoung | gyee, I don;t know if he has submitted it there or not | 01:24 |
| ayoung | gyee, he has no outstanding reviews to oslo | 01:25 |
| morganfainberg | ayoung, the commit message says it's supposed to go to oslo | 01:25 |
| morganfainberg | "soon" | 01:25 |
| morganfainberg | sorry "near future" | 01:25 |
| ayoung | morganfainberg, right. But then, once it does, it has to be merged down again, a system I am starting to think is not right for code that can have security implications | 01:26 |
| morganfainberg | ayoung, i'm not a huge fan of it. | 01:26 |
| *** dolphm has quit IRC | 01:26 | |
| ayoung | anyway...it seems pretty straightforward. | 01:27 |
| morganfainberg | ayoung, i know we shun eventlet, but isn't the reason make_dirs etc is farmed out to shutils is so eventlet can patch/non-blocking run? or am i mis-remembering things? | 01:27 |
| morganfainberg | *shun might be a strong word | 01:28 |
| *** freedomhui has quit IRC | 01:28 | |
| ayoung | morganfainberg, I shun it...but let me see if I can find a reference | 01:29 |
| ayoung | morganfainberg, but..that is probably reason enough not to push this one in tonight | 01:29 |
| ayoung | let me add that this is a bug fix, so we can probably get it in after tonight, assuming we straighten out the bug report | 01:29 |
| morganfainberg | ayoung, i might be totally mis-remembering. but i'd rather be 2x sure rather than having to revert. | 01:30 |
| morganfainberg | ayoung, exactly, bug fix = easier than blueprint on this timeline. | 01:30 |
| ayoung | jdennis1, you here? | 01:30 |
| gyee | morganfainberg, but this is CLI change, why eventlet concern? | 01:31 |
| morganfainberg | gyee oh | 01:31 |
| morganfainberg | gyee, no concern for CLi | 01:31 |
| ayoung | Yeah, this is in Keystone, but will only be invoked by keystone manage | 01:32 |
| morganfainberg | gyee, good point | 01:32 |
| morganfainberg | ayoung, then strike that last concern. | 01:32 |
| ayoung | *whack* | 01:32 |
| ayoung | concern struck. Hard. | 01:32 |
| morganfainberg | ayoung, i think you killed it | 01:32 |
| gyee | heh | 01:32 |
| *** freedomhui has joined #openstack-dev | 01:32 | |
| ayoung | *whack* | 01:33 |
| ayoung | just to be sure | 01:33 |
| gyee | he's not breathing | 01:33 |
| ayoung | *whack* | 01:33 |
| *** markwash has joined #openstack-dev | 01:33 | |
| ayoung | he might be really good at holding his breath | 01:33 |
| morganfainberg | lol | 01:33 |
| ayoung | OK, anything else? We've got 3.5 hours until midnight | 01:34 |
| stevemar | ayoung: says you, morgan has 6.5 hrs til then | 01:34 |
| morganfainberg | ayoung: https://review.openstack.org/#/c/44776/ might be worth sneaking in | 01:35 |
| morganfainberg | httpd not working, kindof bad imo | 01:35 |
| *** bfschott has joined #openstack-dev | 01:35 | |
| morganfainberg | but it's a bug, so, could delay | 01:35 |
| ayoung | morganfainberg, heh https://review.openstack.org/#/c/40444/ | 01:35 |
| stevemar | morganfainberg: seems like a no brainer | 01:35 |
| *** martine has joined #openstack-dev | 01:36 | |
| morganfainberg | ayoung, that was the one that was blocking until credential table stuff right? | 01:36 |
| ayoung | O hells yeah | 01:36 |
| ayoung | morganfainberg, yeah, by me | 01:36 |
| morganfainberg | ayoung, right. | 01:36 |
| *** martine is now known as Guest154 | 01:36 | |
| ayoung | I was slacking in updating the review | 01:37 |
| morganfainberg | yeah that one by jdennis1 looks good, approving. | 01:38 |
| stevemar | ayoung, morganfainberg: are you guys looking for easy to approve things to get in? | 01:38 |
| morganfainberg | stevemar, not so much easy, things that should get in | 01:38 |
| stevemar | ah | 01:38 |
| ayoung | morganfainberg, I know that 40444 is needed by heat | 01:39 |
| ayoung | like to get that in | 01:39 |
| stevemar | well, easy AND should go in: https://review.openstack.org/#/c/44543/ | 01:39 |
| morganfainberg | looking now | 01:39 |
| *** litong has joined #openstack-dev | 01:39 | |
| morganfainberg | stevemar, only thing i was curious on for the man page was date/version | 01:39 |
| morganfainberg | stevemar other than that, it looks good to me | 01:39 |
| stevemar | oh yeah, you mentioned that during the meeting | 01:40 |
| ayoung | ok, can you create a token from a key pair? YOu can...this might be a no-go | 01:40 |
| *** SumitNaiksatam has joined #openstack-dev | 01:40 | |
| *** HenryG has joined #openstack-dev | 01:42 | |
| gyee | ayoung, you can | 01:42 |
| ayoung | morganfainberg, can you check my logic on that | 01:42 |
| morganfainberg | ayoung, looking at it now, | 01:42 |
| morganfainberg | ayoung, it's looking like you can. | 01:42 |
| ayoung | So that is the problem with delegation....you can never lose track of the fact that you are delegating. | 01:43 |
| morganfainberg | yep. | 01:43 |
| ayoung | Damnit, wish he had pinged me earlier | 01:43 |
| *** jasdeepH has quit IRC | 01:43 | |
| *** tjones has quit IRC | 01:44 | |
| ayoung | shardy_afk, you around? | 01:44 |
| *** zz_Guest12172 is now known as leif | 01:45 | |
| *** leif is now known as Guest84103 | 01:45 | |
| *** pabelanger has joined #openstack-dev | 01:46 | |
| *** nati_ueno has quit IRC | 01:48 | |
| *** nati_ueno has joined #openstack-dev | 01:48 | |
| *** adalbas has quit IRC | 01:51 | |
| *** ayoung has quit IRC | 01:52 | |
| *** insanidade has joined #openstack-dev | 01:52 | |
| *** Guest84103 is now known as zz_Guest84103 | 01:55 | |
| *** galstrom_zzz is now known as galstrom | 01:55 | |
| *** giroro_ has joined #openstack-dev | 01:56 | |
| *** alexxu has joined #openstack-dev | 01:57 | |
| *** trey_h has joined #openstack-dev | 01:57 | |
| *** freedomhui has quit IRC | 01:57 | |
| *** Ruetobas has quit IRC | 01:58 | |
| *** Ruetobas has joined #openstack-dev | 01:59 | |
| *** d34dh0r53 has joined #openstack-dev | 01:59 | |
| *** tjones has joined #openstack-dev | 01:59 | |
| *** yaguang has joined #openstack-dev | 01:59 | |
| *** dolphm has joined #openstack-dev | 02:00 | |
| *** giroro_ has quit IRC | 02:00 | |
| *** d34dh0r53 has quit IRC | 02:01 | |
| *** freedomhui has joined #openstack-dev | 02:01 | |
| *** Guest154 is now known as martine_ | 02:03 | |
| *** jbresnah_ has quit IRC | 02:04 | |
| *** colinmcnamara has joined #openstack-dev | 02:04 | |
| *** gyee has quit IRC | 02:04 | |
| *** Ruetobas has quit IRC | 02:07 | |
| *** trey_h has quit IRC | 02:08 | |
| *** stevemar has quit IRC | 02:08 | |
| *** trey_h has joined #openstack-dev | 02:08 | |
| *** Ruetobas has joined #openstack-dev | 02:10 | |
| *** dolphm has quit IRC | 02:10 | |
| *** galstrom is now known as galstrom_zzz | 02:11 | |
| *** ayoung has joined #openstack-dev | 02:11 | |
| *** dperaza1 has quit IRC | 02:12 | |
| *** trey_h has quit IRC | 02:13 | |
| *** mdenny has joined #openstack-dev | 02:13 | |
| *** Ruetobas has quit IRC | 02:14 | |
| *** tjones has quit IRC | 02:14 | |
| *** MIDENN_ has quit IRC | 02:17 | |
| *** briancurtin has quit IRC | 02:17 | |
| *** kushal has joined #openstack-dev | 02:17 | |
| *** pixelbeat has quit IRC | 02:17 | |
| *** melwitt1 has quit IRC | 02:19 | |
| *** novas0x2a|laptop has quit IRC | 02:22 | |
| *** rcleere has joined #openstack-dev | 02:22 | |
| *** gongysh has joined #openstack-dev | 02:23 | |
| *** markmcclain has joined #openstack-dev | 02:23 | |
| *** dstanek has quit IRC | 02:29 | |
| *** colinmcnamara has quit IRC | 02:29 | |
| *** briancurtin has joined #openstack-dev | 02:29 | |
| *** edmund has joined #openstack-dev | 02:30 | |
| *** jasdeepH has joined #openstack-dev | 02:30 | |
| *** stevemar has joined #openstack-dev | 02:33 | |
| stevemar | morganfainberg: is the keystone talk over? | 02:34 |
| morganfainberg | stevemar, it's been a bit quiet for a bit | 02:34 |
| stevemar | morganfainberg: mind if i ask you a q? | 02:35 |
| morganfainberg | stevemar, not at all | 02:35 |
| stevemar | morganfainberg: is it possible to 'rename' a table column? in a sql migration? | 02:35 |
| *** schwicht has quit IRC | 02:36 | |
| morganfainberg | stevemar, sure it is, but sqlite support iirc you need to do a new table, select into it, and rename the table. | 02:36 |
| stevemar | initially, i had the oauth table column "requested_roles", dolphm is asking if i can rename it t just roles | 02:36 |
| *** sungju has joined #openstack-dev | 02:36 | |
| stevemar | darn, not 'easy' :P | 02:37 |
| *** sungju has quit IRC | 02:37 | |
| morganfainberg | stevemar, https://sqlalchemy-migrate.readthedocs.org/en/latest/changeset.html#column-alter | 02:37 |
| stevemar | morganfainberg: i thought you were going to give me a link to github of an example in nova, you seem to know every line of that project :P | 02:38 |
| morganfainberg | stevemar, no i'm just good at searching ;) | 02:38 |
| *** kushal has quit IRC | 02:38 | |
| morganfainberg | stevemar, i used to work mostly with nova and horizon last year. | 02:38 |
| *** admiyo has joined #openstack-dev | 02:38 | |
| *** admiyo has quit IRC | 02:39 | |
| *** jasdeepH has quit IRC | 02:40 | |
| *** nati_ueno_2 has joined #openstack-dev | 02:42 | |
| *** nati_ueno has quit IRC | 02:46 | |
| *** nosnos_ has joined #openstack-dev | 02:47 | |
| *** shang has joined #openstack-dev | 02:47 | |
| *** nosnos has quit IRC | 02:49 | |
| *** dims has quit IRC | 02:51 | |
| *** freedomhui has quit IRC | 02:53 | |
| *** freedomhui has joined #openstack-dev | 02:53 | |
| *** danwent has quit IRC | 02:54 | |
| *** radez is now known as radez_g0n3 | 02:55 | |
| *** nosnos_ has quit IRC | 02:55 | |
| *** nosnos has joined #openstack-dev | 02:55 | |
| *** anteaya has quit IRC | 02:58 | |
| *** kushal has joined #openstack-dev | 03:00 | |
| *** otherwiseguy has quit IRC | 03:01 | |
| *** freedomhui has quit IRC | 03:01 | |
| stevemar | morganfainberg: give https://review.openstack.org/#/c/44509/ a review? | 03:02 |
| ayoung | stevemar, there is a helper function | 03:02 |
| stevemar | ayoung: whatcha mean? one already exists? | 03:03 |
| ayoung | stevemar, I think so | 03:03 |
| ayoung | looking | 03:03 |
| stevemar | ayoung ... really? i'll change things up if thats the case | 03:04 |
| ayoung | stevemar, sorry, it is for renaming a table | 03:04 |
| stevemar | ayoung: oh, whoops, the last link was for a patch to change tests stuff | 03:05 |
| ayoung | stevemar, we need to stop doing sql Alchemy migrations for sqlite....messing us up. | 03:05 |
| ayoung | OK I'm off to bed. gnight | 03:05 |
| *** ayoung has quit IRC | 03:05 | |
| stevemar | man he types fast | 03:05 |
| *** admiyo has joined #openstack-dev | 03:06 | |
| *** kiall has quit IRC | 03:07 | |
| morganfainberg | stevemar, ayoung? | 03:07 |
| morganfainberg | stevemar, yeah he does. | 03:07 |
| stevemar | morganfainberg: yeah, he wrote 4 lines in 1 second | 03:07 |
| *** mkollaro has quit IRC | 03:08 | |
| stevemar | i was half way through writing a reply and he came back with 3 replies | 03:08 |
| stevemar | anywho | 03:08 |
| morganfainberg | stevemar, i make an effort not to type that fast if I can avoid it. tends to be too much for people to take in. | 03:08 |
| morganfainberg | stevemar, _generate_pase_config to tests.core eh? | 03:08 |
| stevemar | thanks for being nice to slow folks like myself | 03:08 |
| stevemar | morganfainberg: yeah, noticed it was something fabio and i had in common | 03:09 |
| stevemar | and i was a comment gyee had made | 03:09 |
| morganfainberg | stevemar, slow? no, hard to follow via text in IM/IRC when you type crazy fast. | 03:09 |
| stevemar | nah, mostly slow | 03:09 |
| morganfainberg | nah. | 03:09 |
| stevemar | hehe | 03:09 |
| morganfainberg | you end up with 3 conversations going at once. not fun. | 03:09 |
| *** jasdeepH has joined #openstack-dev | 03:10 | |
| *** martine_ has quit IRC | 03:11 | |
| *** paragan has joined #openstack-dev | 03:12 | |
| *** dstanek has joined #openstack-dev | 03:12 | |
| *** xchu has joined #openstack-dev | 03:12 | |
| *** tsmithgc has quit IRC | 03:14 | |
| *** jimfehlig has joined #openstack-dev | 03:20 | |
| *** jbresnah_ has joined #openstack-dev | 03:25 | |
| *** nati_ueno_2 has quit IRC | 03:25 | |
| *** nati_ueno has joined #openstack-dev | 03:26 | |
| morganfainberg | stevemar, commented on your _generate_paste_config patchset | 03:26 |
| *** tjones has joined #openstack-dev | 03:28 | |
| *** kiall has joined #openstack-dev | 03:29 | |
| stevemar | morganfainberg: cool, i get what you mean, i'll try and make it happen | 03:31 |
| morganfainberg | stevemar, since we're going for less duplicated code in this instance, why not get the rest of the duplicated stuff out of there? ;) | 03:32 |
| *** freedomhui has joined #openstack-dev | 03:34 | |
| *** rcleere has quit IRC | 03:34 | |
| *** basha has joined #openstack-dev | 03:37 | |
| *** basha has quit IRC | 03:40 | |
| *** novas0x2a|laptop has joined #openstack-dev | 03:47 | |
| stevemar | morganfainberg http://paste.openstack.org/show/45715 | 03:54 |
| stevemar | i think i managed to get rid of just about all duplicated code :P | 03:54 |
| *** phschwartz has quit IRC | 03:59 | |
| *** clarkb has quit IRC | 03:59 | |
| *** mikal has quit IRC | 03:59 | |
| *** aloga has quit IRC | 03:59 | |
| *** toabctl has quit IRC | 03:59 | |
| *** garyk has joined #openstack-dev | 03:59 | |
| *** mikal_ has joined #openstack-dev | 03:59 | |
| *** clarkb has joined #openstack-dev | 03:59 | |
| *** toabctl has joined #openstack-dev | 04:00 | |
| *** phschwartz has joined #openstack-dev | 04:00 | |
| *** garyk1 has quit IRC | 04:01 | |
| *** aditirav has joined #openstack-dev | 04:02 | |
| *** aditirav has joined #openstack-dev | 04:02 | |
| *** mikal_ is now known as mikal | 04:03 | |
| *** gongysh has quit IRC | 04:03 | |
| *** prekarat has joined #openstack-dev | 04:04 | |
| *** aloga has joined #openstack-dev | 04:05 | |
| *** prekarat has quit IRC | 04:06 | |
| *** prekarat1 has joined #openstack-dev | 04:06 | |
| *** noorul``` is now known as noorul | 04:06 | |
| morganfainberg | stevemar, https://review.openstack.org/#/c/43611/2 merge conflict on that one. | 04:09 |
| stevemar | morganfainberg: cool, i'll fix that one up | 04:09 |
| stevemar | morganfainberg, something isn't working when I try to push too much into core.py (in my other patch) | 04:10 |
| morganfainberg | stevemar, how so? | 04:10 |
| stevemar | not sure, the app isn't being served | 04:11 |
| *** markmcclain has quit IRC | 04:11 | |
| stevemar | when i try to put a breakpoint in core.py, it's not stopping :P | 04:11 |
| morganfainberg | stevemar, you putting a breakpoint in your new IF-Statement? | 04:12 |
| *** sridevi has joined #openstack-dev | 04:12 | |
| morganfainberg | stevemar, i'd need to see the code to help more than very limitedly. | 04:13 |
| *** nati_ueno_2 has joined #openstack-dev | 04:14 | |
| *** Mandell has joined #openstack-dev | 04:14 | |
| stevemar | morganfainberg updated: https://review.openstack.org/#/c/43611/2 | 04:17 |
| *** dstanek has quit IRC | 04:17 | |
| *** nati_ueno has quit IRC | 04:17 | |
| stevemar | morganfainberg, breakpoint is outside the IF :P - don't worry, just griping about it :P | 04:18 |
| morganfainberg | stevemar, ok, let me know if you need any help | 04:18 |
| stevemar | probably not working because it's past midnight | 04:18 |
| stevemar | machines turn against me at that time | 04:18 |
| *** radez_g0n3 has quit IRC | 04:19 | |
| morganfainberg | lol | 04:19 |
| *** radez_g0n3 has joined #openstack-dev | 04:20 | |
| *** sridevi has quit IRC | 04:21 | |
| *** dstanek has joined #openstack-dev | 04:22 | |
| stevemar | morganfainberg one of your patches failed jenkins | 04:23 |
| stevemar | http://logs.openstack.org/76/44776/2/gate/gate-tempest-devstack-vm-full/fef150e/console.html | 04:23 |
| stevemar | the bug! https://bugs.launchpad.net/tempest/+bug/1182679 | 04:23 |
| uvirtbot | Launchpad bug 1182679 in tempest "tempest.thirdparty.boto tests fail with TearDownException" [High,Incomplete] | 04:23 |
| *** sridevi has joined #openstack-dev | 04:23 | |
| stevemar | morganfainberg, damn, you already knew | 04:24 |
| morganfainberg | stevemar, though i copy/pasta failed, i've typo'd reverify a couple times :P copied the one w/ "no bug" instead of the one i wanted. *shrug* oh well. | 04:24 |
| *** SergeyLukjanov has joined #openstack-dev | 04:25 | |
| *** basha has joined #openstack-dev | 04:25 | |
| *** mars has joined #openstack-dev | 04:30 | |
| *** mars has joined #openstack-dev | 04:30 | |
| *** zz_Guest84103 is now known as zz_zz_Guest84103 | 04:34 | |
| *** vartom7 has joined #openstack-dev | 04:35 | |
| *** reed has quit IRC | 04:38 | |
| *** vipul is now known as vipul-away | 04:44 | |
| *** yongli has joined #openstack-dev | 04:44 | |
| *** dstanek has quit IRC | 04:45 | |
| *** vipul-away is now known as vipul | 04:50 | |
| *** Ryan_Lane has joined #openstack-dev | 04:51 | |
| *** boris-42 has joined #openstack-dev | 04:51 | |
| stevemar | morganfainberg: for the tests related patch, i managed to move some of the logic to the test_v3 file... that okie? | 04:53 |
| stevemar | also, it seems to have really slowed down the tests..., maybe it's just my system | 04:54 |
| morganfainberg | stevemar, i don't see a problem with it, i just didn't want each extension to have to run that specific code. | 04:54 |
| stevemar | fair enough | 04:54 |
| *** vbian has joined #openstack-dev | 04:54 | |
| morganfainberg | stevemar, hrm, not sure how an if check on something that should be none in most cases wouild cause a significant slow-down | 04:54 |
| stevemar | just ran fabio's tests, and they are taking about 10s each | 04:54 |
| morganfainberg | stevemar, wow. | 04:55 |
| stevemar | running mine now, and it's going much slower than usual | 04:55 |
| morganfainberg | stevemar, post up your change (maybe WIP) and i'll try it on my devstack? | 04:55 |
| stevemar | i've got a check in there to make sure it doesn't run unless EXTENSION_TO_ADD is specified | 04:55 |
| morganfainberg | stevemar, still seems odd that it's that much slower | 04:56 |
| stevemar | uploading them now | 04:57 |
| stevemar | bah, need to rebase | 04:57 |
| morganfainberg | stevemar, http://paste.openstack.org/show/45717/ is what i see before your patchset. | 04:58 |
| morganfainberg | so slow(ish) compared to some of the other tests. | 04:59 |
| morganfainberg | but nothing that is far different than the previous runs of those tests | 04:59 |
| *** dstanek has joined #openstack-dev | 05:01 | |
| *** radsy has quit IRC | 05:01 | |
| *** openmike has joined #openstack-dev | 05:02 | |
| *** aditirav has quit IRC | 05:02 | |
| *** aditirav has joined #openstack-dev | 05:02 | |
| *** nati_ueno has joined #openstack-dev | 05:02 | |
| stevemar | morganfainberg https://review.openstack.org/#/c/44509/2 | 05:05 |
| *** nati_ueno_2 has quit IRC | 05:05 | |
| stevemar | took me 640seconds | 05:05 |
| stevemar | it usually took ~30 seconds | 05:05 |
| stevemar | (for oauth tests) | 05:05 |
| morganfainberg | stevemar, py27? or just a limited test set | 05:06 |
| stevemar | i dunno | 05:06 |
| stevemar | going to try and find out how other tests are running | 05:06 |
| morganfainberg | stevemar, also are you using live sql tests? or just stock like tox -epy27? | 05:06 |
| stevemar | cause this is painfully slow | 05:06 |
| morganfainberg | or you're using run_tests? | 05:06 |
| stevemar | run_tests | 05:06 |
| *** AnilV4 has joined #openstack-dev | 05:08 | |
| morganfainberg | stevemar, you're also causing a sizable amount of failures in things based on test_v3 it looks like: AttributeError: 'VersionTestCase' object has no attribute 'EXTENSION_TO_ADD' | 05:10 |
| stevemar | morganfainberg: yep, i'm seeing that now... guess that happens when I only run the tests i'm interested in | 05:10 |
| morganfainberg | stevemar, i need to look at something regarding how teardown works. | 05:11 |
| morganfainberg | stevemar, might need to move the cache region cleanup into setUp(). | 05:12 |
| morganfainberg | stevemar, not sure what is up in your env, but tests seem to be running fine here. | 05:15 |
| stevemar | morganfainberg booourns | 05:15 |
| morganfainberg | stevemar, maybe grinding into swap? | 05:16 |
| *** vartom7 has quit IRC | 05:16 | |
| stevemar | morganfainberg i looked at that, but swap seems fairly clear | 05:17 |
| stevemar | i dunno, i'm going to do run_tests.sh -u, seems like some libraries were upgraded | 05:17 |
| morganfainberg | stevemar, fyi i use tox -epy27 or tox -epep8 for my tests these days | 05:18 |
| morganfainberg | stevemar, iirc that is what jenkins uses. | 05:18 |
| stevemar | probably a good day | 05:18 |
| morganfainberg | also, when checking for failed unit tests: tox -epy27 -- -x | 05:18 |
| morganfainberg | (stops on first failed test) | 05:19 |
| *** dstanek_ has quit IRC | 05:23 | |
| *** danwent has joined #openstack-dev | 05:23 | |
| *** briancurtin has quit IRC | 05:24 | |
| stevemar | morganfainberg: yep, knew that | 05:24 |
| *** o_petit has joined #openstack-dev | 05:24 | |
| morganfainberg | stevemar, anyways. i'm going to go home. | 05:24 |
| morganfainberg | stevemar, 1hr till midnight. | 05:24 |
| morganfainberg | time to sleep (or some such) | 05:24 |
| stevemar | morganfainberg: have fun | 05:24 |
| stevemar | :P | 05:25 |
| *** sushils has joined #openstack-dev | 05:25 | |
| *** jbresnah_ has quit IRC | 05:25 | |
| *** dstanek has quit IRC | 05:28 | |
| *** vipul is now known as vipul-away | 05:29 | |
| *** Adri2000 has joined #openstack-dev | 05:31 | |
| *** Adri2000 has quit IRC | 05:31 | |
| *** Adri2000 has joined #openstack-dev | 05:31 | |
| *** gongysh has joined #openstack-dev | 05:32 | |
| *** jbresnah_ has joined #openstack-dev | 05:35 | |
| *** colinmcnamara has joined #openstack-dev | 05:35 | |
| *** sumanthns has joined #openstack-dev | 05:36 | |
| *** vbian has quit IRC | 05:36 | |
| *** openstack has joined #openstack-dev | 14:54 | |
| *** openstackstatus has joined #openstack-dev | 14:54 | |
| *** romcheg has quit IRC | 14:54 | |
| *** jmontemayor has quit IRC | 14:56 | |
| *** anniec has joined #openstack-dev | 14:57 | |
| *** trey_h_ has joined #openstack-dev | 14:57 | |
| *** imsurit has joined #openstack-dev | 14:58 | |
| *** basha has joined #openstack-dev | 15:00 | |
| *** salv-mobile has joined #openstack-dev | 15:00 | |
| bourke | does anyone know why when I base my change off another change in gerrit, typing git review also seems to update the dependant change? | 15:01 |
| *** trey_h has quit IRC | 15:01 | |
| *** edmund has joined #openstack-dev | 15:01 | |
| *** jmontemayor has joined #openstack-dev | 15:01 | |
| *** reidrac has quit IRC | 15:02 | |
| dolphm | stevemar: i am! i'm second guessing holding off on your change though | 15:02 |
| *** aeperezt has quit IRC | 15:02 | |
| dolphm | stevemar: A) it's pre-featureproposalfreeze and B) it looks like it'd be nice to have to make backporting tests to stable/havana easier | 15:02 |
| *** aeperezt has joined #openstack-dev | 15:03 | |
| dolphm | bknudson: ayoung: wishlist item for tests that could merge today... https://review.openstack.org/#/c/44509/ | 15:03 |
| *** jprovazn has quit IRC | 15:03 | |
| stevemar | dolphm: i was wondering why it was put on the short list for icehouse | 15:03 |
| ayoung | dolphm, looking | 15:03 |
| *** spzala has joined #openstack-dev | 15:03 | |
| *** jprovazn has joined #openstack-dev | 15:04 | |
| *** yjiang5 has joined #openstack-dev | 15:04 | |
| *** basha has quit IRC | 15:04 | |
| dolphm | stevemar: only because it's unnecessary for havana | 15:05 |
| ayoung | stevemar, dolphm, service_v3? That doesn't belong in core. Why was that carried along? | 15:05 |
| stevemar | dolphm: understandable | 15:05 |
| *** d34dh0r53 has joined #openstack-dev | 15:05 | |
| *** sthaha has quit IRC | 15:06 | |
| stevemar | ayoung: oh referring to: https://review.openstack.org/#/c/44509/4/keystone/tests/core.py | 15:06 |
| dolphm | ayoung: what do you mean? that's just a line from the pipeline, right? | 15:06 |
| *** jprovazn has quit IRC | 15:06 | |
| *** xga has joined #openstack-dev | 15:06 | |
| *** xga_ has quit IRC | 15:06 | |
| *** bdpayne has joined #openstack-dev | 15:07 | |
| stevemar | dolphm: yes | 15:07 |
| ayoung | dolphm, I think it is specific to *a* generated paste config, but should not be there for all of them | 15:07 |
| ayoung | right? | 15:07 |
| ayoung | it is specific to extensions | 15:08 |
| *** jimfehlig has joined #openstack-dev | 15:08 | |
| dolphm | ayoung: it's in the default v3 pipeline, and that's it | 15:08 |
| *** trey_h_ has quit IRC | 15:08 | |
| *** senk has joined #openstack-dev | 15:08 | |
| dolphm | ayoung: it's the v3 application factory | 15:08 |
| stevemar | ayoung: it's just changing this line https://github.com/openstack/keystone/blob/master/etc/keystone-paste.ini#L67 | 15:08 |
| *** datsun180b_ has joined #openstack-dev | 15:08 | |
| *** jasdeepH has joined #openstack-dev | 15:08 | |
| stevemar | to be $extension_name service_v3 | 15:08 |
| *** trey_h has joined #openstack-dev | 15:09 | |
| dolphm | i.e. injecting the extension to be tested as the last extension in the pipeline | 15:09 |
| ayoung | dolphm, I'm just not sure that only extensions should be using this mechanism, but it can stay as is...I might be scope creeping it | 15:09 |
| ayoung | dolphm, there are many tests that could benefit from test specific config files. | 15:09 |
| *** datsun180b has quit IRC | 15:09 | |
| *** datsun180b_ is now known as datsun180b | 15:09 | |
| *** d34dh0r53 has quit IRC | 15:10 | |
| ayoung | But this is a good start | 15:10 |
| dolphm | ayoung: you mean test-specific pipelines? | 15:10 |
| dolphm | ayoung: since our conf is separate now, we already test that differently | 15:11 |
| *** bdpayne has quit IRC | 15:11 | |
| *** martyntaylor has quit IRC | 15:11 | |
| *** senk1 has joined #openstack-dev | 15:11 | |
| ayoung | dolphm, I was thinking more along the lines of things like LDAP and SQL Urls | 15:11 |
| ayoung | we have a lot of duplication. | 15:12 |
| *** xga_ has joined #openstack-dev | 15:12 | |
| *** xchu has joined #openstack-dev | 15:12 | |
| *** jtomasek has quit IRC | 15:12 | |
| *** drewland_ has quit IRC | 15:12 | |
| *** senk has quit IRC | 15:13 | |
| *** reed has joined #openstack-dev | 15:13 | |
| *** henrynash has joined #openstack-dev | 15:13 | |
| *** gargya has joined #openstack-dev | 15:14 | |
| *** xga has quit IRC | 15:14 | |
| dolphm | ayoung: ah, that should be handled through the conf hierarchy thing we do | 15:15 |
| dolphm | ayoung: conf extending conf | 15:15 |
| *** salv-mobile has quit IRC | 15:16 | |
| *** bdpayne has joined #openstack-dev | 15:16 | |
| *** otherwiseguy has joined #openstack-dev | 15:17 | |
| *** drewlander has joined #openstack-dev | 15:18 | |
| *** romcheg has joined #openstack-dev | 15:21 | |
| dolphm | ayoung: (did you want to +1/-1 on that?) | 15:22 |
| *** kmartin has quit IRC | 15:22 | |
| *** datsun180b has quit IRC | 15:23 | |
| *** datsun180b has joined #openstack-dev | 15:23 | |
| *** kiall has quit IRC | 15:24 | |
| *** mlavalle has quit IRC | 15:24 | |
| *** dkehn has quit IRC | 15:25 | |
| *** gyee has joined #openstack-dev | 15:25 | |
| *** trey_h has quit IRC | 15:25 | |
| *** trey_h has joined #openstack-dev | 15:26 | |
| *** jtomasek has joined #openstack-dev | 15:27 | |
| *** nati_ueno has quit IRC | 15:27 | |
| *** shanewang is now known as shanewang_away | 15:28 | |
| *** jimfehlig has quit IRC | 15:28 | |
| *** nati_ueno has joined #openstack-dev | 15:28 | |
| *** danwent has joined #openstack-dev | 15:28 | |
| *** kmartin has joined #openstack-dev | 15:30 | |
| *** mlavalle has joined #openstack-dev | 15:30 | |
| *** edmund1 has joined #openstack-dev | 15:31 | |
| *** edmund has quit IRC | 15:31 | |
| dolphm | ayoung: this is another "big" bug fix that should merge today https://review.openstack.org/#/c/42826/ | 15:31 |
| *** pmathews has joined #openstack-dev | 15:31 | |
| *** ruhe has joined #openstack-dev | 15:31 | |
| *** sumanthns has joined #openstack-dev | 15:32 | |
| *** nati_ueno has quit IRC | 15:32 | |
| *** dkehn has joined #openstack-dev | 15:34 | |
| *** galstrom_zzz is now known as galstrom | 15:34 | |
| *** guardian_ has quit IRC | 15:35 | |
| *** sumanthns has quit IRC | 15:37 | |
| *** basha has joined #openstack-dev | 15:37 | |
| *** mestery_ has joined #openstack-dev | 15:37 | |
| *** prekarat has quit IRC | 15:38 | |
| stevemar | dolphm: ping | 15:38 |
| *** bswartz has joined #openstack-dev | 15:39 | |
| *** gimps_ has joined #openstack-dev | 15:40 | |
| *** mestery has quit IRC | 15:40 | |
| *** dkehn has quit IRC | 15:40 | |
| ayoung | dolphm, why should changing my default membership remove me from a project? | 15:41 |
| *** dkehn has joined #openstack-dev | 15:41 | |
| *** dolphm_ has joined #openstack-dev | 15:41 | |
| ayoung | dolphm, I think the bug report is wrong | 15:43 |
| dolphm_ | ayoung: how so? | 15:43 |
| *** mestery_ is now known as mestery | 15:43 | |
| ayoung | dolphm_, what they are changin in the user object is the default tenant. | 15:43 |
| ayoung | Say I want to change my default tenant from p1 to p2...I should not be removed from p1 | 15:43 |
| *** vipul is now known as vipul-away | 15:44 | |
| ayoung | I think they are right in saying "we should create an user-project role association for p2 if it does not exist" | 15:44 |
| dolphm_ | ayoung: you should lose _member_ on p1 if it's called through the v2 api | 15:44 |
| *** kiall has joined #openstack-dev | 15:44 | |
| ayoung | but not in saying "in spite of changing successfully tenant/project id for a existing user, the user still belongs to the original tenant / project. That means, updating tenant / project id for a user is invalid." | 15:45 |
| *** gargya_ has joined #openstack-dev | 15:45 | |
| ayoung | dolphm, you can be a member of more than one tenant, even in v2, no? | 15:45 |
| dolphm_ | ayoung: yes, but not via a user's default tenancy | 15:45 |
| *** dolphm has quit IRC | 15:45 | |
| *** basha has quit IRC | 15:45 | |
| *** zz_zz_Guest84103 is now known as leif | 15:46 | |
| *** leif is now known as Guest61470 | 15:46 | |
| ayoung | dolphm_, So you think that changing a default tenancy should remove the _member_ role in the old project? Is that documented? Or was that a side effect of how it used to work when we had multiple ways to associated | 15:47 |
| *** paragan has quit IRC | 15:47 | |
| ayoung | associate users with projects | 15:47 |
| *** gargya has quit IRC | 15:47 | |
| *** gargya_ is now known as gargya | 15:47 | |
| *** freedomhui has quit IRC | 15:47 | |
| dolphm_ | ayoung: yeah, that was the point of the _member_ role... to provide an explicit role assignment on an implicit v2 role grant (default tenancy) | 15:47 |
| dolphm_ | ayoung: we don't need the same contruct on v2 because default_project_id doesn't implicitly grant authz | 15:48 |
| *** bdpayne has quit IRC | 15:49 | |
| *** vipul-away is now known as vipul | 15:50 | |
| *** yaguang has quit IRC | 15:50 | |
| *** changbl has joined #openstack-dev | 15:51 | |
| *** imsurit has quit IRC | 15:51 | |
| ayoung | dolphm_, it wasn't just on default tenancy. THe tenant used to have a list of user_memebers. I think that was separate from the default_tenant_id as well. I don't know if that relationship would have been modified by changing the user's default tenant_id, I think not. But I can look back in git to see... | 15:52 |
| *** FunnyLookinHat has quit IRC | 15:52 | |
| *** gimps has quit IRC | 15:53 | |
| *** dhellmann is now known as dhellmann_ | 15:55 | |
| *** gimps has joined #openstack-dev | 15:56 | |
| *** openfly has left #openstack-dev | 15:57 | |
| *** basha has joined #openstack-dev | 15:58 | |
| *** briancurtin has quit IRC | 15:59 | |
| *** jistr has quit IRC | 15:59 | |
| *** markwash has quit IRC | 15:59 | |
| *** jmontemayor has quit IRC | 16:00 | |
| *** dina_belova has quit IRC | 16:00 | |
| *** jimfehlig has joined #openstack-dev | 16:00 | |
| *** tstevenson has joined #openstack-dev | 16:00 | |
| *** jcoufal has quit IRC | 16:00 | |
| dstanek | dolphm_: re: dogpile cache, i was thinking that on_arguments could be replaced with a lazy implementation | 16:00 |
| *** datsun180b_ has joined #openstack-dev | 16:01 | |
| *** Ruetobas has quit IRC | 16:01 | |
| ayoung | dolphm_, so, looks like we never did anything with user membership based on what was passed in for an update https://github.com/openstack/keystone/blob/b1bfca2501ad11a861c9064b97b7fa06fc6d958e/keystone/identity/backends/sql.py#L710 | 16:01 |
| ayoung | that is the last commit prior to my change which added _member_ | 16:01 |
| dolphm_ | dstanek: ++, but it looked like more? (there's three imports from dogpile and several are used at startup) | 16:02 |
| henrynash | dolphm: ping me when you are OK to talk filtering….I am sure we can get this done and resolved today | 16:02 |
| *** xchu has quit IRC | 16:02 | |
| dolphm_ | ayoung: that's the bug, no? | 16:02 |
| dolphm_ | henrynash: have an update? | 16:03 |
| *** datsun180b has quit IRC | 16:03 | |
| *** datsun180b_ is now known as datsun180b | 16:03 | |
| henrynash | dolphm: so just wanted to talk though your idea of passing a list of dicts | 16:03 |
| *** briancurtin has joined #openstack-dev | 16:03 | |
| *** vartom7 has joined #openstack-dev | 16:03 | |
| ayoung | dolphm_, I don't think there is any compelling reason to remove a user from a project just because we change their default project. It is counterintuitive. I can see that setting a users default project automatcially enrolls them in that project. BUt not the reverse. | 16:04 |
| dstanek | dolphm_: the only other one that's used at import time is proxy and that can be lazy as well | 16:04 |
| *** Ruetobas has joined #openstack-dev | 16:04 | |
| *** xga_ has quit IRC | 16:04 | |
| *** briancurtin has quit IRC | 16:04 | |
| dolphm_ | dstanek: hmm.. i thought i saw more; want to attempt an impl? | 16:04 |
| dstanek | that's not the right long term solution, but it could be a temp fix | 16:04 |
| *** briancurtin has joined #openstack-dev | 16:05 | |
| dstanek | dolphm_: sure thing | 16:05 |
| *** xga has joined #openstack-dev | 16:05 | |
| dolphm_ | ayoung: you have to be more specific than that... you certainly wouldn't remove every role assignment, or just any random role assignment | 16:05 |
| dolphm_ | ayoung: the only reason they have the _member_ role is because they were created with default tenancy | 16:05 |
| *** devvesa has quit IRC | 16:05 | |
| *** jaybuff has joined #openstack-dev | 16:05 | |
| dolphm_ | ayoung: so they should lose the _member_ assignemnt when default tenancy changes | 16:05 |
| henrynash | dolphm: so larger issue I had with this was whether we need to differentiate between filters and other things that we might pass (e.g. paginating directives) as well being able to tell the driver which filters we actually support | 16:06 |
| *** briancurtin has quit IRC | 16:06 | |
| henrynash | dolphm: and while I know that pagination may get booted off the table etc. | 16:06 |
| henrynash | dolphm: the attempt was to design something that had legs and could be adapted | 16:06 |
| *** briancurtin has joined #openstack-dev | 16:06 | |
| ayoung | dolphm_, no, they were created to replace the direct member_of value of the tenant. Not for default | 16:06 |
| jaybuff | mordred: what happened to cookiecutter-openstack under openstack-dev? I found it on https://github.com/emonty/cookiecutter-openstack is that the best one to use? | 16:07 |
| insanidade | question: I'd like to restart the services in my Openstack installation (devstack). I've used unstack.sh first. What should I do next? stack.sh ? | 16:07 |
| dolphm_ | henrynash: on your first point, do you mean like domain_id in list_projects()? | 16:07 |
| mordred | jaybuff: it hasn't gotten merged yet | 16:07 |
| *** briancurtin has left #openstack-dev | 16:07 | |
| mordred | jaybuff: https://review.openstack.org/42530 | 16:07 |
| mordred | jaybuff: but - feel free to use the emonty one for now | 16:07 |
| dolphm_ | henrynash: if so, i agree. but i would think the solution is two discrete driver methods ... list_projects(hints=None) and list_projects_in_domain(domain_id, hints=None) | 16:08 |
| jaybuff | mordred: ah, cool, so it was never on openstack-dev. Thanks, I used it and it worked great. | 16:08 |
| mordred | jaybuff: woot! | 16:08 |
| mordred | jaybuff: as soon as that lands, I'll send out a mailing list thing letting people know about it | 16:08 |
| *** yjiang5 has left #openstack-dev | 16:08 | |
| *** SergeyLukjanov has quit IRC | 16:08 | |
| *** tstevenson has quit IRC | 16:08 | |
| *** mmagr has quit IRC | 16:08 | |
| dolphm_ | henrynash: treating domain_id as optional to the driver was bad design on my part (at least, i *think* i wrote that)... but anyway, your change illustrates that pretty well | 16:09 |
| *** Ruetobas has quit IRC | 16:09 | |
| henrynash | dolphm: which is interesting, since the driver method used to take domain_id=None…i.e. an optional driver id, which I folded into the hints/directves | 16:09 |
| *** nati_ueno has joined #openstack-dev | 16:09 | |
| henrynash | dolphm: add | 16:09 |
| henrynash | dolphm: ahh, right | 16:09 |
| *** yassine has quit IRC | 16:09 | |
| *** zbitter has joined #openstack-dev | 16:09 | |
| henrynash | dolphm: the other thing is that, do we let any name, value pair that arrives in the url go into the hints? | 16:10 |
| *** diogogmt has joined #openstack-dev | 16:11 | |
| *** ayoung has quit IRC | 16:11 | |
| primeministerp | dansmith: ping | 16:11 |
| henrynash | dolphm: i.e. we could just let them all be in hints, and if a driver finds it doesn't udnertand one of them it just skips it….. | 16:11 |
| *** odyssey4me has quit IRC | 16:12 | |
| dolphm_ | henrynash: i'd rather explicitly build 'hints' (or whatever) in the controller, rather than passing any random garbage down to drivers | 16:12 |
| *** FunnyLookinHat has joined #openstack-dev | 16:13 | |
| dolphm_ | henrynash: domain_id in the query string is basically a different call to the driver, and should have it's own driver method | 16:13 |
| *** zaneb has quit IRC | 16:13 | |
| dolphm_ | henrynash: *or* if it's low cost to implement it in the controller and not have the driver worry about it, we can do that | 16:13 |
| *** hartsocks has joined #openstack-dev | 16:14 | |
| *** Ruetobas has joined #openstack-dev | 16:14 | |
| *** hartsocks has left #openstack-dev | 16:14 | |
| henrynash | dolphm: oh sure, not saying that….but do you say the url has tentsize=100 (where tentszie is not a known field) to so still build it into hints? | 16:14 |
| *** o_petit has quit IRC | 16:14 | |
| dolphm_ | henrynash: i wouldn't | 16:14 |
| *** boris-42 has quit IRC | 16:14 | |
| *** vuil has joined #openstack-dev | 16:14 | |
| dolphm_ | henrynash: sounds like a problem for the controller to either implement or ignore? | 16:14 |
| dolphm_ | henrynash: and maybe ignore.. because it's not a contract attribute | 16:15 |
| *** mlavalle has quit IRC | 16:15 | |
| *** tsmithgc has joined #openstack-dev | 16:15 | |
| henrynash | dolphm: which I wasn't….until gyee popped up and said that he really wanted access to the other items from the url…so rather than pass the whoel querystting down, I, but any url items into a separate dict that could also be iterated | 16:16 |
| henrynash | dolphm:…all part of the ListDierective | 16:16 |
| dolphm_ | henrynash: side note... i thought it was interesting that (it appears) you implemented (maybe unintentionally) support for name__startswith=Hen&name__contains=nr ... e.g. multiple filters on the same field | 16:16 |
| dolphm_ | henrynash: if so, add a test to ensure that works :) | 16:16 |
| *** zbitter is now known as zaneb | 16:17 | |
| *** angdraug has joined #openstack-dev | 16:17 | |
| dolphm_ | henrynash: it sounds like he should be implementing an api extension, not a custom driver | 16:17 |
| henrynash | dolphm: that was certainly the design, I haven't fully fleshed out the inexact testing since it is switched off | 16:17 |
| *** amotoki has quit IRC | 16:17 | |
| *** o_petit has joined #openstack-dev | 16:17 | |
| *** basha has quit IRC | 16:17 | |
| dolphm_ | gyee: (i've only heard you ask for a specific solution rather than actually presenting a use case) | 16:17 |
| *** vartom8 has joined #openstack-dev | 16:17 | |
| gyee | dolphm_, I have customer drivers that make use of the query parameters | 16:18 |
| dolphm_ | gyee: that's still not a use case, that's just saying "i'm implementing an API in the driver" | 16:19 |
| dolphm_ | which is ... broken | 16:19 |
| gyee | much easier to pass the query string down the driver than having to write a bunch of boilerplate code in contrib | 16:19 |
| dolphm_ | gyee: still not presenting a use case | 16:19 |
| *** o_petit has quit IRC | 16:20 | |
| gyee | dolphm_, what's the easier way to implement GET /domains?HP-<filtername>=... | 16:20 |
| *** safchain_ has quit IRC | 16:20 | |
| gyee | either customer driver or contrib | 16:20 |
| *** vartom7 has quit IRC | 16:20 | |
| *** topshare has joined #openstack-dev | 16:20 | |
| dolphm_ | gyee: that looks like a really small piece of middleware at first glance, but i still don't know what you're doing | 16:21 |
| *** ruhe has quit IRC | 16:21 | |
| gyee | dolphm_, we support more filter options than standard keystone provides | 16:21 |
| *** MaxV has quit IRC | 16:21 | |
| gyee | I need to figure out a way to do that without having to reimplement keystone itself | 16:22 |
| gyee | I can write a bunch of boilerplate code in contrib or simple make use of what's already there in my custom driver | 16:23 |
| dolphm_ | gyee: middleware has access to both the query string and the response from keystone - that would be the lowest cost way to implement a new filter in terms of long term maintenance cost | 16:23 |
| gyee | dolphm_, but I need a way to pass them down to the driver | 16:23 |
| *** corXi has quit IRC | 16:23 | |
| *** jaybuff has quit IRC | 16:24 | |
| gyee | I would rather not having to modify/monkeypatch the interface | 16:24 |
| dolphm_ | gyee: it sounds like you're literally writing an api extension and entirely implementing in a driver, which is broken and unmaintainable | 16:25 |
| gyee | its a simple change in henry's patch, just save the query string and make them accessible in the driver | 16:25 |
| dolphm_ | gyee: you don't need to push your broken architecture back into henry's patch | 16:25 |
| dstanek | just created a new devstack instance, checked out master in /opt/stack/keystone and have failing tests :-( | 16:25 |
| dolphm_ | dstanek: can you paste the failure? | 16:26 |
| gyee | dolphm_, how would you implement custom filters? | 16:26 |
| *** shinylasers has quit IRC | 16:26 | |
| dolphm_ | gyee: intercept calls to the resource you care about 'GET' '/domains' on the pipeline you care about 'api_v3' and intercept and rewrite the wsgi response based on the query parameters in the wsgi environment | 16:27 |
| dstanek | dolphm_: there are a few hundred failures, but http://paste.openstack.org/show/45750/ happens a lot | 16:27 |
| dolphm_ | gyee: that's pretty much how json<-->xml middleware works | 16:27 |
| gyee | dolphm_, I need to pass them into the database drivers | 16:28 |
| dolphm_ | gyee: it's a little more complex than you need because it's rewriting both requests AND responses | 16:28 |
| gyee | for performance reasons | 16:28 |
| *** fbo is now known as fbo_away | 16:28 | |
| dolphm_ | gyee: then the next step would be to call your custom driver | 16:29 |
| gyee | dolphm_, I would have to implement/extend the core interfaces | 16:29 |
| dolphm_ | gyee: i.e. intercept the request rather than the response | 16:29 |
| *** jpich has quit IRC | 16:29 | |
| *** shinylasers has joined #openstack-dev | 16:29 | |
| dolphm_ | gyee: yes, that's totally catered by design | 16:29 |
| gyee | dolphm_, I can easiler intercept the request and response | 16:29 |
| gyee | problem is I need to pass the query parameters down to the database drivers | 16:30 |
| *** kiall has quit IRC | 16:30 | |
| gyee | via the Manager interface | 16:30 |
| dstanek | i remember seing somethere that _ is magically put in a module's namespace...what does that? | 16:30 |
| dolphm_ | gyee: please stop trying to push your broken architecture as some sort of magic solution to a problem it sounds like you don't actually understand... we have better things to do today | 16:32 |
| *** JordanP has quit IRC | 16:32 | |
| *** MaxV has joined #openstack-dev | 16:33 | |
| *** atiwari has joined #openstack-dev | 16:33 | |
| *** senk1 has quit IRC | 16:33 | |
| dolphm_ | dstanek: that just changed within the past week.. i think keystone.__init__ ? or keystone-all | 16:33 |
| dolphm_ | bknudson: i think you refactored how _() is created? | 16:34 |
| *** pmathews has quit IRC | 16:35 | |
| gyee | dolphm_, I am trying to a solution, no magic | 16:35 |
| henrynash | dolphm: can we also agree what we call this thing we are debating (I only want to change the name ONE more time :-) ). hints? list_hints? list_directives? | 16:36 |
| gyee | to fine a solution | 16:36 |
| *** sridevi has joined #openstack-dev | 16:36 | |
| *** sridevi has quit IRC | 16:36 | |
| *** jruzicka has quit IRC | 16:37 | |
| *** stevemar has quit IRC | 16:37 | |
| *** MaxV has quit IRC | 16:37 | |
| *** networkstatic has joined #openstack-dev | 16:38 | |
| *** dkehn has quit IRC | 16:38 | |
| dolphm_ | henrynash: sure... | 16:38 |
| dolphm_ | henrynash: between hints and list_hints... do you only expect list_* methods to consume hints? | 16:39 |
| dolphm_ | henrynash: (or directives vs list_directives) | 16:39 |
| henrynash | dolphm: I can't quite think of a non-list case, to be honest…all others should be real api parameters... | 16:40 |
| dolphm_ | henrynash: agree | 16:40 |
| henrynash | dolphm_: so list_hints? | 16:41 |
| *** AlexR has joined #openstack-dev | 16:41 | |
| dolphm_ | henrynash: i don't know whether that should affect anything ;) i was just curious about what you were thinking | 16:41 |
| *** kiall has joined #openstack-dev | 16:41 | |
| *** kiall has quit IRC | 16:41 | |
| *** kiall has joined #openstack-dev | 16:41 | |
| *** AlexR has quit IRC | 16:41 | |
| *** senk has joined #openstack-dev | 16:41 | |
| *** AlexR has joined #openstack-dev | 16:41 | |
| dolphm_ | henrynash: i'm between just 'hints' or 'directives' or '<some unknown alternative>' | 16:41 |
| *** AlexR has quit IRC | 16:42 | |
| *** pmathews has joined #openstack-dev | 16:42 | |
| *** johnthetubaguy has quit IRC | 16:42 | |
| dolphm_ | henrynash: and i lean towards hints because there's precedent, and it doesn't remind me of hitler | 16:42 |
| henrynash | dolphm: I wasn't 100% sold on directives….since a driver can chose to ignore | 16:42 |
| *** imsurit has joined #openstack-dev | 16:42 | |
| *** colinmcnamara has joined #openstack-dev | 16:42 | |
| dolphm_ | henrynash: just go with 'hints' then | 16:43 |
| *** senk1 has joined #openstack-dev | 16:43 | |
| henrynash | dolphm_: ok, done | 16:43 |
| dolphm_ | henrynash: *if* you need a new module in keystone.common at the end of the day ... 'driver_hints' ? | 16:43 |
| *** jbresnah_ has joined #openstack-dev | 16:43 | |
| henrynash | dolphm_: ok, nice | 16:43 |
| *** AlexR has joined #openstack-dev | 16:43 | |
| dolphm_ | henrynash: but in reality, hints are built by the controller according to a proprietary api, and provided in a standard format to the drivers | 16:44 |
| *** shinylasers has quit IRC | 16:44 | |
| *** nshaikh has left #openstack-dev | 16:44 | |
| henrynash | dolphm_: (agreed)….on the subject of structure, complexity do lists or dicts etc. | 16:44 |
| *** senk1 has quit IRC | 16:44 | |
| dolphm_ | henrynash: so if there's something 'common' about them that's not api-specific, it would be something to read / traverse hints? or validate them? | 16:44 |
| dstanek | dolphm_: bknudson: found it (the _ change) in 2fe72a0 | 16:44 |
| henrynash | dolphm_: so if we have a simple list of dicts (which I did think about), then could do: | 16:45 |
| *** kbrierly has joined #openstack-dev | 16:45 | |
| henrynash | a) All dicts are filters | 16:45 |
| *** senk has quit IRC | 16:45 | |
| *** jhesketh has quit IRC | 16:45 | |
| *** abhishekkr has quit IRC | 16:46 | |
| henrynash | b) We could have a 'type' field in the dict that today said 'filter', but would allow us to add other things that weren't filters in the futre | 16:46 |
| dolphm_ | henrynash: what would be an example of an alternative type? | 16:47 |
| henrynash | which is kind of a simpler way of doing what I did with multiple doctor dicts :-) | 16:47 |
| *** xga has quit IRC | 16:47 | |
| henrynash | dolphm_: pagination hints | 16:47 |
| *** bswartz has quit IRC | 16:47 | |
| AlexR | Hello everyone, could you help me with git? I'm having troubles updating my change. I did this before, but now it does not work. So I made changes in my branch, I did 'commit -a --amend', and when I do 'git review' I get the following message: Errors running git rebase -i remotes/gerrit/master error: could not apply fdb8d8f... Additional variable APACHE_SERVER_NAME added When you have resolved this problem run "git rebase --continue". | 16:47 |
| AlexR | If you would prefer to skip this patch, instead run "git rebase --skip". To check out the original branch and stop rebasing run "git rebase --abort". Could not apply fdb8d8f... Additional variable APACHE_SERVER_NAME added | 16:47 |
| *** asavu has joined #openstack-dev | 16:47 | |
| henrynash | dolphm_: if we do them | 16:47 |
| dolphm_ | henrynash: that sounds good | 16:47 |
| *** tjones has joined #openstack-dev | 16:47 | |
| dolphm_ | henrynash: i also liked your choice of 'comparate' | 16:48 |
| dolphm_ | henrynash: i found that interesting because i spent several weeks on a project fussing over similar terminology before i found & settled on the word 'comparate' (which i didn't know existed before that) | 16:49 |
| *** prekarat has joined #openstack-dev | 16:49 | |
| henrynash | dolphm_: ..and the other thing we *could* do is have type called 'extra' or 'other" which is where we placed the name & value of things from the url that were not filters (or paginators)…..but this goes to the discussion you are having with gyee | 16:50 |
| *** dkehn has joined #openstack-dev | 16:50 | |
| henrynash | dolphm_: (actually I think i used 'comparator'…) | 16:50 |
| *** shinylasers has joined #openstack-dev | 16:50 | |
| dolphm_ | henrynash: bah, that's what i meant | 16:50 |
| dolphm_ | henrynash: yeah, it doens't make sense to pass garbage down to the driver if the controller doesn't know what it is | 16:51 |
| *** kiall has quit IRC | 16:52 | |
| dolphm_ | henrynash: for example, it would make sense for the v2 api to filter on tenantId and it would make sense for the v3 api to filter on default_project_id, but one or both of those is going to be meaningless to the driver... because it's api-specific that doesn't belong at that layer | 16:53 |
| *** derekh has quit IRC | 16:53 | |
| *** johnthetubaguy has joined #openstack-dev | 16:53 | |
| *** ruhe has joined #openstack-dev | 16:54 | |
| henrynash | dolphm_: I agree, I think it is only useful for customer drivers and extensions…. | 16:54 |
| dolphm_ | henrynash: gyee: LDAP bug fix- https://review.openstack.org/#/c/43524/ (bknudson: thank you!) | 16:54 |
| *** johnthetubaguy1 has joined #openstack-dev | 16:54 | |
| *** johnthetubaguy1 has quit IRC | 16:54 | |
| *** topshare has quit IRC | 16:54 | |
| henrynash | dolphm_: at least with this structure we can add something like that if we convince ourselves we need it | 16:54 |
| *** nati_ueno has quit IRC | 16:55 | |
| *** pschaef has quit IRC | 16:55 | |
| *** kiall has joined #openstack-dev | 16:55 | |
| *** marios has quit IRC | 16:55 | |
| henrynash | dolphm_: ok, working on a new patch.... | 16:55 |
| gyee | dolphm_, henrynash, I guess I'll have to monkeypatch listdirectives in contrib | 16:55 |
| *** anniec has quit IRC | 16:55 | |
| *** nati_ueno has joined #openstack-dev | 16:55 | |
| *** anniec has joined #openstack-dev | 16:55 | |
| *** AlexR has quit IRC | 16:56 | |
| *** xga has joined #openstack-dev | 16:57 | |
| *** jaypipes has quit IRC | 16:57 | |
| *** vartom8 has quit IRC | 16:57 | |
| *** colinmcnamara has quit IRC | 16:57 | |
| *** nati_ueno has quit IRC | 16:57 | |
| *** nati_ueno has joined #openstack-dev | 16:57 | |
| *** johnthetubaguy has quit IRC | 16:58 | |
| *** nati_ueno has quit IRC | 16:58 | |
| *** nati_ueno has joined #openstack-dev | 16:59 | |
| *** colinmcnamara has joined #openstack-dev | 16:59 | |
| *** markwash has joined #openstack-dev | 16:59 | |
| *** kaushikc has joined #openstack-dev | 17:00 | |
| *** ruhe has quit IRC | 17:00 | |
| *** jbresnah_ has quit IRC | 17:01 | |
| *** jbresnah_ has joined #openstack-dev | 17:01 | |
| *** danpb has quit IRC | 17:02 | |
| *** imsurit has quit IRC | 17:02 | |
| *** SumitNaiksatam has quit IRC | 17:02 | |
| *** Ryan_Lane has joined #openstack-dev | 17:03 | |
| *** Ryan_Lane has quit IRC | 17:03 | |
| dolphm_ | gyee: ... i'm going to pretend you've already read and understood this code but i'm going to link you to it so i can be more confident in my delusion https://github.com/openstack/keystone/blob/master/keystone/middleware/core.py#L143-L171 | 17:03 |
| *** nati_ueno has quit IRC | 17:03 | |
| *** Shaan7 has quit IRC | 17:04 | |
| *** Shaan7 has joined #openstack-dev | 17:04 | |
| gyee | dolphm_, oh yeah, been there :) | 17:05 |
| *** ayoung has joined #openstack-dev | 17:05 | |
| *** ruhe has joined #openstack-dev | 17:08 | |
| *** boris-42 has joined #openstack-dev | 17:08 | |
| gyee | I have a number of options, catching the response and apply my own filters on top is one of them | 17:08 |
| *** Mandell has joined #openstack-dev | 17:08 | |
| gyee | performance would suck, but yeah, its an option nevertheless | 17:08 |
| *** senk has joined #openstack-dev | 17:10 | |
| dolphm_ | gyee: best to prematurely optimize with an unmaintainable solution then? | 17:11 |
| *** xga has quit IRC | 17:11 | |
| gyee | dolphm_, I need to figure out a solution this is both maintainable and have decent performance | 17:12 |
| gyee | performance comes from drivers | 17:12 |
| morganfainberg | dolphm_, the use of the decorator requires dogpile.cache to be a hard requirement =/ | 17:13 |
| morganfainberg | dolphm_ (sorry just saw the email thread). | 17:13 |
| dolphm_ | dstanek: ^ | 17:13 |
| *** jcoufal has joined #openstack-dev | 17:13 | |
| morganfainberg | dolphm_, i'll respond there as well. | 17:14 |
| morganfainberg | (email) | 17:14 |
| *** sumanthns has joined #openstack-dev | 17:14 | |
| bknudson | I'd think given python's dynamic nature there wouldn't be any hard requirements? | 17:14 |
| dolphm_ | bknudson: AFAICT, it would at least a bit of a pain to avoid | 17:14 |
| morganfainberg | bknudson, decorators are handled at import time, it's really hard to work around. | 17:14 |
| dstanek | morganfainberg: dolphm_: i have a POC for making the decorator and proxy lazy, but i'm waiting to post until i get my devstack issues worked out | 17:15 |
| dstanek | need to make sure the tests pass :-) | 17:15 |
| morganfainberg | though if it is an issue, i could write an abstraction layer to lazy load dogpile today | 17:15 |
| morganfainberg | dolphm_, not sure how much work it would be thouhg. | 17:15 |
| *** jaypipes has joined #openstack-dev | 17:16 | |
| dolphm_ | dstanek: can you post a diff at least? i'm fearing the final solution would be fairly complex for havana vs shipping with dogpile as hard requirement | 17:16 |
| dolphm_ | (and once we ship with dogpile as a hard requirement, we might as well keep it as such?) | 17:16 |
| *** nati_ueno has joined #openstack-dev | 17:17 | |
| morganfainberg | dolphm_, i was hoping to get to a point where we could make cache a default option in I. | 17:17 |
| dolphm_ | morganfainberg: that's what dstanek is tackling right now | 17:17 |
| *** mrunge has quit IRC | 17:17 | |
| dolphm_ | morganfainberg: default to caching enabled? against what backend? | 17:17 |
| *** otherwiseguy has quit IRC | 17:17 | |
| morganfainberg | dolphm_, thats part of the work that would be needed, but i'm not sure yet :P, like i said hope | 17:18 |
| dolphm_ | morganfainberg: well, what are you thinking? defaulting to a more intelligent in-memory backend? | 17:19 |
| morganfainberg | dolphm_, that was the first path to go down. | 17:19 |
| *** stevemar has joined #openstack-dev | 17:19 | |
| dolphm_ | morganfainberg: defaulting to memcached might also be reasonable | 17:19 |
| morganfainberg | dolphm_, work on something that could handle expiring the cache rather than ballooning the memory useage. | 17:19 |
| morganfainberg | dolphm_, that was my second thought | 17:20 |
| dolphm_ | morganfainberg: i'd be curious if any other projects need memcached (or something else suitable) out of the box | 17:20 |
| *** tjones has left #openstack-dev | 17:20 | |
| dolphm_ | (i can't think of any) | 17:20 |
| morganfainberg | dolphm_, AFAIK not needed out of the box. | 17:20 |
| markwash | ttx: around? I'm still trying to put together an understanding of how clients release schedules need to coincide with the openstack release cycle | 17:21 |
| stevemar | anyone know if you can edit a summit sessions after it's been created? | 17:22 |
| *** SumitNaiksatam has joined #openstack-dev | 17:22 | |
| morganfainberg | dstanek, if you need any help with making the dogpile.cache decorator a bit more lazy, let me know. | 17:22 |
| *** SumitNaiksatam has quit IRC | 17:23 | |
| *** kiall has quit IRC | 17:23 | |
| *** jruzicka has joined #openstack-dev | 17:23 | |
| dstanek | dolphm_: sure, let me stash all my other experiments | 17:24 |
| *** SumitNaiksatam has joined #openstack-dev | 17:24 | |
| *** SumitNaiksatam has quit IRC | 17:24 | |
| dstanek | morganfainberg: thanks, i'll let you know when my code is posted | 17:24 |
| *** feleouet has quit IRC | 17:24 | |
| dolphm_ | markwash: i don't follow any schedule at all.. just release as deemed necessary / beneficial | 17:25 |
| *** jmontemayor has joined #openstack-dev | 17:25 | |
| dolphm_ | markwash: at the same time, i'm aware that distros will probably stick with whatever happens to be available at the time of a service release | 17:25 |
| markwash | dolphm_: okay cool. . I was hoping that during FF I could focus some people on a client cleanup push | 17:25 |
| morganfainberg | dstanek, for what it's worth, it should be in the requirements.txt for keystone? | 17:26 |
| dolphm_ | markwash: fwiw, i had the same thought last cycle but it didn't work out due to RC bugs + FFE taking up more time than expected | 17:26 |
| *** pmathews has quit IRC | 17:27 | |
| *** prekarat has quit IRC | 17:28 | |
| markwash | dolphm_: hmm that seems likely to be something I'll have to contend with as well :-) | 17:29 |
| *** drewlander has quit IRC | 17:29 | |
| *** pmathews has joined #openstack-dev | 17:30 | |
| dolphm_ | stevemar: ping | 17:30 |
| *** kiall_ has joined #openstack-dev | 17:31 | |
| *** kiall_ has quit IRC | 17:31 | |
| *** kiall_ has joined #openstack-dev | 17:31 | |
| *** bdpayne has joined #openstack-dev | 17:32 | |
| *** jcoufal has quit IRC | 17:33 | |
| dolphm_ | dtroyer: stevemar: this merged scary fast the other day https://github.com/openstack/python-keystoneclient/commit/661d6cf783d1586e435196f38ffb1b2361b8fe25 and i was wondering if i should revise to be 'pending deprecation' rather than actually a hard dep (in other words, *should* people actually be directed to openstackclient yet?) | 17:33 |
| dolphm_ | s/hard dep/full deprecation/ | 17:33 |
| *** jruzicka has quit IRC | 17:34 | |
| stevemar | dolphm: pong | 17:34 |
| *** epim has joined #openstack-dev | 17:35 | |
| *** dolphm_ has quit IRC | 17:35 | |
| stevemar | dolphm: that did get merged scary fast | 17:35 |
| *** Guest61470 is now known as zz_Guest61470 | 17:35 | |
| *** dolphm has joined #openstack-dev | 17:36 | |
| *** bdpayne_ has joined #openstack-dev | 17:36 | |
| dolphm | stevemar: (sorry, i d/c for a minute) | 17:36 |
| stevemar | dolphm: np, was saying that it did get merged scary fast | 17:36 |
| *** bdpayne has quit IRC | 17:36 | |
| dolphm | stevemar: a 'pending deprecation' warning is long overdue :( | 17:37 |
| stevemar | dolphm, true, want to change it to 'pending' ? | 17:37 |
| dolphm | stevemar: you tell me if it should, or leaving it as-is just means python-keystoneclient shouldn't see a release until <target date> | 17:38 |
| *** radez_g0n3 is now known as radez | 17:39 | |
| dolphm | stevemar: at the moment, this is a bit of a blocker http://pasteraw.com/9tn8my5hhe3e0bumbvgd1at69xli9r | 17:39 |
| *** dkehn has quit IRC | 17:40 | |
| *** dkehn has joined #openstack-dev | 17:40 | |
| stevemar | dolphm: nice | 17:40 |
| dolphm | stevemar: https://bugs.launchpad.net/python-openstackclient/+bug/1220831 | 17:41 |
| uvirtbot | Launchpad bug 1220831 in python-openstackclient "AttributeError: 'module' object has no attribute 'BasicFileKeyring'" [Undecided,New] | 17:42 |
| *** rcrit has quit IRC | 17:42 | |
| stevemar | dolphm: well, using openstack client for v3 stuff is still 'blocked' by keystoneclient v3 auth work | 17:43 |
| dstanek | dolphm: morganfainberg: here is the general idea: https://review.openstack.org/#/c/45086/ | 17:43 |
| stevemar | dolphm: but as fas as v2 goes, it's pretty solid | 17:43 |
| stevemar | dolphm: so the message is fine in that regard | 17:43 |
| dstanek | will need a bit more more to make sure everything still works and the tests pass | 17:43 |
| *** alop has joined #openstack-dev | 17:44 | |
| dolphm | stevemar: all i'm looking for is parity with keystoneclient's v2 shell | 17:44 |
| dolphm | stevemar: everything else is a bonus | 17:45 |
| stevemar | dolphm: then i think the message is safe | 17:45 |
| dolphm | dstanek: is that a draft review? i get a 404 | 17:45 |
| *** kaushikc has quit IRC | 17:45 | |
| dstanek | dolphm: yes, can only i see those? | 17:46 |
| dolphm | dstanek: you have to invite people to them | 17:46 |
| dolphm | dstanek: alternatively Work In Progress means it won't merge as-is, but it's public | 17:46 |
| *** eglynn is now known as eglynn-is-hungry | 17:46 | |
| *** jaypipes has quit IRC | 17:48 | |
| stevemar | dolphm: can you edit design sessions after you create them? | 17:48 |
| dolphm | stevemar: yes | 17:48 |
| stevemar | dolphm: excellent, i will create one for .... talking about federated identity .... dun dun dun | 17:49 |
| dolphm | stevemar: dun dun dun dun dun ... | 17:49 |
| *** jaypipes has joined #openstack-dev | 17:49 | |
| dolphm | stevemar: should probably make that an end-of-day thing so we can continue in "alternative venue" | 17:50 |
| *** krtaylor has quit IRC | 17:50 | |
| dolphm | stevemar: FYI, keystone session should be broken up over several days :D | 17:50 |
| dstanek | dolphm: morganfainberg: http://bit.ly/17zNECN | 17:51 |
| stevemar | dolphm: not just 1! yay! | 17:51 |
| dstanek | is there an OpenStack shortener? | 17:51 |
| dolphm | dstanek: not that i'm aware of, but that's not a bad idea | 17:51 |
| *** jruzicka has joined #openstack-dev | 17:51 | |
| dolphm | dstanek: would be interesting to make link stats public | 17:51 |
| *** sumanthns has quit IRC | 17:51 | |
| dolphm | dstanek: maybe required ATC to create links... | 17:52 |
| *** jruzicka has quit IRC | 17:52 | |
| *** jruzicka has joined #openstack-dev | 17:52 | |
| *** lucasagomes has quit IRC | 17:53 | |
| *** jaypipes has quit IRC | 17:53 | |
| *** yjiang5_away is now known as yjiang5 | 17:53 | |
| *** ruhe has quit IRC | 17:53 | |
| stevemar | dolphm: wondering if I should create a session for web ui for oauth... | 17:54 |
| *** gargya has quit IRC | 17:54 | |
| dolphm | stevemar: for horizon? | 17:54 |
| stevemar | dolphm: it would have to live there right? | 17:54 |
| dolphm | stevemar: if you're talking about a horizon UI | 17:54 |
| *** eharney has quit IRC | 17:55 | |
| *** rcrit has joined #openstack-dev | 17:55 | |
| stevemar | dolphm: m'yeah | 17:55 |
| *** jtomasek has quit IRC | 17:55 | |
| *** jaypipes has joined #openstack-dev | 17:56 | |
| *** tstevenson has joined #openstack-dev | 17:57 | |
| *** openmike has joined #openstack-dev | 17:57 | |
| dtroyer | dolphm: what stevemar said. I'm not ready to tell the general user base to use OSC yet as I think it needs more shakedown, but I do want it to get some use where we can get constructive feedback and bug reports. | 17:57 |
| stevemar | ++ | 17:58 |
| dtroyer | dolphm: That error is fairly new (to me anyway), I'll tackle it next, once I get the requirements properly updated. | 17:58 |
| dolphm | dtroyer: i'll suggest revising to 'pending deprecation' then, which should direct adventurous users in your direction | 17:59 |
| *** xqueralt is now known as xqueralt-afk | 17:59 | |
| dstanek | in Keystone there is a pattern of having logic in a core module and importing core.* into the package init | 18:00 |
| dstanek | is there anything wrong with having some logic in the package init? | 18:01 |
| *** jprovazn has joined #openstack-dev | 18:01 | |
| *** Ryan_Lane has joined #openstack-dev | 18:02 | |
| *** Ryan_Lane has quit IRC | 18:03 | |
| TheSheep | dstanek: hard to import it from other submodules | 18:03 |
| *** ndipanov is now known as ndipanov_afk | 18:03 | |
| dolphm | dstanek: some projects do it, but a lot of developers find it confusing/surprising to see anything actually implemented there | 18:03 |
| *** Ryan_Lane has joined #openstack-dev | 18:03 | |
| TheSheep | dstanek: you can get a circular dependence | 18:03 |
| henrynash | dolphm: new patch posted | 18:07 |
| henrynash | dolphm: still a bit of cleanup to do…but essentially all there | 18:08 |
| dstanek | morganfainberg: it looks like some of the functions in keystone.common.cache.core exist just for tests - is that a correct observation? | 18:09 |
| *** mlavalle has joined #openstack-dev | 18:10 | |
| *** Mandell has quit IRC | 18:10 | |
| *** dina_belova has joined #openstack-dev | 18:11 | |
| *** SergeyLukjanov has joined #openstack-dev | 18:11 | |
| notmyname | russellb: the core swift devs on your review stats page is not up-to-date | 18:11 |
| russellb | https://github.com/openstack-infra/reviewstats/blob/master/projects/swift.json | 18:12 |
| russellb | notmyname: just update that file and 'git review' it | 18:12 |
| russellb | maybe someday i will make it get team info from gerrit, heh | 18:13 |
| dolphm | henrynash: looking | 18:13 |
| *** romcheg has quit IRC | 18:13 | |
| *** romcheg has joined #openstack-dev | 18:14 | |
| notmyname | russellb: done, thanks https://review.openstack.org/#/c/45088/ | 18:14 |
| stevemar | dolphm, lbragstad: https://review.openstack.org/#/c/43610/ | 18:14 |
| *** vuil has quit IRC | 18:14 | |
| *** xqueralt-afk is now known as xqueralt | 18:15 | |
| russellb | notmyname: approved thanks | 18:15 |
| lbragstad | stevemar: cool, checking | 18:15 |
| *** jruzicka has quit IRC | 18:15 | |
| *** novas0x2a|laptop has quit IRC | 18:15 | |
| *** dina_belova has quit IRC | 18:16 | |
| *** melwitt has joined #openstack-dev | 18:16 | |
| dolphm | henrynash: better! | 18:17 |
| dolphm | henrynash: why does case_sensitive default to False? that conflicts with comparator='equals' a bit | 18:17 |
| *** Mandell has joined #openstack-dev | 18:17 | |
| *** basha has joined #openstack-dev | 18:20 | |
| *** dubsquared has joined #openstack-dev | 18:20 | |
| ayoung | stevemar, personally, I want to run Horizon with mod_auth_kerb and S4U2Proxy for it to fetch its tokens. | 18:21 |
| morganfainberg | dstanek, have you tried running with cache on with your diff? | 18:23 |
| ayoung | stevemar, BTW: https://review.openstack.org/#/c/29105/ I need to rewrite that as an extension | 18:24 |
| morganfainberg | dstanek, because the on_arguments decorator has some side effects such as setting .invalidate, .set, and .refresh on the decorated methods. | 18:24 |
| stevemar | ayoung: that sounds insanely cool, what happened to it? | 18:24 |
| *** zaitcev has joined #openstack-dev | 18:26 | |
| *** comay has joined #openstack-dev | 18:26 | |
| *** asavu has quit IRC | 18:26 | |
| *** nati_ueno has quit IRC | 18:27 | |
| *** nati_ueno has joined #openstack-dev | 18:27 | |
| *** jaypipes has quit IRC | 18:28 | |
| dstanek | morganfainberg: no not yet, but i did realize that this reaches deeper than i thought | 18:29 |
| *** anniec has quit IRC | 18:29 | |
| *** hartsocks has joined #openstack-dev | 18:29 | |
| dstanek | for example keystone.service expects a cache region to have already been created | 18:29 |
| *** hartsocks has left #openstack-dev | 18:29 | |
| morganfainberg | dstanek, yep. because the decorator does other magic. I needed to ensure that the decorator could do the work (wrap the methods) at import time. | 18:30 |
| dstanek | i think that could be removed with the lazy impl, but there are probably others | 18:30 |
| dstanek | morganfainberg: does the cache region really need to exist before the managers are created? | 18:31 |
| morganfainberg | dstanek, yes, since the decorator has side effects. | 18:31 |
| *** eglynn-is-hungry has quit IRC | 18:31 | |
| morganfainberg | dstanek, and the decorator is tied to the cache region. | 18:31 |
| *** anniec has joined #openstack-dev | 18:32 | |
| morganfainberg | dstanek: https://bitbucket.org/zzzeek/dogpile.cache/src/67665a304cf284d127f323eab4d6c827032a08f7/dogpile/cache/region.py?at=master#cl-759 | 18:32 |
| *** xqueralt is now known as xqueralt-afk | 18:32 | |
| morganfainberg | dstanek, specifically the locally defined methods invalidate, set, refresh. we would need to re-implement a lot of that logic | 18:32 |
| *** bdpayne_ has quit IRC | 18:33 | |
| morganfainberg | dstanek, not opposed, but i'm not sure how beneficial it would be to re-impl that vs. just writing our own caching system/layer | 18:33 |
| *** cdub has quit IRC | 18:33 | |
| *** cdub_ has quit IRC | 18:33 | |
| *** cdub has joined #openstack-dev | 18:34 | |
| *** cdub_ has joined #openstack-dev | 18:34 | |
| *** utlemming has quit IRC | 18:34 | |
| dstanek | morganfainberg: i was hoping to create the region in a lazy way - really i would create a fake version of the decorator the implements invalidate, set and refresh | 18:34 |
| *** nati_ueno has quit IRC | 18:35 | |
| *** kiall_ has quit IRC | 18:35 | |
| dstanek | morganfainberg: is having our own caching layer the direction we are going? | 18:35 |
| *** nati_ueno has joined #openstack-dev | 18:36 | |
| morganfainberg | dstanek, i am hoping to move the keystone.common.cache stuff into oslo eventually, i think it's important to have caching that we control | 18:36 |
| morganfainberg | we as in, implementation within openstack | 18:36 |
| ayoung | stevemar, -2 by dolph, I updated it, but never pinged him to remove the -2 and was abandonded. I'll bring back for Icehouse | 18:36 |
| *** jculp has joined #openstack-dev | 18:36 | |
| *** novas0x2a|laptop has joined #openstack-dev | 18:36 | |
| dolphm | morganfainberg: dstanek: catching up... this sounds like waaay more trouble than explaining to people that "yes, you need to install dogpile even if you don't want to cache anything" | 18:36 |
| *** utlemming has joined #openstack-dev | 18:37 | |
| morganfainberg | dolphm, that was my thought when I based this on the original example implementation we were talking about. | 18:37 |
| ayoung | stevemar, what will make it ultra cool, though, it HTML forms that explain what values are required and why. We bascially need a GET for each post. | 18:37 |
| *** Shaan7 has quit IRC | 18:38 | |
| dolphm | ayoung: you could build an entire standalone project (a la python-openstackclient) to build HTML forms for all of openstack's web services | 18:38 |
| dolphm | ayoung: you should name it horizon | 18:38 |
| ayoung | dolphm, nope | 18:38 |
| *** kiall has joined #openstack-dev | 18:38 | |
| dolphm | ;) | 18:38 |
| ayoung | dolphm, I do not believe in server side scripting | 18:38 |
| morganfainberg | dstanek, dolphm, i think it is easier to just leave it as a hard dep. the overhead is fairly minimal for using it, it does need ot be installed though. | 18:39 |
| *** aditirav has quit IRC | 18:39 | |
| morganfainberg | ayoung, so… not a fan of nodejs? | 18:39 |
| ayoung | dolphm, Horizon should be a web service...a facade for the rest of openstack...with an optional Javascript file that provides the web functionality | 18:39 |
| dstanek | morganfainberg: i would agree | 18:39 |
| *** aditirav has joined #openstack-dev | 18:39 | |
| dolphm | ayoung: that's ... kind of what it is | 18:39 |
| ayoung | morganfainberg, nodejs is exactly the opposite of what I like | 18:39 |
| *** nati_ueno has quit IRC | 18:40 | |
| morganfainberg | dstanek, i also checked, the overhead of using it is fairly minimal with the default configuration. the no-op backend does a lot to help with that. there is some serialization, but keystone is largely single-threaded so that doesn't matter. | 18:40 |
| bknudson | I thought keystone was a big wsgi script? | 18:40 |
| ayoung | dolphm, but Horizon should get its forms from the other projects...which we will start doing in Icehouse I think | 18:40 |
| *** sarob has joined #openstack-dev | 18:40 | |
| morganfainberg | bknudson, lol | 18:40 |
| ayoung | bknudson, it is | 18:41 |
| *** jcoufal has joined #openstack-dev | 18:41 | |
| *** jaypipes has joined #openstack-dev | 18:43 | |
| ayoung | stevemar, I really don't like the way I was generating the HTML in that patch. Strings concatination is not the right approach. I started using lxml, but never finished it | 18:44 |
| *** welldannit has joined #openstack-dev | 18:44 | |
| morganfainberg | dstanek / dolphm, so we're good with just telling people "you should install dogpile.cache | 18:46 |
| morganfainberg | " | 18:46 |
| morganfainberg | even if not caching? | 18:46 |
| *** dhellmann_ is now known as dhellmann | 18:46 | |
| *** mlavalle has quit IRC | 18:48 | |
| *** yolanda has quit IRC | 18:49 | |
| ttx | markwash: pong | 18:49 |
| markwash | ttx: I think dolphm might have answered my questions | 18:50 |
| ttx | stevemar: yes you can (edit sessions once they are submitted) | 18:50 |
| ttx | markwash: ok then | 18:50 |
| stevemar | ttx: thx, dolphm helped me out! | 18:50 |
| ttx | dolphm: I see I can take a long vacation and you'll cover for me :) | 18:51 |
| dolphm | ttx: o/ | 18:51 |
| * dolphm runs away now that ttx is back | 18:51 | |
| dstanek | morganfainberg: dolphm: i think we have to be | 18:51 |
| dolphm | morganfainberg: sounds like that ^ | 18:52 |
| dolphm | morganfainberg: i don't think it's a big deal, but obviously some subset of users will be annoyed | 18:52 |
| *** colinmcnamara has quit IRC | 18:52 | |
| dolphm | there will always be one | 18:53 |
| morganfainberg | dolphm, ok. that is enough for me. | 18:53 |
| dolphm | dstanek: so, once you get your tests going, you're going to move keystone to testr, right :D | 18:53 |
| *** jpr has joined #openstack-dev | 18:54 | |
| dolphm | dstanek: https://bugs.launchpad.net/keystone/+bug/1177924 | 18:54 |
| uvirtbot | Launchpad bug 1177924 in keystone "Use testr instead of nose as the unittest runner." [Wishlist,Confirmed] | 18:54 |
| jpr | hi. I have a basic question | 18:54 |
| jpr | i'm trying to compare the block storage API features in Essex and Grizzly. | 18:55 |
| *** fbo_away is now known as fbo | 18:55 | |
| dolphm | dstanek: AFAIK, the pain there is how keystone reads in conf files from keystone/tests/ and expects to be able to access etc/ in a similar fashion | 18:55 |
| *** reed has quit IRC | 18:55 | |
| jpr | I can find all the block storage v2.0 info online at api.openstack.org for Grizzly info but I can't seem to find any legacy documentation for Essex. | 18:56 |
| jpr | Is there a way to get at the old docs. | 18:57 |
| *** zz_Guest61470 is now known as leif | 18:58 | |
| *** leif is now known as Guest42194 | 18:59 | |
| *** gyee has quit IRC | 19:00 | |
| *** eglynn-is-hungry has joined #openstack-dev | 19:01 | |
| jpr | sorry going to #openstack | 19:01 |
| *** SumitNaiksatam has joined #openstack-dev | 19:04 | |
| *** epim has quit IRC | 19:05 | |
| *** epim has joined #openstack-dev | 19:06 | |
| *** dstanek has quit IRC | 19:07 | |
| *** SergeyLukjanov has quit IRC | 19:07 | |
| *** xga has joined #openstack-dev | 19:07 | |
| *** otherwiseguy has joined #openstack-dev | 19:08 | |
| *** dina_belova has joined #openstack-dev | 19:11 | |
| *** xga has quit IRC | 19:11 | |
| *** basha has quit IRC | 19:12 | |
| *** kiall has quit IRC | 19:12 | |
| *** sandeepr_ltp has joined #openstack-dev | 19:13 | |
| *** openmike has quit IRC | 19:14 | |
| *** vipul is now known as vipul-away | 19:14 | |
| *** vipul-away is now known as vipul | 19:14 | |
| *** corXi has joined #openstack-dev | 19:15 | |
| *** dina_belova has quit IRC | 19:16 | |
| *** shinylasers has quit IRC | 19:17 | |
| *** jpr1 has joined #openstack-dev | 19:17 | |
| *** shinylasers has joined #openstack-dev | 19:17 | |
| *** jpr has quit IRC | 19:20 | |
| *** jpr1 has left #openstack-dev | 19:21 | |
| *** garyk has quit IRC | 19:24 | |
| *** amscanne_ has joined #openstack-dev | 19:25 | |
| *** dstanek has joined #openstack-dev | 19:27 | |
| *** amscanne has quit IRC | 19:28 | |
| *** networkstatic has quit IRC | 19:29 | |
| *** vipul is now known as vipul-away | 19:29 | |
| *** NobodyCam has joined #openstack-dev | 19:29 | |
| *** anniec has quit IRC | 19:31 | |
| *** sdague has left #openstack-dev | 19:31 | |
| *** eharney has joined #openstack-dev | 19:32 | |
| *** bauzas has quit IRC | 19:32 | |
| *** jculp has quit IRC | 19:32 | |
| *** pasquier-s has quit IRC | 19:33 | |
| *** yjiang5 is now known as yjiang5_away | 19:33 | |
| *** vipul-away is now known as vipul | 19:34 | |
| *** sdague has joined #openstack-dev | 19:36 | |
| *** anniec has joined #openstack-dev | 19:37 | |
| *** nrs_ has quit IRC | 19:37 | |
| *** redbo has quit IRC | 19:38 | |
| *** dosaboy_ has joined #openstack-dev | 19:40 | |
| *** redbo has joined #openstack-dev | 19:40 | |
| *** nrs_ has joined #openstack-dev | 19:40 | |
| *** asavu has joined #openstack-dev | 19:41 | |
| *** mmagr has joined #openstack-dev | 19:41 | |
| dolphm | stevemar: looking | 19:41 |
| *** dosaboy has quit IRC | 19:42 | |
| *** kiall has joined #openstack-dev | 19:43 | |
| dolphm | stevemar: dammit, in the sql driver they're just role_id's right? what's kind of data are you persisting there? | 19:43 |
| stevemar | dolphm... making me remember now... 1 sec | 19:44 |
| *** ociuhandu has quit IRC | 19:44 | |
| *** jcoufal has quit IRC | 19:46 | |
| *** pasquier-s has joined #openstack-dev | 19:47 | |
| stevemar | dolphm: it's a jsondump of a subset of self.identity_api.list_roles() | 19:47 |
| dolphm | stevemar: entire role objects?! | 19:48 |
| ayoung | mhu, the problem is not in your code for https://review.openstack.org/#/c/40444/ | 19:48 |
| *** jasdeepH has quit IRC | 19:48 | |
| *** bauzas has joined #openstack-dev | 19:49 | |
| stevemar | dolphm: yes, that way when the keystone token is issued, it can just use that part | 19:50 |
| stevemar | dolphm: but now i'm guessing that is bad, since you included an exclamation mark | 19:50 |
| dolphm | stevemar: !! | 19:50 |
| *** jvrbanac has joined #openstack-dev | 19:51 | |
| dolphm | stevemar: it's a completely utter nightmare how much garbage is tossed into the token table. this will have the exact same fate | 19:51 |
| dolphm | stevemar: and you're already having to second guess what you put in there, so it's already begun | 19:51 |
| stevemar | dolphm: change it to just ids? | 19:52 |
| dolphm | stevemar: i was expecting a json-encoded list of strings, worst case... role_ids = ["123", "234", "abc"] | 19:52 |
| dolphm | stevemar: at least then, there's not too much room for mis-interpretation or bloat | 19:53 |
| stevemar | dolphm: easy enough change, i'll change the column name to role_ids too | 19:53 |
| ayoung | stevemar, dolphm, check me on this...I might just be paranoid...if this patch goes through, will the EC2 API allow an end run around security : token for ec2 creds not indicating delegation? https://review.openstack.org/#/c/40444/ | 19:54 |
| *** colinmcnamara has joined #openstack-dev | 19:55 | |
| *** anniec has quit IRC | 19:55 | |
| dolphm | ayoung: looking... | 19:55 |
| ayoung | dolphm, I think I was being paranoid...I'd like to let that one get in, as the HEAT team needs it | 19:56 |
| *** jprovazn has quit IRC | 19:56 | |
| dolphm | ayoung: they need a new api today? | 19:56 |
| ayoung | dolphm, this was languishing for a while | 19:56 |
| ayoung | I had -2ed it, and never saw that they updateed it | 19:57 |
| *** shinylasers has quit IRC | 19:57 | |
| dolphm | ayoung: OS-EC2-FROM-TOKEN? | 19:57 |
| ayoung | last night, I gave it a once over, but then got paranoid | 19:57 |
| ayoung | yeah, that is wrong | 19:57 |
| *** jruzicka has joined #openstack-dev | 19:58 | |
| ayoung | dolphm, I think I am going to remove the -2 | 19:58 |
| dolphm | ayoung: i think someone asked about this on list... i didn't understand why the use case required a new api | 19:58 |
| ayoung | dolphm, they need ec2 Credentials for access to vms, and the vms use the credentials to get tokens to perform actions on behalf of the users... | 19:59 |
| ayoung | its all "set up this rule and let it execute at 2am" type stuff | 19:59 |
| *** portante is now known as portante|afk | 20:00 | |
| *** portante|afk is now known as portante | 20:00 | |
| *** rcleere has joined #openstack-dev | 20:02 | |
| *** romcheg has quit IRC | 20:02 | |
| *** gyee has joined #openstack-dev | 20:02 | |
| dolphm | ayoung: that doesn't explain the new api | 20:02 |
| ayoung | dolphm, they need to create an ec2 credential from a trust. | 20:03 |
| *** mlavalle has joined #openstack-dev | 20:03 | |
| *** yjiang5 has joined #openstack-dev | 20:04 | |
| ayoung | they do that by way of a token | 20:04 |
| ayoung | dolphm, yeah...looks like they could have just extended the create | 20:05 |
| *** shinylasers has joined #openstack-dev | 20:05 | |
| *** basha has joined #openstack-dev | 20:05 | |
| *** basha has quit IRC | 20:05 | |
| insanidade | would anyone please take a look at this problem? http://paste.openstack.org/show/45770/ | 20:06 |
| *** nati_ueno has joined #openstack-dev | 20:06 | |
| insanidade | that was copied from teh l3 agent screen during a stack.sh in devstack. | 20:06 |
| insanidade | is that a serious problem ? | 20:07 |
| sdague | jgriffith: you have to stop +Aing this patch before it goes through jenkins..... :P | 20:07 |
| sdague | https://review.openstack.org/#/c/45093/ | 20:07 |
| ayoung | shardy, is there any reason that should have a separate API? Seems gratuitous to me? | 20:07 |
| *** maheshp has quit IRC | 20:09 | |
| dolphm | ayoung: if the token has impersonation, then the existing api looks like it'll do the same thing | 20:12 |
| *** henrynash has quit IRC | 20:12 | |
| *** dina_belova has joined #openstack-dev | 20:12 | |
| *** henrynash has joined #openstack-dev | 20:13 | |
| morganfainberg | dolphm, also, it seems odd that they would use the same trust (explicit) token, seems to me that they would want the keypair to persist longer than the time on the token (something like the length of the trust?) | 20:13 |
| morganfainberg | dolphm, or am i mis-understanding the usecase | 20:13 |
| jgriffith | sdague: so we gate on README files now? | 20:14 |
| jgriffith | sdague: makes a lot of sense | 20:14 |
| sdague | jgriffith: commit message | 20:14 |
| *** radez is now known as radez_g0n3 | 20:14 | |
| dolphm | morganfainberg: where are you seeing that issue? | 20:14 |
| jgriffith | sdague: fair enough | 20:14 |
| jgriffith | sorry bout that | 20:14 |
| sdague | it failed hacking in the commit message, so it was blocking everything behind it | 20:14 |
| morganfainberg | dolphm, line 104 https://review.openstack.org/#/c/40444/10/keystone/contrib/ec2/controllers.py | 20:15 |
| morganfainberg | dolphm, if creds have the token_id, it returns that speciifc token? | 20:15 |
| morganfainberg | dolphm, that would raise TokenNotFound if the token is expired. | 20:15 |
| dolphm | morganfainberg: ah, good catch | 20:15 |
| sdague | no worries, just trying to catch you before it happens again, because I actually fixed it once, then there was a patch 3 which broke it again | 20:16 |
| dolphm | ... every line in this patch seems unnecessary if impersonation=True | 20:16 |
| *** dvarga has quit IRC | 20:16 | |
| stevemar | dolphm: would [{'id':"123"}, {'id': "234"}] be okay to save ? instead of the one you mentioned? | 20:16 |
| jgriffith | sdague: yeah, it's a good thing we block on things like a period at the end of the first line in a commit mesg | 20:17 |
| *** dina_belova has quit IRC | 20:17 | |
| jgriffith | sdague: sorry... | 20:17 |
| jgriffith | sdague: I'm in a bit of a funk today | 20:17 |
| sdague | jgriffith: no worries :) | 20:17 |
| sdague | you can turn that off in cinder if you like | 20:17 |
| jgriffith | sdague: nah, I like the consistency | 20:17 |
| jgriffith | sdague: just in fowl mood. Think I'll get lunch and all will be better | 20:18 |
| sdague | ok, cool beans | 20:18 |
| jgriffith | sdague: in all seriousness, thanks for catching it and fixing it up | 20:18 |
| *** garyk has joined #openstack-dev | 20:18 | |
| sdague | no prob | 20:18 |
| *** jaypipes has quit IRC | 20:19 | |
| sdague | just trying to maximize everyone's throughput on the gate queue today, as it's been a busy one | 20:19 |
| ayoung | shardy, can you address morganfainberg 's questions? | 20:19 |
| ayoung | sdake, or any any other HEAT devs ? | 20:20 |
| *** nati_uen_ has joined #openstack-dev | 20:20 | |
| dolphm | stevemar: that still has the same problem as tokens | 20:20 |
| dolphm | stevemar: ... room for bloat to slip in | 20:21 |
| stevemar | m'alright | 20:21 |
| stevemar | alright alright | 20:21 |
| *** amscanne_ has quit IRC | 20:22 | |
| morganfainberg | stevemar, why not just store the "ID" in a list? seems like un-needed overhead to do '{'id:' <blah}' if you're only storing id | 20:24 |
| stevemar | morganfainberg, cause it came in that way via the request, didn't want to loop through it again | 20:24 |
| stevemar | morganfainberg, no biggie though | 20:24 |
| morganfainberg | stevemar: as in didn't want to loop through in a for x in blah? | 20:25 |
| stevemar | yeah, | 20:25 |
| stevemar | but i think i'm doing it anyway, when checking that the user has the role | 20:25 |
| stevemar | so it's actually a non-issue | 20:25 |
| morganfainberg | stevemar, perhaps. maybe use a set if you want a better lookup on 'if X in blah' | 20:25 |
| morganfainberg | but it's not a huge win unless you're iterating a lot | 20:25 |
| * stevemar nods | 20:26 | |
| dolphm | morganfainberg: ++ | 20:26 |
| stevemar | i'm already using a set, to compare with user's roles | 20:26 |
| dolphm | stevemar: ++ for that | 20:26 |
| stevemar | so i'm just calling list() | 20:26 |
| morganfainberg | stevemar, sets are awesome :) | 20:26 |
| stevemar | indeed | 20:26 |
| dolphm | stevemar: i liked your issubset() | 20:26 |
| dolphm | (or whatever that method is called) | 20:26 |
| stevemar | is was pretty neat | 20:26 |
| morganfainberg | i like using sets for difference sets. | 20:27 |
| morganfainberg | get me everything in x that isn't in y | 20:27 |
| morganfainberg | but haven't had a usecase for that in keystone yet | 20:27 |
| morganfainberg | (personal usecase that is) | 20:27 |
| *** nati_ue__ has joined #openstack-dev | 20:28 | |
| dolphm | stevemar: one of the problems with persisting role objects and then just regurgitating them is that everything but the ID is mutable, so in effect you're just storing stale data in a table right next to the latest data | 20:28 |
| stevemar | dolphm: alright, i think i've got the impl down | 20:29 |
| stevemar | dolphm: leaving soon to head out for the evening, and i want to add some tests for this, so can I have it ready for you tmrw morning? | 20:30 |
| *** nati_uen_ has quit IRC | 20:30 | |
| dolphm | stevemar: not acceptable | 20:30 |
| stevemar | dolphm: i've been under the assumption there is a deadline of friday for this? | 20:30 |
| dolphm | stevemar: put what you have up for review, deadline is today | 20:30 |
| stevemar | d'oh | 20:30 |
| *** Guest85374 has quit IRC | 20:31 | |
| dolphm | stevemar: more accurately, i'm done with havana in an hour :) | 20:31 |
| morganfainberg | dolphm, hehe. | 20:31 |
| dolphm | stevemar: https://wiki.openstack.org/wiki/Havana_Release_Schedule FeatureFreeze (Sep 4) | 20:31 |
| *** fbo is now known as fbo_away | 20:31 | |
| stevemar | oh jeez | 20:32 |
| dolphm | stevemar: you can probably fix it faster than me, but i'm motivated to get this in if you need to go | 20:32 |
| stevemar | dolphm: no, it's fine, just double checking things before i submit | 20:33 |
| dolphm | stevemar: sweeet | 20:33 |
| dolphm | henrynash: how far out is a revision? 39 things in gate right now means we need to join that queue ASAP to make havana :) | 20:35 |
| *** jasdeepH has joined #openstack-dev | 20:35 | |
| henrynash | working on it right now…30 mins | 20:35 |
| *** sushils has quit IRC | 20:35 | |
| *** ayoung is now known as ayoung-afk | 20:36 | |
| *** waa has quit IRC | 20:36 | |
| henrynash | dolphm: how important is it to move the hints build out of the __init__ and into a function in the V3 controller class? | 20:37 |
| dolphm | bknudson: i saw you commented on henry's review... i should go read those? | 20:37 |
| dolphm | s/i should/should i/ | 20:37 |
| dolphm | henrynash: is that not an easy change? | 20:38 |
| henrynash | dolphm: mechanical…just got to change all the places it is called…I can do it... | 20:38 |
| bknudson | dolphm: I wouldn't bother reading my comments if I was you | 20:39 |
| dolphm | bknudson: lol ok | 20:39 |
| *** portante is now known as portante|afk | 20:39 | |
| *** mmagr has quit IRC | 20:40 | |
| *** sgordon has quit IRC | 20:42 | |
| stevemar | dolphm: i sincerely hope this works, running tests now, had to make a last minute change to /access_tokens/{id}/roles and /roles/{role_id} | 20:43 |
| dolphm | henrynash: automechanical! find keystone -type f -exec sed -i 's/driver_hints\.Hints/self.build_driver_hints/g' {} \; | 20:44 |
| dolphm | henrynash: disclaimer- probably made a typo | 20:45 |
| henrynash | dolphm: processing stuff now.. | 20:45 |
| *** Ryan_Lane has quit IRC | 20:49 | |
| *** waa has joined #openstack-dev | 20:50 | |
| *** Ryan_Lane has joined #openstack-dev | 20:50 | |
| *** Ryan_Lane has quit IRC | 20:51 | |
| *** Ryan_Lane has joined #openstack-dev | 20:51 | |
| *** dstanek has quit IRC | 20:52 | |
| *** tmclaugh[work] has quit IRC | 20:52 | |
| *** dims has quit IRC | 20:53 | |
| *** romcheg has joined #openstack-dev | 20:53 | |
| *** jcoufal has joined #openstack-dev | 20:54 | |
| shardy | ayoung-afk, morganfainberg: sorry in meetings | 20:54 |
| morganfainberg | shardy, no worries. had a question about the EC2 keypair from tokens | 20:54 |
| stevemar | dolphm: new patch! | 20:54 |
| dolphm | morganfainberg: https://review.openstack.org/#/c/43610/ | 20:55 |
| shardy | morganfainberg: yes, Heat wants to be able to create an ec2 keypair from a trust, which doesn't expire, so we can deploy credentials in instances, for authentication with Heat APIs | 20:55 |
| *** gimps_ has quit IRC | 20:55 | |
| *** nati_ue__ has quit IRC | 20:55 | |
| *** dims has joined #openstack-dev | 20:55 | |
| morganfainberg | shardy, then the current patch needs work, max lifespan on the keypair is life of the token. so default of 86400 | 20:55 |
| *** nati_uen_ has joined #openstack-dev | 20:55 | |
| morganfainberg | dolphm, looking at stevemar's patchset now | 20:56 |
| *** rcleere has quit IRC | 20:56 | |
| shardy | morganfainberg: essentially we want something derived from the user owning the stack, but which has no real roles to do anything, ie implicitly untrusted in-instance ec2 keypair, which is used for signing requests to our CFN API | 20:56 |
| morganfainberg | shardy, makes perfect sense. i was pretty sure that was the intent | 20:56 |
| shardy | morganfainberg: ok, cool | 20:57 |
| *** maheshp has joined #openstack-dev | 20:57 | |
| *** danwent has quit IRC | 20:57 | |
| dolphm | stevemar: sort of out of scope now, but if you want to change it for icehouse, you can totally drop _format_role_entity() | 20:58 |
| dolphm | stevemar: dropping certain attributes made sense when they were provided as a nested object, but since they have their own first class api resource /roles ... it makes sense to include them | 20:58 |
| stevemar | dolphm: yeah, agreed, i wanted to drop it | 20:59 |
| dolphm | stevemar: i wouldn't argue if you did it now, but it'd also be api backwards compatible to do it later | 21:00 |
| *** noslzzp has quit IRC | 21:00 | |
| *** xqueralt-afk is now known as xqueralt | 21:00 | |
| dolphm | stevemar: whoa, you're getting the user's roles from the wrong place user_token['metadata'].copy() | 21:01 |
| *** asavu has quit IRC | 21:01 | |
| stevemar | dolphm: what? how is that wrong? that's where i've been getting it this whole time | 21:02 |
| dolphm | stevemar: that *might* work, but it'll certainly produce bugs as that's some of the untested garbage that gets inexplicably jammed into the token backend and then regurgitated back out the api | 21:02 |
| dolphm | stevemar: i haven't caught it until now :( | 21:03 |
| stevemar | where should I be getting it from? | 21:03 |
| dolphm | stevemar: assignment_api | 21:03 |
| *** dprince has quit IRC | 21:03 | |
| dolphm | stevemar: you're also requiring the authorizing user's token to be carrying the roles they're delegating... which is interesting, but undocumented | 21:04 |
| morganfainberg | stevemar, it's more sane to always ask for the authoritative data. | 21:04 |
| stevemar | so get user_id from there, and call get users roles? | 21:04 |
| morganfainberg | dolphm, i think there is a comment in token provider about not doing that (roles in metatdata) for at least one type of token. | 21:04 |
| morganfainberg | might be v2 | 21:04 |
| dolphm | stevemar: list_roles_something_something(user_id, project_id) ? | 21:04 |
| dolphm | stevemar: i guess it's still get_metadata() | 21:04 |
| dolphm | morganfainberg: the source of a lot of these problems is that a lot of our api tests check for the presence of data, but almost none check for the lack of unexpected data... so this kind of stuff constantly slips through | 21:06 |
| *** romcheg has left #openstack-dev | 21:06 | |
| *** pcm_ has quit IRC | 21:06 | |
| morganfainberg | dolphm, stevemar: assignment_api.get_roles_for_user_and_project(user_id, tenant_id) | 21:06 |
| dolphm | and then people like stevemar assumes it's part of the api. | 21:06 |
| stevemar | yep | 21:06 |
| dolphm | morganfainberg: oh sweet | 21:06 |
| stevemar | stevemar assumes nothing! | 21:06 |
| dolphm | stevemar: it's not your fault :( | 21:06 |
| morganfainberg | dolphm, i'll toss a BP up to clean that up and/or work on tests. | 21:06 |
| dolphm | morganfainberg: i don't even know how to go about tackling it, other than doing a major api revision at the same time :( | 21:07 |
| morganfainberg | dolphm, at least we could implement tests for it | 21:07 |
| morganfainberg | lack of data that is. | 21:07 |
| *** kiall has quit IRC | 21:07 | |
| dolphm | morganfainberg: asserting it exists? or asserting we don't make more mistakes | 21:07 |
| morganfainberg | maybe better commenting. | 21:08 |
| stevemar | dolphm: give me 1 sec | 21:08 |
| stevemar | dolphm: then i'll update with new patch | 21:08 |
| morganfainberg | dolphm, i'll think about that. if it is assumed to exist for _some_ things, likely it should always exist. | 21:08 |
| dolphm | https://wiki.openstack.org/wiki/Governance/Proposed/APIManagement-sampleGuidelines | 21:09 |
| *** bdpayne has joined #openstack-dev | 21:09 | |
| dolphm | Backwards-incompatible changes (e.g., no longer supporting a resource, changing representation formats) are introduced by incrementing the version identifier. Bumping the version should be done very rarely. When the version is bumped, the most recent version must be supported for at least [n] releases afterwards, preferably longer, and the sunset of the API version needs to be widely announced. | 21:09 |
| *** Ryan_Lane has quit IRC | 21:10 | |
| bknudson | [n] releases... that's a long time. | 21:10 |
| *** Ryan_Lane has joined #openstack-dev | 21:10 | |
| dolphm | henrynash: 30 minutes is up! ;) | 21:10 |
| bknudson | when can we drop v2? | 21:10 |
| henrynash | dolphm; Nearly there | 21:10 |
| *** martine_ has quit IRC | 21:10 | |
| *** anniec has joined #openstack-dev | 21:11 | |
| stevemar | dolphm: so trying to get the project id... from the token, but just a tenant object is showing, that normal? | 21:11 |
| dolphm | bknudson: i'd like to consider everything but v2 auth deprecated as of ... now? last release? | 21:11 |
| dolphm | bknudson: i imagine we can easily (and should) support v2 auth via middleware for a while longer | 21:11 |
| dolphm | bknudson: (just rewrite v2 requests into v3 auth requests, and vice versa) | 21:12 |
| bknudson | dolphm: I keep telling people to use the openstack client, but then it turns out that nobody ships it. | 21:12 |
| morganfainberg | dolphm ++, deprecate v2 would be nice. | 21:12 |
| *** dubsquared has quit IRC | 21:12 | |
| *** dina_belova has joined #openstack-dev | 21:12 | |
| bknudson | I can't convince our own team to ship it until other distros do. | 21:12 |
| dolphm | stevemar: uhh... well... a v2 token would have a tenant... | 21:13 |
| stevemar | ... but i'm running v3 tests, so why would it there be a tenant | 21:13 |
| dolphm | bknudson: soon https://github.com/openstack/python-keystoneclient/commit/661d6cf783d1586e435196f38ffb1b2361b8fe25 | 21:13 |
| *** NobodyCam_ has joined #openstack-dev | 21:14 | |
| bknudson | dolphm: I'm hoping that will convince somebody. | 21:14 |
| *** NobodyCam_ has quit IRC | 21:14 | |
| *** sushils has joined #openstack-dev | 21:14 | |
| *** alunduil has quit IRC | 21:14 | |
| *** ndipanov_afk is now known as ndipanov | 21:15 | |
| *** amscanne has joined #openstack-dev | 21:16 | |
| dolphm | bknudson: although it should be relaxed a tiny bit... https://review.openstack.org/#/c/45127/ | 21:16 |
| dolphm | bknudson: only because openstackclient isn't ready for primetime today | 21:17 |
| stevemar | dolphm: this is what i'm seeing http://paste.openstack.org/raw/45779/ | 21:17 |
| *** dina_belova has quit IRC | 21:17 | |
| *** Ryan_Lane has quit IRC | 21:17 | |
| *** Ryan_Lane1 has joined #openstack-dev | 21:17 | |
| *** marun has quit IRC | 21:17 | |
| dolphm | stevemar: all i see is redundant garbage | 21:18 |
| *** garyk1 has joined #openstack-dev | 21:19 | |
| dolphm | stevemar: indented that.. http://pasteraw.com/7gzms6aj90orpnlgxvn12xsozdzysxm | 21:19 |
| *** Ryan_Lane1 is now known as Ryan_Lane | 21:20 | |
| *** garyk has quit IRC | 21:20 | |
| *** Ryan_Lane has quit IRC | 21:20 | |
| *** Ryan_Lane has joined #openstack-dev | 21:20 | |
| *** maheshp has quit IRC | 21:20 | |
| morganfainberg | dolphm, i think the token spec might be a little loose, we should probably look at building a test that validates we don't add more misc data to the tokens as is | 21:20 |
| stevemar | dolphm: thx, i was trying to do that | 21:20 |
| morganfainberg | without intent | 21:20 |
| dolphm | stevemar: jsonutils | 21:20 |
| dolphm | morganfainberg: +++ | 21:20 |
| *** maheshp has joined #openstack-dev | 21:20 | |
| morganfainberg | erm, duplicated/misc/unneeded data | 21:20 |
| stevemar | dolphm: so, should I dig into token data? or just ... stick with tenant? | 21:20 |
| dolphm | stevemar: d['token_data']['token']['project']['id'] indicates the token's scope | 21:21 |
| stevemar | dolphm: or check project in token data first? then default to tenant if it doesn't exist | 21:21 |
| stevemar | gotcha, so no | 21:21 |
| morganfainberg | dolphm, reminds me. needed to ask you a quick question. | 21:21 |
| dolphm | stevemar: (where are you putting this project id?) | 21:21 |
| dolphm | morganfainberg: o/ | 21:22 |
| stevemar | dolphm: i'm not putting it anywhere, i'm trying to get the project id for the list_users_roles_.... | 21:22 |
| *** dstanek has joined #openstack-dev | 21:22 | |
| dolphm | stevemar: then you'd use the project ID from the request token, right? | 21:22 |
| dolphm | stevemar: you want to know what roles the authorizing user has on the requested project id | 21:22 |
| stevemar | ffs, this is what i get for rushing | 21:23 |
| *** marun has joined #openstack-dev | 21:23 | |
| dolphm | stevemar: lol i think it's *interesting* that you're using the authorizing user's supplied token, but then you'd need to check that the token is scoped to the requested_project_id | 21:24 |
| dolphm | stevemar: i think you can safely assume the token is valid | 21:24 |
| stevemar | ugh, embarassing, i'm rushing like crazy | 21:24 |
| *** jcoufal has quit IRC | 21:25 | |
| *** maheshp has quit IRC | 21:25 | |
| dolphm | stevemar: interesting does not mean wrong, it means my brain is unable to comprehend consequences at the moment :P | 21:25 |
| stevemar | dolphm: re-running the tests once more | 21:26 |
| *** SumitNaiksatam has quit IRC | 21:26 | |
| *** jbresnah_ is now known as jbresnah | 21:28 | |
| *** amscanne has quit IRC | 21:29 | |
| *** amscanne has joined #openstack-dev | 21:30 | |
| *** maheshp has joined #openstack-dev | 21:32 | |
| *** jamielennox|away is now known as jamielennox | 21:32 | |
| *** epim has quit IRC | 21:33 | |
| *** epim has joined #openstack-dev | 21:33 | |
| *** sarob has quit IRC | 21:34 | |
| *** sarob has joined #openstack-dev | 21:35 | |
| *** anniec has quit IRC | 21:36 | |
| dolphm | henrynash: *poke* | 21:36 |
| henrynash | dolphm: running unittests | 21:37 |
| *** anniec has joined #openstack-dev | 21:38 | |
| *** radsy has joined #openstack-dev | 21:39 | |
| *** sarob has quit IRC | 21:39 | |
| *** osphy has joined #openstack-dev | 21:40 | |
| *** portante|afk is now known as portante | 21:40 | |
| *** kiall has joined #openstack-dev | 21:40 | |
| iccha | clarkb: https://review.openstack.org/#/c/43368/ jenkins not rerunning tests, anything I can do? | 21:40 |
| iccha | anteaya: ^ | 21:41 |
| stevemar | dolphm: new patch | 21:41 |
| henrynash | dolphm: so have posted new patch | 21:42 |
| dolphm | yay | 21:42 |
| clarkb | iccha: http://status.openstack.org/zuul/ the jobs are running | 21:42 |
| henrynash | dolphm: there is still an issue that mark_filter_satisfied doesn't really do what we want | 21:42 |
| anteaya | iccha: looking | 21:42 |
| henrynash | dolphm: the consequence is that although the filter will get implemented in the SQL (for example), the controller will still reapply the filter (with no effect, of course) | 21:43 |
| iccha | thanks clarkb anteaya | 21:43 |
| henrynash | dolphm: I need to fix that, but mayeb that could be a fix applied after | 21:43 |
| *** topol has quit IRC | 21:43 | |
| *** jruzicka has quit IRC | 21:43 | |
| *** xqueralt is now known as xqueralt-afk | 21:44 | |
| dolphm | henrynash: ++ | 21:44 |
| dolphm | henrynash: kinda sorta late, but what if we didn't fall back on the controller to do the work? | 21:44 |
| clarkb | iccha: the gate has a higher priority than the check queue | 21:45 |
| anteaya | iccha: should finish in ~10 minutes | 21:45 |
| dolphm | henrynash: either the driver satisfies the hints or the client gets more data than expected | 21:45 |
| henrynash | dolphm: meaning, if the driver doesn't do it, no filtering? | 21:45 |
| clarkb | iccha: you may see that check tests happen more slowly as the gate queue is currently busy | 21:45 |
| dolphm | henrynash: right | 21:45 |
| henrynash | dolphm: kind of an odd thing from an api point of view…the filter *might* work | 21:46 |
| henrynash | dolphm: what's the advantage? | 21:46 |
| *** stevemar has quit IRC | 21:46 | |
| ayoung-afk | henrynash, I'm ok with that approach | 21:47 |
| henrynash | ayoung-afk: with the filter not working depending on the driver (which would be the effect)? | 21:48 |
| *** dstanek has quit IRC | 21:49 | |
| ayoung-afk | henrynash, yeah..it would actually hide less, and thus performance issues would be clearer...might even help prevent some | 21:49 |
| *** BStokes has joined #openstack-dev | 21:49 | |
| dolphm | morganfainberg: https://review.openstack.org/#/c/43610/ | 21:50 |
| *** neoXsys has quit IRC | 21:50 | |
| dolphm | henrynash: it's just a thought that i'm certainly not opposed to | 21:51 |
| *** SumitNaiksatam has joined #openstack-dev | 21:51 | |
| *** ewindisch has joined #openstack-dev | 21:51 | |
| *** ewindisch is now known as ericw | 21:51 | |
| morganfainberg | dolphm, +2'd | 21:51 |
| *** mrodden has quit IRC | 21:52 | |
| *** ericw is now known as ewindisch | 21:52 | |
| morganfainberg | dolphm, didn't approve because jenkins hadn't responded yet. | 21:52 |
| dolphm | morganfainberg: ah shit | 21:52 |
| *** portante is now known as portante|afk | 21:52 | |
| dolphm | morganfainberg: unset approved! | 21:52 |
| *** lbragstad has quit IRC | 21:52 | |
| morganfainberg | dolphm, i didn't approve ;) | 21:52 |
| morganfainberg | oh | 21:52 |
| morganfainberg | you haha | 21:52 |
| dolphm | morganfainberg: i did | 21:52 |
| morganfainberg | yeah =/ | 21:52 |
| dolphm | morganfainberg: i think that only kills everything after it in the gate, which was nothing | 21:53 |
| *** shardy is now known as shardy_afk | 21:53 | |
| *** ewindisch is now known as ewindisch- | 21:53 | |
| morganfainberg | it still runs a verify thought the gate iirc | 21:53 |
| morganfainberg | just wont merge. | 21:53 |
| *** datsun180b has quit IRC | 21:54 | |
| *** sdake_ has quit IRC | 21:54 | |
| henrynash | dolphm: well it's certainly a one line change to stop the controller filtering…although I still worry about the api | 21:54 |
| henrynash | (the impact that sometimes it filters) | 21:54 |
| dolphm | henrynash: i'm just thinking it's not unreasonable to put the burden on the driver to satisfy hints -- or not | 21:55 |
| *** guardian_ has joined #openstack-dev | 21:56 | |
| *** kbringard has quit IRC | 21:56 | |
| dolphm | henrynash: bunch of inconsistency over where hints are required or not | 21:57 |
| dolphm | henrynash: every other method signature in https://review.openstack.org/#/c/43257/36/keystone/assignment/core.py | 21:57 |
| henrynash | dolphm: well, all the list ones define it... | 21:58 |
| henrynash | dolphm: where is it inconstant? | 21:58 |
| dolphm | henrynash: list_projects(self, hints) vs list_user_projects(self, user_id, hints=None) | 21:58 |
| dolphm | henrynash: required in the first, optional kwarg in the second | 21:59 |
| henrynash | dolphm: hmm, damn, yes the call INTO the manager should have it optional, the call to the driver should require it | 22:00 |
| dolphm | henrynash: that's reasonable, as long as it's consistent! | 22:00 |
| dolphm | henrynash: that's the only file i've seen so far with inconsistency in that regard | 22:01 |
| dolphm | henrynash: typo in line 49 https://review.openstack.org/#/c/43257/36/keystone/tests/filtering.py | 22:04 |
| dolphm | s/Big/Bug/ | 22:04 |
| *** trey_h has quit IRC | 22:05 | |
| henrynash | dolphmL got it | 22:05 |
| *** trey_h has joined #openstack-dev | 22:06 | |
| dolphm | henrynash: why do you have to copy the query string here? https://review.openstack.org/#/c/43257/36/keystone/trust/controllers.py | 22:06 |
| henrynash | dolphm: I can undo those changes now, yes | 22:07 |
| dolphm | henrynash: your current change failed py27 tests | 22:08 |
| henrynash | dolphm: yep, got it | 22:08 |
| *** gmurphy_gone is now known as gmurphy | 22:08 | |
| *** spzala has quit IRC | 22:09 | |
| *** asavu has joined #openstack-dev | 22:10 | |
| *** trey_h has quit IRC | 22:11 | |
| *** flaper87 is now known as flaper87|afk | 22:11 | |
| *** dina_belova has joined #openstack-dev | 22:13 | |
| *** anniec has quit IRC | 22:13 | |
| dolphm | henrynash: Hints.__init__ still exists | 22:16 |
| *** ndipanov has quit IRC | 22:16 | |
| *** jayg is now known as jayg|g0n3 | 22:16 | |
| henrynash | dolphm: oops, sort meant to delete, will do | 22:16 |
| *** tstevenson has quit IRC | 22:16 | |
| *** anniec has joined #openstack-dev | 22:17 | |
| *** anniec has quit IRC | 22:17 | |
| *** dina_belova has quit IRC | 22:18 | |
| dolphm | henrynash: i'm being asked to walk out the door :P | 22:18 |
| *** boris-42 has quit IRC | 22:19 | |
| henrynash | dolphm: ? | 22:19 |
| dolphm | henrynash: dinner time! | 22:19 |
| *** boris-42 has joined #openstack-dev | 22:19 | |
| *** jmontemayor has quit IRC | 22:19 | |
| henrynash | dolphm: ok!!! new patch in about 10 mins | 22:19 |
| dolphm | henrynash: it looks like you still have a bunch of calls to driver_hints.Hints() | 22:20 |
| henrynash | dolphm: is that bad? | 22:20 |
| *** anniec has joined #openstack-dev | 22:20 | |
| dolphm | well, that's the code you're about to delete, right? | 22:21 |
| dolphm | henrynash: i mean, not empty inits | 22:21 |
| henrynash | hmmm | 22:21 |
| henrynash | I guess I need to remove the () | 22:21 |
| dolphm | henrynash: err, let me comment on them in the review to illustrate.. | 22:22 |
| henrynash | dolphm: ok... | 22:22 |
| *** donaldh has quit IRC | 22:23 | |
| *** burt has quit IRC | 22:25 | |
| dolphm | henrynash: http://pasteraw.com/hs2coq5qksvh4yvd1vwyldxl4la29ek | 22:25 |
| *** sarob has joined #openstack-dev | 22:25 | |
| henrynash | dolphm: ok | 22:27 |
| henrynash | dolphm: thx | 22:27 |
| *** changbl has quit IRC | 22:27 | |
| *** otherwiseguy has quit IRC | 22:28 | |
| *** eglynn-is-hungry is now known as eglynn | 22:28 | |
| *** eglynn is now known as eglynn-is-tired | 22:28 | |
| *** danwent has joined #openstack-dev | 22:29 | |
| *** sarob has quit IRC | 22:30 | |
| *** prad_ has quit IRC | 22:31 | |
| *** jhesketh has joined #openstack-dev | 22:32 | |
| *** eharney has quit IRC | 22:32 | |
| *** jhesketh__ has joined #openstack-dev | 22:32 | |
| insanidade | anyone around who could give some basic hints on the way an agent works? | 22:32 |
| *** thedodd has quit IRC | 22:32 | |
| *** guardian_ has quit IRC | 22:33 | |
| *** sarob has joined #openstack-dev | 22:33 | |
| *** FunnyLookinHat has quit IRC | 22:35 | |
| *** atiwari has quit IRC | 22:38 | |
| *** sarob has quit IRC | 22:41 | |
| *** sarob has joined #openstack-dev | 22:41 | |
| *** sheeprine has quit IRC | 22:43 | |
| *** dhellmann is now known as dhellmann_ | 22:43 | |
| *** nati_ueno has quit IRC | 22:44 | |
| *** shakayumi has joined #openstack-dev | 22:44 | |
| *** sarob has quit IRC | 22:45 | |
| *** Guest42194 is now known as zz_Guest42194 | 22:45 | |
| *** maheshp has quit IRC | 22:45 | |
| *** networkstatic has joined #openstack-dev | 22:47 | |
| *** bdpayne has quit IRC | 22:51 | |
| dolphm | henrynash: i'm off to dinner | 22:52 |
| *** sarob has joined #openstack-dev | 22:53 | |
| *** galstrom is now known as galstrom_zzz | 22:53 | |
| henrynash | dolphm: ok, final unit tests running now…will post in a few | 22:53 |
| henrynash | dolphm: have fixed up all issues raised since last patch | 22:54 |
| *** asavu has quit IRC | 22:54 | |
| *** gordc has quit IRC | 22:54 | |
| *** eglynn-is-tired has quit IRC | 22:55 | |
| *** sushils has quit IRC | 22:55 | |
| *** sushils has joined #openstack-dev | 22:55 | |
| *** sarob_ has joined #openstack-dev | 22:56 | |
| *** sdake_ has joined #openstack-dev | 22:56 | |
| henrynash | dolphm: patch posted | 22:58 |
| *** _cerberus_ has quit IRC | 22:58 | |
| *** sarob has quit IRC | 22:59 | |
| *** sheeprine has joined #openstack-dev | 23:00 | |
| *** sheeprine has joined #openstack-dev | 23:00 | |
| *** sarob_ is now known as sarob | 23:02 | |
| *** fbo_away is now known as fbo | 23:03 | |
| henrynash | morganfainberg, bknudson: if your around and good look at https://review.openstack.org/#/c/43257/37 - have fixed up most comments, unless it is major, happy to follow up with fixes once we get it in | 23:05 |
| henrynash | morganfainberg, bknudson: looking for at least one +2 and then dolphm can do the rest | 23:06 |
| *** zz_Guest42194 is now known as leif | 23:07 | |
| *** reed has joined #openstack-dev | 23:07 | |
| *** leif is now known as Guest28385 | 23:07 | |
| *** jbresnah has quit IRC | 23:09 | |
| morganfainberg | henrynash, only 37 patchsets? I don't know… | 23:09 |
| morganfainberg | henrynash, looking at it now. | 23:09 |
| *** edmund1 has quit IRC | 23:10 | |
| bknudson | it's only 1000 lines of code. | 23:10 |
| morganfainberg | henrynash, but i might get pushed into a meeting here. | 23:10 |
| henrynash | morganfainberg: I know, getting a boy to do a man's job etc. etc. | 23:10 |
| *** jbresnah has joined #openstack-dev | 23:12 | |
| *** dina_belova has joined #openstack-dev | 23:13 | |
| *** Ruetobas has quit IRC | 23:14 | |
| *** jimfehlig has quit IRC | 23:14 | |
| *** fbo is now known as fbo_away | 23:15 | |
| *** MaxV has joined #openstack-dev | 23:16 | |
| *** mlavalle has left #openstack-dev | 23:17 | |
| *** dina_belova has quit IRC | 23:18 | |
| *** Ruetobas has joined #openstack-dev | 23:19 | |
| *** alunduil has joined #openstack-dev | 23:23 | |
| *** alexxu has joined #openstack-dev | 23:23 | |
| *** MaxV has quit IRC | 23:24 | |
| *** pmathews has quit IRC | 23:24 | |
| *** sdake_ has quit IRC | 23:24 | |
| *** MaxV has joined #openstack-dev | 23:26 | |
| *** mdenny has quit IRC | 23:27 | |
| *** dims has quit IRC | 23:29 | |
| *** hemna is now known as hemnafk | 23:30 | |
| *** diogogmt has quit IRC | 23:33 | |
| *** gongysh has joined #openstack-dev | 23:36 | |
| *** welldannit has quit IRC | 23:37 | |
| *** kbrierly has quit IRC | 23:38 | |
| *** jbresnah has quit IRC | 23:38 | |
| henrynash | morganfainberg: me, being the boy, in that expression (in case you were confused, or insulted!) | 23:39 |
| *** michchap has joined #openstack-dev | 23:40 | |
| henrynash | bknudson: if you are around, if you think https://review.openstack.org/#/c/43257/37 is in good enough shape then if you could help push it along with a +2 that would be great | 23:41 |
| bknudson | henrynash: I'll take a look at it. | 23:41 |
| henrynash | bknudson: its just about to pop out of jenkins | 23:42 |
| henrynash | bknudson: there are still a few fixes we will want to apply after H3, but in general I think it is in good shape | 23:43 |
| morganfainberg | henrynash, i see at first pass a doc nit. | 23:43 |
| *** jbresnah has joined #openstack-dev | 23:44 | |
| *** jhesketh has quit IRC | 23:44 | |
| morganfainberg | henrynash, nothing is standing out, but i've been busy in meetings so i can only really +1 it at the moment. | 23:44 |
| henrynash | morganfainberg: every little helps :-) | 23:44 |
| morganfainberg | still back looking at it, but if bknudson beats me to +2 .. so be it :) | 23:44 |
| henrynash | morganfainberg: 2 x +2s would be even better of course :-) | 23:45 |
| bknudson | morganfainberg: I'm not a fast reviewer | 23:45 |
| morganfainberg | bknudson, i know but with the amount of distraction on this end, i might not get through it until tomorrow :P | 23:45 |
| henrynash | morganfainberg: we wouldn't hold it for nits (would follow up and fix) | 23:45 |
| morganfainberg | which would be bad. | 23:45 |
| morganfainberg | henrynash, aye i published the +1 and nit, just so you see it | 23:46 |
| henrynash | morganfainberg: ok, thx | 23:46 |
| *** dims has joined #openstack-dev | 23:47 | |
| *** alop has quit IRC | 23:48 | |
| *** MaxV has quit IRC | 23:52 | |
| *** jasdeepH has quit IRC | 23:53 | |
| *** eglynn-is-tired has joined #openstack-dev | 23:56 | |
| henrynash | dolphm: you back? | 23:58 |
| *** epim has quit IRC | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!