| ayoung | morganfainberg, I was just looking at that. I think the logic there is wrong | 00:06 |
|---|---|---|
| ayoung | it fetches the roles out of the backend, which means if the roles change between token creation and conversion, you would get different values. It pre-supposed revoke_by_id semantics | 00:07 |
| ayoung | morganfainberg, you know how you asked to see what the dictionary looks like for 'is_revoked" | 00:07 |
| ayoung | token_data dictionary | 00:07 |
| morganfainberg | ayoung, aye | 00:07 |
| ayoung | its built in the Provider code... | 00:08 |
| morganfainberg | ayoung, ah. | 00:08 |
| ayoung | and I don't like it | 00:08 |
| morganfainberg | ayoung, hey this is why i ask these questions ;) | 00:08 |
| ayoung | link in a sec | 00:08 |
| morganfainberg | ayoung, cool. | 00:08 |
| ayoung | https://review.openstack.org/#/c/55908/63/keystone/token/provider.py line 118 | 00:09 |
| ayoung | for v2 (which has a bug I found) | 00:09 |
| ayoung | and then line 162 | 00:09 |
| ayoung | Now...I know what you are thinking: | 00:09 |
| ayoung | seems wrong to not just pass in the token.... | 00:09 |
| morganfainberg | ayoung, yeah. | 00:09 |
| ayoung | I would need to do "flatten" | 00:10 |
| morganfainberg | ayoung, but if that is the way we need to do it (only good option atm), so be it... | 00:10 |
| ayoung | but I really don';t like that the v2 and v3 code paths are different | 00:10 |
| morganfainberg | flatten would be, imo, a bit cleaner | 00:10 |
| morganfainberg | but i can see where that might be icky | 00:10 |
| morganfainberg | ayoung, ++ | 00:10 |
| ayoung | yeah, icky | 00:10 |
| morganfainberg | ayoung, i really want to see that "problem" go away | 00:10 |
| ayoung | I want to see v2 tokens go away, too | 00:10 |
| morganfainberg | ayoung, might need to wait until K to "really" be resolved (bye bye v2) | 00:10 |
| ayoung | I suspect we could issue nothing be v3 tokens and be fine | 00:11 |
| ayoung | nothing *but* | 00:11 |
| morganfainberg | actually... | 00:11 |
| morganfainberg | maybe that is the real fix. | 00:11 |
| morganfainberg | option "only issue v3 tokens" | 00:11 |
| ayoung | yep | 00:12 |
| morganfainberg | revocation events require that, same as PKI | 00:12 |
| morganfainberg | ayoung, this might make things too easy ... :P | 00:12 |
| ayoung | not really...this logic works...but I need to do more testing...I think I need to have a set of data, and do "v2 token is revoked " and "v3 data is revoked" | 00:12 |
| ayoung | dolphm, can we stop issuing v2 tokens? | 00:13 |
| ayoung | actually, we really can't | 00:13 |
| ayoung | since the token data is what is returned from the validate call, there might be 3rd party apps using that data | 00:13 |
| morganfainberg | no, we can't well not unless we make it an optional thing | 00:13 |
| morganfainberg | correct | 00:13 |
| ayoung | ok, not to be solved today | 00:14 |
| ayoung | I don';t think I'm going to do "flatten" here either...but I might when I port this code to the client | 00:14 |
| ayoung | cuz auth token is going to need it | 00:14 |
| ayoung | so this just needs to work. | 00:14 |
| morganfainberg | ayoung, fair enough | 00:15 |
| morganfainberg | ayoung, maybe in J we offer an "issue v3 only option" to help people transition more cleanly | 00:15 |
| ayoung | ++ | 00:15 |
| morganfainberg | even if they still use some V2 keystone api calls | 00:15 |
| morganfainberg | ayoung, sounds good to me | 00:15 |
| ayoung | morganfainberg, we need the v2 v3 interop from the client we talked about. Then transition should be straightforward. Did that make it in? | 00:16 |
| morganfainberg | ayoung, no, think thats going to be J | 00:16 |
| ayoung | No the client fix is not aligned with Ice | 00:16 |
| morganfainberg | ayoung, ooh | 00:16 |
| morganfainberg | ayoung, uhm. | 00:16 |
| morganfainberg | ayoung, not sure. | 00:16 |
| * ayoung looks | 00:17 | |
| ayoung | https://review.openstack.org/#/c/75731/ | 00:18 |
| morganfainberg | oh that change is pretty simple | 00:19 |
| ayoung | morganfainberg, ^^ is the simplest hack to make it work | 00:19 |
| morganfainberg | yeah | 00:19 |
| ayoung | logic looks a touch convoluted even for a simple patch like that | 00:19 |
| ayoung | looks like there is a loophole in there where hasattr(self, '_management_url'): is true but _management_url is None | 00:20 |
| ayoung | needs to and those two checks together | 00:21 |
| *** krsna has quit IRC | 00:30 | |
| *** gokrokve_ has quit IRC | 00:32 | |
| *** david-lyle has joined #openstack-keystone | 00:41 | |
| *** gokrokve has joined #openstack-keystone | 00:41 | |
| *** browne1 has quit IRC | 00:52 | |
| *** richm has quit IRC | 00:55 | |
| *** gokrokve has quit IRC | 00:56 | |
| *** ayoung has quit IRC | 01:01 | |
| *** ayoung has joined #openstack-keystone | 01:08 | |
| *** devlaps1 has joined #openstack-keystone | 01:20 | |
| *** devlaps has quit IRC | 01:21 | |
| *** devlaps1 has quit IRC | 01:22 | |
| *** devlaps has joined #openstack-keystone | 01:23 | |
| *** devlaps has quit IRC | 01:23 | |
| *** ChanServ sets mode: +o dolphm | 01:33 | |
| *** achampion has joined #openstack-keystone | 01:43 | |
| *** gokrokve has joined #openstack-keystone | 01:46 | |
| *** lnxnut has joined #openstack-keystone | 02:10 | |
| *** lnxnut has quit IRC | 02:52 | |
| *** lnxnut has joined #openstack-keystone | 02:55 | |
| *** marcoemorais has quit IRC | 03:00 | |
| *** dolphm is now known as dolphm_503 | 03:00 | |
| *** devlaps has joined #openstack-keystone | 03:03 | |
| *** devlaps has quit IRC | 03:28 | |
| *** amcrn has quit IRC | 03:29 | |
| *** lbragstad has joined #openstack-keystone | 03:31 | |
| *** dolphm_503 is now known as dolphm | 03:39 | |
| *** dolphm is now known as dolphm_503 | 03:52 | |
| *** lnxnut has quit IRC | 04:18 | |
| *** lnxnut has joined #openstack-keystone | 04:22 | |
| *** harlowja is now known as harlowja_away | 04:31 | |
| *** dolphm_503 is now known as dolphm | 04:43 | |
| *** chandan_kumar has joined #openstack-keystone | 04:48 | |
| *** wchrisj has quit IRC | 04:49 | |
| ayoung | morganfainberg, any chance you are around and can help me with a dogpile question? I need to use it to syncronize the revoke tree | 04:51 |
| morganfainberg | ayoung, hehe sure | 04:51 |
| ayoung | so... | 04:51 |
| ayoung | https://review.openstack.org/#/c/55908/63/keystone/contrib/revoke/core.py | 04:52 |
| ayoung | line 179 | 04:52 |
| * morganfainberg is unfortunately not out having a drink yet, but fortunately able to help | 04:52 | |
| * ayoung had scotch hours ago... | 04:52 | |
| morganfainberg | i might pour some of the cognac here that is on my desk | 04:52 |
| ayoung | I think this is last thing before I can repost. | 04:52 |
| morganfainberg | but i'm debugging some sql migrate issues *production* | 04:52 |
| morganfainberg | and. | 04:53 |
| ayoung | Noice! | 04:53 |
| morganfainberg | i might need a new phone now | 04:53 |
| morganfainberg | sigh | 04:53 |
| ayoung | ouch | 04:53 |
| morganfainberg | just poured coffee all over my desk and phone | 04:53 |
| morganfainberg | *sigh* | 04:53 |
| ayoung | anyway...if I want to use dogpile to syncronize the built Revoke Tree between greenthreads, what should I look at as an example | 04:54 |
| morganfainberg | hmm | 04:54 |
| morganfainberg | kvs is sufficient yes? | 04:54 |
| morganfainberg | in-mem that is | 04:54 |
| morganfainberg | brb getting towel | 04:55 |
| *** dolphm is now known as dolphm_503 | 04:55 | |
| morganfainberg | ok back | 04:57 |
| ayoung | Ooh, that just happend? I thought you were talking about something that happened earlier tonight. | 04:57 |
| ayoung | Sorry | 04:57 |
| ayoung | kvs...sufficient to start. | 04:57 |
| morganfainberg | no i mean i went to type and poured coffee on myself. | 04:58 |
| ayoung | probably should be memcached also in the future | 04:58 |
| morganfainberg | you could probably just use the same backend you're using for the events | 04:58 |
| ayoung | nope | 04:58 |
| morganfainberg | just use another key | 04:58 |
| morganfainberg | wrap in a lock | 04:58 |
| morganfainberg | no? | 04:58 |
| ayoung | they might be stored in sql | 04:58 |
| morganfainberg | oh oh | 04:59 |
| morganfainberg | right | 04:59 |
| morganfainberg | hmm. | 04:59 |
| morganfainberg | same mechanics as you're doing now, probably not configurable to begin with | 04:59 |
| ayoung | not sure how cleanly they are going to serialize.... | 04:59 |
| morganfainberg | just hard set openstack.kvs.memory as the driver | 04:59 |
| morganfainberg | this is a case where i wish we could toggle off the copy mechanism in that backend | 05:00 |
| ayoung | with kvs, it just stores python objects, right? No serialization? | 05:00 |
| morganfainberg | correct | 05:00 |
| ayoung | hmmm | 05:00 |
| morganfainberg | it's an in-memory dict | 05:00 |
| morganfainberg | but get/set does a copy.deepcopy | 05:00 |
| morganfainberg | let me check something | 05:00 |
| ayoung | OK...I can do that...probably will wait until tomorrow | 05:00 |
| morganfainberg | you should eb able to use dogpile.memory.MemoryBackend | 05:01 |
| morganfainberg | it wont do copy/deepcopy so changes will affect the cached values | 05:01 |
| morganfainberg | but you can still use the locking semantics i added to the keystone kvs stuff | 05:02 |
| morganfainberg | it _should_ work, but i haven't tested it extensively with greenthreads/threading | 05:02 |
| morganfainberg | it relies on threading.lock which iirc is patched | 05:03 |
| ayoung | I'll write it up and post...its pretty simple, I think | 05:05 |
| ayoung | Oooh...I need to be able to overide the backing store for Event storage...any reason not do open it up to storage other than In Memory? | 05:06 |
| morganfainberg | ayoung, hm. nah. it's just for syncronization between greenthreads | 05:07 |
| morganfainberg | i don't think it'll make a huge difference/benefit to share outside / across processes | 05:07 |
| ayoung | for events, it will be the actual backing store...I would like to get persistance | 05:07 |
| morganfainberg | ayoung, right. which you will get with Redis or mongo or whatever | 05:08 |
| morganfainberg | and yes that should be configurable | 05:08 |
| ayoung | for the tree...I'll leave that to later, but I might make it a config option as well. It might be useful for Apache based, to not have to rebuild the tree everytime | 05:08 |
| morganfainberg | but i don't see a need to make the tree part more than in-mem | 05:08 |
| ayoung | so memcached too | 05:08 |
| morganfainberg | aye memcached is... ok | 05:08 |
| morganfainberg | :P | 05:08 |
| *** wchrisj__ has joined #openstack-keystone | 05:16 | |
| ayoung | dagnabit...just using the cache buys me nothing unless I copy | 05:20 |
| ayoung | Ok..headed to bed...I need to think about this. | 05:20 |
| *** stevemar has joined #openstack-keystone | 05:42 | |
| *** ChanServ sets mode: +v stevemar | 05:42 | |
| *** dolphm_503 is now known as dolphm | 05:46 | |
| *** dolphm is now known as dolphm_503 | 05:56 | |
| *** chandan_kumar has quit IRC | 06:01 | |
| *** chandan_kumar has joined #openstack-keystone | 06:13 | |
| *** marcoemorais has joined #openstack-keystone | 06:19 | |
| *** wchrisj__ has quit IRC | 06:22 | |
| *** chandan_kumar has quit IRC | 06:28 | |
| *** chandan_kumar has joined #openstack-keystone | 06:40 | |
| *** dstanek has quit IRC | 06:46 | |
| *** dolphm_503 is now known as dolphm | 06:47 | |
| *** chandan_kumar has quit IRC | 06:48 | |
| *** dolphm is now known as dolphm_503 | 06:56 | |
| *** chandan_kumar has joined #openstack-keystone | 06:58 | |
| *** stevemar has quit IRC | 06:59 | |
| *** chandan_kumar has quit IRC | 07:22 | |
| *** dolphm_503 is now known as dolphm | 07:47 | |
| *** marcoemorais1 has joined #openstack-keystone | 07:50 | |
| *** marcoemorais has quit IRC | 07:50 | |
| *** dolphm is now known as dolphm_503 | 07:57 | |
| *** marcoemorais1 has quit IRC | 08:00 | |
| *** morganfainberg is now known as morganfainberg_Z | 08:33 | |
| *** gokrokve has quit IRC | 08:38 | |
| *** gokrokve_ has joined #openstack-keystone | 08:41 | |
| *** gokrokv__ has joined #openstack-keystone | 08:43 | |
| *** gokrokve_ has quit IRC | 08:45 | |
| *** dolphm_503 is now known as dolphm | 08:48 | |
| *** dolphm is now known as dolphm_503 | 08:58 | |
| *** david_lyle_ has joined #openstack-keystone | 09:10 | |
| *** david-lyle has quit IRC | 09:13 | |
| *** david-lyle has joined #openstack-keystone | 09:13 | |
| *** david_lyle_ has quit IRC | 09:15 | |
| *** gokrokv__ has quit IRC | 09:17 | |
| *** gokrokve has joined #openstack-keystone | 09:48 | |
| *** dolphm_503 is now known as dolphm | 09:49 | |
| *** gokrokve_ has joined #openstack-keystone | 09:49 | |
| *** chandan_kumar has joined #openstack-keystone | 09:50 | |
| *** gokrokve has quit IRC | 09:52 | |
| *** dolphm is now known as dolphm_503 | 09:59 | |
| *** chandan_kumar has quit IRC | 10:02 | |
| *** david_lyle_ has joined #openstack-keystone | 10:19 | |
| *** david-lyle has quit IRC | 10:23 | |
| *** chandan_kumar has joined #openstack-keystone | 10:24 | |
| *** dolphm_503 is now known as dolphm | 10:50 | |
| *** dolphm is now known as dolphm_503 | 11:00 | |
| *** chandan_kumar has quit IRC | 11:36 | |
| *** dolphm_503 is now known as dolphm | 11:50 | |
| *** chandan_kumar has joined #openstack-keystone | 11:53 | |
| *** dolphm is now known as dolphm_503 | 12:00 | |
| *** chandan_kumar has quit IRC | 12:38 | |
| *** dolphm_503 is now known as dolphm | 12:51 | |
| *** dolphm is now known as dolphm_503 | 13:01 | |
| *** dstanek has joined #openstack-keystone | 13:33 | |
| *** ChanServ sets mode: +v dstanek | 13:33 | |
| *** dolphm_503 is now known as dolphm | 13:35 | |
| *** dolphm is now known as dolphm_503 | 14:17 | |
| *** dstanek has quit IRC | 14:22 | |
| *** dstanek has joined #openstack-keystone | 14:36 | |
| *** ChanServ sets mode: +v dstanek | 14:36 | |
| *** dolphm_503 is now known as dolphm | 14:36 | |
| *** lnxnut has quit IRC | 14:58 | |
| *** lnxnut has joined #openstack-keystone | 15:04 | |
| *** dolphm is now known as dolphm_503 | 15:05 | |
| *** dolphm_503 is now known as dolphm | 15:25 | |
| *** lnxnut has quit IRC | 15:27 | |
| *** dstanek has quit IRC | 15:28 | |
| *** lnxnut has joined #openstack-keystone | 15:35 | |
| *** dolphm is now known as dolphm_503 | 15:37 | |
| *** lnxnut has quit IRC | 15:39 | |
| *** dolphm_503 is now known as dolphm | 15:56 | |
| *** dstanek has joined #openstack-keystone | 16:28 | |
| *** ChanServ sets mode: +v dstanek | 16:28 | |
| *** lnxnut has joined #openstack-keystone | 16:35 | |
| *** lnxnut has quit IRC | 16:40 | |
| *** dolphm is now known as dolphm_503 | 16:40 | |
| *** lnxnut has joined #openstack-keystone | 17:00 | |
| *** wchrisj has joined #openstack-keystone | 17:22 | |
| *** wchrisj has quit IRC | 17:26 | |
| *** nkinder has quit IRC | 17:26 | |
| *** wchrisj has joined #openstack-keystone | 17:28 | |
| *** wchrisj has quit IRC | 17:31 | |
| *** dolphm_503 is now known as dolphm | 17:35 | |
| *** marcoemorais has joined #openstack-keystone | 17:35 | |
| *** marcoemorais has quit IRC | 17:39 | |
| *** gokrokve_ has quit IRC | 17:43 | |
| *** gokrokve has joined #openstack-keystone | 17:43 | |
| *** thedodd has joined #openstack-keystone | 17:49 | |
| *** lnxnut has quit IRC | 17:52 | |
| *** lnxnut has joined #openstack-keystone | 17:53 | |
| *** lnxnut has quit IRC | 18:20 | |
| *** lnxnut has joined #openstack-keystone | 18:21 | |
| *** lnxnut has quit IRC | 18:21 | |
| *** lnxnut has joined #openstack-keystone | 18:22 | |
| *** lnxnut has quit IRC | 18:26 | |
| *** wchrisj has joined #openstack-keystone | 18:39 | |
| *** wchrisj has quit IRC | 18:48 | |
| *** leseb has joined #openstack-keystone | 18:50 | |
| *** wchrisj has joined #openstack-keystone | 18:56 | |
| *** ayoung has quit IRC | 18:59 | |
| *** dolphm is now known as dolphm_503 | 19:00 | |
| *** wchrisj has quit IRC | 19:06 | |
| *** leseb has quit IRC | 19:08 | |
| *** dolphm_503 is now known as dolphm | 19:20 | |
| *** lnxnut has joined #openstack-keystone | 19:23 | |
| *** leseb has joined #openstack-keystone | 19:24 | |
| *** leseb has quit IRC | 19:24 | |
| *** lnxnut has quit IRC | 19:27 | |
| *** dolphm is now known as dolphm_503 | 19:29 | |
| *** lnxnut has joined #openstack-keystone | 19:35 | |
| *** lnxnut has quit IRC | 19:40 | |
| *** topol has joined #openstack-keystone | 20:04 | |
| *** topol_ has joined #openstack-keystone | 20:06 | |
| *** topol has quit IRC | 20:09 | |
| *** topol_ is now known as topol | 20:09 | |
| *** topol has quit IRC | 20:20 | |
| *** dolphm_503 is now known as dolphm | 20:20 | |
| *** leseb has joined #openstack-keystone | 20:22 | |
| *** lnxnut has joined #openstack-keystone | 20:35 | |
| *** rwsu has quit IRC | 20:40 | |
| *** lnxnut has quit IRC | 20:40 | |
| *** ayoung has joined #openstack-keystone | 20:45 | |
| *** leseb has quit IRC | 20:51 | |
| *** dolphm is now known as dolphm_503 | 20:58 | |
| *** thedodd has quit IRC | 21:05 | |
| *** leseb has joined #openstack-keystone | 21:18 | |
| *** wchrisj has joined #openstack-keystone | 21:34 | |
| *** lnxnut has joined #openstack-keystone | 21:35 | |
| *** lnxnut has quit IRC | 21:40 | |
| *** leseb has quit IRC | 21:43 | |
| *** leseb has joined #openstack-keystone | 21:44 | |
| *** leseb has quit IRC | 21:48 | |
| *** dolphm_503 is now known as dolphm | 21:49 | |
| *** wchrisj has quit IRC | 21:54 | |
| *** dolphm is now known as dolphm_503 | 21:58 | |
| *** leseb has joined #openstack-keystone | 22:09 | |
| *** dolphm_503 is now known as dolphm | 22:14 | |
| *** dolphm is now known as dolphm_503 | 22:17 | |
| *** ayoung has quit IRC | 22:20 | |
| *** dolphm_503 is now known as dolphm | 22:27 | |
| *** ayoung has joined #openstack-keystone | 22:33 | |
| *** lnxnut has joined #openstack-keystone | 22:35 | |
| *** lnxnut has quit IRC | 22:40 | |
| *** dolphm is now known as dolphm_503 | 23:29 | |
| *** dolphm_503 is now known as dolphm | 23:30 | |
| *** lnxnut has joined #openstack-keystone | 23:35 | |
| *** leseb has quit IRC | 23:39 | |
| *** dolphm is now known as dolphm_503 | 23:39 | |
| *** leseb has joined #openstack-keystone | 23:39 | |
| *** lnxnut has quit IRC | 23:40 | |
| *** leseb has quit IRC | 23:44 | |
| *** leseb has joined #openstack-keystone | 23:44 | |
| *** dolphm_503 is now known as dolphm | 23:49 | |
| *** dolphm is now known as dolphm_503 | 23:59 | |
| *** leseb has quit IRC | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!