| *** mokas has joined #openstack-dev | 00:01 | |
| *** mnewby has quit IRC | 00:02 | |
| *** spiffxp has joined #openstack-dev | 00:03 | |
| *** steveb_ has joined #openstack-dev | 00:04 | |
| *** anniec has quit IRC | 00:08 | |
| *** issackelly has quit IRC | 00:09 | |
| *** thingee has quit IRC | 00:10 | |
| *** johnpostlethwait has quit IRC | 00:11 | |
| *** anniec has joined #openstack-dev | 00:16 | |
| *** johnpostlethwait has joined #openstack-dev | 00:24 | |
| *** Mandell has quit IRC | 00:25 | |
| *** matwood has quit IRC | 00:26 | |
| *** halfss has quit IRC | 00:27 | |
| *** novas0x2a|lapto1 has quit IRC | 00:27 | |
| *** johnpostlethwait has quit IRC | 00:28 | |
| *** Aaton is now known as Aaton_off | 00:34 | |
| *** s0mik has joined #openstack-dev | 00:35 | |
| *** gongys has quit IRC | 00:36 | |
| *** utlemming has joined #openstack-dev | 00:39 | |
| *** adjohn has joined #openstack-dev | 00:40 | |
| *** anniec has quit IRC | 00:40 | |
| *** s0mik has quit IRC | 00:43 | |
| *** pixelbeat has quit IRC | 00:44 | |
| *** johnpostlethwait has joined #openstack-dev | 00:47 | |
| *** anniec has joined #openstack-dev | 00:48 | |
| *** johnpostlethwait has quit IRC | 00:49 | |
| *** rods_ has quit IRC | 00:51 | |
| *** mokas has quit IRC | 00:52 | |
| *** andrewsmedina has joined #openstack-dev | 00:54 | |
| *** matwood has joined #openstack-dev | 00:55 | |
| *** s0mik has joined #openstack-dev | 00:59 | |
| *** dachary has quit IRC | 01:02 | |
| *** dachary has joined #openstack-dev | 01:02 | |
| *** spiffxp has quit IRC | 01:03 | |
| *** andrewsmedina has quit IRC | 01:06 | |
| *** nunosantos has quit IRC | 01:10 | |
| *** jog0 has quit IRC | 01:17 | |
| *** andrewsmedina has joined #openstack-dev | 01:18 | |
| *** jdurgin has quit IRC | 01:19 | |
| *** anniec has quit IRC | 01:22 | |
| *** s0mik has quit IRC | 01:24 | |
| *** danwent has quit IRC | 01:32 | |
| *** anniec has joined #openstack-dev | 01:50 | |
| *** zhuadl has joined #openstack-dev | 01:53 | |
| *** johnpostlethwait has joined #openstack-dev | 01:57 | |
| *** johnpostlethwait has quit IRC | 02:00 | |
| *** anniec_ has joined #openstack-dev | 02:02 | |
| *** anniec_ has quit IRC | 02:03 | |
| *** mokas has joined #openstack-dev | 02:04 | |
| *** anniec has quit IRC | 02:04 | |
| *** xchu has joined #openstack-dev | 02:07 | |
| *** sniperd has quit IRC | 02:10 | |
| *** sacharya has joined #openstack-dev | 02:12 | |
| *** littleidea has quit IRC | 02:14 | |
| *** danwent has joined #openstack-dev | 02:18 | |
| *** littleidea has joined #openstack-dev | 02:24 | |
| *** mestery has joined #openstack-dev | 02:31 | |
| *** roge has joined #openstack-dev | 02:36 | |
| *** JordanRinke has quit IRC | 02:38 | |
| *** JordanRinke has joined #openstack-dev | 02:38 | |
| *** troytoman-away is now known as troytoman | 02:40 | |
| *** xchu has quit IRC | 02:42 | |
| *** roge has quit IRC | 02:51 | |
| *** JordanRinke has joined #openstack-dev | 02:52 | |
| *** JordanRinke has quit IRC | 02:53 | |
| *** JordanRinke has joined #openstack-dev | 02:53 | |
| *** bencherian has quit IRC | 02:57 | |
| *** dachary has quit IRC | 03:02 | |
| *** dachary has joined #openstack-dev | 03:02 | |
| *** littleidea has quit IRC | 03:09 | |
| *** markmcclain has quit IRC | 03:23 | |
| *** ayoung_ has quit IRC | 03:24 | |
| *** adjohn has quit IRC | 03:26 | |
| *** adjohn has joined #openstack-dev | 03:27 | |
| *** eglynn_ has joined #openstack-dev | 03:38 | |
| *** eglynn has quit IRC | 03:41 | |
| *** andrewsmedina has quit IRC | 03:50 | |
| *** littleidea has joined #openstack-dev | 03:52 | |
| *** blamar has quit IRC | 03:57 | |
| *** dachary has quit IRC | 04:00 | |
| *** dachary has joined #openstack-dev | 04:00 | |
| *** littleidea has quit IRC | 04:07 | |
| *** zhuadl has quit IRC | 04:09 | |
| *** adjohn has quit IRC | 04:10 | |
| *** Ryan_Lane has joined #openstack-dev | 04:14 | |
| *** spiffxp has joined #openstack-dev | 04:16 | |
| *** littleidea has joined #openstack-dev | 04:17 | |
| *** danwent has quit IRC | 04:18 | |
| *** xchu has joined #openstack-dev | 04:23 | |
| *** Mandell has joined #openstack-dev | 04:24 | |
| *** sacharya has quit IRC | 04:30 | |
| *** anderstj has joined #openstack-dev | 04:33 | |
| *** Mandell has quit IRC | 04:33 | |
| *** littleidea has quit IRC | 04:35 | |
| vishy | jeblair, mtaylor, LinuxJedi: looks like the update to use cinder in ci broke stable-essex merge: https://jenkins.openstack.org/job/gate-tempest-devstack-vm/4888/consoleFull | 04:36 |
|---|---|---|
| vishy | ideas? | 04:36 |
| *** dtroyer is now known as dtroyer_zzz | 04:39 | |
| *** jcooley has joined #openstack-dev | 04:40 | |
| clarkb | which is expected because cinder isn't part of openstack in essex? Maybe we can have devstack-gate determine what the enabled services should be based on GERRIT_BRANCH | 04:40 |
| clarkb | vishy: ^ | 04:42 |
| vishy | clarkb: we might have to have a stable/essex branch of devstack-gate | 04:43 |
| vishy | in fact I'm surprised we haven't needed one before. | 04:43 |
| vishy | clarkb: we have one for devstack already. | 04:43 |
| *** salv-orlando has joined #openstack-dev | 04:45 | |
| clarkb | I don't think I have the permissions to add a branch to devstack-gate, but that is a good idea. Then we can simply checkout GERRIT_BRANCH of devstack-gate when cleaning up the workspace | 04:49 |
| jgriffith | well that's disappointing | 04:49 |
| clarkb | jeblair: mtaylor ^ | 04:49 |
| *** bencherian has joined #openstack-dev | 04:59 | |
| *** steveb_ has quit IRC | 05:01 | |
| *** spiffxp has quit IRC | 05:04 | |
| *** dtroyer_zzz is now known as dtroyer | 05:04 | |
| *** dachary has quit IRC | 05:06 | |
| *** dachary has joined #openstack-dev | 05:06 | |
| *** mokas has quit IRC | 05:07 | |
| *** danwent has joined #openstack-dev | 05:07 | |
| mtaylor | clarkb: aroo? | 05:08 |
| mtaylor | vishy: EXCELLENT | 05:08 |
| mtaylor | vishy: wait! what? you're telling me that we didnt' have cinder in stable/essex??? | 05:08 |
| mtaylor | vishy: dammit man! | 05:09 |
| *** littleidea has joined #openstack-dev | 05:09 | |
| *** asakhnov has joined #openstack-dev | 05:10 | |
| clarkb | I want an easy button that says "aroo?!" | 05:10 |
| mtaylor | jgriffith: I don't think we can add a branch... I think we need to backport cinder to stable/essex | 05:10 |
| mtaylor | clarkb: weechat plugin? :) | 05:10 |
| jgriffith | mtaylor: say whaaaaa | 05:10 |
| mtaylor | jgriffith: wait, you don't think my idea is better than clarkb's ? | 05:10 |
| jgriffith | mtaylor: I'm working on it now | 05:11 |
| jgriffith | mtaylor: I'm certain it's the best solution | 05:11 |
| jgriffith | Done... cp -R nova-folsom nova-essex-stable | 05:12 |
| clarkb | that was easy | 05:13 |
| clarkb | :) | 05:13 |
| jgriffith | pheww... lots of work averted | 05:13 |
| *** zhuadl has joined #openstack-dev | 05:14 | |
| *** zaitcev has quit IRC | 05:14 | |
| mtaylor | jgriffith: thanks! that's so helpful | 05:14 |
| jgriffith | mtaylor: no problem... any time | 05:15 |
| jgriffith | mtaylor: Seriously though... we can add a gate branch right? | 05:16 |
| jgriffith | I'm out | 05:18 |
| jgriffith | mtaylor: lemme know if I can help with anything (other than backporting cinder to Essex that is) | 05:19 |
| clarkb | mtaylor: if you need inspiration http://www.youtube.com/watch?v=Igi1LSkASyE | 05:21 |
| *** garyk has joined #openstack-dev | 05:24 | |
| mtaylor | clarkb: thank you :) | 05:30 |
| *** dtroyer is now known as dtroyer_zzz | 05:35 | |
| *** mokas has joined #openstack-dev | 05:36 | |
| *** matiu has quit IRC | 05:36 | |
| *** asakhnov has quit IRC | 05:38 | |
| *** markmc has joined #openstack-dev | 05:40 | |
| *** troytoman is now known as troytoman-away | 05:56 | |
| *** hattwick has quit IRC | 06:01 | |
| *** johnpostlethwait has joined #openstack-dev | 06:07 | |
| *** GheRivero has joined #openstack-dev | 06:18 | |
| *** anderstj has quit IRC | 06:22 | |
| *** sniperd has joined #openstack-dev | 06:31 | |
| *** dachary has quit IRC | 06:32 | |
| *** dachary has joined #openstack-dev | 06:33 | |
| *** littleidea has quit IRC | 06:33 | |
| *** dachary has quit IRC | 06:33 | |
| *** alex88 has joined #openstack-dev | 06:49 | |
| *** Ryan_Lane has quit IRC | 06:57 | |
| *** rohitk has joined #openstack-dev | 07:07 | |
| alex88 | hi guys, someone ever tried to live-migrate a vm with floating ip attached? i've tried and live migration works, but the DNAT rules are still on the old host | 07:10 |
| *** Ryan_Lane has joined #openstack-dev | 07:13 | |
| *** amotoki has joined #openstack-dev | 07:19 | |
| ttx | alex88: did the CI guys solve your CLA issue ? | 07:26 |
| alex88 | ttx: they did, thanks! | 07:26 |
| alex88 | i've filling a bug about the above problem | 07:27 |
| ttx | (apparently yesterday I bet in the one that was in vacation, sorry about that) | 07:27 |
| alex88 | ttx: np, they told me there was a bug in the cron job | 07:27 |
| *** erikzaadi has joined #openstack-dev | 07:31 | |
| *** lianhao has quit IRC | 07:31 | |
| *** Ryan_Lane has quit IRC | 07:32 | |
| *** Mandell has joined #openstack-dev | 07:33 | |
| *** djinni has quit IRC | 07:35 | |
| *** sulochan has joined #openstack-dev | 07:35 | |
| *** sdake has quit IRC | 07:36 | |
| *** sulochan_ has joined #openstack-dev | 07:36 | |
| *** djinni has joined #openstack-dev | 07:37 | |
| *** Mandell_ has joined #openstack-dev | 07:38 | |
| *** Mandell has quit IRC | 07:38 | |
| *** sulochan has quit IRC | 07:40 | |
| *** sulochan_ is now known as sulochan | 07:40 | |
| *** jcooley has quit IRC | 07:41 | |
| *** jcooley has joined #openstack-dev | 07:42 | |
| *** sdake has joined #openstack-dev | 07:45 | |
| *** steveb_ has joined #openstack-dev | 07:52 | |
| *** bencherian has quit IRC | 07:53 | |
| *** toey has joined #openstack-dev | 07:56 | |
| *** dubsquared has joined #openstack-dev | 08:04 | |
| *** johnpostlethwait has quit IRC | 08:11 | |
| *** robert-vanleeuwe has joined #openstack-dev | 08:13 | |
| *** derekh has joined #openstack-dev | 08:13 | |
| GheRivero | hi people | 08:14 |
| *** almaisan-away is now known as al-maisan | 08:21 | |
| *** dubsquared has quit IRC | 08:27 | |
| *** pixelbeat has joined #openstack-dev | 08:28 | |
| *** sdake has quit IRC | 08:29 | |
| *** darraghb has joined #openstack-dev | 08:31 | |
| salv-orlando | GheRivero: hi! | 08:31 |
| *** toey has quit IRC | 08:36 | |
| *** danpb has joined #openstack-dev | 08:41 | |
| *** danwent has quit IRC | 08:43 | |
| *** hattwick has joined #openstack-dev | 08:43 | |
| *** sdake has joined #openstack-dev | 08:44 | |
| *** apevec has joined #openstack-dev | 08:50 | |
| *** dachary has joined #openstack-dev | 08:56 | |
| *** anniec has joined #openstack-dev | 08:59 | |
| *** anniec_ has joined #openstack-dev | 09:00 | |
| comstud | File "tools/hacking.py", line 389, in once_git_check_commit_title | 09:02 |
| comstud | title = subprocess.check_output('git log --pretty=%s -1', shell=True) | 09:02 |
| comstud | AttributeError: 'module' object has no attribute 'check_output' | 09:02 |
| comstud | someone committed a 2.7 only thing | 09:02 |
| *** anniec has quit IRC | 09:03 | |
| *** anniec_ is now known as anniec | 09:03 | |
| *** robert-vanleeuwe has quit IRC | 09:05 | |
| comstud | mtaylor, jeblair: https://review.openstack.org/#/c/9994/ <- may want to check on how that passed python 2.6 gate. subprocess.check_output was added in 2.7 | 09:06 |
| *** Mandell_ has quit IRC | 09:31 | |
| *** steveb_ has quit IRC | 09:37 | |
| *** amotoki has quit IRC | 09:42 | |
| *** matwood has quit IRC | 09:46 | |
| *** dachary has quit IRC | 09:48 | |
| *** thovden has quit IRC | 09:53 | |
| *** matwood has joined #openstack-dev | 09:54 | |
| *** trapni has quit IRC | 10:07 | |
| *** trapni has joined #openstack-dev | 10:07 | |
| *** xchu has quit IRC | 10:17 | |
| *** ttx has quit IRC | 10:21 | |
| *** ttx has joined #openstack-dev | 10:21 | |
| *** ttx has quit IRC | 10:21 | |
| *** ttx has joined #openstack-dev | 10:21 | |
| *** nunosantos has joined #openstack-dev | 10:22 | |
| *** nunosantos_ has joined #openstack-dev | 10:22 | |
| *** rods_ has joined #openstack-dev | 10:24 | |
| *** thovden has joined #openstack-dev | 10:25 | |
| *** matwood has quit IRC | 10:41 | |
| *** Vek has quit IRC | 11:02 | |
| *** toey has joined #openstack-dev | 11:04 | |
| *** al-maisan is now known as almaisan-away | 11:10 | |
| garyk | salv-orlando: ping | 11:16 |
| *** andrea_ has joined #openstack-dev | 11:19 | |
| *** toey has quit IRC | 11:26 | |
| *** arosen has quit IRC | 11:31 | |
| *** zhuadl has quit IRC | 11:33 | |
| *** markvoelker has joined #openstack-dev | 11:37 | |
| *** wiliam has joined #openstack-dev | 11:41 | |
| *** arosen has joined #openstack-dev | 11:50 | |
| *** halfss has joined #openstack-dev | 11:57 | |
| salv-orlando | garyk: pong | 11:58 |
| garyk | salv-orlando, hey, how are you? can you help me for a few minutes? i am writing tests for the rpc. it is a bit of a nightmare | 11:59 |
| salv-orlando | sure. what's the problem? | 11:59 |
| garyk | salv-orlando, from wjhat i understand i am unable to test the entire system, so i just need to check the callbacks and thier methods. is this correct? | 12:00 |
| salv-orlando | yes, the actual rpc class should be either stubbed out or mocked | 12:00 |
| salv-orlando | I meant calls not class | 12:01 |
| garyk | salv-orlando, ok, thanks, i'll try that | 12:01 |
| *** thovden has quit IRC | 12:05 | |
| *** epic has quit IRC | 12:09 | |
| *** epic has joined #openstack-dev | 12:09 | |
| *** mdomsch has joined #openstack-dev | 12:20 | |
| *** salgado has joined #openstack-dev | 12:25 | |
| *** dhellmann_ has joined #openstack-dev | 12:28 | |
| *** dhellmann_ has quit IRC | 12:29 | |
| *** salgado has quit IRC | 12:30 | |
| andrea_ | I am working an a change that require a new script for the SQLAlchemy migration tool, is there a process to follow to get a kind of reservation for a migration number? | 12:32 |
| *** dhellmann has quit IRC | 12:33 | |
| markmc | andrea_, just pick the next number | 12:36 |
| markmc | andrea_, if someone else does the same and their patch gets merged first, you'll just need to rebase and use the next number | 12:36 |
| andrea_ | markmc: ok thanks | 12:38 |
| andrea_ | I asked because I found an old message in the ML talking about a proposal for a reservation process (https://lists.launchpad.net/openstack/msg02680.html). | 12:39 |
| andrea_ | thanks again | 12:39 |
| *** toey has joined #openstack-dev | 12:45 | |
| *** sacharya has joined #openstack-dev | 12:47 | |
| *** dachary has joined #openstack-dev | 12:51 | |
| *** littleidea has joined #openstack-dev | 12:51 | |
| markmc | dansmith, hey, there? | 12:52 |
| *** kbringard has joined #openstack-dev | 12:52 | |
| *** kbringard has joined #openstack-dev | 12:53 | |
| *** blamar has joined #openstack-dev | 12:53 | |
| *** roge has joined #openstack-dev | 12:53 | |
| mtaylor | comstud: we don't run hacking/pep8 on both 27 and 26 - only on 27 | 12:53 |
| mtaylor | comstud: normally it wouldn't make any sense to run pep8 in both - but obviously nova wraps pep8 | 12:54 |
| mtaylor | perhaps we should put in a unit test which tests hacking? | 12:54 |
| *** rkukura has quit IRC | 12:55 | |
| comstud | hrm | 12:55 |
| *** rkukura has joined #openstack-dev | 12:55 | |
| comstud | tests would not even start for me | 12:55 |
| comstud | until the hacking thing was fixed | 12:55 |
| comstud | for whatever reason.. i didn't look at why | 12:56 |
| *** toey has quit IRC | 12:56 | |
| comstud | although | 12:56 |
| comstud | i must have run with only -p or something | 12:56 |
| * comstud checks | 12:56 | |
| comstud | ah i lied | 12:56 |
| mtaylor | oh, you're still running run_tests.sh probably? | 12:57 |
| comstud | so Yeah.. I guess maybe unit tests should test hacking ;) | 12:57 |
| comstud | yeah i am | 12:57 |
| comstud | 2 pep8 failures snuck in to trunk somehow too | 12:57 |
| mtaylor | right pep8 version? | 12:57 |
| comstud | yes | 12:57 |
| comstud | someone fixed it already | 12:57 |
| comstud | merged | 12:58 |
| comstud | last merge for nova | 12:58 |
| *** GheRivero has quit IRC | 12:58 | |
| mtaylor | looking | 12:58 |
| *** sacharya has quit IRC | 12:59 | |
| *** lorin1 has joined #openstack-dev | 13:02 | |
| mtaylor | HAHAHAH | 13:02 |
| *** Vek has joined #openstack-dev | 13:02 | |
| mtaylor | ok. that's funny | 13:03 |
| *** japage has quit IRC | 13:03 | |
| *** japage has joined #openstack-dev | 13:04 | |
| mtaylor | comstud: we had an exclude for pep8 for "openstack" - to avoid re-checking openstack/common | 13:05 |
| *** blamar has quit IRC | 13:05 | |
| markmc | mtaylor, did you just had an orgasm looking at code in your terminal? | 13:05 |
| * markmc searching for evidence of this | 13:05 | |
| markmc | http://www.cloudave.com/21107/two-years-of-openstack-looking-from-the-other-side/ | 13:05 |
| markmc | "Open source world is dominated by developers who had dedicated their life to geekdom. They get their orgasm seeing the code in a terminal window." | 13:06 |
| mtaylor | comstud: turns out, that matches nova/api/openstack/compute/contrib/console_output.py | 13:06 |
| *** gargya has joined #openstack-dev | 13:06 | |
| mtaylor | wow. we do? | 13:06 |
| markmc | apparently | 13:06 |
| * markmc feels he's missing out | 13:06 | |
| mtaylor | wait - we've dedicated our life to geekdom itself? | 13:06 |
| markmc | that too | 13:06 |
| mtaylor | always seemed like a by-product to me, but whatevs | 13:06 |
| comstud | mtaylor: aha | 13:08 |
| comstud | so all of nova/api/openstack is/was not being tested? | 13:08 |
| *** rkukura has quit IRC | 13:09 | |
| *** vanchester has joined #openstack-dev | 13:09 | |
| mtaylor | yup | 13:11 |
| vanchester | Hi. I am using swift + keystone + s3. Is it possible to authenticate with access_key='tenant:user' like in examples, cause it seems like it isn't? | 13:12 |
| vanchester | I look at the code, it splits access_key with ':'. But it's many samples with 'tenant:user' access, maybe i do something wrong? | 13:14 |
| *** dtroyer_zzz is now known as dtroyer | 13:18 | |
| *** galvint_ has joined #openstack-dev | 13:19 | |
| *** ayoung has joined #openstack-dev | 13:19 | |
| *** thovden has joined #openstack-dev | 13:20 | |
| *** mnewby has joined #openstack-dev | 13:21 | |
| *** dprince has joined #openstack-dev | 13:24 | |
| *** GheRivero has joined #openstack-dev | 13:24 | |
| mtaylor | comstud: working on a fix - turns out pep8 does its exclude testing in the most assinine manner known to man | 13:25 |
| comstud | haha | 13:25 |
| comstud | no worries | 13:25 |
| *** matiu_ is now known as matiu | 13:26 | |
| *** thovden has quit IRC | 13:26 | |
| *** littleidea has quit IRC | 13:27 | |
| *** thovden has joined #openstack-dev | 13:28 | |
| *** Hien has quit IRC | 13:30 | |
| dansmith | markmc: I am now | 13:32 |
| mtaylor | comstud: https://review.openstack.org/10340 | 13:32 |
| markmc | dansmith, some comments in your review | 13:32 |
| mtaylor | comstud: that's the best I got - not crazy about monkey patching, but what can you do? | 13:32 |
| markmc | dansmith, I'm going to post a couple of checks_instance_lock patches now too | 13:32 |
| *** Hien has joined #openstack-dev | 13:33 | |
| dansmith | markmc: hmm, so checks_instance_lock() returns False if locked, that's the behavior that the stuff above expects | 13:34 |
| comstud | mtaylor: you have a pep8 error in your pep8 fix | 13:34 |
| dansmith | ...or it did, I didn't realize russellb's stuff had made it in yet | 13:34 |
| comstud | :) | 13:34 |
| comstud | pep8.excluded=excluded | 13:34 |
| markmc | dansmith, where is the return value used? | 13:34 |
| comstud | spaces | 13:34 |
| mtaylor | comstud: hahaha | 13:34 |
| mtaylor | thanks | 13:34 |
| comstud | haha | 13:34 |
| markmc | dansmith, I don't see any usage, apart from one unit test | 13:34 |
| comstud | no problem! | 13:34 |
| dansmith | markmc: on the other end of the rpc layer.. it's very obscure | 13:35 |
| *** sacharya has joined #openstack-dev | 13:35 | |
| dansmith | there was a comment too, let me see if I can find it | 13:35 |
| comstud | pep8.input_dir=input_dir also | 13:35 |
| mtaylor | comstud: fixed | 13:36 |
| comstud | thnx! | 13:38 |
| dansmith | markmc: hmm, well, I'm not spotting it.. maybe I'm just thinking of my prior attempt to catch the False and do something at the api layer | 13:38 |
| markmc | dansmith, yeah, eglynn has a patch where he does that for set_admin_password() | 13:39 |
| markmc | dansmith, which is what lead me down this path | 13:39 |
| dansmith | markmc: right | 13:39 |
| markmc | dansmith, patches coming soon, should explain all | 13:39 |
| *** andrewsmedina has joined #openstack-dev | 13:40 | |
| dansmith | markmc: are you going to do the state reversal, or just do cleanup upon which I should base another iteration? | 13:40 |
| comstud | ok, almost 7am | 13:41 |
| comstud | time for a nap | 13:41 |
| * comstud & bbl | 13:41 | |
| mtaylor | comstud: https://github.com/jcrocholl/pep8/pull/111 | 13:41 |
| markmc | dansmith, just the cleanup - see https://github.com/markmc/nova/commits/checks_instance_lock | 13:41 |
| markmc | dansmith, not done thinking over the state reversal thing, but here's a thought | 13:42 |
| markmc | dansmith, the client side is setting the task state | 13:42 |
| markmc | dansmith, and on the server side you're saying which states can be reverted | 13:42 |
| markmc | dansmith, definitely looks like a case of repeating ourselves | 13:42 |
| markmc | dansmith, why not e.g. | 13:42 |
| markmc | with revert_task_state(instance, task_states.REBOOTING): | 13:43 |
| markmc | self.compute_rpcapi.reboot_instance(instance) | 13:43 |
| markmc | dansmith, that's on the client side | 13:44 |
| *** utlemming has quit IRC | 13:44 | |
| markmc | dansmith, ah, because we're talking about casts - these are async actions | 13:44 |
| dansmith | markmc: right, I already went down that path :) | 13:44 |
| dansmith | it seems horrifically imbalanced to me, to do the set on one side and the revert on the other, | 13:45 |
| dansmith | but c'est la via (or something like that) | 13:45 |
| dansmith | er, vie | 13:45 |
| markmc | dansmith, ok, now I'm thinking that specifying the state that may be reverted on the server side is overkill | 13:45 |
| markmc | dansmith, when the task succeeds, it clears the state irrespective of what it currently is | 13:46 |
| markmc | dansmith, why do different on error? | 13:46 |
| dansmith | does it? where is that? | 13:46 |
| markmc | dansmith, well, e.g. in reboot_instance we just do: | 13:46 |
| markmc | self._instance_update(context, instance['uuid'], | 13:47 |
| markmc | ... task_state=None) | 13:47 |
| *** djoyner has joined #openstack-dev | 13:47 | |
| markmc | dansmith, there's no check that task_state is currently REBOOTING | 13:47 |
| *** asakhnov has joined #openstack-dev | 13:48 | |
| dansmith | oh, I see, I misunderstood you above.. I was thinking that the error case should be more cautious because we don't know why it failed necessarily, | 13:48 |
| dansmith | and if it succeeded, then blowing away the task state is more appropriate | 13:49 |
| dansmith | however, | 13:49 |
| dansmith | it'll make it smaller and make jaypipes happier | 13:49 |
| dansmith | so that's fiiiiiiiiiine with me :) | 13:49 |
| markmc | cool | 13:49 |
| *** mdomsch_ has joined #openstack-dev | 13:51 | |
| *** mdomsch has quit IRC | 13:52 | |
| *** utlemming has joined #openstack-dev | 13:53 | |
| markmc | dansmith, also, two separate decorators | 13:56 |
| markmc | dansmith, where it's basically not valid to use one without the other | 13:56 |
| markmc | dansmith, hmm, I guess you can use @revert_task_state without @checks_instance_lock, ok | 13:57 |
| markmc | dansmith, since it catches all exceptions | 13:57 |
| *** bencherian has joined #openstack-dev | 13:57 | |
| dansmith | yeah | 13:58 |
| garyk | markmc, can you let me know if you have a minute or two to discuss --config-file | 14:01 |
| markmc | dansmith, on a call now, sorry | 14:02 |
| *** djoyner has left #openstack-dev | 14:04 | |
| * dansmith forwards packets from markmc to garyk | 14:06 | |
| *** utlemming has quit IRC | 14:06 | |
| *** rods_ has quit IRC | 14:08 | |
| markmc | dansmith, thanks for looking at this btw, hadn't realized how much of a mess it was | 14:08 |
| *** utlemming has joined #openstack-dev | 14:10 | |
| dansmith | markmc: heh, well, I clearly wasn't making it any better, I just wanted it to not be broken at least :) | 14:12 |
| dansmith | markmc: I haven't quite gotten over the hump to feeling like it's okay to substantially refactor stuff like that.. :) | 14:12 |
| mtaylor | dansmith: http://i.imgur.com/wGUTG.gif | 14:13 |
| markmc | dansmith, just assume everything is completely broken :) | 14:13 |
| markmc | mtaylor, I love that! seriously, best gif ever | 14:13 |
| mtaylor | markmc: clarkb gave that to me two days ago. it's been in its own browser tab ever since :) | 14:14 |
| *** dachary has quit IRC | 14:14 | |
| markmc | mtaylor, I tweeted it months ago | 14:14 |
| dansmith | hehe | 14:14 |
| *** dachary has joined #openstack-dev | 14:15 | |
| *** andrewbogott has quit IRC | 14:16 | |
| markmc | mtaylor, you clearly don't pay close enough attention to me on twitter | 14:16 |
| *** dachary1 has joined #openstack-dev | 14:16 | |
| *** Mandell has joined #openstack-dev | 14:16 | |
| dansmith | heh, I say that to my wife too often | 14:17 |
| dansmith | "Wife: Oh, I didn't know you..." | 14:17 |
| dansmith | "Me: Well, if you _read_ my blog now and then..." | 14:17 |
| *** dachary1 has quit IRC | 14:17 | |
| *** matiu has quit IRC | 14:18 | |
| *** matiu has joined #openstack-dev | 14:18 | |
| *** dachary has quit IRC | 14:20 | |
| *** markmcclain has joined #openstack-dev | 14:20 | |
| *** zhuadl has joined #openstack-dev | 14:22 | |
| markmc | dprince, any idea what's going on with these smokestack failures? https://review.openstack.org/#/c/10247/ | 14:23 |
| *** Mandell has quit IRC | 14:24 | |
| dprince | markmc: let me have a look. | 14:25 |
| dprince | markmc: Let me re-kick it. | 14:26 |
| markmc | dprince, thanks | 14:26 |
| *** dtroyer is now known as dtroyer_zzz | 14:28 | |
| *** japage has quit IRC | 14:33 | |
| *** japage has joined #openstack-dev | 14:33 | |
| *** maoy has joined #openstack-dev | 14:34 | |
| *** blamar has joined #openstack-dev | 14:35 | |
| *** zhuadl has quit IRC | 14:36 | |
| *** maoy has quit IRC | 14:38 | |
| *** dtroyer_zzz is now known as dtroyer | 14:42 | |
| *** nunosantos_ has quit IRC | 14:52 | |
| *** nunosantos_ has joined #openstack-dev | 14:53 | |
| *** mnewby has quit IRC | 14:53 | |
| eglynn_ | markmc: are we on the same page at all regarding https://review.openstack.org/10130? | 14:54 |
| eglynn_ | markmc: i.e. ... am I getting the wrong end of the stick re. the silent failure of checks_instance_lock, and the potential consequences of that? | 14:54 |
| markmc | eglynn_, right, we have a serious issue with task_state vs locks and dansmith's revert_task_state fixes that | 14:55 |
| *** dhellmann has joined #openstack-dev | 14:55 | |
| markmc | eglynn_, I don't get the background on the silent failure, but I suspect it may be intentional | 14:56 |
| markmc | eglynn_, may be, totally unclear to me | 14:56 |
| *** gargya has quit IRC | 14:56 | |
| markmc | eglynn_, it's too obviously silent for it to be unintentional | 14:56 |
| markmc | eglynn_, idea might that users should know their VM is locked? seems weird | 14:56 |
| eglynn_ | makrmc: hmmm, it being intentional doesn't make sense to me ... kinda like "sleath-locking" | 14:57 |
| *** andrewbogott has joined #openstack-dev | 14:57 | |
| *** andrewbogott has joined #openstack-dev | 14:57 | |
| eglynn_ | markmc ^^^ | 14:57 |
| markmc | eglynn_, but wrt to your patch, it should just now become turning the silent not-implemented failure in set_admin_password to returning an error | 14:57 |
| markmc | eglynn_, sure, doesn't make much sense to me either - addressing that is a much wider thing, though | 14:58 |
| jeblair | vishy, jgriffith: sorry about that, this should fix it: https://review.openstack.org/10344 | 14:58 |
| markmc | eglynn_, not something that should be done piecemeal for a single action | 14:58 |
| eglynn_ | markmc: fair nuff, I'll file a bug to address the silent failure issue in a more general way | 14:58 |
| markmc | eglynn_, cool stuff | 14:59 |
| *** datsun180b has joined #openstack-dev | 14:59 | |
| markmc | cripes, looking back on pre-git history in git is painful | 15:00 |
| markmc | eglynn_, ok, this is the origin - https://code.launchpad.net/~tr3buchet/nova/lock/+merge/44874 | 15:01 |
| markmc | eglynn_, looks silent there too, at first glance | 15:02 |
| dansmith | markmc, eglynn_: I had the same realization a week ago, thinking "how can it possibly be right to ignore all these failures and locked states?" | 15:02 |
| dansmith | but with casting those actions from the api, I don't see how we can change that | 15:02 |
| dansmith | what I'd like to see, | 15:02 |
| GheRivero | 0 | 15:02 |
| dansmith | is a "get errors for instance" method, | 15:03 |
| markmc | eglynn_, dansmith, blueprint - https://blueprints.launchpad.net/nova/+spec/xs-lock | 15:03 |
| dansmith | that would return a sanitized event log or something for an instance, | 15:03 |
| dansmith | so that you had a possible way to tell why something isn't working | 15:03 |
| markmc | tr3buchet, can you shed some light on this? | 15:04 |
| markmc | tr3buchet, if an instance is locked and a user attempts to run some actions, they fail without the user getting an error | 15:04 |
| markmc | tr3buchet, was that intentional? | 15:04 |
| eglynn_ | markmc, dansmith: looking at the comments trail on that original locking branch, I'm not sure the generally-understood intent was to fail silently if locked (e.g. the comment on reporting a 405 status) | 15:05 |
| markmc | eglynn_, dansmith makes a good point - these are mostly async calls, failure happens after user has got e.g. 202 Accepted | 15:06 |
| dansmith | hmm, is that blueprint the same sort of lock? | 15:06 |
| dansmith | so if a user attempts to reboot a suspended instance, only an admin can unlock it? | 15:06 |
| markmc | dansmith, yeah, it's the blueprint associated with the code that added this | 15:07 |
| markmc | dansmith, see https://code.launchpad.net/~tr3buchet/nova/lock/+merge/44874 | 15:07 |
| dansmith | yeah, looking, I just don't understand the business logic | 15:07 |
| tr3buchet | markmc: hmmmmm | 15:08 |
| tr3buchet | markmc: i wrote that so very long ago | 15:08 |
| tr3buchet | markmc: at the time the idea was for a service provider to be able to lock an instance so a user could not alter it | 15:09 |
| *** anderstj has joined #openstack-dev | 15:10 | |
| dansmith | right, _that_ I understand | 15:10 |
| *** anderstj has quit IRC | 15:10 | |
| eglynn_ | markmc: so if its a 202 Accepted, fire-and-forget-stylee, I would expect the RPC to generally be a cast not a call | 15:10 |
| eglynn_ | markmc: ... so there's no conduit to report a failure from checks_instance_lock anyway | 15:10 |
| dansmith | eglynn_: and that's true, right? | 15:10 |
| eglynn_ | markmc: but that's not known within checks_instance_lock | 15:10 |
| markmc | eglynn_, yeah, so e.g. reboot_instance() is a cast | 15:11 |
| eglynn_ | dansmith: yep, but for any RPC-call-based paths into checks_instance_lock, it would make sense to report right? | 15:11 |
| markmc | eglynn_, probably, yes | 15:12 |
| dansmith | eglynn_: report locked state? yes, from the user perspective, but not from the performance perspective, AIUI | 15:12 |
| markmc | eglynn_, in fact, definitely yes | 15:12 |
| dansmith | I got shot down for making reboot, start, and stop synchronous calls so that I could report the result | 15:12 |
| eglynn_ | markmc: cool | 15:12 |
| markmc | eglynn_, which would mean revert_task_state() shouldn't swallow the except but let it be sent to the remote side | 15:12 |
| *** koolhead17 has joined #openstack-dev | 15:12 | |
| markmc | eglynn_, no need to use booleans for this | 15:13 |
| markmc | eglynn_, (note dansmith's patch adds an exception) | 15:13 |
| eglynn_ | markmc: yeah, cool, I'm down with reporting via an exception instead of a boolean | 15:13 |
| * dansmith is confused | 15:14 | |
| eglynn_ | markmc: (as long as its not just silently ignored when it needn't/shouldn't be ...) | 15:14 |
| *** wiliam has quit IRC | 15:14 | |
| markmc | eglynn_, there's probably a case for making some of these casts calls too - how long running are some of them? | 15:14 |
| markmc | dansmith, for any of the calls we do, we should be able to report the locked error to the user | 15:14 |
| *** sstent has quit IRC | 15:15 | |
| markmc | dansmith, so we need the instance locked exception to be serialized and caught by the client side | 15:15 |
| eglynn_ | markmc: in the unsual case, many of them are pretty quick, not sure though if there are edges where they might block ... | 15:15 |
| dansmith | markmc: I totally agree, but that seems to be contrary to the intended performance aspect | 15:15 |
| *** sstent has joined #openstack-dev | 15:15 | |
| dansmith | perhaps we should ask russellb and ewindisch to chime in | 15:15 |
| dansmith | because of: https://review.openstack.org/#/c/9922/ | 15:15 |
| markmc | dansmith, I'm not saying we change all casts to calls | 15:15 |
| markmc | dansmith, initially, we just do this for the existing calls | 15:15 |
| dansmith | I'm not sure why we wouldn't just shoot for consistent behavior of "don't expect to hear back" for all of them, | 15:15 |
| dansmith | especially if there is a "get_event_log()" method to check on why something failed | 15:16 |
| eglynn_ | surely we just need to change selected actions, where its important to report a definite outcome ... | 15:16 |
| markmc | dansmith, if it's a sync call, there's no reason to discard the error which is useful to users | 15:16 |
| dansmith | eglynn_: are start/stop/reboot not in that set of "important" ? | 15:16 |
| dansmith | markmc: how does the user (of the API) know if it's a sync call and to expect a meaningful status? | 15:17 |
| eglynn_ | dansmith: its debatable, since we can poll the server state | 15:17 |
| markmc | dansmith, 200 OK vs 202 Accepted, I guess | 15:17 |
| dansmith | mkay | 15:17 |
| eglynn_ | dansmith: but that poll wouldn't tell for example if a passwd change actually happened | 15:17 |
| dansmith | okay, categorizing actions into things we can poll for and not makes sense, although it seems overly complicated to me | 15:18 |
| dansmith | but I still wonder how the user knows that the state hasn't changed because of a lock vs. some other error | 15:18 |
| eglynn_ | dansmith: good point, now they can only trawl the logs I guess | 15:19 |
| *** rkukura has joined #openstack-dev | 15:19 | |
| *** rkukura has quit IRC | 15:19 | |
| dansmith | right, so it seems like this might be an opportunity to improve things for all those scenarios | 15:20 |
| *** rkukura has joined #openstack-dev | 15:20 | |
| eglynn_ | dansmith: yep, I hear ya ... so say we do change to call accross the board (from cast) | 15:20 |
| *** rnirmal has joined #openstack-dev | 15:21 | |
| dansmith | in the CIM world, I think it's rather common to generate and return a request ID, throw that back to the user, | 15:21 |
| markmc | dansmith, eglynn_, tried to summarise in https://review.openstack.org/#/c/10130/ | 15:21 |
| dansmith | and then let them check on the status of that request later | 15:21 |
| markmc | dansmith, eglynn_, 4 separate issues | 15:21 |
| * eglynn_ looking ... | 15:21 | |
| dansmith | am I making trouble again? :) | 15:21 |
| markmc | no, it's all good stuff | 15:21 |
| ewindisch | dansmith: right, an async model. | 15:22 |
| markmc | serious issues which really do need addressing | 15:22 |
| dansmith | markmc: since we're setting and reverting task_state in two places, would it be reasonable to check the lock in the api and avoid the attempt in the first place? | 15:22 |
| dansmith | that doesn't solve everything, but it would reduce some of the problem | 15:22 |
| markmc | dansmith, hmm, what two places? | 15:22 |
| ewindisch | it is also possible, at times, to use cast+callbacks in an actor model, rather than using call… | 15:22 |
| eglynn_ | dansmith: yep, that would simplify | 15:22 |
| *** cp16net is now known as cp16net|away | 15:22 | |
| dansmith | compute/api and compute/manager | 15:22 |
| markmc | dansmith, I don't think we'd revert task_state in compute/api | 15:23 |
| *** cp16net|away is now known as cp16net | 15:23 | |
| dansmith | no, | 15:23 |
| markmc | dansmith, at least on the road we're currently on | 15:23 |
| markmc | dansmith, ah, got you | 15:23 |
| ewindisch | Are you talking now about the revert-task-state patch, or the password patch? | 15:23 |
| markmc | dansmith, setting in api, unsetting in manager | 15:23 |
| dansmith | markmc: right now, we set task_state = STARTING in api, and then set task_state=None in manager when it's complete | 15:23 |
| dansmith | right | 15:23 |
| markmc | dansmith, doing the lock checking on the client side is worth thinking about, yeah | 15:23 |
| dansmith | markmc: client site == compute/api.py, right? | 15:24 |
| markmc | dansmith, there'll be backwards compat to think about, but we consider that later | 15:24 |
| markmc | dansmith, right, yes | 15:24 |
| dansmith | okay, I can cook something up for that if you want | 15:24 |
| *** alex88 has quit IRC | 15:24 | |
| markmc | dansmith, certainly sounds like it's worth trying | 15:25 |
| markmc | eglynn_, what do you think? | 15:25 |
| eglynn_ | dansimth: great that would be a good rationalization, the divsion of responsibility as currently stands is needlessly confusing I think | 15:25 |
| markmc | cool | 15:25 |
| eglynn_ | (division between api and manager that is ...) | 15:25 |
| dansmith | alrighty | 15:25 |
| dprince | pixelbeat: I just missed this one when I approved your qemu branch: Don't import libvirt_utils in disk api. https://review.openstack.org/10347 | 15:26 |
| markmc | dansmith, wait, wait | 15:27 |
| *** andrewbogott has quit IRC | 15:27 | |
| markmc | dansmith, that can't work for async | 15:27 |
| * markmc wonders is he getting confused | 15:27 | |
| markmc | dansmith, I think we talked this through already :) | 15:27 |
| markmc | dansmith, no, sorry - I did get confused | 15:28 |
| markmc | dansmith, carry on | 15:28 |
| dansmith | markmc: yeah, I'm talking about the preemtive check, not the reversion | 15:28 |
| markmc | dansmith, right | 15:28 |
| markmc | dansmith, I understood that and then forgot it 5 seconds later :) | 15:28 |
| dansmith | how is that possible? not like this is really complicated or anything | 15:28 |
| dansmith | wait, what's my name again? | 15:28 |
| dansmith | :) | 15:29 |
| markmc | screw it, let's all just become cloud analysts | 15:29 |
| pixelbeat | dprince, err OK. I'm wondering why I didn't get any test failures, nor were any flagged in the commit | 15:30 |
| markmc | dansmith, you still need the task_state reversion on the server side, right? | 15:30 |
| markmc | dansmith, for random other errors | 15:30 |
| dprince | pixelbeat: I hit this on XenServer only... | 15:30 |
| *** littleidea has joined #openstack-dev | 15:30 | |
| dansmith | markmc: yeah | 15:30 |
| markmc | dansmith, on the client side it would probably be a context manager for taking the lock | 15:30 |
| markmc | dansmith, and would need to pass a new "lock already acquired" flag in the RPC call for backwards compat | 15:31 |
| markmc | dansmith, i.e. old clients which aren't doing locking should still have locking done on the server side | 15:31 |
| dansmith | markmc: uhm, why? we don't need to acquire the lock before we do things, just check if it's locked so we can report to the user, right? | 15:31 |
| dansmith | this is an admin lock, not a thread lock, unless I'm very confused | 15:32 |
| markmc | dansmith, ok, so only set task_state after checking the lock - so no need for reversion | 15:32 |
| dansmith | er s/thread lock/mutex/ | 15:32 |
| dansmith | right | 15:32 |
| markmc | dansmith, ok, cool | 15:32 |
| dansmith | in fact, | 15:32 |
| eglynn_ | markmc, dansmith: its not much a lock aquistion, as a check-for-unlocked, no? | 15:32 |
| dansmith | just don't ever run the handler in the first place if locked | 15:32 |
| dansmith | eglynn_: yeah | 15:32 |
| eglynn_ | dansmith: how about just reversing the current scenario: check-for-unlocked in compute/api ... set task_state and revert in compute/manager? | 15:33 |
| dansmith | eglynn_: I don't think that's the point of task state, | 15:34 |
| eglynn_ | dansmith: no ignore that, you want to set the task state as the earliest possible point, right? | 15:34 |
| dansmith | I think the idea is to set and check that in api.py so that 100 calls to, say, start_instance(), transltates to only a single hit to manager.py | 15:34 |
| dansmith | right | 15:34 |
| eglynn_ | cool, gotcha | 15:34 |
| *** mnaser has joined #openstack-dev | 15:35 | |
| markmc | well now, that was fun | 15:36 |
| eglynn_ | ... and enlightening | 15:36 |
| annegentle | ttx: around? I have a root wrap question about "when" you configure nova to use it - are there pre-reqs like you must have run db sync? Or is root_helper added by packagers and deployers just need to know of its existence? | 15:37 |
| ewindisch | markmc: those links in your review (revert task state) are bad | 15:37 |
| *** spiffxp has joined #openstack-dev | 15:37 | |
| ttx | annegentle: no, should be deploy before you try to run nova. | 15:37 |
| annegentle | ttx: ok, makes sense | 15:37 |
| markmc | ewindisch, bah | 15:38 |
| annegentle | I'm mostly taking your wiki page and figuring out where to insert it | 15:38 |
| mnaser | anyone ever ran into the strange issue of nova-compute not refreshing/running any firewall rule updates? modifying security groups i see no nova.virt.firewall logs (firewall_driver=nova.virt.xenapi.firewall.Dom0IptablesFirewallDriver) | 15:38 |
| *** apevec is now known as apevec_ | 15:38 | |
| ttx | annegentle: no prerequisites | 15:38 |
| annegentle | ttx: well-done document, by the way | 15:38 |
| *** danwent has joined #openstack-dev | 15:38 | |
| mnaser | i do see the do_refresh_security_group_rules calls in the debug log but nothing for the nova.virt.firewall where it creates the sec group iptables rules | 15:39 |
| *** apevec_ is now known as apevec__ | 15:40 | |
| *** halfss has quit IRC | 15:40 | |
| *** apevec__ is now known as apevec | 15:40 | |
| jgriffith | eglynn_: Are you able to port your fix for https://bugs.launchpad.net/nova/+bug/1027281 to Cinder as well? | 15:40 |
| uvirtbot | Launchpad bug 1027281 in nova "In the nova-volume api a GET to /types returns extra db fields for json response" [Undecided,In progress] | 15:40 |
| jgriffith | eglynn_: Or should I grab it later this week? | 15:40 |
| jgriffith | err.. next week | 15:40 |
| * jgriffith can't believe it's Thursday | 15:40 | |
| eglynn_ | jgriffith: yep, I'll grab the cinder aspect too | 15:42 |
| jgriffith | eglynn_: thanks, I'll go ahead and assign it to you in Cinder as well then | 15:42 |
| *** andrewbogott has joined #openstack-dev | 15:43 | |
| *** andrewbogott has joined #openstack-dev | 15:43 | |
| sdague | has anyone been using stacktach for debugging? I was starting to look at it after russellb pointed it out, but my django-fu is weak enough that I'm not quite sure how I get it up on django. | 15:44 |
| *** jaypipes has joined #openstack-dev | 15:45 | |
| * russellb hasn't used it | 15:45 | |
| *** littleidea has quit IRC | 15:47 | |
| *** thovden has quit IRC | 15:48 | |
| *** littleidea has joined #openstack-dev | 15:48 | |
| *** rohitk has quit IRC | 15:52 | |
| jaypipes | russellb: lol, I'm starting to get all your "send a full instance via RPC" patches confused.. :) | 15:53 |
| markmc | they all rock, that's all we need to remember :) | 15:55 |
| markmc | says he who hasn't gotten around to reviewing any of them yet :( | 15:55 |
| *** matwood has joined #openstack-dev | 15:56 | |
| *** zul has quit IRC | 15:56 | |
| *** zul has joined #openstack-dev | 15:57 | |
| jaypipes | markmc: :) | 15:57 |
| *** matwood has quit IRC | 16:00 | |
| *** matwood has joined #openstack-dev | 16:00 | |
| vanchester | Hi. I'm using swift + keystone + s3. Is it possible to authenticate with access_key='tenant:user' like i.e. http://docs.openstack.org/diablo/openstack-object-storage/admin/content/configuring-openstack-object-storage-with-s3_api.html, cause it seems like it isn't? I look at the code, it splits access_key with ':'. Who's wrong? :) | 16:00 |
| *** zaitcev has joined #openstack-dev | 16:02 | |
| mtaylor | markmc, jaypipes, bcwaldon: here are the dependent lib changes that pip causes when devstack is run | 16:02 |
| mtaylor | markmc, jaypipes, bcwaldon: so, basically, this is the current defactor state of the global requirements list - the + lines are things that pip winds up installing after devstack attempts to install all of our needs via packages | 16:02 |
| mtaylor | so the + lines are the ones zul should care about, because they are the divergence from precise | 16:03 |
| jaypipes | you mean zuul or zul? | 16:03 |
| mtaylor | I mean zul | 16:03 |
| jaypipes | k | 16:03 |
| * jaypipes slaps jeblair for naming his project like chucky. | 16:03 | |
| *** jtran has joined #openstack-dev | 16:03 | |
| *** matwood has quit IRC | 16:04 | |
| mtaylor | because anything wiht a + is somehting our code needs that isn't installed via apt | 16:04 |
| *** SuperSonicSound has joined #openstack-dev | 16:04 | |
| jaypipes | k | 16:04 |
| markmc | mtaylor, nice special casing of ubuntu there again :) | 16:04 |
| mtaylor | zul: or, would it be helpful if I added a depends ppa to the host I'm doing this on? | 16:04 |
| mtaylor | markmc: well, I've got to start with something | 16:04 |
| markmc | mtaylor, sure | 16:04 |
| mtaylor | markmc: I'll do rh next | 16:04 |
| *** matwood has joined #openstack-dev | 16:05 | |
| zul | mtaylor: otp | 16:05 |
| markmc | mtaylor, point is just that there's wider interest in new deps than the ubuntu packagers | 16:05 |
| mtaylor | markmc: totally :) | 16:05 |
| markmc | mtaylor, but you know that, so I'm just being a dork :) | 16:05 |
| *** mokas has quit IRC | 16:05 | |
| *** thingee has joined #openstack-dev | 16:05 | |
| *** e1mer has quit IRC | 16:06 | |
| jaypipes | markmc: lol | 16:06 |
| *** garyk has quit IRC | 16:06 | |
| ameade | can someone rekick jenkins verify on this: https://review.openstack.org/#/c/10170/ | 16:06 |
| timello | hey jaypipes :)! Whenever you have a time! Please take a look into the new patch I submitted! I would appreciate to see your valuable comments! https://review.openstack.org/9666 Thanks! | 16:07 |
| mtaylor | ameade: you can | 16:07 |
| mtaylor | ameade: just leave a comment that says "recheck" | 16:07 |
| jaypipes | timello: on my long list of things today :) | 16:08 |
| jaypipes | timello: got a bunch of tempest/QA and Glance things ahead of you but I should be able to get there by end of day | 16:09 |
| timello | jaypipes: awesome! | 16:09 |
| timello | jaypipes: that would be great! Thank you! | 16:09 |
| zaitcev | chmouel: poke - not sure if you're on the list already - http://lists.openstack.org/pipermail/openstack-dev/2012-July/000216.html | 16:09 |
| jaypipes | np! :) | 16:09 |
| *** asakhnov has quit IRC | 16:10 | |
| *** galvint_ has quit IRC | 16:12 | |
| *** anotherjesse_zz is now known as anotherjesse | 16:13 | |
| mtaylor | markmc: https://review.openstack.org/10351 - I used that modification to devstack. on a clean machine, if you set TRACK_DEPENDS, you should get an /opt/stack/requires.diff file after running stack.sh | 16:16 |
| *** alex88 has joined #openstack-dev | 16:16 | |
| *** alex88 has joined #openstack-dev | 16:16 | |
| markmc | mtaylor, ah, cool | 16:17 |
| *** mnaser has quit IRC | 16:17 | |
| mtaylor | markmc: I figure if we're going to start down the path of a global reqs list, we should probably have some reports to let us know what's up :) | 16:17 |
| markmc | mtaylor, totally | 16:17 |
| markmc | mtaylor, nice work | 16:18 |
| mtaylor | markmc: thanks! | 16:18 |
| *** s0mik has joined #openstack-dev | 16:18 | |
| *** thovden has joined #openstack-dev | 16:18 | |
| mtaylor | markmc: (devstack doesn't fully properly run yet with the virtualenv enabled ... perhaps I should just have TRACK_DEPENDS do an exit after doing the final diff instead of trying to run...) | 16:18 |
| mtaylor | yeah, I think that's more sensible than trying to get everything to actually run from a venv :) | 16:19 |
| *** erikzaadi has quit IRC | 16:20 | |
| russellb | jaypipes: heh, not surprising, i've literally been copying/pasting the commit message and changing the method name. they're all the same change applied to a different method of the compute rpc api. | 16:23 |
| jeblair | jaypipes, zul: i'm going to call my next project "James Page" | 16:24 |
| jaypipes | jeblair: :) | 16:24 |
| *** troytoman-away is now known as troytoman | 16:25 | |
| mtaylor | jeblair: don't you mean JaamesPaage? | 16:28 |
| *** Mandell has joined #openstack-dev | 16:29 | |
| jeblair | mtaylor: you named a package without _any_ vowels! | 16:29 |
| mtaylor | jeblair: that's because you took them all! | 16:29 |
| jaypipes | bcwaldon, ameade: around? | 16:31 |
| *** andrewbogott has quit IRC | 16:31 | |
| jaypipes | want to chat about https://review.openstack.org/#/c/10160/4 | 16:32 |
| *** andrewbogott has joined #openstack-dev | 16:32 | |
| jaypipes | ameade, bcwaldon: so I am uneasy about the solution in that patch | 16:32 |
| jaypipes | ameade, bcwaldon: I think I would prefer instead to have the metadata and members still be marked deleted, but have the call to delete_image() return the metadata and member info that was marked deleted so that notifications can have that information | 16:33 |
| *** anderstj has joined #openstack-dev | 16:34 | |
| *** anderstj_ has joined #openstack-dev | 16:38 | |
| *** thovden has quit IRC | 16:38 | |
| *** anotherjesse is now known as anotherjesse_zz | 16:40 | |
| *** anderstj has quit IRC | 16:41 | |
| *** koolhead17 has quit IRC | 16:42 | |
| *** ewindisch_ has joined #openstack-dev | 16:48 | |
| *** ewindisch has quit IRC | 16:50 | |
| *** ewindisch_ is now known as ewindisch | 16:50 | |
| mtaylor | jeblair, bcwaldon: https://review.openstack.org/10354 | 16:51 |
| dansmith | markmc, eglynn_: just pushed it to gerrit and added you two | 16:51 |
| *** rohitk has joined #openstack-dev | 16:51 | |
| * eglynn_ looking ... | 16:51 | |
| dansmith | one of the reasons I started down the path in the first place was bug 872541, which wanted positive nak when instances are locked | 16:52 |
| uvirtbot | Launchpad bug 872541 in openstack-qa "nova.compute.api don't raise exception when an instance is locked" [High,In progress] https://launchpad.net/bugs/872541 | 16:52 |
| dansmith | I never touched that after the discussion with russellb and ewindisch about async casts, but this will resolve it | 16:52 |
| bcwaldon | jaypipes: hey hey | 16:53 |
| bcwaldon | mtaylor: ! | 16:53 |
| mtaylor | bcwaldon: ! | 16:53 |
| jeblair | uvirtbot: ! | 16:53 |
| uvirtbot | jeblair: Error: "!" is not a valid command. | 16:53 |
| *** anotherjesse_zz is now known as anotherjesse | 16:53 | |
| *** anotherjesse is now known as anotherjesse_zz | 16:54 | |
| ewindisch | dansmith, what is "this" ? | 16:55 |
| *** koolhead17 has joined #openstack-dev | 16:56 | |
| dansmith | ewindisch: https://review.openstack.org/#/c/10355/ | 16:56 |
| *** markmcclain has quit IRC | 16:58 | |
| ewindisch | dansmith: this replaces the revert- change too? | 16:59 |
| dansmith | ewindisch: no, just cuts out about 75% of the cases where you might need state reversion | 17:00 |
| dansmith | (note that figure is completely fabricated) | 17:00 |
| *** jdurgin has joined #openstack-dev | 17:02 | |
| *** dachary has joined #openstack-dev | 17:02 | |
| *** anniec has quit IRC | 17:03 | |
| bcwaldon | mtaylor: can you just do a git revert of the recent commit I put in to relax the req from 0.6 to 0.6.x? | 17:04 |
| mtaylor | bcwaldon: sure | 17:04 |
| *** cp16net is now known as cp16net|away | 17:05 | |
| mtaylor | bcwaldon: or, should we relax the depend on novaclient too? | 17:05 |
| bcwaldon | mtaylor: I would prefer that | 17:06 |
| mtaylor | bcwaldon: let's do that - the real problem is merely that their requirements conflict | 17:06 |
| *** sulochan has quit IRC | 17:06 | |
| bcwaldon | mtaylor: kk | 17:06 |
| bcwaldon | mtaylor: go for it | 17:06 |
| mtaylor | bcwaldon: awesome | 17:06 |
| *** sulochan has joined #openstack-dev | 17:06 | |
| kbringard | I have a quick question about the nova leases file. in Diablo the leases were written out as MAC,server-name,IP | 17:06 |
| *** koolhead17 has quit IRC | 17:06 | |
| *** jcooley has quit IRC | 17:07 | |
| kbringard | in essex it looks like they're being set as MAC,host-$fixed_ip,IP | 17:07 |
| *** mnewby has joined #openstack-dev | 17:07 | |
| kbringard | so like, 00:00:00,host-10.0.0.1,10.0.0.1 | 17:07 |
| kbringard | is that expected, or is something else going on in my env? | 17:07 |
| mtaylor | bcwaldon: https://review.openstack.org/10358 | 17:08 |
| *** wiliam has joined #openstack-dev | 17:08 | |
| *** issackel_ has joined #openstack-dev | 17:09 | |
| *** vanchester has quit IRC | 17:10 | |
| *** sulochan has quit IRC | 17:10 | |
| *** danwent has quit IRC | 17:12 | |
| *** reed has joined #openstack-dev | 17:13 | |
| *** dhellmann has quit IRC | 17:13 | |
| *** danwent has joined #openstack-dev | 17:15 | |
| *** derekh has quit IRC | 17:15 | |
| *** Aaton_off is now known as Aaton | 17:18 | |
| *** apevec has quit IRC | 17:19 | |
| *** koolhead17 has joined #openstack-dev | 17:19 | |
| *** rkukura has quit IRC | 17:20 | |
| ameade | jaypipes: just getting all of the deleted metadata will also return any metadata that was deleted seperate from deleting the image | 17:20 |
| *** Ryan_Lane has joined #openstack-dev | 17:20 | |
| *** jakedahn_zz is now known as jakedahn | 17:21 | |
| ameade | jaypipes: so if metadata was added then changed, the notification could have duplicate metadata with different values | 17:22 |
| ameade | jaypipes: at least this was the case in nova | 17:22 |
| ameade | for instances | 17:22 |
| jaypipes | ameade: one sec, finishing up QA meeting.. | 17:23 |
| *** mokas has joined #openstack-dev | 17:24 | |
| *** lorin1 has quit IRC | 17:28 | |
| *** jog0 has joined #openstack-dev | 17:30 | |
| *** ayoung has quit IRC | 17:31 | |
| *** ayoung has joined #openstack-dev | 17:34 | |
| dprince | jkoelker: around? | 17:36 |
| *** s0mik has left #openstack-dev | 17:37 | |
| *** garyk has joined #openstack-dev | 17:38 | |
| dprince | jkoelker: I'm suspicous of this code in openstack-common: https://github.com/openstack/openstack-common/blob/master/openstack/common/log.py#L360 | 17:38 |
| dprince | jkoelker: My first complaint is this seems like a Nova specific thing to do... (based on the comment) | 17:38 |
| dprince | jkoelker: second. This code seems to be the cause of some rather bad side effects like... nova-manage not telling you why it fails. | 17:39 |
| dprince | jkoelker: When most of the nova-manage commands fail they actually say they log to the log file... however they don't. | 17:40 |
| dprince | Same could happen for services too. | 17:40 |
| *** gyee has joined #openstack-dev | 17:45 | |
| *** johnpostlethwait has joined #openstack-dev | 17:45 | |
| jaypipes | ameade, bcwaldon: ok, done with qa now | 17:46 |
| *** danpb has quit IRC | 17:46 | |
| bcwaldon | jaypipes: I'm almost there | 17:47 |
| jaypipes | k | 17:47 |
| *** nunosantos_ has quit IRC | 17:47 | |
| *** alex88 has quit IRC | 17:47 | |
| *** andrewbogott has quit IRC | 17:47 | |
| *** nunosantos_ has joined #openstack-dev | 17:47 | |
| *** rohitk has quit IRC | 17:57 | |
| jkoelker | dprince: yea that was all blindly copied from nova, | 18:01 |
| jkoelker | i'm looking now to see how the manage commands use the logging | 18:01 |
| dprince | jkoelker: Would you concur that it can just go away then? | 18:02 |
| dprince | jkoelker: https://bugs.launchpad.net/nova/+bug/1029577 | 18:02 |
| uvirtbot | Launchpad bug 1029577 in nova "nova-manage isn't logging errors" [High,In progress] | 18:02 |
| jkoelker | i would love for it to go away | 18:02 |
| dprince | jkoelker: Removing that code seems to fix the issue :) | 18:02 |
| dprince | jkoelker: Also. In setup we already clear out the root logger anyway at the top of the function. | 18:02 |
| *** jdurgin has quit IRC | 18:03 | |
| *** lorin1 has joined #openstack-dev | 18:03 | |
| dprince | jkoelker: The difference seems to be that one uses 'product_name' where as the stuff at the bottom of setup seems to use the default 'unknown' logger. | 18:03 |
| jkoelker | dprince: part of the issue is that nova-manage sets up it own logging | 18:03 |
| jkoelker | yea | 18:03 |
| jkoelker | i'm still looking at it, been a while since i looked at all the logging stuffs | 18:04 |
| *** rohitk has joined #openstack-dev | 18:04 | |
| *** mdomsch_ is now known as mdomsch | 18:07 | |
| *** dhellmann has joined #openstack-dev | 18:09 | |
| jkoelker | dprince: ah, I remeber this now | 18:09 |
| jkoelker | yea so this was for backwards compatability with the way logging used to be setup before we allowed real file config | 18:09 |
| dprince | jkoelker: meaning it can go away? | 18:10 |
| jkoelker | well the question is when its removed do you get multiple log statements? | 18:10 |
| jkoelker | due to the basicConfig in nova/__init__.py still sending messages? | 18:11 |
| zul | mtaylor: ping what were you talking about earlier? | 18:11 |
| dprince | jkoelker: I'll double check... | 18:11 |
| *** rkukura has joined #openstack-dev | 18:11 | |
| *** cp16net|away is now known as cp16net | 18:12 | |
| *** PotHix has joined #openstack-dev | 18:13 | |
| mtaylor | zul: deps stuff | 18:14 |
| *** markmcclain has joined #openstack-dev | 18:14 | |
| zul | mtaylor: eh? | 18:14 |
| *** markmcclain has quit IRC | 18:14 | |
| *** markmcclain has joined #openstack-dev | 18:14 | |
| *** anniec has joined #openstack-dev | 18:14 | |
| *** alop has joined #openstack-dev | 18:14 | |
| mtaylor | zul: http://paste.openstack.org/show/19632 | 18:15 |
| mtaylor | zul: i've been poking at global depends - as part of that, I wanted to see what pip was installing when we ran devstack | 18:15 |
| mtaylor | zul: since we theoretically install all the deps from apt during our devstack gating runs | 18:16 |
| jkoelker | dprice specifically to nova-manage I'm thinking what needs to happen is to grab a logger then in the excepts down at the bottom call LOG.exception so the exception gets logged | 18:16 |
| *** markmcclain has quit IRC | 18:16 | |
| mtaylor | zul: the diff I just posted is a diff of what's on the system before I let pip do anyhting, and after all of the projects get python setup.py develop'd | 18:16 |
| mtaylor | zul: fwiw | 18:16 |
| *** markmcclain has joined #openstack-dev | 18:16 | |
| jkoelker | otherwise re the bug it never hits the nova module code so no log messages are generated | 18:16 |
| zul | mtaylor: whats cmd2? | 18:17 |
| mtaylor | zul: no clue | 18:17 |
| mtaylor | looks like there is a deb package for it already... | 18:17 |
| *** jdurgin has joined #openstack-dev | 18:20 | |
| *** rohitk has left #openstack-dev | 18:21 | |
| clarkb | mtaylor: A + is something added by pip that wasn't installed by apt? | 18:21 |
| Ryan_Lane | ayoung: so, I was able to get keystone with the ldap backend working | 18:22 |
| Ryan_Lane | but there's a bug with regards to roles | 18:22 |
| Ryan_Lane | the fix is actually sitting in a todo, with the code commented out | 18:23 |
| *** ewindisch has quit IRC | 18:24 | |
| *** ewindisch has joined #openstack-dev | 18:24 | |
| mtaylor | clarkb: yeah | 18:25 |
| * adam_g really wishes for a hard freeze on new dependencies at some point in OS release cycles | 18:26 | |
| ayoung | Ryan_Lane, reference? | 18:29 |
| Ryan_Lane | gimme a sec | 18:30 |
| Ryan_Lane | ayoung: https://github.com/openstack/keystone/blob/stable/essex/keystone/identity/backends/ldap/core.py line 97 | 18:31 |
| Ryan_Lane | though it needs to be: metadata_ref = {'roles': self.get_metadata(user_id, tenant_id)} | 18:31 |
| ayoung | Ryan_Lane, let me try it out | 18:32 |
| * Ryan_Lane nods | 18:32 | |
| Ryan_Lane | if I run "nova credentials" without that change, I get an empty list of roles | 18:32 |
| Ryan_Lane | with the change, I get: [{u'id': u'sysadmin', u'name': u'sysadmin'}, {u'id': u'netadmin', u'name': u'netadmin'}] | 18:33 |
| *** darraghb has quit IRC | 18:33 | |
| Ryan_Lane | all this requires that I have roles defined in the roles ou that match roles set in the tenants | 18:33 |
| *** nunosantos_ has quit IRC | 18:34 | |
| ayoung | Ryan_Lane, hmm...that is on essex stable? | 18:34 |
| Ryan_Lane | yes | 18:34 |
| ayoung | significantly different in Folsom | 18:34 |
| Ryan_Lane | indeed | 18:34 |
| Ryan_Lane | but... | 18:34 |
| Ryan_Lane | I need this to work in production | 18:34 |
| ayoung | but it looks like the Folsom change is wrong | 18:34 |
| *** nunosantos has quit IRC | 18:36 | |
| *** armaan has joined #openstack-dev | 18:36 | |
| ayoung | Ryan_Lane, I think you submit that as a patch against stable | 18:38 |
| Ryan_Lane | ok. will do | 18:39 |
| Ryan_Lane | ayoung: https://review.openstack.org/10365 | 18:44 |
| Ryan_Lane | otherwise it's working, with the changes to hardcoded attributes | 18:44 |
| ayoung | Ryan_Lane, ah..lets address that too | 18:45 |
| ayoung | the change to fix that is .... | 18:45 |
| Ryan_Lane | it's fairly hard to know if it's working based on the documentation, since most of the functions in the documentation are unimplemented | 18:45 |
| ayoung | BTW, did you run that against the unit tests? | 18:45 |
| ayoung | I think it will fail | 18:45 |
| Ryan_Lane | oh | 18:45 |
| Ryan_Lane | :( | 18:45 |
| Ryan_Lane | I didn't | 18:45 |
| ayoung | I had to make it more like this: | 18:45 |
| ayoung | http://fpaste.org/LK0W/ | 18:46 |
| Ryan_Lane | ah | 18:46 |
| ayoung | not saying that is more correct, mind you, but that is using the same fixture as the other identity tests | 18:46 |
| ayoung | so only set roles if you have some.... | 18:47 |
| ayoung | for the username change... | 18:47 |
| ayoung | https://bugs.launchpad.net/keystone/+bug/997700/+attachment/3152375/+files/keystone-admiyo-0005-configuration-option-for-user-name.patch | 18:48 |
| uvirtbot | Launchpad bug 997700 in keystone "LDAP should not check username on "sn" field" [Medium,Triaged] | 18:48 |
| Ryan_Lane | is that diff against master? | 18:48 |
| ayoung | probably | 18:48 |
| ayoung | the diff in the fpaste, yes | 18:48 |
| * Ryan_Lane nods | 18:48 | |
| ayoung | Ryan_Lane, did you have to modify any fields besides user_name? | 18:49 |
| Ryan_Lane | I'll have to look. | 18:49 |
| Ryan_Lane | the tenant one as well | 18:50 |
| Ryan_Lane | name is defined as ou | 18:50 |
| ayoung | OK, I'll spin up a patch for that as well. | 18:51 |
| *** jakedahn is now known as jakedahn_zz | 18:51 | |
| *** johnpostlethwait has quit IRC | 18:53 | |
| Ryan_Lane | awesome. thanks | 18:54 |
| *** jakedahn_zz is now known as jakedahn | 18:54 | |
| Ryan_Lane | hm. how to run tests against stable.... | 18:55 |
| Ryan_Lane | shouldn't the tests run from the change submission? | 18:56 |
| *** mnewby has quit IRC | 18:56 | |
| ayoung | Ryan_Lane, start with the unit tests | 18:56 |
| ayoung | ./run_tests.sh | 18:56 |
| ayoung | in keystone git checkout dir | 18:56 |
| Ryan_Lane | right, but I have no way of running them | 18:56 |
| Ryan_Lane | can devstack use the stable branch? | 18:57 |
| ayoung | Ryan_Lane, yes you do. | 18:58 |
| ayoung | Do you havea a git checkout? | 18:58 |
| Ryan_Lane | on my mac | 18:59 |
| Ryan_Lane | which is missing all of the dependencies | 18:59 |
| ayoung | I'll take that as a "No" | 18:59 |
| ayoung | Luddite | 18:59 |
| Ryan_Lane | well, I tend to use devstack for development | 18:59 |
| ayoung | OK... | 18:59 |
| ayoung | devstack will work fine | 18:59 |
| Ryan_Lane | I have a devstack instance, but it uses master | 18:59 |
| ayoung | OK, go to that machine | 19:00 |
| Ryan_Lane | and its dependencies are for master | 19:00 |
| ayoung | and the keystone checkout is in /opt/stack | 19:00 |
| ayoung | kill all the devstack stuff you have running | 19:00 |
| ayoung | and get an essex tracking branch: | 19:00 |
| ayoung | git branch --track essex gerrit/stable/essex | 19:00 |
| ayoung | git checkout essex | 19:00 |
| ayoung | apply your patch | 19:00 |
| ayoung | ./run_tests.sh | 19:01 |
| ayoung | the real fun is running the live ldap tests...but one step at a time | 19:01 |
| *** epim has joined #openstack-dev | 19:01 | |
| *** anniec has quit IRC | 19:04 | |
| *** anniec has joined #openstack-dev | 19:04 | |
| Ryan_Lane | meh. my vm isn't behaving. this may take a while. I'll get the tests going and let you know how it goes | 19:04 |
| *** EmilienM has joined #openstack-dev | 19:06 | |
| *** EmilienM has left #openstack-dev | 19:07 | |
| *** EmilienM has joined #openstack-dev | 19:08 | |
| *** andrewbogott has joined #openstack-dev | 19:08 | |
| *** andrewbogott has joined #openstack-dev | 19:08 | |
| *** fc__ has quit IRC | 19:10 | |
| *** novas0x2a|laptop has joined #openstack-dev | 19:14 | |
| sdague | jgriffith: any further thoughts on the storewise driver - https://review.openstack.org/#/c/9840/ ? | 19:15 |
| Ryan_Lane | ayoung: in devstack in essex, the stable branch doesn't pass tests | 19:15 |
| ayoung | right | 19:16 |
| ayoung | author? | 19:16 |
| Ryan_Lane | ? | 19:16 |
| ayoung | for me it is the authors file check that fails | 19:16 |
| ayoung | Ryan_Lane, nothing to do with LDAP | 19:16 |
| Ryan_Lane | I'm talking about without my patch | 19:16 |
| ayoung | yes | 19:16 |
| ayoung | so to run just the LDAP unit tests... | 19:16 |
| Ryan_Lane | I have 18 failures | 19:17 |
| *** johnpostlethwait has joined #openstack-dev | 19:17 | |
| ayoung | ./run_tests.sh test_backend_ldap | 19:17 |
| ayoung | Ryan_Lane, right now we only care about the ldap backend tests | 19:18 |
| ayoung | Ryan_Lane, so run ./run_tests.sh test_backend_ldap | 19:18 |
| jgriffith | sdague: I'm good with it, just hoping to get some other folks to at least look at it | 19:18 |
| * Ryan_Lane nods | 19:19 | |
| Ryan_Lane | they're running | 19:19 |
| jgriffith | sdague: I'll see if I can get someone to look at it today, otherwise I'll push it through tomorrow for you | 19:19 |
| *** mnewby has joined #openstack-dev | 19:21 | |
| *** dachary has quit IRC | 19:21 | |
| *** dachary has joined #openstack-dev | 19:22 | |
| Ryan_Lane | ugh. pip failure | 19:22 |
| *** mokas has quit IRC | 19:23 | |
| *** andrewbogott_ has joined #openstack-dev | 19:26 | |
| *** andrewbogott_ has joined #openstack-dev | 19:26 | |
| *** andrewbogott has quit IRC | 19:26 | |
| *** andrewbogott_ is now known as andrewbogott | 19:26 | |
| ayoung | Ryan_Lane, anyway, let me test out your change on my VM | 19:27 |
| sdague | jgriffith: cool, don't want to push anything outside of the process, but some extra reviews would be great. Just don't want to see it lost in review limbo. | 19:27 |
| *** vanchester has joined #openstack-dev | 19:28 | |
| jgriffith | sdague: Promise I won't let it get lost. :) | 19:28 |
| jgriffith | sdague: F3 is my deadline for all of the drivers, then I bribe people or just approve them :) | 19:29 |
| sdague | :) | 19:29 |
| jgriffith | sdague: There ya have it... | 19:29 |
| sdague | especially if nova-volume is still going to exist in folsom, as the team wants to get it into nova if that's still the case | 19:29 |
| sdague | but I told them it should land in cinder first | 19:30 |
| jgriffith | sdague: Just FYI there's nothing preventing you from doing them in paralell | 19:30 |
| sdague | jgriffith: ok, I'll let them know | 19:30 |
| ayoung | Ryan_Lane, I get failures on test_authenticate and test_authenticate_no_metadata | 19:30 |
| ayoung | I'm going -1 the patch as is. Please submit one with the modification we discussed | 19:31 |
| *** pixelbeat has quit IRC | 19:31 | |
| Ryan_Lane | I already have | 19:31 |
| Ryan_Lane | patchset 2 | 19:31 |
| *** thovden has joined #openstack-dev | 19:31 | |
| Ryan_Lane | it's also failing tests, though | 19:31 |
| Ryan_Lane | ah. pep8 failure | 19:32 |
| *** koolhead17 has quit IRC | 19:32 | |
| Ryan_Lane | and https://jenkins.openstack.org/job/gate-keystone-python26/648/testReport/test_backend_ldap/LDAPIdentity/test_authenticate_invalid_tenant/Â | 19:33 |
| *** dtroyer is now known as dtroyer_zzz | 19:34 | |
| jeblair | vishy: okay, i have the no-cinder-on-stable-branch problem straightened out; this change passes now (but needs code review): https://review.openstack.org/#/c/10321/ | 19:35 |
| jeblair | vishy, jgriffith, adam_g: ^ and sorry about the trouble. | 19:35 |
| Ryan_Lane | ah. that exception doesn't exist in this version. I can just remove that except | 19:35 |
| *** wiliam has quit IRC | 19:36 | |
| Ryan_Lane | this patchset should pass | 19:36 |
| *** arosen has quit IRC | 19:36 | |
| *** arosen has joined #openstack-dev | 19:37 | |
| *** armaan has left #openstack-dev | 19:39 | |
| ayoung | Ryan_Lane, can you give a first pass to https://review.openstack.org/#/c/10370/ as well | 19:45 |
| Ryan_Lane | ayoung: this changes the default that's currently in stable | 19:46 |
| Ryan_Lane | for tenant | 19:46 |
| ayoung | Ryan_Lane, your change looks good. | 19:48 |
| Ryan_Lane | cool | 19:48 |
| ayoung | at least, against the fake ldap back end | 19:48 |
| * Ryan_Lane nods | 19:49 | |
| ayoung | I need to do a little hack to run against live ldap...1 sec | 19:49 |
| *** datsun180b_ has joined #openstack-dev | 19:50 | |
| *** mokas has joined #openstack-dev | 19:51 | |
| ayoung | Ryan_Lane, self.get_metadata(user_id, tenant_id) throws a different exception, I think | 19:52 |
| *** cp16net is now known as cp16net|away | 19:52 | |
| *** datsun180b has quit IRC | 19:53 | |
| *** datsun180b_ is now known as datsun180b | 19:53 | |
| ayoung | Ryan_Lane, I think you need an Except: pass for that | 19:53 |
| Ryan_Lane | that exception doesn't exist | 19:53 |
| Ryan_Lane | in this version | 19:53 |
| Ryan_Lane | let me look at the function | 19:53 |
| Ryan_Lane | no exception thrown | 19:54 |
| Ryan_Lane | it just returns an empty dictionary | 19:54 |
| ayoung | er..yeah...loking through, looks like I swallowed all the exceptions. THat must be why I was putting on weight | 19:54 |
| Ryan_Lane | :D | 19:54 |
| *** datsun180b has quit IRC | 19:55 | |
| ayoung | Ryan_Lane, I do think you duplicated the metadata_ref = {} call | 19:56 |
| ayoung | its on line 93 and 83 | 19:56 |
| Ryan_Lane | ah. I did indeed | 19:56 |
| ayoung | that is what you get for copying my sloppiness | 19:56 |
| Ryan_Lane | there's a failed test too. let me check that | 19:56 |
| ayoung | You also don't need to put it in the Excepiont block | 19:56 |
| ayoung | Ryan_Lane, on test_rename_duplicate_role_name_fails I think that is a pre-existing condition | 19:57 |
| Ryan_Lane | Excepiont? | 19:57 |
| ayoung | the rest of the live ldap tests pass for me | 19:57 |
| *** n0ano has joined #openstack-dev | 19:57 | |
| Ryan_Lane | ah. the extra metadata_ref = {} | 19:57 |
| *** rnirmal has quit IRC | 19:57 | |
| ayoung | except exception.TenantNotFound: | 19:58 |
| ayoung | tenant_ref = None | 19:58 |
| ayoung | metadata_ref = {} | 19:58 |
| Ryan_Lane | I need the block itself, right? | 19:58 |
| ayoung | You really, really want to set that to {} don't you? | 19:58 |
| Ryan_Lane | heh | 19:58 |
| Ryan_Lane | I like to set the same thing over and over :D | 19:58 |
| ayoung | yeah, just remove all but one of the default assignements | 19:58 |
| * Ryan_Lane nods | 19:58 | |
| *** thovden_ has joined #openstack-dev | 20:01 | |
| *** thovden has quit IRC | 20:03 | |
| *** thovden_ is now known as thovden | 20:03 | |
| *** nikhil has quit IRC | 20:05 | |
| ayoung | Ryan_Lane, +1ed | 20:06 |
| *** ecarlin has joined #openstack-dev | 20:07 | |
| ayoung | Ryan_Lane, added a few review requests to the other core devs | 20:07 |
| Ryan_Lane | sweet. thanks | 20:10 |
| *** ecarlin has quit IRC | 20:10 | |
| andrewbogott | dhellmann, markmc, russellb, could I get a +2 from one of you on https://review.openstack.org/#/c/10035/? It's been vetted by _cerberus_ who is the expert in that area. | 20:10 |
| ayoung | Ryan_Lane, new version of mine pushed. | 20:11 |
| ayoung | posted, rather | 20:11 |
| Ryan_Lane | +1'd | 20:11 |
| *** anniec has quit IRC | 20:12 | |
| _cerberus_ | expert, eh? O_o | 20:13 |
| andrewbogott | Didn't you write the notification code in the first place? | 20:13 |
| _cerberus_ | yeah ;-) | 20:13 |
| _cerberus_ | Just refuting this supposed "expert" status | 20:14 |
| andrewbogott | Serves you right for adding a new feature, now your the expert for life. | 20:14 |
| andrewbogott | you're | 20:14 |
| _cerberus_ | Don't I know it :-/ | 20:14 |
| _cerberus_ | :-D | 20:14 |
| *** cp16net|away is now known as cp16net | 20:14 | |
| *** johnpostlethwait has quit IRC | 20:15 | |
| clarkb | devcamca-: what do you think about adding https://jenkins.openstack.org/job/gate-horizon-selenium/ as a gate test to horizon? It has been running as a check test and I think we have ironed out the kinks | 20:16 |
| *** dprince has quit IRC | 20:20 | |
| *** andrewsmedina has quit IRC | 20:22 | |
| *** nikhil has joined #openstack-dev | 20:23 | |
| *** mnewby has quit IRC | 20:26 | |
| ayoung | Ryan_Lane, what are your LDAP objectclasses again? | 20:29 |
| ayoung | posix user ? | 20:29 |
| Ryan_Lane | inetorgperson, posixuser, person | 20:29 |
| Ryan_Lane | tenants are groupofnames and posixgroup | 20:29 |
| ayoung | OK... let me test that out with the liveldap tests and a modified schema. | 20:30 |
| Ryan_Lane | actually: person, inetorgperson, organizationalPerson, ldapPublicKey, shadowaccount, posixaccount | 20:30 |
| Ryan_Lane | heh | 20:30 |
| Ryan_Lane | I use uid as naming attribute, and CN for a user's full name | 20:31 |
| ayoung | I think I only need to name the end of the hierarchy | 20:32 |
| ayoung | so | 20:32 |
| ayoung | user_objectclass=posixaccount | 20:32 |
| ayoung | user_name_attribute=cn? | 20:33 |
| *** anderstj_ has quit IRC | 20:33 | |
| Ryan_Lane | I'm using inetorgperson and uid | 20:33 |
| Ryan_Lane | and uid for user_id_attribute | 20:33 |
| ayoung | user_objectclass=inetorgperson | 20:34 |
| ayoung | user_name_attribute=cn | 20:34 |
| ayoung | user_id_attribute=uid | 20:34 |
| ayoung | then for tenant | 20:34 |
| Ryan_Lane | apparently if you use cn for the name attribute, it will do searches with the username for both uid and cn | 20:34 |
| Ryan_Lane | I'm using uid for both name and id | 20:34 |
| Ryan_Lane | for tenant I'm using cn for both name and id | 20:35 |
| *** thovden_ has joined #openstack-dev | 20:35 | |
| ayoung | Hmm.ok | 20:35 |
| Ryan_Lane | for similar reasons | 20:35 |
| *** anniec has joined #openstack-dev | 20:35 | |
| Ryan_Lane | my tenants and users are all unique, so it's not a problem | 20:35 |
| ayoung | Ryan_Lane, http://fpaste.org/D8MG/ | 20:35 |
| ayoung | but with spaces, of course | 20:36 |
| Ryan_Lane | I'm using groupofnames for the tenant objectclass | 20:36 |
| ayoung | I thought you said posixgroup? | 20:36 |
| Ryan_Lane | both | 20:36 |
| Ryan_Lane | so I keep the default in the config | 20:36 |
| ayoung | when that adds a new one, it will only add it with the specified objectclass | 20:36 |
| Ryan_Lane | I'm using posixgroup as well, so that I can extend the groups to the instances | 20:37 |
| Ryan_Lane | I don't use keystone to manage the dit | 20:37 |
| *** thovden has quit IRC | 20:37 | |
| *** thovden_ is now known as thovden | 20:37 | |
| Ryan_Lane | keystone is basically read-only | 20:37 |
| ayoung | right...let me try it this way and see. | 20:37 |
| Ryan_Lane | that said, keystone adds tenants incorrectly. I already tested that | 20:37 |
| Ryan_Lane | it adds the uuid as the naming attribute, then tries to look up the tenant by dn, using the name, not the uuid | 20:38 |
| ayoung | object class 'posixGroup' requires attribute 'gidNumber' | 20:38 |
| Ryan_Lane | yep | 20:38 |
| Ryan_Lane | I have to leave. back in like 30-45 | 20:38 |
| ayoung | cuz it uses a UUID | 20:38 |
| ayoung | NP | 20:38 |
| ayoung | 'inetOrgPerson' requires attribute 'sn'", 'desc': 'Object class violation' | 20:39 |
| ayoung | hmmmm | 20:39 |
| *** mnewby has joined #openstack-dev | 20:42 | |
| *** littleidea_ has joined #openstack-dev | 20:42 | |
| *** littleidea has quit IRC | 20:43 | |
| *** littleidea_ is now known as littleidea | 20:43 | |
| soren | vishy: It's rather manual. | 20:45 |
| soren | vishy: I've always made sure to adjust the ordering to account for people having swapped their turns, so that they don't have two review days really close to each other. | 20:48 |
| soren | vishy: Also, a particular subset of the nova-core group has refused to show their e-mail addresses on Launchpad. | 20:49 |
| soren | vishy: ...and subset of that subset has refused to give anything resembling a real name, making it even harder to figure out what the heck is going on and who's who and what's what. | 20:50 |
| soren | vishy: So I'm afraid this can't be automated. | 20:50 |
| soren | vishy: ..without building a separate system for it with separate tracking of these details, of course. There's certainly prior art in this project to building stuff like that. | 20:52 |
| *** anniec_ has joined #openstack-dev | 20:54 | |
| *** anniec has quit IRC | 20:55 | |
| *** anniec_ is now known as anniec | 20:55 | |
| *** jk0 has quit IRC | 20:59 | |
| *** lhindman has joined #openstack-dev | 20:59 | |
| *** GheRivero has quit IRC | 21:01 | |
| *** jk0 has joined #openstack-dev | 21:02 | |
| *** jk0 has joined #openstack-dev | 21:02 | |
| *** ChanServ sets mode: +v jk0 | 21:02 | |
| *** anniec has quit IRC | 21:03 | |
| *** mdomsch has quit IRC | 21:05 | |
| *** mnewby_ has joined #openstack-dev | 21:09 | |
| *** aryan has quit IRC | 21:09 | |
| *** aryan has joined #openstack-dev | 21:09 | |
| *** bhuvan_ has quit IRC | 21:09 | |
| *** mnewby has quit IRC | 21:12 | |
| *** mnewby_ is now known as mnewby | 21:12 | |
| *** markvoelker has quit IRC | 21:14 | |
| *** apevec has joined #openstack-dev | 21:15 | |
| *** SuperSonicSound has quit IRC | 21:19 | |
| apevec | ayoung, re. https://review.openstack.org/#/c/10359/ | 21:21 |
| apevec | "This was already merged." - on master yes | 21:21 |
| Ryan_Lane | ayoung: I'm back | 21:21 |
| apevec | this is for stable/essex | 21:21 |
| ayoung | apevec, ah | 21:21 |
| *** heckj has joined #openstack-dev | 21:21 | |
| apevec | https://bugs.launchpad.net/keystone/+bug/1020569/comments/4 | 21:22 |
| uvirtbot | Launchpad bug 1020569 in keystone "wsgi code ties all middleware consumers to Eventlet" [Medium,Fix committed] | 21:22 |
| ayoung | I'm not sure that this change needs to be backported.... | 21:22 |
| apevec | ayoung, see above comment | 21:22 |
| apevec | in LP | 21:22 |
| ayoung | looking | 21:22 |
| Ryan_Lane | ayoung: the approach I usually take for populating values is to duplicate one, if it isn't provided | 21:22 |
| Ryan_Lane | for required attributes | 21:22 |
| Ryan_Lane | sn is an annoying one. | 21:23 |
| Ryan_Lane | I usually just replicate cn | 21:23 |
| ayoung | Ryan_Lane, yeah, that works for testing. I have to think about what I would recommend for a live deployment. | 21:23 |
| *** heckj has quit IRC | 21:23 | |
| ayoung | Ryan_Lane, What we have works for read only just fine, I suspect | 21:23 |
| Ryan_Lane | yep. it does | 21:23 |
| Ryan_Lane | I've already tested that | 21:23 |
| ayoung | and then the question is, would anyone really want to use LDAP except as read only | 21:23 |
| Ryan_Lane | some people might | 21:24 |
| Ryan_Lane | especially for tenants and roles | 21:24 |
| ayoung | if so...do we need a way to expand the user_add etc APIs to handle the additional values | 21:24 |
| Ryan_Lane | I like a template approach for this kind of thing | 21:24 |
| Ryan_Lane | let me give an example | 21:24 |
| ayoung | apevec, I think I am going to nix applying my patch to essex. It is not technically necessary, unless you want to avoid the eventlet dependencies in running Keystone in Apache HTTPD. Even then, it will work. I'd say I'm ok with leaving the change for Folsom | 21:26 |
| *** lorin1 has quit IRC | 21:26 | |
| ayoung | unless you have some strong opinion why it needs to be in stable | 21:26 |
| ayoung | And I have to say, I am still suspect of monkey patching Thread. It smells like a mistake to me. | 21:27 |
| Ryan_Lane | ayoung: http://pastebin.com/FiDUUxhk | 21:27 |
| apevec | ayoung, my only justification is that the patch which prompted you to move monkeypatching is already in stable | 21:27 |
| apevec | and I do would like to try running essex in httpd | 21:28 |
| ayoung | apevec, yes, and it needs to be there to deal with an issue in memcached. My fix just makes it possible to work around for debuggin purposes | 21:28 |
| *** eglynn_ has quit IRC | 21:28 | |
| Ryan_Lane | then user creation would use the object classes and attributes in the template, and could also have the possibility of custom variables via extensions | 21:28 |
| *** steveb_ has joined #openstack-dev | 21:29 | |
| ayoung | Ryan_Lane, it makes sense, the question is should we default values, or provde a way to expand the APIs to let the end user supply them. I am more prone toward the latter | 21:29 |
| Ryan_Lane | well, the template as shown in the pastebin would allow users to supply them | 21:30 |
| Ryan_Lane | since the template would be user created | 21:30 |
| ayoung | Ryan_Lane, so we drop something like that into /etc/keystone for each of the entities? | 21:30 |
| Ryan_Lane | like instance templates are in nova | 21:30 |
| Ryan_Lane | ayoung: yeah | 21:30 |
| ayoung | Ryan_Lane, write it up as a blueprint. We can thrash it around at the summit | 21:31 |
| Ryan_Lane | I'd actually consider using keystone as the ldap interface if I could customize it enough :) | 21:31 |
| apevec | ayoung, ok, I'll abandon the change, after checking why tempest test failed | 21:31 |
| ayoung | apevec, do you needto test the released version of essex in HTTPD, or is it OK to hack it, just to see it work? | 21:31 |
| *** vanchester has quit IRC | 21:32 | |
| apevec | hack is fine | 21:32 |
| ayoung | apevec, then just remove the whole monkeypatch line | 21:33 |
| jaypipes | _cerberus_: lol "the increasingly large decorator storm". no shit! :) | 21:33 |
| _cerberus_ | ;-) | 21:33 |
| ayoung | apevec, I recently commited a change to master that gives you the rest of the files you need | 21:33 |
| apevec | yep, saw it, very good | 21:33 |
| ayoung | 9b31383c7d8665cf99003974d98df147471e61e1 | 21:34 |
| ayoung | apevec, looks like the tempest failure was unrelated | 21:34 |
| apevec | yeah, I did "recheck" to confirm | 21:35 |
| apevec | nova-volume failing | 21:35 |
| *** thovden_ has joined #openstack-dev | 21:36 | |
| *** anniec has joined #openstack-dev | 21:37 | |
| Ryan_Lane | ayoung: https://blueprints.launchpad.net/keystone/+spec/ldap-object-templates | 21:37 |
| ayoung | Ryan_Lane, looking | 21:37 |
| *** anniec_ has joined #openstack-dev | 21:37 | |
| *** mnewby has quit IRC | 21:38 | |
| dansmith | markmc: I'm sure you saw it, but the instance lock check thingy just needs one more little green check on it :D | 21:38 |
| ayoung | Ryan_Lane so how would we expose that to the end user? python-keystone client would have to be smart enough to query Keystone to know to ask for and allow those values | 21:39 |
| Ryan_Lane | oh. sorry. this would be on the server side config | 21:39 |
| Ryan_Lane | let me adjust | 21:39 |
| *** dachary has quit IRC | 21:39 | |
| *** thovden has quit IRC | 21:39 | |
| *** thovden_ is now known as thovden | 21:39 | |
| ayoung | Ryan_Lane, right, necessary but not sufficient | 21:39 |
| ayoung | we need to then let the CLI and WebUI (and other apps) know about the new mandatory fields | 21:40 |
| Ryan_Lane | ah. true | 21:40 |
| ayoung | In visual HTML, you would add them to a form. We don't do anything like that in JSON, thoough | 21:40 |
| ayoung | although we should. HATEOAS and all that | 21:40 |
| Ryan_Lane | the extensions would need to extend the api and cli | 21:41 |
| *** anniec has quit IRC | 21:41 | |
| *** anniec_ is now known as anniec | 21:41 | |
| Ryan_Lane | the plugin framework could be used to provide variables, and extend the api | 21:42 |
| ayoung | Ryan_Lane, I would like it to be something like this: if you do a GET on a certain resource, you get the Metadata needed to create a new instance of that resource. | 21:42 |
| Ryan_Lane | andrewbogott: ^^ correct? | 21:42 |
| ayoung | like if you do GET http://host/v3/users/new | 21:42 |
| Ryan_Lane | ah, that would provide back a set of required fields? | 21:43 |
| * andrewbogott catches up | 21:43 | |
| ayoung | if you set a Media type of HTML, it would give you a form, and the default is JSON | 21:43 |
| Ryan_Lane | hm. the only downside to that is that each implementation of keystone could require a different set of things | 21:43 |
| Ryan_Lane | and clients would have no clue based on discovery | 21:43 |
| ayoung | No, that is an upside | 21:43 |
| andrewbogott | Ryan_Lane: Yeah, a plugin can extend the commandline. It would most likely happen via python-openstackclient rather than python-keystoneclient. | 21:44 |
| ayoung | yes, they would | 21:44 |
| Ryan_Lane | wouldn't it be better if these were formalized as extensions, so that clients could do discovery? | 21:44 |
| ayoung | ah, you mean that the client would have no A-Priori knowledge | 21:44 |
| Ryan_Lane | and then they could claim compatibility with different extensions | 21:44 |
| ayoung | Ryan_Lane, not if you buy into REST | 21:44 |
| ayoung | I think that I like this approach slightly better | 21:45 |
| Ryan_Lane | I'm just thinking consistency with how the api is extended now :) | 21:45 |
| ayoung | its more like Progressive enhancement... | 21:45 |
| * Ryan_Lane nods | 21:46 | |
| Ryan_Lane | it's a good topic for the summit | 21:46 |
| ayoung | Anyway, we have your blueprint. I think we should discuss at summite | 21:46 |
| ayoung | Heh, you are ahead of me as usual | 21:46 |
| ayoung | Ryan_Lane, http://adam.younglogic.com/2012/03/hateoas-openstack-keystone/ | 21:47 |
| Ryan_Lane | ok. I've updated the blueprint some | 21:47 |
| ayoung | Ryan_Lane, looks good. I have to catch a bus shortly. I'll muse on this some more in the next few days, and try to work up a fuller spec | 21:49 |
| Ryan_Lane | ayoung: ah. I see | 21:49 |
| Ryan_Lane | based on discovery of attributes, the application would dynamically create forms | 21:49 |
| Ryan_Lane | great. thanks | 21:49 |
| *** darkhorse2go has quit IRC | 21:50 | |
| * andrewbogott spent a while thinking this was about 'Hate as a service' | 21:51 | |
| Ryan_Lane | :D | 21:51 |
| Ryan_Lane | I could use it as a proxy for every technology I hate | 21:52 |
| Ryan_Lane | which is like all of them | 21:52 |
| ayoung | andrewbogott, sorry it is Hate O as a service, so it is only useful for Hating O things. | 21:53 |
| Ryan_Lane | :( | 21:53 |
| ayoung | http://en.wikipedia.org/wiki/HATEOAS | 21:53 |
| ayoung | http://roy.gbiv.com/untangled/2008/rest-apis-must-be-hypertext-driven | 21:53 |
| andrewbogott | I read (and, I think, mostly understood) your blog post. Is there anything in openstack that currently does that? Returns a form in order to define a future interface? | 21:54 |
| ayoung | andrewbogott, I would say that Horizon does | 21:54 |
| ayoung | what I would like is for Horizon to be far more of a passthrough from the remote services | 21:55 |
| andrewbogott | Hm... | 21:55 |
| ayoung | andrewbogott, I have to leave in 2 minutes. I will send you a write up of what I would like it to be like in an ideal world | 21:55 |
| ayoung | fully acking that I am dreaming | 21:55 |
| *** jmeredit has joined #openstack-dev | 21:56 | |
| andrewbogott | ayoung: ok! I'm not very involved in horizon but would be interested in thinking about it as a lesson in HATEOAS. | 21:56 |
| *** ayoung has quit IRC | 21:57 | |
| *** sacharya has quit IRC | 22:03 | |
| *** johnpostlethwait has joined #openstack-dev | 22:04 | |
| *** apevec has quit IRC | 22:04 | |
| *** dhellmann has quit IRC | 22:05 | |
| *** kbringard has quit IRC | 22:10 | |
| Ryan_Lane | damn. no ayoung. found another ldap bug | 22:13 |
| Ryan_Lane | authenticating with no tenant never returns | 22:13 |
| *** Gordonz_ has quit IRC | 22:20 | |
| *** markmcclain has quit IRC | 22:25 | |
| reed | is this team existing? http://wiki.openstack.org/Governance/Proposed/OpenStack%20Security%20Group | 22:26 |
| clarkb | adam_g: fyi (also don't hate me) but horizon depends on django 1.4 now and precise has 1.3.1 | 22:28 |
| *** heckj has joined #openstack-dev | 22:28 | |
| jmeredit | searching for approval....https://review.openstack.org/#/c/10361/ | 22:30 |
| *** pixelbeat has joined #openstack-dev | 22:31 | |
| *** heckj has quit IRC | 22:33 | |
| *** anderstj has joined #openstack-dev | 22:35 | |
| *** nati_ueno has joined #openstack-dev | 22:37 | |
| *** e1mer has joined #openstack-dev | 22:37 | |
| *** anniec has quit IRC | 22:38 | |
| adam_g | clarkb: thanks | 22:41 |
| *** andrewsmedina has joined #openstack-dev | 22:44 | |
| *** salv-orlando has quit IRC | 22:45 | |
| *** rods_ has joined #openstack-dev | 22:47 | |
| *** rods has quit IRC | 22:50 | |
| *** rods has joined #openstack-dev | 22:51 | |
| *** rods_ has quit IRC | 22:51 | |
| *** rods_ has joined #openstack-dev | 22:51 | |
| *** mnewby has joined #openstack-dev | 22:51 | |
| *** rods has quit IRC | 22:52 | |
| *** rods has joined #openstack-dev | 22:52 | |
| *** rods has joined #openstack-dev | 22:53 | |
| *** rods_ has quit IRC | 22:54 | |
| *** rods has joined #openstack-dev | 22:54 | |
| *** rods_ has joined #openstack-dev | 22:54 | |
| *** thovden has quit IRC | 22:55 | |
| *** troytoman is now known as troytoman-away | 22:57 | |
| *** rods_ has quit IRC | 22:57 | |
| *** rods_ has joined #openstack-dev | 22:57 | |
| *** rods has quit IRC | 23:03 | |
| *** rods has joined #openstack-dev | 23:03 | |
| *** zaitcev has quit IRC | 23:03 | |
| *** rods_ has quit IRC | 23:03 | |
| *** rods_ has joined #openstack-dev | 23:04 | |
| *** rods_ has left #openstack-dev | 23:05 | |
| *** rods__ has joined #openstack-dev | 23:05 | |
| *** rods has quit IRC | 23:05 | |
| *** rods has joined #openstack-dev | 23:05 | |
| *** lhindman has quit IRC | 23:06 | |
| *** rods_ has joined #openstack-dev | 23:08 | |
| *** rods has quit IRC | 23:08 | |
| *** rods_ is now known as rods | 23:09 | |
| *** rods` has joined #openstack-dev | 23:09 | |
| *** rods has quit IRC | 23:09 | |
| *** rods has joined #openstack-dev | 23:10 | |
| *** halfss has joined #openstack-dev | 23:12 | |
| *** anniec_ has joined #openstack-dev | 23:12 | |
| *** andrewbogott has quit IRC | 23:20 | |
| *** rods has quit IRC | 23:22 | |
| *** danwent has quit IRC | 23:23 | |
| *** dachary has joined #openstack-dev | 23:25 | |
| *** anniec_ has quit IRC | 23:33 | |
| Ryan_Lane | crap. in keystone you need to authenticate a user for every single tenant? | 23:34 |
| Ryan_Lane | keystone is really, really painful for my usecase :( | 23:36 |
| arosen | Ryan_Lane: you can use admin if you want. | 23:38 |
| Ryan_Lane | in which way? | 23:38 |
| arosen | i think if you connect via port 5000 | 23:38 |
| Ryan_Lane | for my use case, a user can be a member of multiple projects | 23:38 |
| Ryan_Lane | when they log into the web interface, they should get keystone credentials that allow them to do actions in any project they are a member of | 23:39 |
| *** mnaser has joined #openstack-dev | 23:39 | |
| Ryan_Lane | but they should only ever need to provide their password on login | 23:39 |
| Ryan_Lane | I want to ensure they are accessing services as their user, though, for auditing reasons | 23:40 |
| arosen | Sorry, I'm not sure off hand hopefully someone else can chime in. | 23:40 |
| *** pixelbeat has quit IRC | 23:40 | |
| *** anderstj has quit IRC | 23:41 | |
| *** Adri2000 has quit IRC | 23:45 | |
| *** nunosantos has joined #openstack-dev | 23:50 | |
| *** timello has quit IRC | 23:53 | |
| *** koolhead17 has joined #openstack-dev | 23:53 | |
| *** anniec has joined #openstack-dev | 23:55 | |
| *** anniec_ has joined #openstack-dev | 23:56 | |
| *** jaypipes has quit IRC | 23:58 | |
| *** dhellmann has joined #openstack-dev | 23:59 | |
| *** anniec has quit IRC | 23:59 | |
| *** anniec_ is now known as anniec | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
