| *** j05h has left #openstack-dev | 00:02 | |
| *** maplebed has joined #openstack-dev | 00:04 | |
| *** flaviamissi_ has joined #openstack-dev | 00:04 | |
| *** zzed_ has quit IRC | 00:04 | |
| *** anotherjesse_zz is now known as anotherjesse | 00:05 | |
| *** flaviamissi has quit IRC | 00:06 | |
| *** flaviamissi_ has quit IRC | 00:07 | |
| *** hub_cap has quit IRC | 00:08 | |
| *** dtroyer has joined #openstack-dev | 00:09 | |
| *** dolphm has quit IRC | 00:11 | |
| littleidea | comstud: on that return for the exception, would you get rid of the try? I was just trying to preserve behavior | 00:11 |
|---|---|---|
| *** j05h has joined #openstack-dev | 00:11 | |
| *** sleepsonzzz has joined #openstack-dev | 00:12 | |
| comstud | littleidea: there was another exception there to keep i think.. looking | 00:13 |
| *** jakedahn_zz has joined #openstack-dev | 00:13 | |
| comstud | littleidea: ah, i guess not | 00:13 |
| comstud | hmmm | 00:13 |
| *** sleepsonthefloo has left #openstack-dev | 00:14 | |
| comstud | littleidea: looks like technically you should also try to set ERROR i guess | 00:15 |
| comstud | littleidea: there's 1 other call for instances that could hit that case... which is 'live_migration' | 00:16 |
| comstud | littleidea: so I guess the return should change to trying to set the instance to ERROR | 00:16 |
| littleidea | comstud: that is kind of a big change of behavior, nothing could set it before but 'run_instances'. I don't have enough context to advocate one way or another. | 00:16 |
| comstud | (like you did for prep_resize) | 00:16 |
| comstud | if you feel more comfortable just dropping the 'try' then.. that's fine | 00:17 |
| comstud | technically live_migration could fail and we'd want to do something there... | 00:17 |
| comstud | but yeah, it'd change the behavior | 00:17 |
| comstud | we can fix that in a future patch | 00:17 |
| comstud | (i would probably add a 'live_migration' method in the manager in the future) | 00:18 |
| littleidea | comstud: I could go either way, I'll add it to set an error and leave the reraise | 00:18 |
| comstud | personally, i'd like to get rid of the __getattr_ and _schedule()... but.. another day. | 00:18 |
| comstud | littleidea: ok | 00:19 |
| *** dwalleck has joined #openstack-dev | 00:20 | |
| *** flaviamissi has joined #openstack-dev | 00:20 | |
| *** dwalleck has quit IRC | 00:21 | |
| *** adjohn has joined #openstack-dev | 00:21 | |
| *** flaviamissi has quit IRC | 00:25 | |
| *** paulormg has quit IRC | 00:25 | |
| *** flaviamissi has joined #openstack-dev | 00:26 | |
| *** dwalleck has joined #openstack-dev | 00:27 | |
| *** rkukura has joined #openstack-dev | 00:29 | |
| *** thingee has quit IRC | 00:31 | |
| *** flaviamissi has quit IRC | 00:36 | |
| *** flaviamissi has joined #openstack-dev | 00:37 | |
| *** dwalleck has quit IRC | 00:38 | |
| *** nati has joined #openstack-dev | 00:41 | |
| *** dolphm has joined #openstack-dev | 00:42 | |
| *** flaviamissi has quit IRC | 00:43 | |
| *** asdfasdf has quit IRC | 00:44 | |
| *** pixelbeat has quit IRC | 00:51 | |
| *** dolphm has quit IRC | 00:56 | |
| littleidea | comstud: fire in the hole | 00:56 |
| comstud | ok | 00:56 |
| comstud | littleidea: bueno :) only one little thing i just spotted.. i like to leave task_state alone when setting vm_state to ERROR... so you can see what task_state the instance was last in.. but it's not a huge deal | 01:01 |
| *** ncode has quit IRC | 01:02 | |
| littleidea | easy to fix | 01:15 |
| littleidea | comstud: changed | 01:18 |
| vishy | comstud: I think you misread something on my review. Commented | 01:19 |
| anotherjesse | dtroyer: thoughts on https://bugs.launchpad.net/keystone/+bug/947011 | 01:23 |
| uvirtbot` | Launchpad bug 947011 in keystone "users should be able to get ec2 credentials" [Undecided,New] | 01:23 |
| *** rkukura has left #openstack-dev | 01:23 | |
| comstud | vihsy: I did... | 01:23 |
| comstud | vihsy: There's a useful test missing that would have helped | 01:23 |
| comstud | but I switched to +2 | 01:23 |
| comstud | ah | 01:26 |
| comstud | littleidea: you have a test failure | 01:26 |
| *** PotHix has quit IRC | 01:26 | |
| jeblair | annegentle, anotherjesse: https://jenkins.openstack.org/job/openstack-docs-site/ <-- should run each time master branch of openstack-manuals is updated. | 01:33 |
| anotherjesse | jeblair: awesome - we are about to have a test case! | 01:33 |
| jeblair | great! | 01:33 |
| anotherjesse | once heckj approves https://review.openstack.org/#change,4892 | 01:33 |
| *** x86brandon has quit IRC | 01:34 | |
| *** anotherjesse is now known as anotherjesse_zz | 01:37 | |
| *** zykes- has quit IRC | 01:38 | |
| heckj | anotherjesse: just approverized | 01:39 |
| *** maplebed has quit IRC | 01:39 | |
| *** zykes- has joined #openstack-dev | 01:40 | |
| *** martine has joined #openstack-dev | 01:40 | |
| *** reed has quit IRC | 01:44 | |
| littleidea | comstud, vishy: can you look at the failing test with me. it looks like it is using the ChanceScheduler, which has schedule_run_instances defined, so I'm confused why it blows up | 01:45 |
| *** danwent has quit IRC | 01:47 | |
| *** rkukura has joined #openstack-dev | 01:48 | |
| dtroyer | anotherjesse: yeah, I'd expect that to work | 01:49 |
| jeblair | anotherjesse_zz, heckj, annegentle: yay! docs.openstack.org looks different now. :) | 01:57 |
| *** benner has quit IRC | 02:00 | |
| *** vincentricci has quit IRC | 02:01 | |
| *** novas0x2a|laptop has joined #openstack-dev | 02:03 | |
| comstud | littleidea: yea | 02:06 |
| comstud | littleidea: sec | 02:06 |
| comstud | littleidea: ah, the methods need added to the multi scheduler :-/ | 02:08 |
| comstud | littleidea: try this: http://paste.openstack.org/show/6683/ | 02:10 |
| *** j05h has quit IRC | 02:13 | |
| *** benner has joined #openstack-dev | 02:18 | |
| *** hattwick has quit IRC | 02:21 | |
| *** mattray has joined #openstack-dev | 02:24 | |
| *** hattwick has joined #openstack-dev | 02:24 | |
| *** martine_ has joined #openstack-dev | 02:28 | |
| *** martine has quit IRC | 02:29 | |
| *** martine_ has quit IRC | 02:29 | |
| *** martine has joined #openstack-dev | 02:30 | |
| *** heckj has quit IRC | 02:30 | |
| *** ches has quit IRC | 02:30 | |
| *** ches has joined #openstack-dev | 02:31 | |
| *** vizsla has joined #openstack-dev | 02:34 | |
| *** j05h has joined #openstack-dev | 02:35 | |
| *** x86brandon has joined #openstack-dev | 02:37 | |
| *** j05h has quit IRC | 02:39 | |
| *** littleidea has quit IRC | 02:40 | |
| *** littleidea has joined #openstack-dev | 02:45 | |
| *** novas0x2a|laptop has quit IRC | 02:46 | |
| *** j05h has joined #openstack-dev | 02:46 | |
| *** Ryan_Lane has quit IRC | 02:46 | |
| *** novas0x2a|laptop has joined #openstack-dev | 02:47 | |
| *** dalang has left #openstack-dev | 02:47 | |
| *** ayoung has quit IRC | 02:47 | |
| *** shevek_ has quit IRC | 02:48 | |
| *** hattwick has quit IRC | 02:52 | |
| *** hattwick has joined #openstack-dev | 02:55 | |
| *** reed has joined #openstack-dev | 02:57 | |
| *** j05h has quit IRC | 03:06 | |
| *** spiffxp has quit IRC | 03:07 | |
| *** j05h has joined #openstack-dev | 03:08 | |
| *** j05h has quit IRC | 03:09 | |
| *** anotherjesse_zz is now known as anotherjesse | 03:09 | |
| *** asdfasdf has joined #openstack-dev | 03:11 | |
| *** AlanClark has quit IRC | 03:12 | |
| *** nati has quit IRC | 03:13 | |
| *** AlanClark has joined #openstack-dev | 03:16 | |
| *** anotherjesse is now known as anotherjesse_zz | 03:19 | |
| *** littleidea has quit IRC | 03:31 | |
| *** anotherjesse_zz is now known as anotherjesse | 03:32 | |
| *** littleidea has joined #openstack-dev | 03:33 | |
| *** anotherjesse is now known as anotherjesse_zz | 03:36 | |
| littleidea | comstud: raising those exceptions is getting a bit twisty | 03:37 |
| *** cp16net has joined #openstack-dev | 03:39 | |
| *** sandywalsh has quit IRC | 03:39 | |
| *** asdfasdf has quit IRC | 03:40 | |
| *** sandywalsh has joined #openstack-dev | 03:40 | |
| *** openstackjenkins has quit IRC | 03:45 | |
| *** openstackjenkins has joined #openstack-dev | 03:46 | |
| *** mattray has quit IRC | 03:47 | |
| *** bencherian_ has joined #openstack-dev | 03:50 | |
| *** bencherian has quit IRC | 03:50 | |
| *** bencherian_ is now known as bencherian | 03:50 | |
| *** bencherian has quit IRC | 03:54 | |
| *** novas0x2a|laptop has quit IRC | 03:58 | |
| *** littleidea has quit IRC | 04:02 | |
| *** littleidea has joined #openstack-dev | 04:03 | |
| *** AlanClark has quit IRC | 04:04 | |
| mtaylor | anybody around who feels brave want to review https://review.openstack.org/#change,4536 ? | 04:04 |
| *** jdurgin has quit IRC | 04:06 | |
| *** bencherian has joined #openstack-dev | 04:12 | |
| mtaylor | annegentle: ping | 04:14 |
| *** anotherjesse_zz is now known as anotherjesse | 04:18 | |
| anotherjesse | mtaylor: sleepsonthefloor did some of the work on that | 04:20 |
| littleidea | does XenAPIDom0IptablesFirewallTestCase.test_do_refresh_security_group_rules hang sometimes for anyone else? | 04:22 |
| annegentle | mtaylor: pong | 04:23 |
| mtaylor | anotherjesse: awesome. I'll bug him about it | 04:25 |
| mtaylor | sleepsonzzz: ^^^ | 04:25 |
| mtaylor | annegentle: the css files in glance and nova's sphinx docs are different | 04:25 |
| mtaylor | annegentle: oddly enough, adding the css files from glance to nova fixes the bug there docs in nova don't build inside of a virtualenv | 04:26 |
| anotherjesse | mtaylor: jakedahn is going to be exploring that tomorrow | 04:26 |
| mtaylor | anotherjesse: the autodoc patch thing? | 04:26 |
| anotherjesse | no | 04:26 |
| anotherjesse | improving the css theme to be more openstacky | 04:26 |
| mtaylor | ah, great! | 04:26 |
| annegentle | mtaylor: do you really think there are 89 files to be transferred wowsza | 04:27 |
| mtaylor | anotherjesse: well - if you could have him look at https://review.openstack.org/#change,4951 | 04:27 |
| annegentle | er probably jeblair would be more interested in my 89 files comment :) | 04:27 |
| mtaylor | anotherjesse: I defer to everyone else on the css content - but something about adding files to nova's css causes python setup.py build_sphinx to start working in a venv | 04:28 |
| anotherjesse | k - let's it in there sooner | 04:28 |
| *** anotherjesse is now known as anotherjesse_zz | 04:32 | |
| *** hattwick has quit IRC | 04:33 | |
| annegentle | how come no book file for dev doc? | 04:34 |
| *** danwent has joined #openstack-dev | 04:39 | |
| *** yamahata has quit IRC | 04:41 | |
| *** yamahata__ has quit IRC | 04:42 | |
| *** sandywalsh has quit IRC | 04:47 | |
| *** sandywalsh has joined #openstack-dev | 04:48 | |
| *** anotherjesse_zz is now known as anotherjesse | 04:55 | |
| anotherjesse | annegentle: ? | 04:55 |
| annegentle | anotherjesse: yes? | 04:56 |
| annegentle | anotherjesse: oh. when I look at docs.openstack.org, no book icon | 04:56 |
| anotherjesse | ahh, cloudsites cache | 04:56 |
| annegentle | annegentle: firefox on mac | 04:56 |
| anotherjesse | hit refresh a few times | 04:56 |
| annegentle | anotherjesse: ah got it | 04:56 |
| anotherjesse | need to wait for cache invalidation to occur :( | 04:56 |
| annegentle | anotherjesse: cool, thanks! | 04:56 |
| annegentle | anotherjesse: hrm… guess I didn't really Need to see a book icon then :) | 04:57 |
| anotherjesse | annegentle: jake is working on making sphinx/rst look like docbook html output | 05:00 |
| annegentle | anotherjesse: what do you want to do about comments? | 05:01 |
| anotherjesse | annegentle: do you mean the fact that docbook splits it into lots of individual pages (per subsection)? | 05:01 |
| anotherjesse | and so we just put a discus block at each? | 05:01 |
| annegentle | anotherjesse: you can chunk it where you think appropriate, for comments, or have no comments. Main thing I think is to have a disqus short name for each "dev guide" | 05:02 |
| annegentle | anotherjesse: so that devs will moderate dev doc comments | 05:02 |
| annegentle | anotherjesse: (if that doesn't make sense let's talk it through) | 05:02 |
| anotherjesse | annegentle: it does … we can make it so we add a directive into the RST .. discus "foo" -- which will expand to "docname-foo" and put a discuss section inline | 05:04 |
| anotherjesse | or make it so we build multiple pages | 05:04 |
| anotherjesse | each with a discuss block at bottom | 05:04 |
| annegentle | anotherjesse: seems like sphinx already does chunking as well, per "page" but also enables mini-tocs in a page. | 05:04 |
| annegentle | anotherjesse: docname-foo makes sense | 05:05 |
| *** hattwick has joined #openstack-dev | 05:05 | |
| *** martine has quit IRC | 05:12 | |
| *** x86brandon has quit IRC | 05:17 | |
| *** anotherjesse is now known as anotherjesse_zz | 05:22 | |
| *** danwent has quit IRC | 05:23 | |
| *** yamahata has joined #openstack-dev | 05:24 | |
| *** danwent has joined #openstack-dev | 05:24 | |
| *** yamahata__ has joined #openstack-dev | 05:24 | |
| *** cp16net has quit IRC | 05:38 | |
| *** anotherjesse_zz is now known as anotherjesse | 05:45 | |
| *** novas0x2a|laptop has joined #openstack-dev | 05:45 | |
| *** cp16net has joined #openstack-dev | 05:58 | |
| *** zaitcev has quit IRC | 06:03 | |
| *** dayou has joined #openstack-dev | 06:06 | |
| *** danwent has quit IRC | 06:09 | |
| *** journeeman has joined #openstack-dev | 06:09 | |
| *** reed has quit IRC | 06:10 | |
| *** hattwick has quit IRC | 06:32 | |
| *** hashar has joined #openstack-dev | 06:33 | |
| *** bepernoot has joined #openstack-dev | 06:35 | |
| *** mnewby_ has joined #openstack-dev | 06:36 | |
| *** mnewby_ has quit IRC | 06:37 | |
| *** Ryan_Lane has joined #openstack-dev | 06:55 | |
| *** bepernoot has quit IRC | 07:00 | |
| *** cp16net has quit IRC | 07:02 | |
| *** anotherjesse is now known as anotherjesse_zz | 07:05 | |
| *** hashar has quit IRC | 07:09 | |
| *** shevek_ has joined #openstack-dev | 07:27 | |
| *** dtroyer has quit IRC | 07:33 | |
| *** zykes- has quit IRC | 07:46 | |
| *** zykes- has joined #openstack-dev | 07:52 | |
| *** mihgen has joined #openstack-dev | 07:52 | |
| *** zykes- has quit IRC | 07:53 | |
| *** bepernoot has joined #openstack-dev | 08:04 | |
| ttx | ohnoimdead: around now | 08:15 |
| *** dubsquared has joined #openstack-dev | 08:15 | |
| *** dubsquared has quit IRC | 08:16 | |
| *** n0ano has quit IRC | 08:22 | |
| *** bencherian has quit IRC | 08:25 | |
| *** Ryan_Lane has quit IRC | 08:25 | |
| *** reidrac has joined #openstack-dev | 08:27 | |
| *** dneary has joined #openstack-dev | 08:31 | |
| *** darraghb has joined #openstack-dev | 08:32 | |
| *** adjohn has quit IRC | 08:33 | |
| *** pixelbeat has joined #openstack-dev | 08:34 | |
| ttx | ohnoimdead: replied by email | 08:34 |
| *** zykes- has joined #openstack-dev | 08:42 | |
| *** Mkenneth has joined #openstack-dev | 08:43 | |
| zykes- | soren: around ? :) | 08:43 |
| *** ches has quit IRC | 08:45 | |
| *** ches has joined #openstack-dev | 08:45 | |
| *** apevec has joined #openstack-dev | 08:55 | |
| *** hattwick has joined #openstack-dev | 08:59 | |
| *** Mkenneth has quit IRC | 09:00 | |
| *** Mkenneth has joined #openstack-dev | 09:01 | |
| eglynn | ttx: quick question? | 09:01 |
| ttx | eglynn: sure | 09:01 |
| eglynn | ttx: any idea when the folsom design summit will be open for session proposals? | 09:01 |
| eglynn | ttx: or should I be addressing the question to stef? | 09:02 |
| ttx | eglynn: I started the discussion with the newly-elected PTLs yesterday, on how they want the summit to look like | 09:02 |
| ttx | eglynn: we should have something ready in ~1 week, I'd say | 09:02 |
| eglynn | ttx: great thanks! | 09:02 |
| soren | zykes-: *I* have been here all along. You're the one who keeps wandering in and out :) | 09:04 |
| zykes- | ;) | 09:18 |
| zykes- | congrats to all PTLs :) | 09:19 |
| *** mihgen has quit IRC | 09:29 | |
| *** Mkenneth has left #openstack-dev | 09:30 | |
| *** mihgen has joined #openstack-dev | 09:30 | |
| *** zykes- has quit IRC | 09:30 | |
| *** mihgen has quit IRC | 09:35 | |
| *** mihgen has joined #openstack-dev | 09:36 | |
| *** derekh has joined #openstack-dev | 09:43 | |
| *** hashar has joined #openstack-dev | 09:45 | |
| *** dayou has quit IRC | 09:50 | |
| *** eglynn has quit IRC | 09:53 | |
| *** eglynn has joined #openstack-dev | 09:54 | |
| *** dayou has joined #openstack-dev | 10:03 | |
| *** maploin has joined #openstack-dev | 10:12 | |
| *** maploin has quit IRC | 10:12 | |
| *** maploin has joined #openstack-dev | 10:12 | |
| *** adiantum has quit IRC | 10:23 | |
| *** pixelbeat has quit IRC | 10:25 | |
| *** paulormg has joined #openstack-dev | 10:31 | |
| *** novas0x2a|laptop has quit IRC | 10:41 | |
| *** bepernoo1 has joined #openstack-dev | 10:48 | |
| *** bepernoot has quit IRC | 10:50 | |
| *** pixelbeat has joined #openstack-dev | 10:53 | |
| *** bepernoot has joined #openstack-dev | 10:53 | |
| *** bepernoo1 has quit IRC | 10:54 | |
| *** LinuxJedi has quit IRC | 11:05 | |
| *** LinuxJedi has joined #openstack-dev | 11:06 | |
| *** mancdaz has joined #openstack-dev | 11:07 | |
| *** eglynn has quit IRC | 11:28 | |
| *** mihgen has quit IRC | 11:34 | |
| *** eglynn has joined #openstack-dev | 11:36 | |
| *** adiantum has joined #openstack-dev | 11:38 | |
| *** sdague_ has joined #openstack-dev | 11:48 | |
| *** sdague has quit IRC | 11:49 | |
| *** adiantum has quit IRC | 11:52 | |
| *** Mkenneth has joined #openstack-dev | 12:01 | |
| *** hashar has quit IRC | 12:05 | |
| *** gakott has joined #openstack-dev | 12:15 | |
| *** mancdaz has quit IRC | 12:20 | |
| *** openstackjenkins has quit IRC | 12:20 | |
| *** milner has quit IRC | 12:20 | |
| *** YorikSar has quit IRC | 12:20 | |
| *** gkotton has quit IRC | 12:20 | |
| *** pmyers has quit IRC | 12:20 | |
| *** stokachu has quit IRC | 12:20 | |
| *** bsza has joined #openstack-dev | 12:22 | |
| *** mancdaz has joined #openstack-dev | 12:23 | |
| *** openstackjenkins has joined #openstack-dev | 12:23 | |
| *** milner has joined #openstack-dev | 12:23 | |
| *** YorikSar has joined #openstack-dev | 12:23 | |
| *** pmyers has joined #openstack-dev | 12:23 | |
| *** stokachu has joined #openstack-dev | 12:23 | |
| *** eglynn has quit IRC | 12:43 | |
| *** markvoelker has joined #openstack-dev | 12:44 | |
| *** journeeman has quit IRC | 12:47 | |
| *** andrewsmedina has quit IRC | 12:48 | |
| *** littleidea has quit IRC | 12:53 | |
| *** littleidea has joined #openstack-dev | 12:55 | |
| *** andrewsmedina has joined #openstack-dev | 13:09 | |
| *** jmcboots has joined #openstack-dev | 13:11 | |
| *** dprince has joined #openstack-dev | 13:12 | |
| *** littleidea has quit IRC | 13:17 | |
| *** Mkenneth has quit IRC | 13:25 | |
| *** sniperd has joined #openstack-dev | 13:28 | |
| *** eglynn has joined #openstack-dev | 13:30 | |
| *** j05h has joined #openstack-dev | 13:32 | |
| *** vizsla has quit IRC | 13:32 | |
| *** Mkenneth has joined #openstack-dev | 13:48 | |
| *** dayou has quit IRC | 13:49 | |
| *** markmc has joined #openstack-dev | 13:49 | |
| *** littleidea has joined #openstack-dev | 13:52 | |
| *** adiantum has joined #openstack-dev | 13:53 | |
| *** sdague_ is now known as sdague | 13:58 | |
| *** martine has joined #openstack-dev | 13:59 | |
| *** ecarlin has joined #openstack-dev | 13:59 | |
| *** littleidea has quit IRC | 14:04 | |
| *** lts has joined #openstack-dev | 14:05 | |
| *** oneiroi has joined #openstack-dev | 14:06 | |
| *** littleidea has joined #openstack-dev | 14:06 | |
| *** oneiroi has quit IRC | 14:08 | |
| *** oneiroi has joined #openstack-dev | 14:08 | |
| *** jmcboots_ has joined #openstack-dev | 14:12 | |
| *** kbringard has joined #openstack-dev | 14:13 | |
| *** jmcboots has quit IRC | 14:15 | |
| *** jmcboots_ is now known as jmcboots | 14:15 | |
| *** stuntmachine has joined #openstack-dev | 14:17 | |
| *** stuntmachine has joined #openstack-dev | 14:17 | |
| *** sandywalsh has quit IRC | 14:29 | |
| *** ecarlin has quit IRC | 14:29 | |
| *** mattray has joined #openstack-dev | 14:30 | |
| *** crobinso has joined #openstack-dev | 14:31 | |
| *** dtroyer has joined #openstack-dev | 14:39 | |
| *** ayoung has joined #openstack-dev | 14:41 | |
| *** sandywalsh has joined #openstack-dev | 14:42 | |
| *** LinuxJedi_ has joined #openstack-dev | 14:44 | |
| *** rkukura has quit IRC | 14:45 | |
| *** LinuxJedi has quit IRC | 14:46 | |
| *** LinuxJedi has joined #openstack-dev | 14:47 | |
| *** armaan has joined #openstack-dev | 14:53 | |
| *** Mkenneth has quit IRC | 15:02 | |
| *** LinuxJedi has quit IRC | 15:03 | |
| *** LinuxJedi has joined #openstack-dev | 15:04 | |
| *** jmcboots has quit IRC | 15:07 | |
| *** jmcboots has joined #openstack-dev | 15:07 | |
| *** jmcboots has left #openstack-dev | 15:07 | |
| *** ecarlin has joined #openstack-dev | 15:08 | |
| *** rnirmal has joined #openstack-dev | 15:09 | |
| kbringard | is keystone-light at a place where I should think about using it with diablo-stable, or should I stick with the full fat version? | 15:15 |
| *** Mkenneth has joined #openstack-dev | 15:16 | |
| *** dtroyer has quit IRC | 15:28 | |
| *** armaan has left #openstack-dev | 15:33 | |
| *** ecarlin has quit IRC | 15:34 | |
| *** rkukura has joined #openstack-dev | 15:35 | |
| *** ecarlin has joined #openstack-dev | 15:36 | |
| chmouel | kbringard: you should probably use keystone-legacy with diablo | 15:38 |
| kbringard | chmouel: cool, thanks | 15:39 |
| kbringard | that's what I figured, but wanted to ask | 15:39 |
| *** ecarlin has quit IRC | 15:41 | |
| *** eglynn__ has joined #openstack-dev | 15:41 | |
| *** ecarlin has joined #openstack-dev | 15:42 | |
| *** deshantm has joined #openstack-dev | 15:42 | |
| *** eglynn has quit IRC | 15:43 | |
| *** sandywalsh has quit IRC | 15:45 | |
| *** cp16net has joined #openstack-dev | 15:48 | |
| *** reed has joined #openstack-dev | 15:49 | |
| *** hhoover has joined #openstack-dev | 15:49 | |
| *** zzed has joined #openstack-dev | 15:55 | |
| *** sandywalsh has joined #openstack-dev | 16:00 | |
| *** danwent has joined #openstack-dev | 16:04 | |
| *** dolphm has joined #openstack-dev | 16:06 | |
| *** spiffxp has joined #openstack-dev | 16:12 | |
| *** eglynn has joined #openstack-dev | 16:14 | |
| *** hhoover has left #openstack-dev | 16:14 | |
| *** eglynn__ has quit IRC | 16:15 | |
| *** jdg has joined #openstack-dev | 16:17 | |
| *** reidrac has quit IRC | 16:18 | |
| *** spiffxp has quit IRC | 16:21 | |
| *** maplebed has joined #openstack-dev | 16:22 | |
| *** Mkenneth has left #openstack-dev | 16:22 | |
| *** eglynn_ has joined #openstack-dev | 16:23 | |
| *** eglynn has quit IRC | 16:26 | |
| *** deshantm has quit IRC | 16:27 | |
| *** bepernoot has quit IRC | 16:27 | |
| *** heckj has joined #openstack-dev | 16:27 | |
| *** hashar has joined #openstack-dev | 16:28 | |
| YorikSar | blamar: around? | 16:29 |
| annegentle | feel free to join us in #openstack-docday | 16:29 |
| *** hub_cap has joined #openstack-dev | 16:33 | |
| *** Gordonz_ has quit IRC | 16:48 | |
| *** Gordonz has joined #openstack-dev | 16:48 | |
| *** anotherjesse_zz is now known as anotherjesse | 16:57 | |
| *** reed has quit IRC | 17:05 | |
| chmouel | notmyname: any chance to get this reviewed https://review.openstack.org/#change,4893 asap? as we will need that to get glance+swift/auth2.0 support | 17:06 |
| heckj | chmouel: sorry for being a slacker on that review - +2'd and approved. Should flow through shortly | 17:07 |
| chmouel | heckj: no worries, thanks.. | 17:08 |
| littleidea | comstud: you around? | 17:09 |
| comstud | littleidea; yup | 17:10 |
| comstud | littleidea: you see my response here last night about multi scheduler? | 17:10 |
| littleidea | I didn't but I will read now. | 17:10 |
| comstud | littleidea: ok.. i left you a paste that i think should fix the issue | 17:11 |
| comstud | (untested) | 17:11 |
| comstud | littleidea: http://paste.openstack.org/show/6683/ | 17:11 |
| littleidea | ok, I saw that | 17:12 |
| *** sniperd has quit IRC | 17:12 | |
| littleidea | turns out, that's only a third of it | 17:12 |
| littleidea | and when you do that, the multi tests start breaking | 17:13 |
| *** mikeyp has joined #openstack-dev | 17:13 | |
| littleidea | I also had to pull the unimplemented tests into a different test class because it was breaking the vsa tests | 17:13 |
| comstud | probably just adjustment needed due to the fact that those aren't proxied the same way anymore | 17:14 |
| comstud | hm | 17:14 |
| mtaylor | did anybody sort out the nova test timeout issue? | 17:14 |
| comstud | mtaylor: johannes has a review up for it | 17:14 |
| comstud | mtaylor: 4968 | 17:14 |
| mtaylor | comstud: awesome | 17:14 |
| littleidea | comstud: so the thing I'm looking at now is the test_multi_scheduler proxy failing, as with your changes, no compute methods are actually proxied that way | 17:15 |
| comstud | littleidea: in test_proxy_calls in the multi tests... you can replace: | 17:15 |
| comstud | 82 test_methods = {compute_driver: ['run_instance', 'prep_resize'], | 17:15 |
| comstud | remove run_instance and prep_resize and add 'live_migration' | 17:16 |
| comstud | i think | 17:16 |
| comstud | *looks* | 17:16 |
| *** maploin has quit IRC | 17:16 | |
| littleidea | comstud: tried that already, doesn't work because live_migration is defined in the base class | 17:16 |
| comstud | oooh crap | 17:16 |
| comstud | taht's a bug | 17:16 |
| comstud | or wait | 17:16 |
| littleidea | I was thinking of doing something like this: | 17:16 |
| comstud | the compute driver should have inherited it though | 17:17 |
| littleidea | #no compute methods are proxied at this time | 17:17 |
| littleidea | test_methods = {compute_driver: [], | 17:17 |
| comstud | oh i guess i see | 17:17 |
| comstud | yeah | 17:17 |
| comstud | i think that's appropriate | 17:17 |
| littleidea | the compute driver did inherit it, but so did multi | 17:18 |
| littleidea | since it's base class is driver.Scheduler | 17:18 |
| comstud | yep | 17:18 |
| comstud | got it... fun | 17:19 |
| littleidea | wheee | 17:19 |
| comstud | i think [] for compute_driver in that test is appropriate right now | 17:19 |
| littleidea | test timeouts are so awesome | 17:19 |
| comstud | but this shows there's technically a bug | 17:19 |
| comstud | live_migration should also be added to multi scheduler explicitly | 17:20 |
| comstud | otherwise... it always uses the base live_migration | 17:20 |
| comstud | (which is the only thing that exists right now, but) | 17:20 |
| comstud | in any case...we can address that later | 17:20 |
| littleidea | Is there someway to parameterize the _METHOD_MAP? | 17:20 |
| comstud | not easily | 17:21 |
| littleidea | I don't see one, but my python fu is weak | 17:21 |
| littleidea | not a straight forward one | 17:21 |
| comstud | yeah, we can think about that later | 17:21 |
| littleidea | intended to be used by people who want to kiss their mother | 17:21 |
| comstud | haha | 17:22 |
| comstud | i'm cool with you suggestion for test_proxy_calls() | 17:22 |
| comstud | your | 17:23 |
| comstud | does that address everything? | 17:23 |
| littleidea | seems to | 17:24 |
| annegentle | jeblair: or mtaylor: can someone from CI pop into #openstack-docday to help Alexey with a public key problem? | 17:24 |
| *** PotHix has joined #openstack-dev | 17:25 | |
| *** dtroyer has joined #openstack-dev | 17:25 | |
| *** spiffxp has joined #openstack-dev | 17:25 | |
| *** oneiroi has quit IRC | 17:26 | |
| *** zykes- has joined #openstack-dev | 17:28 | |
| zykes- | soren: ? :) | 17:28 |
| *** vincentricci has joined #openstack-dev | 17:30 | |
| jeblair | annegentle: ok | 17:30 |
| *** ecarlin has quit IRC | 17:32 | |
| *** x86brandon has joined #openstack-dev | 17:32 | |
| *** x86brandon has quit IRC | 17:33 | |
| *** x86brandon has joined #openstack-dev | 17:33 | |
| *** ecarlin has joined #openstack-dev | 17:35 | |
| *** hashar has quit IRC | 17:35 | |
| *** rnirmal has quit IRC | 17:37 | |
| maplebed | Anybody know what version of Swift introduced inter-cluster replication? (for maintaining a failover cluster in a separate data center, for example) (cross-asking from #openstack) | 17:39 |
| maplebed | (http://swift.openstack.org/overview_container_sync.html) | 17:39 |
| notmyname | maplebed: swift 1.4.2 | 17:40 |
| maplebed | I'm running 1.4.3 and I see at least some evidence it exists, but I'm not sure whether it actually works in this version... (evidence == container/server.py knows about the allowed_sync_hosts config variable) | 17:41 |
| maplebed | excellent. | 17:41 |
| *** lloydde has joined #openstack-dev | 17:41 | |
| maplebed | do you know if there were any major changes or bugs fixed between 1.4.3 and 1.4.7 in that section of the code? | 17:41 |
| *** dolphm has quit IRC | 17:42 | |
| *** zykes- has quit IRC | 17:42 | |
| notmyname | maplebed: nothing springs to mind. it should have been noted in the CHANGELOG file if there was something though | 17:42 |
| *** zykes- has joined #openstack-dev | 17:43 | |
| maplebed | I haven't downloaded a copy of 1.4.7 yet to check. | 17:43 |
| *** devcamcar has quit IRC | 17:43 | |
| *** sleepsonthefloor has quit IRC | 17:43 | |
| *** vishy has quit IRC | 17:43 | |
| *** xtoddx has quit IRC | 17:43 | |
| maplebed | I'll do that though. | 17:43 |
| maplebed | thanks! | 17:43 |
| chmouel | maplebed: git log does not seem to mention anything major since July (except some timeout fixes) | 17:43 |
| maplebed | good news. thanks for the check, chmouel | 17:44 |
| *** vishy has joined #openstack-dev | 17:45 | |
| *** dneary has quit IRC | 17:46 | |
| *** AlanClark has joined #openstack-dev | 17:46 | |
| *** derekh has quit IRC | 17:46 | |
| *** rods has joined #openstack-dev | 17:46 | |
| littleidea | comstud: look now, I think I fat fingered an old buffer | 17:48 |
| comstud | littleidea: looks good now, thanks! | 17:48 |
| comstud | littleidea: i'll prob do a little more cleanup after this hits... as live_migration should be added to multi scheduler and so forth. i'll think if there's a better way to do multi in general | 17:49 |
| littleidea | comstud: my impression of multi is it should probably be rethought | 17:50 |
| comstud | i feel like 'multi' shouldn't exist... its support should be folded into the manager. | 17:50 |
| comstud | or something like that. | 17:51 |
| littleidea | that makes more sense | 17:51 |
| littleidea | and really, the map shouldn't just be compute or volume | 17:51 |
| *** gakott has quit IRC | 17:51 | |
| *** sniperd has joined #openstack-dev | 17:51 | |
| comstud | yeah, although those the only 2 things that schedule right now | 17:52 |
| comstud | +are | 17:52 |
| *** dalang has joined #openstack-dev | 17:52 | |
| comstud | i had thrown network in there at one point... as there was a network method... although then I found out it was not actually being used.. so it all got ripped out | 17:52 |
| littleidea | sure, but what if you want to use a certain kind of scheduler for one type of compute action and another for a different one at some point | 17:52 |
| comstud | oh ok | 17:52 |
| comstud | i see what you mean | 17:52 |
| *** oneiroi has joined #openstack-dev | 17:53 | |
| *** oneiroi has joined #openstack-dev | 17:53 | |
| *** oneiroi has quit IRC | 17:56 | |
| *** Ryan_Lane has joined #openstack-dev | 17:56 | |
| *** zaitcev has joined #openstack-dev | 17:57 | |
| *** dolphm has joined #openstack-dev | 17:58 | |
| *** adiantum has quit IRC | 17:59 | |
| *** ecarlin has quit IRC | 18:00 | |
| *** jdurgin has joined #openstack-dev | 18:01 | |
| *** reed has joined #openstack-dev | 18:01 | |
| *** gyee has joined #openstack-dev | 18:08 | |
| *** andrewbogott has quit IRC | 18:09 | |
| vishy | tr3buchet: ping | 18:10 |
| *** joesavak has joined #openstack-dev | 18:10 | |
| tr3buchet | vishy: hiyas | 18:18 |
| vishy | tr3buchet: so there is an error in _teardown | 18:19 |
| vishy | address is not defined | 18:19 |
| tr3buchet | vishy: looking | 18:19 |
| vishy | tr3buchet: I'm trying to figure out the best way to get address from the vif | 18:19 |
| vishy | or if i should change the definition of teardown to optionally take the address | 18:19 |
| *** markvoelker has quit IRC | 18:20 | |
| vishy | tr3buchet: right now I'm thinking do fixed_ips_by_virtual_interface and calling release dhcp on all of them | 18:21 |
| vishy | tr3buchet: does that seem reasonable? | 18:22 |
| tr3buchet | it does | 18:22 |
| tr3buchet | just realized my mistake | 18:22 |
| tr3buchet | you want to me to do it and push up the patchset? | 18:22 |
| *** adjohn has joined #openstack-dev | 18:22 | |
| tr3buchet | actually | 18:22 |
| tr3buchet | it would probably be better to just pass it in | 18:23 |
| vishy | tr3buchet: yeah because it seems weird to call release_ip | 18:23 |
| vishy | and have it release all of the ips | 18:23 |
| vishy | I will pass it in. | 18:23 |
| tr3buchet | yeah only want to do that in the case where that's the goal | 18:24 |
| tr3buchet | as in setup_networks_on_host(teardown=True) | 18:24 |
| tr3buchet | this requires a few changes, since it alters the method signature | 18:24 |
| *** bepernoot has joined #openstack-dev | 18:25 | |
| *** adjohn has quit IRC | 18:27 | |
| *** darraghb has quit IRC | 18:28 | |
| *** andrewbogott has joined #openstack-dev | 18:28 | |
| *** adjohn has joined #openstack-dev | 18:29 | |
| *** bepernoot has quit IRC | 18:31 | |
| vishy | tr3buchet: i just noticed we might be able to get rid of rpc_allocate | 18:31 |
| tr3buchet | vishy: why's that? | 18:31 |
| vishy | tr3buchet: seems like we could just pass the network object instead of the id | 18:31 |
| tr3buchet | network objects can't make it through rpc | 18:31 |
| vishy | tr3buchet: we should fix that :) | 18:32 |
| vishy | tr3buchet: here are my changes http://paste.openstack.org/show/6866/ | 18:32 |
| vishy | pushed up the patchset | 18:32 |
| tr3buchet | i see | 18:35 |
| vishy | tr3buchet: ah missed a couple references to teardown. Hopefully that gets it :) | 18:35 |
| tr3buchet | i have also fixed it | 18:35 |
| tr3buchet | you | 18:35 |
| *** bepernoot has joined #openstack-dev | 18:37 | |
| tr3buchet | vishy: http://pastie.org/3535465 | 18:37 |
| tr3buchet | i'll push that up | 18:37 |
| tr3buchet | it calls teardown with the address when deallocating | 18:38 |
| tr3buchet | and for each address when tearing them all down | 18:38 |
| tr3buchet | vishy: thoughts? | 18:38 |
| vishy | tr3buchet: sure, not sure what the addresses on setup is though | 18:39 |
| tr3buchet | vishy: look at the diff | 18:39 |
| tr3buchet | i just added it as an optional parameter | 18:39 |
| tr3buchet | so when setup_networks_on_host() calls it with the same signature as it does the _teardown, the address is ignored just like vif | 18:41 |
| vishy | tr3buchet: I'm not totally convinced about calling multiple times on teardown | 18:42 |
| vishy | tr3buchet: in any case doesn't address need to be optional on teardown? | 18:43 |
| vishy | or are you passing it in all places now? | 18:43 |
| tr3buchet | all calling it multiple times does it call self.driver.release_dhcp() multiple times | 18:43 |
| tr3buchet | and that's necessary whether the loop is above or below | 18:43 |
| *** novas0x2a|laptop has joined #openstack-dev | 18:43 | |
| tr3buchet | vishy: i am passing it in all places | 18:44 |
| *** bepernoot has quit IRC | 18:44 | |
| *** flaviamissi has joined #openstack-dev | 18:46 | |
| vishy | tr3buchet: ok i'm cool with it | 18:47 |
| tr3buchet | alrighty | 18:47 |
| tr3buchet | when tests finish i'll push | 18:47 |
| tr3buchet | i fixed the if multihost issue as well | 18:47 |
| *** devcamcar has joined #openstack-dev | 18:49 | |
| *** mattstep has quit IRC | 18:51 | |
| *** mattstep has joined #openstack-dev | 18:51 | |
| tr3buchet | vishy: tests passed, pushed | 18:52 |
| vishy | cool i think we might have one or two more bugs in there but we're getting close | 18:52 |
| tr3buchet | vishy: good deal, keep me posted | 18:54 |
| *** markvoelker has joined #openstack-dev | 18:55 | |
| *** dolphm has quit IRC | 18:59 | |
| vishy | tr3buchet: you overwrote my cleanup of your commit message :'( | 19:00 |
| tr3buchet | :( | 19:00 |
| tr3buchet | shit | 19:00 |
| tr3buchet | vishy: sorry i didn't notice it | 19:00 |
| tr3buchet | :( | 19:00 |
| tr3buchet | still have it locally? | 19:01 |
| vishy | tr3buchet: no worries, just added a summary at the top and moved the bug reference into bullets | 19:01 |
| tr3buchet | k | 19:02 |
| vishy | tr3buchet: summary was: Setup and teardown networks during migration | 19:02 |
| jeblair | https://review.openstack.org/#patch,sidebyside,4646,4,/COMMIT_MSG | 19:02 |
| jeblair | vishy, tr3buchet: hard to get rid of anything with git. :) | 19:02 |
| tr3buchet | jeblair: cool | 19:03 |
| tr3buchet | vishy: fixed | 19:03 |
| Ryan_Lane | QueuePool limit of size 10 overflow 10 reached <— do I need to increase my pool in nova, or in mysql? | 19:03 |
| soren | mtaylor: CI meeting? | 19:04 |
| *** dolphm has joined #openstack-dev | 19:05 | |
| vishy | tr3buchet: thanks | 19:06 |
| vishy | Ryan_Lane: haven't seen that one | 19:06 |
| Ryan_Lane | well, I just rebooted 115 instances | 19:06 |
| vishy | but i would guess in mysql? | 19:06 |
| Ryan_Lane | they are all slamming the database, but the database is fine | 19:06 |
| vishy | Ryan_Lane: actually hold on | 19:06 |
| Ryan_Lane | I'm thinking it's sqlalchemy | 19:06 |
| Ryan_Lane | http://docs.sqlalchemy.org/en/latest/core/pooling.html | 19:06 |
| Ryan_Lane | ah | 19:08 |
| Ryan_Lane | FLAGS.sql_max_pool_size | 19:08 |
| *** ecarlin has joined #openstack-dev | 19:09 | |
| vishy | Ryan_Lane: you must be on really old code | 19:10 |
| Ryan_Lane | diablo | 19:10 |
| vishy | that flag only existed temporarily in diablo | 19:10 |
| vishy | i don't think it is even in stable/diablo | 19:10 |
| Ryan_Lane | I'm on stable/diablo | 19:10 |
| Ryan_Lane | so, it's there | 19:10 |
| vishy | Ryan_Lane: incorrect | 19:11 |
| vishy | vishvananda@firefly-2:~/os/nova (stable/diablo)$ ack sql_max_pool | 19:11 |
| vishy | vishvananda@firefly-2:~/os/nova (stable/diablo)$ | 19:11 |
| Ryan_Lane | hm | 19:11 |
| vishy | it was there during the diablo release and removed almost immediately in stable/diablo | 19:11 |
| Ryan_Lane | this brings up something I wanted to talk about :) | 19:12 |
| *** dolphm_ has joined #openstack-dev | 19:12 | |
| Ryan_Lane | is it going to be possible to upgrade to essex from diablo by moving from lucid to precise? | 19:12 |
| Ryan_Lane | I don't have a clue how I'm supposed to do distro version upgrades without using brand new hardware every time | 19:13 |
| Ryan_Lane | and even then, I don't know how it'll be possible | 19:13 |
| Ryan_Lane | will essex and diablo be able to run together? | 19:13 |
| Ryan_Lane | this is why I'm on a weird version of diablo | 19:14 |
| Ryan_Lane | I followed the upgrade instructions in the documentation | 19:14 |
| *** mnewby has joined #openstack-dev | 19:15 | |
| *** mnewby has joined #openstack-dev | 19:15 | |
| *** mattstep has quit IRC | 19:15 | |
| *** mattstep has joined #openstack-dev | 19:16 | |
| Ryan_Lane | hm. upping the pool size doesn't help anyway, because eventlet then dies from using too many files | 19:16 |
| *** ecarlin has quit IRC | 19:17 | |
| *** dolphm has quit IRC | 19:18 | |
| *** gkotton has joined #openstack-dev | 19:18 | |
| adam_g | oh | 19:24 |
| vishy | tr3buchet: ok new issue | 19:25 |
| *** dolphm_ has quit IRC | 19:25 | |
| vishy | tr3buchet: we call setup_network before we've changed the host in the database | 19:25 |
| vishy | tr3buchet: so it doesn't get dhcp updated | 19:26 |
| vishy | tr3buchet: so we could call setup_network again in post migration | 19:26 |
| *** johnpur has joined #openstack-dev | 19:26 | |
| *** ChanServ sets mode: +v johnpur | 19:26 | |
| vishy | tr3buchet: or we could force dnsmasq to include vms that are migrating to the host | 19:26 |
| vishy | tr3buchet: I think i prefer calling setup again after the move because it is simpler | 19:27 |
| vishy | thoughts? | 19:27 |
| tr3buchet | vishy: i agree | 19:27 |
| tr3buchet | vishy: which function are you talking about | 19:28 |
| tr3buchet | i wish i had the diagram i made :( | 19:28 |
| tr3buchet | there were a lot of functions for both resize/migrate and live migrate | 19:28 |
| *** eglynn__ has joined #openstack-dev | 19:28 | |
| *** vizsla has joined #openstack-dev | 19:28 | |
| tr3buchet | vishy: i actually added a host parameter | 19:29 |
| *** gyee has quit IRC | 19:29 | |
| tr3buchet | for the times when the host wasn't available or correct in the db | 19:29 |
| tr3buchet | vishy: did i get one of them wrong? | 19:29 |
| vishy | tr3buchet: no we need to | 19:29 |
| tr3buchet | vishy: wha? | 19:29 |
| vishy | setup on dest host | 19:29 |
| vishy | then migrate | 19:29 |
| vishy | then setup again | 19:29 |
| vishy | the first time to create the net, the second time to update dhcp | 19:30 |
| tr3buchet | why not just setup after migrate? | 19:30 |
| vishy | well you need to create the vlan and bridge before migrate (i think) | 19:30 |
| *** gyee has joined #openstack-dev | 19:31 | |
| vishy | although i guess there is already code in live migrate to do that via vif plug | 19:31 |
| vishy | tr3buchet: so i guess we could do it after? | 19:31 |
| tr3buchet | sounds like it should work | 19:31 |
| *** eglynn_ has quit IRC | 19:31 | |
| tr3buchet | shall i push a patch or do you want to play with a bit first and see if it works? | 19:31 |
| *** x86brandon_ has joined #openstack-dev | 19:32 | |
| vishy | tr3buchet: ah i see, you do the dest setup at the end in resize | 19:32 |
| vishy | tr3buchet: so yeah it looks like it is just in the wrong place | 19:32 |
| vishy | for live migrate | 19:33 |
| *** vizsla has quit IRC | 19:33 | |
| vishy | tr3buchet: let me try moving it first | 19:33 |
| tr3buchet | vishy: well it was a little bit different | 19:33 |
| tr3buchet | resize and live migrate work very differently | 19:33 |
| tr3buchet | which i think is unfortunate given what they do | 19:33 |
| vishy | tr3buchet: yes they need to use some common code :) | 19:33 |
| tr3buchet | :D | 19:34 |
| *** Ryan_Lane has quit IRC | 19:35 | |
| russellb | pixelbeat: around? | 19:36 |
| *** x86brandon has quit IRC | 19:36 | |
| *** x86brandon_ is now known as x86brandon | 19:36 | |
| vishy | tr3buchet: ok we do need it in both places | 19:38 |
| tr3buchet | kk | 19:38 |
| tr3buchet | where is the other place? | 19:38 |
| vishy | tr3buchet: only because multi_host mode doesn't create the bridge in plug_vifs | 19:39 |
| tr3buchet | ah ok | 19:39 |
| vishy | tr3buchet: it waits for the network code to do it | 19:39 |
| vishy | so first call is to create the bridge, setup an ip for dnsmasq etc. | 19:39 |
| vishy | the second call is to update dhcp | 19:39 |
| vishy | tr3buchet: it is kind of nasty that we can't do it in one call | 19:39 |
| tr3buchet | yeah i agree | 19:40 |
| *** ecarlin has joined #openstack-dev | 19:40 | |
| tr3buchet | multi_host introduced some strange behaviors | 19:40 |
| vishy | tr3buchet: 612 if network_ref['multi_host'] and FLAGS.host != instance_ref['host']: | 19:40 |
| vishy | 613 continue | 19:40 |
| vishy | that is the line that is messing things up | 19:41 |
| vishy | so the instance_ref still has the old host set pre-migrate | 19:41 |
| tr3buchet | where is that? | 19:41 |
| vishy | in linux_net update_dhcp | 19:41 |
| tr3buchet | ah ok | 19:41 |
| tr3buchet | :( | 19:41 |
| tr3buchet | yeah someone probably saw multihost and was like "wtf? uh.... skip!" | 19:42 |
| vishy | tr3buchet: I think we have a reference to the migration host right? | 19:42 |
| tr3buchet | yeah they are stored in the migrations table | 19:42 |
| tr3buchet | but i was just using self.host most of the time the live_migration code | 19:42 |
| vishy | tr3buchet: i can extend that check to look for a migration as well | 19:42 |
| tr3buchet | since it's explicit about which functions are run on which host | 19:42 |
| vishy | tr3buchet: kind of ugly | 19:42 |
| tr3buchet | yeah | 19:42 |
| vishy | tr3buchet: I don't want to update the instance host in pre live migrate | 19:43 |
| tr3buchet | i can't remmeber when the host was actually updated | 19:44 |
| tr3buchet | actually wait | 19:44 |
| tr3buchet | i don't htink live migration uses the migrations table | 19:45 |
| *** Ryan_Lane has joined #openstack-dev | 19:45 | |
| tr3buchet | vishy: should be able to just call setup_networks in post_live_migration_at_destination | 19:46 |
| tr3buchet | and pass in self.host | 19:46 |
| Kiall | The migrations take is about database migrations I believe.. not instance migrations | 19:46 |
| Kiall | table* | 19:46 |
| *** jog0 has joined #openstack-dev | 19:47 | |
| vishy | tr3buchet: yeah i'm trying to avoid calling it twice though | 19:47 |
| vishy | tr3buchet: sigh i guess we can call it twice for now and just put in a note | 19:47 |
| vishy | tr3buchet: looking for migrating instance in linux_net seems pretty dirty | 19:47 |
| tr3buchet | Kiall: we are talking about the table migrations, the sqlalchemy model for it is ~637 in nova/db/sqlalchemy/models.py | 19:48 |
| tr3buchet | vishy: it's horribly dirty | 19:48 |
| vishy | tr3buchet: Ok. I guess we have to clean all this up later :) | 19:49 |
| vishy | tr3buchet: We can just call it twice for now | 19:49 |
| *** johnpur has quit IRC | 19:49 | |
| tr3buchet | alright | 19:49 |
| tr3buchet | running tests to make sure nothing breaks | 19:49 |
| *** ecarlin has quit IRC | 19:50 | |
| tr3buchet | vishy: http://pastie.org/3535909 | 19:51 |
| tr3buchet | we're on the same page? | 19:51 |
| hub_cap | Hey im trying to call keystone on behalf of a user token, IE, i have a admin user/pass, and im gettin a NotImplemented Error, as per the code. Shoudl this be implemented? Or am i hitting a path i shoudlnt be? This is whats generated from the novaclient based on the | 19:55 |
| hub_cap | http://paste.openstack.org/show/6887/ | 19:55 |
| hub_cap | *based on the call :D | 19:55 |
| *** GheRivero_ has joined #openstack-dev | 19:57 | |
| *** andrewbogott has quit IRC | 19:57 | |
| *** stuntmachine has quit IRC | 19:58 | |
| hub_cap | dolph updated this code, whats his name on irc? | 19:59 |
| *** hashar has joined #openstack-dev | 19:59 | |
| *** bepernoot has joined #openstack-dev | 20:00 | |
| *** rkukura has quit IRC | 20:00 | |
| *** dolphm has joined #openstack-dev | 20:00 | |
| *** andrewbogott has joined #openstack-dev | 20:00 | |
| hub_cap | it was changed in this rev | 20:01 |
| hub_cap | https://github.com/openstack/keystone/commit/510061e0807d17234d533a83579074f7d0a42a77 | 20:01 |
| *** gkotton has quit IRC | 20:01 | |
| hub_cap | and was apparently "broken and redundant" based on his commit | 20:01 |
| *** dolphm has quit IRC | 20:04 | |
| *** dolphm has joined #openstack-dev | 20:05 | |
| *** bepernoot has quit IRC | 20:09 | |
| *** dolphm has quit IRC | 20:11 | |
| *** ecarlin has joined #openstack-dev | 20:14 | |
| *** ecarlin has quit IRC | 20:15 | |
| *** novas0x2a|laptop has quit IRC | 20:15 | |
| *** ecarlin has joined #openstack-dev | 20:15 | |
| cp16net | apparently reverting that code from hub_cap(keystone commit url) allows the authorization to work | 20:19 |
| cp16net | so it doesnt look like it was broken | 20:19 |
| hub_cap | it was neither broken nor redundant apparently cp16net ;) | 20:19 |
| cp16net | not sure what the reasoning was for the change to begin with | 20:20 |
| hub_cap | heckj: mr keystone we have a issue w/ keystone | 20:20 |
| hub_cap | maybe u can help us out, or point us in the right direction | 20:21 |
| *** dolphm has joined #openstack-dev | 20:26 | |
| *** shevek_ has quit IRC | 20:27 | |
| hub_cap | dolphm: ping, ive got a keystone Q | 20:27 |
| dolphm | hub_cap: hit me | 20:28 |
| hub_cap | so you removed some code here | 20:28 |
| hub_cap | https://github.com/openstack/keystone/commit/510061e0807d17234d533a83579074f7d0a42a77 | 20:28 |
| dolphm | hub_cap: yeah | 20:28 |
| hub_cap | python novaclient is calling that token/BLAH/endpoints to get authorization on behalf of a proxy_token | 20:28 |
| dolphm | hub_cap: it was broken, and the content was redundant with GET /tokens/{token_id} | 20:29 |
| hub_cap | so when i try to use proxy_token in novaclient, im getting a borked'd'd response | 20:29 |
| hub_cap | ok so it should in fact just be calling that? | 20:29 |
| hub_cap | instead of ..../endpoints | 20:29 |
| dolphm | hub_cap: yes, or it should already have the data from /endpoints | 20:30 |
| dolphm | hub_cap: can you point me to where it's being used? | 20:30 |
| hub_cap | sure sec | 20:31 |
| *** kbringard has quit IRC | 20:31 | |
| *** GheRivero_ has quit IRC | 20:31 | |
| *** kbringard has joined #openstack-dev | 20:32 | |
| hub_cap | dolphm: http://paste.openstack.org/show/6893/ | 20:33 |
| hub_cap | its in _fetch_endpoints_from_auth() | 20:33 |
| *** bengrue has quit IRC | 20:34 | |
| *** anotherjesse is now known as anotherjesse_zz | 20:34 | |
| hub_cap | dolphm: http://paste.openstack.org/show/6894/ | 20:35 |
| hub_cap | thats the call from python-novaclient when using proxy_token | 20:35 |
| dolphm | hub_cap: i'm looking through code, give me a few | 20:37 |
| hub_cap | sure thang | 20:37 |
| hub_cap | basically dolphm, if you have a proxy token, it calls that dude | 20:37 |
| *** anotherjesse_zz is now known as anotherjesse | 20:38 | |
| *** mikeyp has quit IRC | 20:38 | |
| *** roge has joined #openstack-dev | 20:40 | |
| dolphm | hub_cap: yeah, i don't see why this couldn't just call GET /tokens/{token_id} | 20:42 |
| hub_cap | so dolphm the old code added service catalog stuff to the endpoints call, but the non endpoints (tokens/tok_id) does not | 20:43 |
| hub_cap | thats why ^ ^ | 20:43 |
| hub_cap | its relying on teh service_catalog being _in_ the response | 20:43 |
| hub_cap | which has the endpoints | 20:43 |
| hub_cap | but tokens/{token_id} only returns a subset, let me paste | 20:43 |
| hub_cap | http://paste.openstack.org/show/6896/ | 20:43 |
| hub_cap | omg thats ugly paste | 20:43 |
| hub_cap | but thats what the tokens/token_id returns, if yo look at the code you rremoved it also gets the service_catalog | 20:44 |
| hub_cap | i do see the reduncency in the token_api.get_token code tho | 20:44 |
| dolphm | hub_cap: how is it a subset? | 20:44 |
| hub_cap | dolphm: return {'token': {'serviceCatalog': self._format_catalog(catalog_ref)}} | 20:44 |
| hub_cap | there is no serviceCatalog in that anymroe right? | 20:44 |
| hub_cap | thats the return from the code you removed (https://github.com/openstack/keystone/commit/510061e0807d17234d533a83579074f7d0a42a77) | 20:45 |
| hub_cap | it gets the token AND the serviceCatalog and returns them together | 20:45 |
| dolphm | wtf there should be | 20:45 |
| hub_cap | the tokens/token_id only gets the token_api.get_token | 20:45 |
| *** Ghe_Rivero has joined #openstack-dev | 20:46 | |
| hub_cap | looking @ the format_token method (the last bit of code from the validate_token method), it does not mention a serviceCatalog at all | 20:47 |
| hub_cap | looks like the serviceCatalog is _only_ in authenticate now | 20:48 |
| dolphm | hub_cap: i have no idea why that is | 20:48 |
| hub_cap | ie _format_catalog is only called from teh authenticate method in the service.py dolphm | 20:48 |
| hub_cap | haha thats what im wondering too :D | 20:48 |
| hub_cap | if u tell me how best to fix it ill be glad to offer up a patch | 20:49 |
| cp16net | haha | 20:49 |
| hub_cap | as in, what woudl conform best to the keystone code | 20:49 |
| *** rkukura has joined #openstack-dev | 20:51 | |
| *** ecarlin_ has joined #openstack-dev | 20:53 | |
| *** ecarlin_ has quit IRC | 20:53 | |
| *** mattstep has quit IRC | 20:53 | |
| *** vincentricci has quit IRC | 20:54 | |
| *** mattstep has joined #openstack-dev | 20:54 | |
| *** ecarlin has quit IRC | 20:57 | |
| *** novas0x2a|laptop has joined #openstack-dev | 21:01 | |
| *** jaypipes has joined #openstack-dev | 21:02 | |
| *** nati has joined #openstack-dev | 21:02 | |
| *** nati is now known as nati-ueno | 21:06 | |
| *** adjohn has quit IRC | 21:07 | |
| *** dprince has quit IRC | 21:09 | |
| nati-ueno | ping mtaylor | 21:09 |
| *** cp16net has quit IRC | 21:12 | |
| *** uvirtbot` is now known as uvirtbot | 21:13 | |
| *** Ryan_Lane has quit IRC | 21:13 | |
| vishy | tr3buchet: can you recheck https://review.openstack.org/#change,4824 and update your -2? | 21:14 |
| tr3buchet | sure | 21:14 |
| tr3buchet | vishy: looking over it now | 21:15 |
| anotherjesse | jeblair: raise Exception("No ready nodes") again :( | 21:15 |
| jeblair | drat | 21:16 |
| tr3buchet | vishy: what's your opinion on fixed_ip_ref.network_id vs fixed_ip_ref['network_id'] ? | 21:16 |
| mtaylor | nati-ueno: pong | 21:16 |
| jeblair | https://jenkins.openstack.org/job/devstack-launch-vms/10395/console | 21:16 |
| jeblair | anotherjesse: ^ rackspace doesn't love us anymore | 21:17 |
| mtaylor | pvo: ^^ | 21:17 |
| pvo | jeblair: of course we do. Let me see what I can do. | 21:18 |
| * mtaylor hands pvo beer | 21:18 | |
| jeblair | :) | 21:18 |
| jeblair | pvo: it looks like we have exactly 100 servers | 21:19 |
| pvo | directing to side channel... | 21:19 |
| *** bepernoot has joined #openstack-dev | 21:20 | |
| heckj | hub_cap: just getting back - reading backlog | 21:23 |
| hub_cap | heckj: thats fine ive got a fix im about to propose | 21:24 |
| hub_cap | im also talking w/ termie | 21:24 |
| hub_cap | here is my bug report | 21:24 |
| hub_cap | heckj: https://bugs.launchpad.net/keystone/+bug/948439 | 21:25 |
| uvirtbot | Launchpad bug 948439 in keystone "belongs_to token calls need a serviceCatalog returned" [Undecided,In progress] | 21:25 |
| hub_cap | heey that one! | 21:25 |
| heckj | hub_cap: heh, sounds like you don't need me then! :-) I'll mark it up confirmed and target -> rc1. | 21:25 |
| hub_cap | word | 21:25 |
| tr3buchet | vishy: also, is it even possible for fixed_ips to not have a network? i'm not aware of it being possible | 21:25 |
| heckj | hub_cap - you good on how to solve? | 21:25 |
| hub_cap | heckj: http://paste.openstack.org/show/6903/ | 21:26 |
| hub_cap | i believe so | 21:26 |
| vishy | tr3buchet: in db layer +1 to . | 21:26 |
| *** adjohn has joined #openstack-dev | 21:26 | |
| heckj | hub_cap: cool - thanks | 21:26 |
| tr3buchet | vishy: right | 21:26 |
| vishy | otherwise [] | 21:27 |
| *** dubsquared has joined #openstack-dev | 21:27 | |
| vishy | tr3buchet: a fixed ip should always have a network afaik | 21:27 |
| tr3buchet | yeah | 21:27 |
| tr3buchet | so i'm thinking this whole if not fixed_ip_ref.network_id section is not needed | 21:27 |
| johan_-_ | gate-integration-tests-devstack-vm appears to be broken again | 21:28 |
| mtaylor | johan_-_: yes. it's being looked in to | 21:29 |
| johan_-_ | last 11 jobs all failed with "No ready nodes" | 21:29 |
| mtaylor | johan_-_: we've hit a usage cap on rackspace cloud | 21:29 |
| kbringard | has anyone seen this error before? | 21:29 |
| kbringard | http://paste.openstack.org/show/6906/ | 21:29 |
| hub_cap | so does anyone _own_ python-novaclient persay? | 21:29 |
| mtaylor | hub_cap: it's a sub-project of nova | 21:29 |
| hub_cap | ok so ping vishy and co to get approval i assume? | 21:30 |
| mtaylor | hub_cap: yup | 21:30 |
| vishy | tr3buchet: lets leave it for now | 21:30 |
| hub_cap | <3 mtaylor | 21:30 |
| tr3buchet | vishy: ok i'll just approve it then | 21:30 |
| vishy | tr3buchet: just because there was a point when fixed_ips could be in a pool that wasn't associated to a network | 21:30 |
| tr3buchet | k | 21:30 |
| vishy | tr3buchet: i think that is gone now, but i don't want to risk breaking something | 21:30 |
| *** anotherjesse is now known as anotherjesse_zz | 21:30 | |
| tr3buchet | vishy: yeah we can look at it later | 21:31 |
| tr3buchet | although i'll surely forget ;) | 21:31 |
| *** adjohn has quit IRC | 21:31 | |
| tr3buchet | tickets purchased for sfo! | 21:31 |
| *** adjohn has joined #openstack-dev | 21:31 | |
| *** vladimir3p has joined #openstack-dev | 21:32 | |
| *** bhall has joined #openstack-dev | 21:32 | |
| *** bhall has quit IRC | 21:32 | |
| *** bhall has joined #openstack-dev | 21:32 | |
| jeblair | i'm still working on recovering the devstack test, it will be a few more minutes before it's ready to go again. | 21:36 |
| *** ecarlin has joined #openstack-dev | 21:36 | |
| *** bepernoot has quit IRC | 21:36 | |
| kbringard | it looks like the sql_max_pool_size may be too small | 21:37 |
| *** ecarlin has quit IRC | 21:37 | |
| *** andrewbogott has quit IRC | 21:41 | |
| *** sniperd has quit IRC | 21:42 | |
| *** sniperd has joined #openstack-dev | 21:43 | |
| termie | is dave walker on here? | 21:43 |
| jeblair | Daviey: ^ | 21:43 |
| *** bepernoot has joined #openstack-dev | 21:44 | |
| *** sandywalsh has quit IRC | 21:45 | |
| Daviey | termie: o/ | 21:46 |
| ayoung | heckj, question on the auth_token code, and anotherjesse is in zz mode... | 21:46 |
| heckj | ayoung: yo! | 21:46 |
| *** bepernoot has quit IRC | 21:46 | |
| ayoung | app_factory is the entry point to the whole use case, right? | 21:46 |
| * heckj looks | 21:47 | |
| ayoung | that is what is called by our wsgi framework to create the handler for the request? | 21:47 |
| vishy | kbringard: that param doesn't exist | 21:47 |
| ayoung | I can't find where that is called, so I am assuming it is from some library that we use | 21:47 |
| Daviey | termie: wassup? | 21:47 |
| vishy | kbringard: I had this discussion with Ryan_Lane earlier | 21:47 |
| heckj | ayoung: it's imported and used in other projects, calling down into __call__() to handle the requests and "do it's thang" | 21:47 |
| heckj | ayoung: glance and nova specifically | 21:48 |
| dolphm | ayoung: app factory is called by wsgi, essentially, and the __call__ method is called when a request gets that far in the pipeline | 21:48 |
| ayoung | AuthProtocall.__call__()? | 21:48 |
| kbringard | vishy: I must have missed it | 21:48 |
| kbringard | grep sql_max_pool_size flags.py | 21:48 |
| kbringard | DEFINE_integer('sql_max_pool_size', 10, | 21:48 |
| kbringard | so that's not used? | 21:48 |
| ayoung | dolphm, so the config object it takes is what? | 21:49 |
| vishy | no it doesn't exist | 21:49 |
| vishy | in stable/diablo or essex | 21:49 |
| kbringard | ooooooooo, stable diablo… what about Vanilla diablo? | 21:49 |
| kbringard | was it removed? | 21:49 |
| vishy | yes | 21:49 |
| kbringard | ah | 21:49 |
| kbringard | that makes sense | 21:49 |
| kbringard | OK | 21:49 |
| dolphm | ayoung: it's populated from the compiled paste config | 21:49 |
| kbringard | well, assuming I can't upgrade | 21:49 |
| vishy | because the db_pool was mega broken | 21:49 |
| apevec | ayoung, it's api-paste | 21:49 |
| heckj | ayoung: yeah, one of those classes is instantiated, and that instance gets the __call__() | 21:49 |
| kbringard | do you have any suggestions? | 21:49 |
| kbringard | or is the only route to upgrade to stable? | 21:49 |
| vishy | sure, set it higher :) | 21:49 |
| ayoung | dolphm, I'm trying to figure out a limited something to pass to return AuthProtocol(None, conf) | 21:49 |
| vishy | but we ran into lots of issues with that code | 21:50 |
| kbringard | yea, fair enough | 21:50 |
| *** davlap has joined #openstack-dev | 21:50 | |
| kbringard | thanks :-) | 21:50 |
| heckj | soren: o/ | 21:50 |
| kbringard | vishy: ^^ | 21:50 |
| ayoung | Oh, wait....we have a TokenAuthMiddlewareTest class in test_middleware.py | 21:50 |
| ayoung | I take it I should be modifying that | 21:51 |
| heckj | ayoung: or making another like it and extending it, yeah. | 21:51 |
| dolphm | ayoung: not sure what you mean | 21:51 |
| soren | heckj: o/ | 21:51 |
| soren | heckj: 'sup? | 21:51 |
| heckj | ayoung: jaypipes mentioned he had some relevant tests you might be able to rip off from Glance, haven't read/found something easy to steal there though | 21:51 |
| ayoung | dolphm, I'm trying to handle https://bugs.launchpad.net/keystone/+bug/943722 | 21:51 |
| uvirtbot | Launchpad bug 943722 in keystone "token auth needs unit tests" [Critical,Confirmed] | 21:52 |
| heckj | soren: re: https://review.openstack.org/#change,4923, what do you (zul, whomever) want to do to work around the additional testing we've enabled that needs internet access. | 21:52 |
| dolphm | ayoung: ah cool | 21:52 |
| heckj | soren: did you want to do something where you're presetting the environ, or would you prefer a flag to disable that additional integration testing specifically for your building | 21:52 |
| ttx | anticw: ping | 21:53 |
| dolphm | ayoung: most of the other middleware is tested in isolation because they support no app being passed in, and then the tests just call methods directly | 21:53 |
| dolphm | ayoung: for testing, you don't need to call the app factory | 21:53 |
| ayoung | dolphm, nah, I was trying to figure out what the config object was supposed to look like | 21:54 |
| ttx | anticw: sudo sanitizes PYTHON_PATH before executing nova-rootwrap | 21:54 |
| dolphm | ayoung: AuthMiddleware(None, {}) might get you somewhere | 21:54 |
| soren | heckj: The mechanics for adjusting the behaviour isn't really important. Environment variable, command line option or whatever else is equally easy to apply. | 21:54 |
| dolphm | ayoung: it's a dict, ultimately | 21:54 |
| Daviey | heckj: I really can't that integration testing belongs as a 'unit test' | 21:55 |
| heckj | soren: Okay - to be clear - you're not pre-manipulating an environment, you need something from the keystone project to turn off that testing for your use, yes? | 21:55 |
| heckj | Daviey: totally agree it's not a unit test, we're just using the unit test framework to run it. | 21:56 |
| *** markmc has quit IRC | 21:56 | |
| Daviey | heckj: And you can't see that it smells bad? | 21:56 |
| *** joearnold has joined #openstack-dev | 21:56 | |
| soren | heckj: I'm not entirely sure what "pre-manipulating an environment" means :) | 21:56 |
| heckj | Daviey: totally want to see this in Tempest in the future, but in general we're not quite ready to gate commits on tempest as yet, and it was critical for us to have inline for gating. | 21:56 |
| heckj | Daviey: actually, additional testing of any form always smells great to me. I'm not applying the semantics of unit-only when invoking run_tests, but I think you are. Not sure why. | 21:57 |
| ayoung | dolphm, OK, that got me started, thanks | 21:57 |
| vishy | jeblair: when devstack is back up, is there any easy way to figure out the branches that need to be retriggered? | 21:57 |
| heckj | soren: you don't want to pre-load keystoneclient into the right place for the tests - they don't check out if the repo already exists in the right location. | 21:58 |
| dolphm | ayoung: ping me if you need more help on that bug | 21:58 |
| vishy | johan_-_: in regards to the logging race condition. Does nose use a threadpool to run tests? | 21:58 |
| ayoung | dolphm, what specific functionality should I be looking to test? | 21:58 |
| Daviey | heckj: My requirement is that when I run ./run_tests.sh -N <-- it happes offline. If that means you make it go down a different path, that is ok. | 21:58 |
| dolphm | ayoung: the three basic flows: request comes in without X-Auth-Token, request comes in with valid X-Auth-Token, request comes in with invalid X-Auth-Token | 22:00 |
| *** sandywalsh has joined #openstack-dev | 22:00 | |
| dolphm | ayoung: and then delay_auth_decision=True should be tested in those scenarios | 22:00 |
| heckj | Daviey: you've been clear with what you need, it's the "smells bad" comments that aren't helping. I will get you something that allows you to run the testing completely offline. I don't promise it'll be the "-N", as that's meant to be "no virtualenv", not "entirely offline". Totally get the need, and we'll get it covered for you. | 22:01 |
| dolphm | ayoung: and you could unit test the individual methods in there, but i like to start with business cases | 22:01 |
| *** lts has quit IRC | 22:02 | |
| Daviey | heckj: Okay, the reason i say it 'smells bad'.. is because the test_suite will return different values based on time. | 22:02 |
| *** andrewbogott has joined #openstack-dev | 22:02 | |
| *** andrewbogott has joined #openstack-dev | 22:02 | |
| Daviey | If i want to bisect something in the future, it'll still run the test against what is in current trunk | 22:02 |
| Daviey | heckj: This seems somewhat unsuitable. | 22:03 |
| ayoung | Daviey has been reading too much Martin Fowler | 22:03 |
| heckj | Daviey: that's actually exactly what we intend, but I totally get why you dont ever want to do that in packaging | 22:03 |
| *** andrewbogott has quit IRC | 22:03 | |
| *** roge has quit IRC | 22:04 | |
| Daviey | heckj: a snapshot giving different output based on external libraries development? | 22:04 |
| heckj | Daviey: we've had LOTs (too much) breakage because we didn't enforce integration testing between multiple projects. We're getting better, and our additions to run_tests (while they suck for you need) are totally what we mean to do to drive that farther forward | 22:04 |
| *** flaviamissi has quit IRC | 22:05 | |
| *** roge has joined #openstack-dev | 22:05 | |
| Daviey | heckj: well sure, but it just seems kinda ugly. :/ | 22:07 |
| jeblair | vishy: I will go through the console output on the devstack job looking for the error (https://jenkins.openstack.org/job/gate-integration-tests-devstack-vm/2203/console) and hit retrigger all. | 22:07 |
| jeblair | vishy: https://review.openstack.org/#q,status:open+verified%253D-1+approved%253D1,n,z | 22:07 |
| Daviey | heckj: But anyway, if you can provide a line which does do offline testing, I am happy. | 22:07 |
| heckj | Daviey: Ugly, I'll grant you. | 22:07 |
| jeblair | vishy: ^ that's another way of finding them. | 22:07 |
| heckj | Daviey: just logged https://bugs.launchpad.net/keystone/+bug/948495 to track what you need, and I've assigned it to myself to get it done here in the "really near soon right now" timeframe. | 22:08 |
| uvirtbot | Launchpad bug 948495 in keystone "need to provide totally offline ./run_tests.sh to support downstream packaging" [High,Confirmed] | 22:08 |
| *** rods has quit IRC | 22:08 | |
| *** littleidea has quit IRC | 22:08 | |
| *** Ryan_Lane has joined #openstack-dev | 22:08 | |
| Daviey | heckj: thanks! | 22:09 |
| ayoung | dolphm, is there a standard Mock start_response object I should use? | 22:09 |
| *** joesavak has quit IRC | 22:09 | |
| heckj | dolphm: actually, I'd like to know that one myself… how have we been doing that in other systems? | 22:09 |
| notmyname | ttx: https://github.com/openstack/swift/commit/12b118ee2cc7025d5f8880d01fdc16ec316345eb for 1.4.7 | 22:09 |
| heckj | Daviey: better than a patch you slap in after the fact to remove or deactivate tests. :-) I can't tell you the fun I had packaging the Django 1.4 beta release... | 22:10 |
| dolphm | ayoung: will None work? | 22:10 |
| Daviey | heckj: The racey unit tests? :) | 22:11 |
| ayoung | checking... | 22:11 |
| heckj | Daviey: yeah | 22:11 |
| dolphm | ayoung: or is it referenced/manipulated? i feel like it's just returned | 22:11 |
| Daviey | heckj: yeah, that hit me aswell. | 22:11 |
| ayoung | dolphm, pretty sure it is munged in the response. | 22:11 |
| vishy | jeblair: nifty | 22:12 |
| gyee | heckj, the design summit registration is full | 22:12 |
| gyee | how do I send cash to? :) | 22:12 |
| *** joearnold has quit IRC | 22:13 | |
| heckj | gyee: lemme see what I can do | 22:14 |
| jeblair | vishy, johan_-_: devstack is running again, and i've retrigged (most of) the jobs | 22:14 |
| ayoung | dolphm, yeah, that is the generated response, should probably be some form of Mock object. | 22:15 |
| johan_-_ | jeblair: thanks! | 22:16 |
| dolphm | ayoung: not sure then :-/ I'd probably refactor the middleware so i don't have to test __call__ lol | 22:16 |
| dolphm | ayoung: there's probably another project with a solution | 22:16 |
| ayoung | OK....I'll find it | 22:16 |
| *** martine has quit IRC | 22:18 | |
| kbringard | vishy: can I just change that value in flags.py, or do I need to do it elsewhere? | 22:18 |
| cloudfly | hrmm what's the point of having the ability to set devstack to go back to older releases in localrc if stack.sh isn't capable of doing it. | 22:18 |
| vishy | just in flags | 22:18 |
| kbringard | and restart compute, right? | 22:19 |
| *** andrewsmedina has quit IRC | 22:19 | |
| *** paulormg has quit IRC | 22:19 | |
| hub_cap | using run_tests.sh is it possible to run a single test? | 22:20 |
| *** internap has joined #openstack-dev | 22:20 | |
| jakedahn | hub_cap: yeah but its kind of weird - i think its somehting like ./run_tests.sh tests.dir.dir.dir.testfile.TestClass | 22:20 |
| hub_cap | nice thx | 22:21 |
| jakedahn | might be dir.dir.dir.testfile:TestClass | 22:21 |
| jakedahn | haven't used it for a really long time | 22:21 |
| mnewby | The latter | 22:21 |
| mnewby | Nose uses a colon (:) do separate the qualified module path from the contents of the module | 22:22 |
| hub_cap | well that worked, but all the tests failed cuz of some funky deps im sure | 22:23 |
| hub_cap | thx | 22:23 |
| sleepsonzzz | cloudfly - you can go back to stable/diablo by checking out devstack's stable/diablo branch, if that is what you are trying to do | 22:23 |
| cloudfly | might have to hop back a bit. | 22:24 |
| cloudfly | seems horizon is busted up in current build | 22:24 |
| dolphm | hub_cap: ./run_tests.sh -N test_content_types:TestClass.test_method | 22:26 |
| hub_cap | dolphm: i did that, it failed miserably | 22:26 |
| *** Ghe_Rivero has quit IRC | 22:26 | |
| *** andrewbogott has joined #openstack-dev | 22:27 | |
| *** andrewbogott has joined #openstack-dev | 22:27 | |
| hub_cap | its looking for stuff it cant find it seems :( | 22:27 |
| hub_cap | anyone know how to use query params in a route? id like to populate a param in my index method…. as in blah?foo=bar with index(…, foo=None) | 22:27 |
| dolphm | hub_cap: works for me- http://pastie.org/3537030 | 22:27 |
| cloudfly | anyone know what the last working rev of devstack was? | 22:28 |
| hub_cap | Ah i was doing test_content_types:CoreApiTests dolphm | 22:28 |
| hub_cap | try that :) | 22:28 |
| dolphm | i can tell you that won't work lol | 22:28 |
| *** hashar has quit IRC | 22:28 | |
| *** davlap has quit IRC | 22:28 | |
| hub_cap | very nice tho now i can test faster | 22:28 |
| hub_cap | haha ya fo sure | 22:28 |
| hub_cap | now if i can only figure out how to acutally populate belongs_to ill be in hog heaven dolphm | 22:29 |
| dolphm | hub_cap: looking for that now | 22:29 |
| hub_cap | im trying query params but even w/ a small test on soem of my personal routes code i cant get it to populate | 22:29 |
| hub_cap | even though belongs_to is like on 3 different methods in keystone, there is not a single mention of it in the tests ;) | 22:31 |
| dolphm | hub_cap: =( i wrote those tests lol | 22:31 |
| hub_cap | hahaah oops! | 22:31 |
| hub_cap | i mean, good work dolphm, keep looking for that answer! | 22:31 |
| dolphm | ;) | 22:31 |
| dolphm | oooh | 22:32 |
| hub_cap | oooh? | 22:32 |
| dolphm | oooh. | 22:32 |
| hub_cap | oooh! | 22:32 |
| dolphm | hub_cap: req.environ['openstack.params'] | 22:33 |
| dolphm | ugh nevermind | 22:34 |
| hub_cap | oh yer looking @ the wsgi stuff to see how its done, ill view too | 22:34 |
| dolphm | wait, no that would work | 22:34 |
| hub_cap | sure ill just look @ all that goodness | 22:35 |
| dolphm | keystone.common.wsgi.Application converts the query string in the env into method params | 22:35 |
| hub_cap | dolphm: openstack.context = {'query_string': {'belongs_to': u'46f5cdde48884db2b5261108e97d4bb1'}, 'token_id': '46f5cdde48884db2b5261108e97d4bb1', 'is_admin': False} | 22:38 |
| hub_cap | thats whats printed to me | 22:38 |
| hub_cap | its never moved into the context tho as a "param" | 22:39 |
| hub_cap | does nova or glance do this? do any of the projects use query strings i wonder | 22:39 |
| dolphm | no clue | 22:40 |
| hub_cap | the arg_dict def does not contain belongs_to | 22:40 |
| cloudfly | anyone know a past revision of horizon + devstack that works? | 22:40 |
| cloudfly | this is getting annoying | 22:41 |
| hub_cap | vishy, pvo: do u know offhand if there is any use of query params in any urls? | 22:41 |
| dolphm | hub_cap: looks like you're right, the context is passed into the method as a whole, including the query_string, but the query_string *should* be tacked onto the additional params | 22:41 |
| dolphm | so | 22:41 |
| hub_cap | def, so the common wsgi can just append those to the params, no? | 22:41 |
| pvo | hub_cap: I think some of the changes since uses query params | 22:42 |
| dolphm | hub_cap: yes | 22:42 |
| hub_cap | kk thx pvo | 22:42 |
| dolphm | there could be a method out there expecting to find the query string in context, but auth/validate definitely expect it to be an argument | 22:42 |
| *** andrewsmedina has joined #openstack-dev | 22:44 | |
| vishy | hub_cap: yes there are query params in the nova_api | 22:46 |
| vishy | filtering on servers | 22:46 |
| vishy | for example | 22:46 |
| ayoung | dolphm, why does the evaluation of the Admin token happen via an HTTP call? | 22:50 |
| ayoung | response, data = self._json_request('GET', | 22:51 |
| ayoung | '/v2.0/tokens/%s' % user_token, | 22:51 |
| ayoung | additional_headers=headers) | 22:51 |
| dolphm | ayoung: as opposed to HTTPS? | 22:52 |
| ayoung | no as opposed to a local call | 22:52 |
| ayoung | is this code executed on other servers? | 22:52 |
| dolphm | ayoung: from the auth_token middleware? | 22:52 |
| ayoung | or is it inside Keystone? | 22:52 |
| ayoung | yes | 22:52 |
| dolphm | auth_token sits in front a nova, swift, etc, and calls out to the remote keystone server | 22:52 |
| ayoung | so to test this we need an http server up and running... | 22:53 |
| *** crobinso has quit IRC | 22:53 | |
| hub_cap | ok victory ive got it firgured out now | 22:53 |
| *** sniperd has quit IRC | 22:53 | |
| dolphm | ayoung: you *could* test it that way, but you'd need a service to put behind it | 22:54 |
| dolphm | ayoung: (i actually have an "echo" service that doesn't do much more than respond with your token & tenant info, but it's not committed) | 22:54 |
| ayoung | dolphm, I'd like to figure out a way to test this without having to make an actual HTTP call.... | 22:55 |
| ayoung | course, I'd also like to write this code so that it doesn't need to do an HTTP call in the first place | 22:56 |
| *** sniperd has joined #openstack-dev | 22:57 | |
| dolphm | ayoung: hmm... so, i guess 3 options: rewrite __call__ so there's nothing there worth testing -- but that looks hard in this case | 22:59 |
| dolphm | 2) do a lot of mocking | 23:00 |
| *** joearnold has joined #openstack-dev | 23:00 | |
| dolphm | i think there's plenty of mocking in test_keystoneclient you could work with | 23:00 |
| dolphm | 3) stand auth_token up in a real pipeline next to keystone and do more of a system test | 23:01 |
| *** sandywalsh has quit IRC | 23:01 | |
| vishy | jeblair: fyi it looks like we are still getting No Ready Nodes from the gate | 23:01 |
| *** markvoelker has quit IRC | 23:02 | |
| *** hub_cap has quit IRC | 23:02 | |
| *** ayoung has quit IRC | 23:03 | |
| *** internap has quit IRC | 23:03 | |
| *** apevec has quit IRC | 23:04 | |
| *** anotherjesse_zz is now known as anotherjesse | 23:05 | |
| *** joearnold has quit IRC | 23:08 | |
| vishy | jeblair: scratch that, one went through so i guess we are on again | 23:10 |
| *** mishesks has joined #openstack-dev | 23:13 | |
| vishy | mtaylor, jeblair: any idea why this is suddenly blowing up on an venv package? | 23:13 |
| vishy | https://review.openstack.org/#change,4380 | 23:13 |
| *** dubsquared has quit IRC | 23:13 | |
| *** sandywalsh has joined #openstack-dev | 23:14 | |
| vishy | mtaylor, jeblair: it is in stable/diablo and here is the console: https://jenkins.openstack.org/job/gate-nova-unittests/1401/console | 23:14 |
| mtaylor | vishy: looking | 23:15 |
| *** zzed has quit IRC | 23:16 | |
| jeblair | vishy: a lot of nodes are coming up with no network, so the script is having trouble keeping up with demand | 23:18 |
| vishy | jeblair: maybe we should switch over to using cloudservers beta :) | 23:18 |
| jeblair | vishy: i think that makes a lot of sense. this script is the worst possible tenant a cloud provider could ever have. it's great for testing. :) | 23:19 |
| jeblair | pvo: ^ what do you think? | 23:19 |
| *** sniperd has quit IRC | 23:20 | |
| *** ywu has joined #openstack-dev | 23:21 | |
| mtaylor | vishy: pile of monkies... how urgent is fixing that? can I get to it in a few hours? | 23:22 |
| vishy | mtaylor: not urgent, just curious how it failed | 23:22 |
| vishy | getting gating to work again in general would be nice though :( | 23:23 |
| mtaylor | vishy: ++ | 23:23 |
| mtaylor | vishy: the break is (I think) my fault ... I'm trying to get nova moved all the way to the new more resilient and more correct way of making venvs... but doing it in baby steps so that we don't break things | 23:24 |
| mtaylor | vishy: as a result, I think I broke stable/diablo venvs. sigh | 23:24 |
| *** mishesks has quit IRC | 23:24 | |
| * vishy flicks mtaylor in the ear | 23:24 | |
| *** mishesks has joined #openstack-dev | 23:24 | |
| vishy | bad dog! | 23:24 |
| mtaylor | vishy: have I mentioned how much I can't wait for all of the stuff to be converged on consistency? | 23:24 |
| mtaylor | OW! | 23:24 |
| * mtaylor tucks tail and hides in corner | 23:24 | |
| vishy | mtaylor: +1 | 23:24 |
| vishy | mtaylor: btw a) how do i run tests with tox | 23:24 |
| mtaylor | vishy: simple answer "tox" | 23:25 |
| vishy | mtaylor: b) are we close to using a nose plugin for fancy output instead of run_tests ? | 23:25 |
| mtaylor | vishy: just run tox, and it will create a venv if one isn't there, and run tests in it - but by default it's going to do pep8, py26 and py27 | 23:25 |
| mtaylor | vishy: so if you want just one, do "tox -epy27" | 23:25 |
| jeblair | vishy, mtaylor: https://review.openstack.org/#change,4571 | 23:26 |
| mtaylor | vishy: and we're REALLY close - there is one weirdness in the nova test suite that breaks under plain nose + plugin that works on run_Tests | 23:26 |
| jeblair | according to that, which is the first change that failed like this, bcwaldon thinks the problem is related to python-novaclient | 23:26 |
| mtaylor | AH | 23:27 |
| vishy | jeblair: ah that one failed with the iso error? | 23:27 |
| vishy | mtaylor: what is the weirdness? | 23:27 |
| jeblair | vishy: weird, no, it didn't, but it did log "cannot import name flavors" | 23:28 |
| jeblair | vishy: https://jenkins.openstack.org/job/gate-nova-unittests/1226/console | 23:28 |
| jeblair | vs https://jenkins.openstack.org/job/gate-nova-unittests/1401/console | 23:29 |
| jeblair | i confess, i don't know why those are different. :/ | 23:29 |
| mtaylor | jeblair: I was looking at the logic in copy-venv - seems like I might have missed something in there | 23:30 |
| *** apevec has joined #openstack-dev | 23:31 | |
| *** mattstep has quit IRC | 23:31 | |
| heckj | did cloudfiles undergo some changes recently? Having some issue with a clear devstack run: http://paste.openstack.org/show/6940/ | 23:32 |
| heckj | looks like it can't find cloudfiles.errors.Error | 23:32 |
| cloudfly | pretty sure devstack is hosed at the moment | 23:32 |
| *** mattstep has joined #openstack-dev | 23:33 | |
| heckj | cloudfly: I'm definitely seeing that… was hoping someone knew more... | 23:33 |
| cloudfly | i am seeing failures in horizon | 23:34 |
| cloudfly | which i see a bug opened for horizon refactor issues | 23:34 |
| cloudfly | but a fix was merged on the 1st | 23:34 |
| cloudfly | so maybe it got rerefactored there | 23:34 |
| cloudfly | not seeing errors on cloudfiles atm | 23:34 |
| kbringard | hey vishy, I know it's not really supported, but any idea why when I up the the sql_max number in flags.py and restart compute I still get errors saying it hit the max of 10? | 23:35 |
| jeblair | mtaylor: it looks like it started with the correct bundle, and is using the venv it built from that; i'm not sure what would be missed. :/ | 23:36 |
| cloudfly | kbringard are you sure you aren't just changing the flag definition then getting it overwritten by the nova flags file? | 23:36 |
| kbringard | cloudfly: I don't think so… I changed it in flags.py, which I assumed was the default | 23:36 |
| cloudfly | well flags.py has definitions but the nova.conf has flags. | 23:37 |
| kbringard | I don't have it set in my nova.conf | 23:37 |
| cloudfly | well then there goes my idea | 23:37 |
| cloudfly | try defining it maybe. =/ | 23:37 |
| kbringard | I was going to try that next, though | 23:37 |
| kbringard | yea | 23:37 |
| creiht | heckj: odd | 23:38 |
| creiht | https://github.com/rackspace/python-cloudfiles/blob/master/cloudfiles/errors.py | 23:38 |
| creiht | Error is there | 23:38 |
| heckj | Wow - that's interesting. I wonder what interesting import didn't allow it to show up? | 23:38 |
| cloudfly | creiht do you have anything in your localrc that might be grabbing from the wrong location? | 23:39 |
| heckj | creiht: just did an "import cloudfiles" and "dir(cloudfiles.errors)" | 23:39 |
| heckj | >>> dir(cloudfiles.errors) | 23:39 |
| heckj | ['AuthenticationError', 'AuthenticationFailed', 'CDNNotEnabled', 'ContainerNotEmpty', 'ContainerNotPublic', 'IncompleteSend', 'InvalidContainerName', 'InvalidMetaName', 'InvalidMetaValue', 'InvalidObjectName', 'InvalidObjectSize', 'InvalidUrl', 'NoSuchContainer', 'NoSuchObject', 'ResponseError', '__builtins__', '__doc__', '__file__', '__name__', '__package__'] | 23:39 |
| heckj | only passwords in my localrc | 23:39 |
| creiht | heckj: what is cloudfiles.__version__ | 23:39 |
| mtaylor | jeblair: where does it copy the bundle in from jenkins_venvs/**/.cache.bundle? | 23:39 |
| mtaylor | jeblair: see - it only does mv jenkins_venvs/$GERRIT_BRANCH/venv.tgz . | 23:40 |
| heckj | 1.7.9.1 - I must have something already installed on this host... | 23:40 |
| creiht | heckj: yeah that is odd | 23:40 |
| cloudfly | heckj might want to try an apt-get purge on all your nova related packages | 23:40 |
| cloudfly | then rerun stack | 23:40 |
| heckj | I'm going to also go make sure I didn't have anything in dist-packages (manually) | 23:41 |
| cloudfly | good call | 23:41 |
| *** mattstep has quit IRC | 23:43 | |
| *** mattstep has joined #openstack-dev | 23:43 | |
| *** hub_cap has joined #openstack-dev | 23:47 | |
| *** apevec has quit IRC | 23:47 | |
| creiht | heckj: also make sure you don't have a random cloudfiles dir somewhere on your path | 23:49 |
| creiht | with python code in it | 23:49 |
| heckj | creiht: yep, just did that. Burned out everything that looked like cloudfiles, and am starting fresh with a devstack run now | 23:50 |
| cloudfly | good luck | 23:50 |
| *** mattray has quit IRC | 23:50 | |
| heckj | creiht: I suspect that was the issue - I found a "cloudfiles" that looks like it was installed directly some time ago. Will know shortly. | 23:51 |
| creiht | cool | 23:51 |
| cloudfly | i'd be interested to know if you run into the horizon problem i am seeing... assuming you are using a current devstack rev | 23:51 |
| heckj | very latest - I'll let you know in a few. Just realized I had a passwd on mysql, so I'm resetting it all again | 23:52 |
| cloudfly | heh i usually just purge mysql and rabbit entirely | 23:52 |
| *** dtroyer has quit IRC | 23:52 | |
| *** roge has quit IRC | 23:53 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!