| *** rkukura has joined #openstack-dev | 00:02 | |
| *** PotHix has quit IRC | 00:03 | |
| *** rha has joined #openstack-dev | 00:04 | |
| *** Ruetobas has quit IRC | 00:06 | |
| *** heckj has quit IRC | 00:06 | |
| *** datsun180b has quit IRC | 00:10 | |
| *** blamar has quit IRC | 00:10 | |
| *** winston-d has joined #openstack-dev | 00:14 | |
| *** danwent_ has quit IRC | 00:14 | |
| *** novas0x2a|laptop has quit IRC | 00:16 | |
| *** Ruetobas has joined #openstack-dev | 00:18 | |
| *** reed has quit IRC | 00:18 | |
| *** winston-d has quit IRC | 00:20 | |
| *** galstrom has joined #openstack-dev | 00:20 | |
| *** anderstj has joined #openstack-dev | 00:22 | |
| *** jakedahn is now known as jakedahn_zz | 00:23 | |
| *** Aaton is now known as Aaton_off | 00:25 | |
| jkoelker | mtaylor: sent in, you want me to bump the version and upload to pypi? | 00:26 |
|---|---|---|
| mtaylor | jkoelker: if you don't mind ... I think seeing the ERROR strings freaks people out | 00:26 |
| mtaylor | jkoelker: I've got a few of the projects tied to 0.7 - I think I should just tie them all to latest, tbh | 00:27 |
| *** dubsquared has joined #openstack-dev | 00:29 | |
| *** anderstj has quit IRC | 00:30 | |
| jkoelker | mtaylor: 0.8 is now on da pypis | 00:32 |
| mtaylor | jkoelker: w00t! | 00:39 |
| mtaylor | jkoelker, bcwaldon: I think I have sorted out logcapture properly for glance now | 00:40 |
| mtaylor | so we it should properly do debug logging for unittests and attach the results to the test result on failure as needed same as nova | 00:40 |
| mtaylor | still don't have it fully for functional yet - but I'll get there | 00:40 |
| *** nati_ueno has quit IRC | 00:40 | |
| *** somik has quit IRC | 00:43 | |
| *** dtroyer is now known as dtroyer_zzz | 00:51 | |
| *** BK_man has joined #openstack-dev | 00:57 | |
| *** sam_wan has joined #openstack-dev | 00:59 | |
| *** rods has quit IRC | 01:00 | |
| *** littleidea has quit IRC | 01:01 | |
| *** torgomatic has quit IRC | 01:05 | |
| *** nati_ueno has joined #openstack-dev | 01:06 | |
| *** nati_ueno has quit IRC | 01:06 | |
| *** Mandell has quit IRC | 01:06 | |
| *** nati_ueno has joined #openstack-dev | 01:06 | |
| *** jdurgin has quit IRC | 01:07 | |
| *** roge has quit IRC | 01:12 | |
| *** matiu has quit IRC | 01:12 | |
| *** roge has joined #openstack-dev | 01:12 | |
| *** matiu has joined #openstack-dev | 01:12 | |
| *** matiu has quit IRC | 01:12 | |
| *** matiu has joined #openstack-dev | 01:12 | |
| *** BK_man has quit IRC | 01:19 | |
| *** Gordonz has joined #openstack-dev | 01:22 | |
| *** adjohn has quit IRC | 01:30 | |
| *** spiffxp has quit IRC | 01:32 | |
| mtaylor | bcwaldon: the logging for the server processes spun up in glance functional testing is a bit tricky | 01:33 |
| mtaylor | I don't think I can hook it in to nose output apropriately without completely reengineering how that proces sowrks | 01:34 |
| mtaylor | or how the process works | 01:34 |
| *** pixelbeat has quit IRC | 01:34 | |
| *** Gordonz has quit IRC | 01:36 | |
| *** wiliam has joined #openstack-dev | 01:40 | |
| *** dtroyer_zzz is now known as dtroyer | 01:45 | |
| *** eglynn__ has joined #openstack-dev | 01:47 | |
| *** eglynn_ has quit IRC | 01:47 | |
| *** koolhead17 has joined #openstack-dev | 01:53 | |
| *** s0mik has joined #openstack-dev | 02:01 | |
| *** dubsquared has quit IRC | 02:03 | |
| *** ayoung has quit IRC | 02:08 | |
| *** adjohn has joined #openstack-dev | 02:13 | |
| *** wiliam has quit IRC | 02:20 | |
| *** agonella is now known as agonella1 | 02:21 | |
| *** agonella1 is now known as agonella | 02:22 | |
| *** adjohn has quit IRC | 02:23 | |
| *** bencherian has quit IRC | 02:24 | |
| *** bencherian has joined #openstack-dev | 02:26 | |
| *** koolhead17 has quit IRC | 02:33 | |
| *** bencherian has quit IRC | 02:34 | |
| *** dubsquared has joined #openstack-dev | 02:36 | |
| *** BLZbubba has quit IRC | 02:50 | |
| *** BLZbubba has joined #openstack-dev | 02:51 | |
| *** adjohn has joined #openstack-dev | 02:53 | |
| *** adjohn has quit IRC | 02:54 | |
| *** adalbas has quit IRC | 02:55 | |
| *** adjohn has joined #openstack-dev | 02:58 | |
| *** johnpostlethwait has quit IRC | 03:00 | |
| *** dubsquared has quit IRC | 03:01 | |
| *** adjohn has quit IRC | 03:06 | |
| *** steveb___ has quit IRC | 03:11 | |
| *** agonella has quit IRC | 03:25 | |
| *** sleepsonzzz is now known as sleepsonthefloor | 03:33 | |
| *** koolhead17 has joined #openstack-dev | 03:33 | |
| *** steveb_ has joined #openstack-dev | 03:34 | |
| *** jakedahn_zz is now known as jakedahn | 03:34 | |
| *** Keoven has joined #openstack-dev | 03:34 | |
| *** littleidea has joined #openstack-dev | 03:45 | |
| *** Keoven has left #openstack-dev | 03:46 | |
| *** Keoven has joined #openstack-dev | 03:47 | |
| *** Keoven has left #openstack-dev | 03:47 | |
| *** littleidea has quit IRC | 03:53 | |
| *** littleidea has joined #openstack-dev | 03:53 | |
| *** adjohn has joined #openstack-dev | 03:59 | |
| *** adjohn has quit IRC | 04:03 | |
| *** Dr_Who has quit IRC | 04:14 | |
| *** timello has quit IRC | 04:17 | |
| *** bencherian has joined #openstack-dev | 04:20 | |
| *** s0mik has quit IRC | 04:23 | |
| *** garyk has quit IRC | 04:36 | |
| *** harlowja has quit IRC | 04:43 | |
| *** Mandell has joined #openstack-dev | 04:45 | |
| *** littleidea has quit IRC | 04:45 | |
| *** edygarcia has quit IRC | 04:51 | |
| *** mdomsch has quit IRC | 05:03 | |
| *** adjohn has joined #openstack-dev | 05:03 | |
| *** nati_uen_ has joined #openstack-dev | 05:04 | |
| *** nati_uen_ has quit IRC | 05:04 | |
| *** jgriffith has quit IRC | 05:04 | |
| *** nati_uen_ has joined #openstack-dev | 05:05 | |
| *** adjohn has quit IRC | 05:05 | |
| *** nati_ueno has quit IRC | 05:07 | |
| *** timello has joined #openstack-dev | 05:15 | |
| *** andrewsben is now known as andrewsben_zz | 05:16 | |
| *** steveb_ has quit IRC | 05:18 | |
| *** e1mer has quit IRC | 05:22 | |
| *** blamar has joined #openstack-dev | 05:24 | |
| *** e1mer has joined #openstack-dev | 05:24 | |
| *** epim has quit IRC | 05:25 | |
| *** zigo has joined #openstack-dev | 05:26 | |
| *** dhellmann has quit IRC | 05:29 | |
| *** dhellmann_ has joined #openstack-dev | 05:29 | |
| *** dtroyer is now known as dtroyer_zzz | 05:33 | |
| *** nati_uen_ has quit IRC | 05:33 | |
| *** matwood has quit IRC | 05:53 | |
| *** garyk has joined #openstack-dev | 05:56 | |
| *** hattwick has quit IRC | 06:01 | |
| *** sleepsonthefloor is now known as sleepsonzzz | 06:03 | |
| *** zaitcev has quit IRC | 06:05 | |
| GheRivero | morning people | 06:06 |
| *** jakedahn is now known as jakedahn_zz | 06:07 | |
| *** Vek has quit IRC | 06:08 | |
| *** dachary has joined #openstack-dev | 06:10 | |
| *** Mandell has quit IRC | 06:12 | |
| *** mindpixel has joined #openstack-dev | 06:13 | |
| *** dachary has quit IRC | 06:15 | |
| *** erikzaadi has joined #openstack-dev | 06:16 | |
| *** mnewby_ has quit IRC | 06:25 | |
| *** matwood has joined #openstack-dev | 06:30 | |
| *** longhudou has joined #openstack-dev | 06:39 | |
| *** danwent has joined #openstack-dev | 06:44 | |
| *** adjohn has joined #openstack-dev | 07:10 | |
| *** zigo has quit IRC | 07:13 | |
| *** zigo has joined #openstack-dev | 07:13 | |
| *** mrunge has joined #openstack-dev | 07:15 | |
| *** reidrac has joined #openstack-dev | 07:19 | |
| *** sulochan has joined #openstack-dev | 07:24 | |
| *** roge has quit IRC | 07:33 | |
| *** adjohn has quit IRC | 07:37 | |
| *** dachary has joined #openstack-dev | 07:43 | |
| garyk | ping gongys | 07:46 |
| *** dhellmann_ has quit IRC | 07:54 | |
| *** dhellmann has joined #openstack-dev | 07:55 | |
| *** hattwick has joined #openstack-dev | 07:58 | |
| garyk | danwent: ping | 07:59 |
| danwent | garyk: hello | 07:59 |
| garyk | danwent: i have a little problem with the stable essex | 08:03 |
| garyk | danwent: https://review.openstack.org/#/c/8384/ | 08:04 |
| *** Vek has joined #openstack-dev | 08:04 | |
| garyk | danwent: by mistake it was based on an abandoned patch | 08:04 |
| *** bencherian has quit IRC | 08:05 | |
| dachary | anotherjesse_zz: ping ? | 08:07 |
| dachary | Does anyone know how to get in touch with anotherjesse_zz ? I need less than 2 minutes of his time, to publish the licensing terms of https://github.com/cloudbuilders/openstack-munin ;-) ( more context here http://lists.alioth.debian.org/pipermail/openstack-devel/2012-June/000989.html ) | 08:08 |
| *** danwent_ has joined #openstack-dev | 08:10 | |
| *** danwent has quit IRC | 08:10 | |
| *** danwent_ is now known as danwent | 08:10 | |
| garyk | danwent: any idea how to recover from this? | 08:15 |
| *** shang has quit IRC | 08:17 | |
| danwent | garyk: sorry, got interrrupted | 08:18 |
| *** shang has joined #openstack-dev | 08:18 | |
| *** darraghb has joined #openstack-dev | 08:18 | |
| *** pixelbeat has joined #openstack-dev | 08:18 | |
| *** dachary has quit IRC | 08:19 | |
| *** dachary has joined #openstack-dev | 08:20 | |
| danwent | garyk: if you look just below the review info on this page (https://review.openstack.org/#/c/8384/), you will see that this review is "stacked" on top of another review. That is, when you pushed this change to gerrit, there was more than one git commit between your change and the current commit of stable/essex | 08:20 |
| danwent | so gerrit won't let you commit this until you've commit the review it was stacked on top of | 08:21 |
| danwent | however, it looks like that review is abandoned. | 08:21 |
| ttx | danwent: coming to help with EU TZ developer numbers ? | 08:21 |
| danwent | so if you don't plan on committing that one, you should push a new version that is not stacked on top of that review | 08:22 |
| danwent | ttx: haha… yes, for one week at least. I'm at a conf in paris | 08:22 |
| danwent | if i can get someone to keep paying me, maybe i'll stay :) | 08:22 |
| ttx | danwent: oh! Next time let me know, I could have come over :) | 08:22 |
| danwent | ttx: i thought you were pretty far from paris? | 08:23 |
| danwent | ttx: aren't you in the south? | 08:23 |
| ttx | danwent: too far to come at the last minute, but not that far. 3 hours away | 08:23 |
| ttx | and my father lives in Paris so it's an easy stay | 08:23 |
| danwent | ttx: ok, will give you a heads up next time. am talking about openstack + quantum at the "carrier cloud summit" | 08:24 |
| * ttx didn't even know there was such a thing. | 08:24 | |
| garyk | danwent to europe: if i restore the abonded task and get that approved will that suffice? in short the fix is a bug fix for the orrignal cherry pick | 08:24 |
| ttx | dansmith: ew, CDG hotel :) | 08:25 |
| danwent | garyk, yes, that would work. | 08:25 |
| danwent | ttx: sign, yes... | 08:25 |
| danwent | sigh | 08:25 |
| danwent | ttx: not too bad getting into the city on the RER, but yeah, not idea. | 08:26 |
| garyk | danwent: thanks | 08:26 |
| ttx | danwent: I know people that never came back to Paris after such an experience | 08:26 |
| danwent | garyk: alternately, you could squash your changes into his, and make it a new commit. It seems a bit odd to backport something that is broken to stable, then have another commit fix it. | 08:27 |
| danwent | ttx: don't worry, i won't hold it against you :) | 08:27 |
| garyk | danwent: oy "too late she cried as she shabed her wooden leg" | 08:28 |
| garyk | saved not shabed | 08:28 |
| danwent | :) | 08:29 |
| danwent | its fine, i don't have a strong preference either way (other on the stable/essex team may though) | 08:29 |
| *** dhellmann has quit IRC | 08:31 | |
| garyk | ok - looks like jenkins is sorted. i'll try and allocate ip's instead of messing with jenkins | 08:31 |
| danwent | garyk: cool | 08:36 |
| zykes- | allocate ips ? ;p | 08:40 |
| *** danwent has quit IRC | 08:43 | |
| *** davidha has joined #openstack-dev | 08:49 | |
| *** apevec has joined #openstack-dev | 08:59 | |
| *** danwent has joined #openstack-dev | 09:11 | |
| garyk | danwent: if you have a few minutes can you please check the mail i sent you | 09:14 |
| danwent | garyk: ack | 09:15 |
| *** danpb has joined #openstack-dev | 09:23 | |
| zykes- | do you ever sleep danwent ? :p | 09:24 |
| *** Ryan_Lane has quit IRC | 09:25 | |
| *** cooper_ has quit IRC | 09:34 | |
| danwent | zykes-: i'm actually in europe this week, hence the odd hours | 09:35 |
| danwent | garyk: responded | 09:35 |
| zykes- | ooooh | 09:38 |
| zykes- | danwent: where you at ? | 09:38 |
| danwent | zykes-: paris | 09:38 |
| danwent | zykes-: at a conference, talking about openstack + quantum | 09:38 |
| zykes- | ah | 09:38 |
| zykes- | come to scandinavia ;p | 09:38 |
| *** mestery has quit IRC | 09:38 | |
| danwent | zykes-: ha, if only i had time. | 09:39 |
| zykes- | I really wonder when openstack is going to take off up here | 09:40 |
| danwent | zykes-: interesting… what do you seem people using instead? | 09:41 |
| *** rods has joined #openstack-dev | 09:41 | |
| zykes- | danwent: take a wild guess. it starts with V | 09:41 |
| zykes- | and another with C | 09:41 |
| zykes- | for all the money basically | 09:41 |
| danwent | zykes-: interesting. Are there key features that folks see openstack lacking vs. cloudstack? | 09:44 |
| danwent | its helpful to have that context. | 09:44 |
| zykes- | danwent: people here are not even sniffing / looking at cloudstack / openstack | 09:44 |
| zykes- | everyone is using homegrown alike solutions or vmware / citrix xenserver | 09:45 |
| zykes- | as far as I know | 09:45 |
| danwent | ah, i see. | 09:45 |
| danwent | well, different markets emerge at different rates. | 09:45 |
| zykes- | ;p | 09:45 |
| zykes- | i'm trying to push it but it's always the argument to use well known / proven technologies instead of testing new stuff :( | 09:45 |
| danwent | zykes-: yeah, for someone to adopt openstack now they have to be forward thinking and making a bet on where the platform will be in the future. | 09:46 |
| *** mgz has joined #openstack-dev | 09:48 | |
| garyk | danwent: thanks for the explanation. i need to think about it a bit. | 09:50 |
| danwent | garyk: cool. if you want, we could skype as well if that would help | 09:51 |
| zykes- | danwent: people here dunno what forwards thinking is. they are like donkeys when they hear it and sit back on their asses. | 09:51 |
| danwent | garyk: i'd like to better understand what you were proposing, as I think it was different from both of my proposals :) | 09:51 |
| danwent | zykes-: hehe… that's unfortunate | 09:51 |
| zykes- | basically if you say this is cool stuff they say it's scary stuff :p | 09:51 |
| zykes- | danwent: is nicira planning on doing GA anytime soon for Medium size enterprises ? | 09:53 |
| *** sam_wan has quit IRC | 09:54 | |
| danwent | zykes-: we've already been GA for almost a year, but yeah, focused many on larger customers. | 09:54 |
| danwent | we pretty much handle customer opportunities on a case-by-base basis. | 09:55 |
| danwent | there is also the open source OVS plugin, which anyone can use. should be sufficient for many enterprises if they are looking for quantum use cases. | 09:56 |
| zykes- | i think most companies here would want a commercial solution (due to support) | 09:57 |
| danwent | zykes-: yes, though I expect that companies that support openstack in general will support quantum with the open source plugins. | 10:01 |
| danwent | (as well as quantum with commercial plugins) | 10:01 |
| garyk | danwent: ok. would be great to discuss. | 10:02 |
| *** Ryan_Lane has joined #openstack-dev | 10:10 | |
| *** camm has quit IRC | 10:20 | |
| *** zigo has quit IRC | 10:22 | |
| *** danwent has quit IRC | 10:23 | |
| *** agonella has joined #openstack-dev | 11:02 | |
| *** rkukura has quit IRC | 11:14 | |
| *** rods has quit IRC | 11:16 | |
| *** wiliam has joined #openstack-dev | 11:19 | |
| *** maploin has joined #openstack-dev | 11:22 | |
| *** maploin has quit IRC | 11:22 | |
| *** maploin has joined #openstack-dev | 11:22 | |
| *** dhellmann has joined #openstack-dev | 11:23 | |
| *** rods has joined #openstack-dev | 11:28 | |
| *** dachary has quit IRC | 11:42 | |
| *** dachary has joined #openstack-dev | 11:42 | |
| *** ywu has joined #openstack-dev | 11:44 | |
| *** berendt has joined #openstack-dev | 11:47 | |
| garyk | danwent: ping | 11:55 |
| *** zigo has joined #openstack-dev | 11:56 | |
| *** Dr_Who has joined #openstack-dev | 11:57 | |
| *** dolphm has joined #openstack-dev | 12:00 | |
| *** danwent has joined #openstack-dev | 12:04 | |
| *** sdague has quit IRC | 12:04 | |
| *** adalbas has joined #openstack-dev | 12:15 | |
| *** salgado has joined #openstack-dev | 12:18 | |
| *** salgado has joined #openstack-dev | 12:18 | |
| *** lts has joined #openstack-dev | 12:20 | |
| *** mrunge has quit IRC | 12:20 | |
| *** alaski has joined #openstack-dev | 12:24 | |
| *** dolphm has quit IRC | 12:24 | |
| *** roge has joined #openstack-dev | 12:32 | |
| *** dolphm has joined #openstack-dev | 12:43 | |
| *** markmc has joined #openstack-dev | 12:54 | |
| *** stuntmachine has joined #openstack-dev | 12:56 | |
| *** Ryan_Lane has quit IRC | 13:02 | |
| *** Ryan_Lane has joined #openstack-dev | 13:03 | |
| *** dachary has quit IRC | 13:03 | |
| *** eglynn__ has quit IRC | 13:06 | |
| *** mestery has joined #openstack-dev | 13:09 | |
| *** edygarcia has joined #openstack-dev | 13:18 | |
| *** eglynn__ has joined #openstack-dev | 13:20 | |
| *** andrewsmedina has joined #openstack-dev | 13:24 | |
| *** Ruetobas has quit IRC | 13:24 | |
| *** ayoung has joined #openstack-dev | 13:25 | |
| *** sandywalsh_ has joined #openstack-dev | 13:31 | |
| *** corXi has joined #openstack-dev | 13:32 | |
| *** dtroyer_zzz is now known as dtroyer | 13:32 | |
| *** dprince has joined #openstack-dev | 13:32 | |
| *** sandywalsh has quit IRC | 13:32 | |
| *** stuntmac_ has joined #openstack-dev | 13:34 | |
| *** stuntmachine has quit IRC | 13:34 | |
| *** iccha has joined #openstack-dev | 13:34 | |
| *** nati_ueno has joined #openstack-dev | 13:35 | |
| *** dhellmann_ has joined #openstack-dev | 13:37 | |
| *** dhellmann_ has quit IRC | 13:37 | |
| *** dhellmann has quit IRC | 13:41 | |
| *** mjfork has joined #openstack-dev | 13:43 | |
| *** sdague has joined #openstack-dev | 13:44 | |
| *** m4xmr has joined #openstack-dev | 13:48 | |
| ttx | sdague: hey! Was wndering if we should consider https://blueprints.launchpad.net/nova/+spec/virt-driver-cleanup fully implemented ? | 13:48 |
| *** rkukura has joined #openstack-dev | 13:49 | |
| ttx | markmc: do you want a spot to discuss 2012.1.1 at the meeting today ? (or want me to mention it, if you can't make it ?) | 13:49 |
| sdague | ttx: it is not yet, still working through parts c, d, e | 13:50 |
| sdague | I'm hoping this week | 13:50 |
| ttx | sdague: ok, thx for the update :) | 13:50 |
| sdague | ttx: no prob | 13:51 |
| *** sstent has joined #openstack-dev | 13:52 | |
| *** jsavak has joined #openstack-dev | 13:56 | |
| *** markvoelker has joined #openstack-dev | 13:58 | |
| *** mattray has joined #openstack-dev | 13:59 | |
| sstent | is there are new dependency for horizon? -- i'm getting -- ImportError: No module named compressor -- on startup | 14:01 |
| *** nati_ueno has quit IRC | 14:02 | |
| *** kbringard has joined #openstack-dev | 14:03 | |
| danpb | does anyone have an idea why i'd see the following error from Nova API service ? | 14:08 |
| danpb | 2012-06-19 15:08:21 CRITICAL nova [-] <module 'nova.api.metadata.handler' from '/usr/lib/python2.7/site-packages/nova/api/metadata/handler.py'> has no 'Versions' attribute | 14:08 |
| *** mindpixel has quit IRC | 14:10 | |
| *** sieutruc has joined #openstack-dev | 14:12 | |
| markmc | ttx, sure, I'll try and be there | 14:13 |
| markmc | ttx, if I'm not, though, you know the status :) | 14:14 |
| *** nikhil has quit IRC | 14:14 | |
| *** nikhil has joined #openstack-dev | 14:15 | |
| markmc | danpb, must be related to smoser's refactoring in 55128d1c5 | 14:15 |
| markmc | danpb, stale api-paste.ini file ? | 14:15 |
| markmc | i.e. this shouldn't be there anymore: | 14:16 |
| markmc | paste.app_factory = nova.api.metadata.handler:Versions.factory | 14:16 |
| *** gholt has quit IRC | 14:16 | |
| *** gholt has joined #openstack-dev | 14:16 | |
| danpb | markmc++ for the win ! updating api-paste.ini solved it | 14:16 |
| *** littleidea has joined #openstack-dev | 14:17 | |
| markmc | cool | 14:17 |
| *** Gordonz has joined #openstack-dev | 14:20 | |
| *** pyw has joined #openstack-dev | 14:20 | |
| *** edygarcia has quit IRC | 14:22 | |
| zykes- | jk0: ping | 14:22 |
| *** nunosantos has joined #openstack-dev | 14:22 | |
| *** maploin has quit IRC | 14:25 | |
| markmc | ttx, re: #1014812 | 14:25 |
| markmc | ttx, good points | 14:25 |
| markmc | ttx, but how does/will rootwrap work with the different users setup? | 14:26 |
| ttx | markmc: it would work well. | 14:26 |
| markmc | ttx, e.g. how to configure sudo to only allow the nova-api user to run rootwrap with a config specific to nova-api? | 14:26 |
| ttx | markmc: different sudoers lines: | 14:26 |
| *** Gordonz has quit IRC | 14:27 | |
| ttx | nova-compute ALL = (root) NOPASSWD: /usr/bin/nova-rootwrap /only/compute/filters.d | 14:27 |
| *** Gordonz has joined #openstack-dev | 14:27 | |
| ttx | nova-api ALL = (root) NOPASSWD: /usr/bin/nova-rootwrap /only/api/filters.d | 14:27 |
| ttx | markmc: Basically if your packaging uses different users... it should also use different rootwrap.d configuration directories | 14:28 |
| markmc | ttx, hmm, you can specify arguments like in sudoers like that? | 14:28 |
| ttx | actually the line should read: | 14:29 |
| ttx | nova-compute ALL = (root) NOPASSWD: /usr/bin/nova-rootwrap /only/compute/filters.d * | 14:29 |
| ttx | so that the first argument is set in stone, but the others are open | 14:29 |
| ttx | markmc: yes | 14:29 |
| *** danwent_ has joined #openstack-dev | 14:29 | |
| *** danwent has quit IRC | 14:29 | |
| *** danwent_ is now known as danwent | 14:29 | |
| *** dtroyer is now known as dtroyer_zzz | 14:30 | |
| ttx | markmc: again, I'm not sure the trade-off is worth it. Added complexity to cover a corner case (multiple roles on a single machine) -- but I leave the choice to distros | 14:30 |
| markmc | ttx, ah, ok - args in sudoers was the bit I was missing | 14:31 |
| markmc | ttx, yeah, I don't think we'd bother in Fedora, just wanted to make sure we weren't missing something | 14:31 |
| markmc | ttx, thanks | 14:31 |
| ttx | markmc: np | 14:31 |
| *** danwent_ has joined #openstack-dev | 14:33 | |
| *** danwent has quit IRC | 14:33 | |
| *** danwent_ is now known as danwent | 14:33 | |
| *** salgado is now known as salgado-afk | 14:33 | |
| * markmc can actually follow stuff happening in launchpad now | 14:34 | |
| markmc | set up a new filter based on: | 14:34 |
| markmc | X-Launchpad-Message-Rationale: Subscriber | 14:34 |
| *** dachary has joined #openstack-dev | 14:34 | |
| *** blamar has quit IRC | 14:34 | |
| markmc | 11k unread emails in my nova launchpad folder | 14:34 |
| *** Ruetobas has joined #openstack-dev | 14:35 | |
| *** AlanClark has joined #openstack-dev | 14:38 | |
| *** mnewby has joined #openstack-dev | 14:41 | |
| jaypipes | davidkranz: I'll have a ML post about the code reviews and negative tests up by noon. | 14:43 |
| davidkranz | jaypipes: Thanks! | 14:44 |
| pyw | why nova db use integer, but Long type? | 14:45 |
| *** pballand_ has joined #openstack-dev | 14:46 | |
| *** maoy has joined #openstack-dev | 14:47 | |
| *** Mandell has joined #openstack-dev | 14:51 | |
| *** blamar has joined #openstack-dev | 14:51 | |
| *** pballand_ has left #openstack-dev | 14:52 | |
| *** bencherian has joined #openstack-dev | 14:53 | |
| *** reed has joined #openstack-dev | 14:57 | |
| *** datsun180b has joined #openstack-dev | 14:57 | |
| *** rnirmal has joined #openstack-dev | 14:58 | |
| *** metral has joined #openstack-dev | 14:59 | |
| *** sstent has quit IRC | 14:59 | |
| smoser | markmc, is there some way that my refactor should have handled that api-paste.ini change better? | 15:00 |
| *** sstent has joined #openstack-dev | 15:00 | |
| markmc | smoser, nope, I don't think so | 15:00 |
| *** dolphm has quit IRC | 15:00 | |
| markmc | smoser, probably danpb is running from git using config from his /etc/nova/ rather than the config from git | 15:01 |
| smoser | right. | 15:01 |
| *** dolphm has joined #openstack-dev | 15:01 | |
| *** dolphm has quit IRC | 15:01 | |
| smoser | config file changes suck. | 15:01 |
| *** dolphm has joined #openstack-dev | 15:01 | |
| markmc | these aren't config files either :) | 15:01 |
| * markmc thinks they should be in /usr/share | 15:01 | |
| markmc | well, once we get the rest of the config out of them | 15:02 |
| danpb | markmc: yep, i was running from git, but had a stale api-paste.ini from a previous RPM | 15:02 |
| markmc | e.g. what glance has - choose your pipeline from glance-api.conf without editing api-paste.ini | 15:02 |
| smoser | thats true, we dont really expect end user editing of those files. | 15:02 |
| markmc | oh, and the auth_token config needs to move too | 15:02 |
| *** mattray1 has joined #openstack-dev | 15:03 | |
| * danpb likes this MeetingBurner site OpenStack uses for conf calls .....so much nicer than Elluminate | 15:04 | |
| *** mattray has quit IRC | 15:04 | |
| *** lloydde has joined #openstack-dev | 15:04 | |
| *** mattray1 is now known as mattray | 15:04 | |
| *** mattray has joined #openstack-dev | 15:04 | |
| *** winston-d has joined #openstack-dev | 15:06 | |
| *** zigo-_- has joined #openstack-dev | 15:10 | |
| *** zigo has quit IRC | 15:11 | |
| *** Mandell has quit IRC | 15:11 | |
| *** mnewby has quit IRC | 15:12 | |
| *** PotHix has joined #openstack-dev | 15:13 | |
| *** ncode has joined #openstack-dev | 15:14 | |
| *** ncode has quit IRC | 15:14 | |
| *** ncode has joined #openstack-dev | 15:14 | |
| *** reidrac has quit IRC | 15:15 | |
| *** dachary has quit IRC | 15:17 | |
| *** dhellmann has joined #openstack-dev | 15:17 | |
| *** galstrom has joined #openstack-dev | 15:18 | |
| *** rkukura has quit IRC | 15:21 | |
| *** s0mik has joined #openstack-dev | 15:23 | |
| *** dachary has joined #openstack-dev | 15:24 | |
| sandywalsh_ | ttx: do we have hotel reservation codes available? | 15:25 |
| ttx | sandywalsh_: heh, I'm not even sure they have the contract signed with the hotel yet. | 15:26 |
| sandywalsh_ | ttx: hah, perfect | 15:26 |
| ttx | sandywalsh_: i didn't think they would announce the hotel name yet | 15:27 |
| sandywalsh_ | ttx: there's tweets on it | 15:27 |
| sandywalsh_ | Manchester Grand Hyatt | 15:27 |
| trapni | Hey. 2 KVM instances in my VLAN enabled network (with a dedicated nova-network gateway) became *unreachable* after sending DHCPREQUEST within the last days, even now, where I upgraded dnsmasq on the gateway to 1.62-git (all nodes Ubuntu 12.04 LTS). dnsmasq (the right one) even responded with DHCPACK, as of the log, but the host(s) became unreachable since then. What could the reason be? | 15:27 |
| trapni | I could fix it with a hard reboot, however. | 15:27 |
| ttx | sandywalsh_: sure, they announced it... was just surprised the deal was closed :) | 15:27 |
| sandywalsh_ | ttx: ah, not sure | 15:28 |
| *** e1mer has quit IRC | 15:31 | |
| *** s0mik has quit IRC | 15:33 | |
| *** e1mer has joined #openstack-dev | 15:34 | |
| *** zaitcev has joined #openstack-dev | 15:36 | |
| *** jgriffith has joined #openstack-dev | 15:37 | |
| *** andrewbogott_ has joined #openstack-dev | 15:37 | |
| *** bencherian has quit IRC | 15:38 | |
| *** dhellmann has quit IRC | 15:41 | |
| *** lloydde has quit IRC | 15:43 | |
| andrewbogott_ | markmc: I'm running into trouble using the common cfg CONF object in nova. (Well, actually, in a different common file which is used in nova.) | 15:44 |
| andrewbogott_ | I can see that my options are defined, but the step where the config file is loaded seems to not be happening. | 15:44 |
| *** japage has joined #openstack-dev | 15:44 | |
| *** andrewbogott_ is now known as andrewbogott | 15:45 | |
| andrewbogott | Currently nova seems to do a hybrid thing where it uses common cfg but wraps it with 'flags.py' which is confusing me :( | 15:45 |
| markmc | andrewbogott, is anything running CONF() ? e.g. nova.flags.parse_args() ? | 15:45 |
| andrewbogott | markmc: Possibly not! Is it safe to run that multiple times in a given process? | 15:45 |
| markmc | andrewbogott, there's not much left in flags.py apart from FLAGS = CONF, parse_args() and DECLARE | 15:46 |
| markmc | andrewbogott, why multiple times? | 15:46 |
| andrewbogott | by 'run it multiple times' I mean 'run it on import'. | 15:46 |
| andrewbogott | Um... the big picture is: I'm working on moving logging into common. | 15:47 |
| kbringard | smoser: you about? | 15:49 |
| smoser | kbringard, here | 15:49 |
| *** adalbas has quit IRC | 15:49 | |
| kbringard | smoser: hey, non-openstack but Ubuntu/udev related question, if you have a moment? | 15:49 |
| andrewbogott | markmc: Most likely my issue is that files are setting up logging before they set up cfg | 15:49 |
| markmc | andrewbogott, you can register options before or after parsing (i.e. calling CONF()) | 15:50 |
| markmc | andrewbogott, but only e.g. bin/nova-api should be calling CONF()/parse_args() | 15:50 |
| *** salgado-afk is now known as salgado | 15:51 | |
| smoser | kbringard, sure | 15:51 |
| kbringard | smoser: I have hardware with 4 nics. I know the PCI addresses of the nics, so once the machines are up I force ethX device assignments based on PCI address | 15:51 |
| kbringard | smoser: but, during PXE it just assigns ethX devices in the order they become ready | 15:52 |
| andrewbogott | markmc: Yeah, and I see that the order in bin/nova-api should work... | 15:52 |
| jaypipes | davidkranz: do you have Jose Idar's email address by any chance | 15:52 |
| jaypipes | ? | 15:52 |
| kbringard | smoser: which causes a problem because my pxelinux.cfg needs to specify which eth device to use to get DHCP, etc | 15:52 |
| kbringard | smoser: but it's not always consistent | 15:52 |
| andrewbogott | markmc: If I reference common/cfg.CONF in nova code, would you expect it to have config values from nova.conf, just like that? | 15:52 |
| davidkranz | jaypipes: Checking... | 15:52 |
| jaypipes | davidkranz: cheers | 15:52 |
| markmc | andrewbogott, yep | 15:52 |
| kbringard | smoser: so my question is, is there a way to force udev assignment in the pxelinux.cfg? or some other way of addressing this issue? | 15:53 |
| jaypipes | smoser: I tried to tell kbringard to stop smoking the wacky, but he's never listened to me. | 15:53 |
| *** piluso has joined #openstack-dev | 15:53 | |
| kbringard | jaypipes: if you said things that make sense I might listen more often :-p | 15:53 |
| smoser | kbringard, hm.. i would have thought that you could somehow tell pxelinux to use the device that got a response on it. | 15:53 |
| kbringard | smoser: possibly… that's what I'm not sure of :-/ I've been googling for like an hour but the preseed/pxe configuration docs aren't super great | 15:54 |
| andrewbogott | markmc: What's happening is that I have a debug line "print "CONF.keys: %s" % CONF.keys()" in my new log.py file. And that statement prints out /only/ keys that are defined inside log.py. And, also, none of the values are defined. It's as though I have a CONF object that's purely local to log.py. | 15:54 |
| kbringard | smoser: no biggie if you're not sure, jaypipes had just suggested that you may know, or may know who I should bug :-D | 15:55 |
| jaypipes | kbringard: :P | 15:55 |
| markmc | andrewbogott, how did you initialize the CONF variable? | 15:55 |
| *** sleepsonzzz is now known as sleepsonthefloor | 15:55 | |
| BLZbubba | one amazon feature that I could use is the ability to right click on an instance and create a new image out of it; is there a way to do this with Horizon? | 15:55 |
| *** n0ano has joined #openstack-dev | 15:55 | |
| andrewbogott | markmc: "CONF = cfg.CONF" at import time | 15:55 |
| andrewbogott | where cfg is 'from openstack.common import cfg' | 15:55 |
| davidkranz | jaypipes: Sorry, don't have it. | 15:55 |
| jaypipes | BLZbubba: you mean create a snapshot of an instance? | 15:55 |
| BLZbubba | jaypipes: if by "snapshot" you mean a new image in glance, yes | 15:56 |
| markmc | andrewbogott, something's screwed then - CONF is an instance of CommonConfigOpts which has a bunch of options registered | 15:56 |
| * andrewbogott nods | 15:56 | |
| smoser | kbringard, let me look a bit. | 15:56 |
| BLZbubba | i want users to be able to tweak images without any intervention from me | 15:56 |
| jaypipes | BLZbubba: well, yes, the snapshot gets stored in Glance :) | 15:56 |
| kbringard | smoser: thanks, I appreciate it… let me know if you want more info about our setup | 15:56 |
| smoser | kbringard, but most certainly you wont have udev rules at that point. | 15:56 |
| jaypipes | BLZbubba: Should be able to snapshot any instance in Horizon... | 15:56 |
| jaypipes | BLZbubba: IIRC, there is an action on the Instances action button thingie. | 15:57 |
| *** erikzaadi has quit IRC | 15:57 | |
| BLZbubba | jaypipes: ok great i'll try it, thanks | 15:57 |
| jaypipes | BLZbubba: sorry, not exactly my area of expertise (Horizon) | 15:57 |
| kbringard | smoser: yea, that's what I was thinking… the main reason I know that's the issue is because once the install starts and the dhcp fails, I can check the order in /etc/udev/rules.d/ via the busybox shell | 15:57 |
| kbringard | it would have to be something I pass via pxelinux.cfg, because obviously without network it can't pull and read the preseed.cfg | 15:58 |
| smoser | oh. wait. you're just trying to tell the installer which nic to use? | 15:58 |
| kbringard | well yes, but | 15:58 |
| BLZbubba | jaypipes: ok hopefully it shuts the machine down nicely first | 15:58 |
| kbringard | I have netcfg/choose_interface set | 15:58 |
| kbringard | but you have to use the ethX device name, right? | 15:58 |
| *** e1mer has quit IRC | 15:59 | |
| kbringard | the problem is that ethX changes physical hardware devices across reboots | 15:59 |
| kbringard | so it could be any of the 4 nics in the system | 15:59 |
| kbringard | and because of how the network is setup, only one of them will actually work… I can't just bring up the DHCP server to respond to any of them | 15:59 |
| *** dachary has quit IRC | 16:02 | |
| *** garyk has quit IRC | 16:03 | |
| *** davidha has quit IRC | 16:03 | |
| jaypipes | BLZbubba: it will :) | 16:04 |
| jaypipes | BLZbubba: if it doesn't, that's a serious bug :) | 16:04 |
| *** lloydde has joined #openstack-dev | 16:06 | |
| *** heckj has joined #openstack-dev | 16:06 | |
| smoser | kbringard, https://bugs.launchpad.net/ubuntu/+source/netcfg/+bug/56679 | 16:08 |
| uvirtbot | Launchpad bug 56679 in netcfg "provide a method to use a specified MAC-address as the installation device" [High,Fix released] | 16:08 |
| smoser | so.. .unfortunately its not that clear there, but 2 things might help you. | 16:09 |
| smoser | you can set 'auto' as the value | 16:09 |
| smoser | it seems that BOOTIF= can be set on the command line possibly with a MAC as the value. | 16:10 |
| *** heckj has left #openstack-dev | 16:10 | |
| BLZbubba | jaypipes: it tried to write the snapshot to /tmp, #$^$#^ | 16:10 |
| *** bencherian has joined #openstack-dev | 16:11 | |
| *** heckj has joined #openstack-dev | 16:11 | |
| smoser | but, as jaypipes said, you really should stop the wacky, at least during working hours :) | 16:11 |
| kbringard | haha, indeed | 16:12 |
| kbringard | too many breakfast beers | 16:12 |
| BLZbubba | jaypipes: lol i just googled and found your question about it | 16:12 |
| jaypipes | BLZbubba: :) | 16:13 |
| kbringard | ok, it sounds like what this is doing is, dynamically appending the MAC of the interface that made the PXE request to the netcfg option to make it use that interface during the install? | 16:13 |
| kbringard | smoser: ^^ | 16:13 |
| jaypipes | BLZbubba: easy fix is to ln -s /some/dir/with/room /tmp and retry it ;) | 16:13 |
| kbringard | cool, that sounds like what I want | 16:14 |
| kbringard | smoser: you are the man | 16:14 |
| smoser | kbringard, well, i think with BOOTIF= you will have to specify the mac | 16:14 |
| kbringard | # This allows the initrd to access to the mac address of the interface | 16:14 |
| kbringard | # which PXE booted and is probably the interface we want to use for DHCP. | 16:14 |
| smoser | the issue is, afaik, there is no connect between pxe and linux. ie, pxe doesn't pass that information of "i got pxe booted off of mac=" in any way to linux. so you have to basically tell linux that. | 16:15 |
| kbringard | right | 16:15 |
| kbringard | https://launchpadlibrarian.net/60443113/S31pxedust | 16:15 |
| smoser | it would seem that a pxe executable coudl be linux-smart and made to dynamically append a mac address to a kernel command line | 16:16 |
| kbringard | that's what it looks like this script is doing | 16:16 |
| smoser | well there ya go. | 16:16 |
| smoser | yeah. | 16:16 |
| kbringard | ah, but yea, you're right, it looks like I have to set it | 16:16 |
| kbringard | err, maybe not | 16:16 |
| kbringard | I dunno, I'll figure it out | 16:16 |
| kbringard | either way, this seems like my solution | 16:16 |
| kbringard | I'll let you know how it goes | 16:16 |
| kbringard | thanks a lot, I don't know that I'd have found this in a timely manner :-D | 16:17 |
| markmc | zul, there? | 16:17 |
| zul | markmc: kind of...in a meeting though | 16:18 |
| zul | markmc: whats up? | 16:18 |
| markmc | zul, permitted_instances = cpu_info.get('permitted_instance_types', None) | 16:18 |
| markmc | zul, isn't cpu_info an int in the xenapi case? | 16:18 |
| andrewbogott | markmc: Confirmed, the cfg.CONF object in flags.py is a different one from the cfg.CONF in my new log.py file. So... now I just have to figure out why :( | 16:19 |
| smoser | kbringard, that is awesome. aparently pxelinux can do exactly what i described :) | 16:19 |
| andrewbogott | Thank you for the lead. | 16:19 |
| zul | markmc: possibly, but the field in the database is text | 16:19 |
| markmc | andrewbogott, np, enjoy :) | 16:19 |
| markmc | zul, well, either way ... it doesn't have a get() method right? | 16:19 |
| zul | markmc: right | 16:20 |
| andrewbogott | markmc: Is it possible that importing the same file but via a different path would produce different objects? | 16:20 |
| zul | markmc: i havent touched xenapi at all | 16:20 |
| kbringard | smoser: yea, it looks like that is exactly what this does. Nice work, you thought of it and it happened | 16:20 |
| smoser | it would seem, though, that the s31pedust is no longer needed. | 16:20 |
| kbringard | smoser: can you think about a million dollars for me, next? | 16:20 |
| markmc | zul, cool, thought I was missing something - just need to add a safety check? | 16:20 |
| markmc | zul, well, it looks like you've broken it, no? :) | 16:20 |
| zul | markmc: eh? | 16:21 |
| markmc | zul, the filter is enabled by default, right? | 16:21 |
| zul | markmc: no | 16:21 |
| markmc | zul, and will traceback in xenapi case when it does get() on a string ... | 16:22 |
| markmc | zul, oh? | 16:22 |
| zul | markmc: it isnt | 16:22 |
| andrewbogott | markmc: I ask because the object in flags.py is a nova.openstack.common.cfg.CommonConfigOpts object, but the one accessed within common is an openstack.common.cfg.CommonConfigOpts object... | 16:22 |
| markmc | zul, hmm, ok - I misunderstood how filters are enabled then | 16:22 |
| zul | markmc: it can be enabled by default but that will break xenapi :) | 16:22 |
| markmc | zul, it isn't in nova.scheduler.filters.standard_filters? | 16:23 |
| zul | markmc: no its set in host_manager.py | 16:23 |
| markmc | zul, what is? | 16:23 |
| zul | scheduler_default_filters | 16:23 |
| smoser | kbringard, where did you see that attachment? | 16:23 |
| smoser | (s31pxedust) | 16:24 |
| markmc | zul, ah, got it - cool | 16:24 |
| markmc | zul, thanks | 16:24 |
| zul | markmc: AvailabilityZoneFilters, RamFilter, and ComputeFilter is on by default | 16:24 |
| BLZbubba | jaypipes: ok cool i just hacked the /etc/init/nova-compute.conf file to do "export TMPDIR=/big/filesystem" right before executing nova-compute | 16:24 |
| smoser | i see it now. | 16:24 |
| BLZbubba | not a good permanent solution of course | 16:24 |
| markmc | zul, it'd be cool to add the safety check, though - filters probably shouldn't traceback with different hypervisors | 16:24 |
| BLZbubba | thanks for the suggestion back in march :P | 16:24 |
| zul | markmc: agreed | 16:24 |
| *** bencherian has quit IRC | 16:24 | |
| zul | markmc: it would also be nice if xenapi stored some useful info in the cpu_info as well :) | 16:25 |
| markmc | zul, it would :) | 16:25 |
| smoser | yeah, kbringard it would seem you do not need the pxedust script anymore. | 16:26 |
| jaypipes | BLZbubba: no worries :) | 16:26 |
| kbringard | even better | 16:26 |
| markmc | zul, ok, stuck it in https://bugs.launchpad.net/nova/+bug/1015190 | 16:28 |
| uvirtbot | Launchpad bug 1015190 in nova "arch_filter tracebacks with xenapi" [Low,Triaged] | 16:28 |
| *** dolphm has quit IRC | 16:28 | |
| zul | markmc: thanks | 16:28 |
| *** dolphm has joined #openstack-dev | 16:28 | |
| kbringard | smoser: yea, but it looks like you still need to specify the mac to bootif, right? or am I misreading the rest of the bug? | 16:31 |
| smoser | no. you're missing the genius | 16:32 |
| smoser | pxelinux knows which mac it booted from. it appends "BOOTIF=the-mac-address-here" to the kernel command line | 16:32 |
| *** dtroyer_zzz is now known as dtroyer | 16:32 | |
| *** dolphm has quit IRC | 16:33 | |
| kbringard | ok, right, that's what the pxedust script did, I didn't see where that functionality got integrated directly into netcfg, or I guess more specifically what I need to set to make it use it | 16:33 |
| smoser | and then if you say "auto" as the value for the netcfg preseed, then it reads that. | 16:33 |
| smoser | so the 2 things yo uave to do | 16:33 |
| *** garyk has joined #openstack-dev | 16:34 | |
| smoser | a.) preseed with netcfg/choose_interface as 'auto' | 16:34 |
| smoser | (which i guess can be specified on the kernel command line) | 16:34 |
| *** m4xmr has quit IRC | 16:34 | |
| smoser | b.) modify pxe linux config to have 'ipappend 2' | 16:34 |
| andrewbogott | markmc: the internet seems to agree that that will get me different objects. But doesn't that raise a big problem with trying to share objects with common? | 16:34 |
| smoser | then pxelinux adds "BOOTIF=" to the linux kernel cmdline, and the installer says "oh, by 'auto', you meant to respect BOOTIF=" | 16:35 |
| markmc | andrewbogott, not following, maybe show a small test case? | 16:35 |
| *** danpb has quit IRC | 16:35 | |
| kbringard | ah, OK, it was the ipappend option I wasn't understanding | 16:36 |
| *** andrewsben_zz is now known as andrewsben | 16:36 | |
| andrewbogott | markmc: nova/flags.py does "from nova.openstack.common import cfg" but common/log.py does "from openstack.common import cfg." Because of the two different import lines, cfg.CONF refers to different objects rather than a single common object. | 16:37 |
| markmc | andrewbogott, ah! | 16:38 |
| andrewbogott | But of course I can't use 'from nova.blahblah' from within common. | 16:38 |
| markmc | andrewbogott, common/log.py should do 'from nova.openstack.common import cfg' | 16:38 |
| andrewbogott | but common/log.py isn't necessarily in nova. It's common! | 16:38 |
| markmc | andrewbogott, cfg itself does 'from nova.openstack.common import iniparse' :) | 16:39 |
| markmc | andrewbogott, update.py changes the imports | 16:39 |
| andrewbogott | really? But then what if cfg is in e.g. glance instead... | 16:39 |
| andrewbogott | really? | 16:39 |
| markmc | yeah :) | 16:39 |
| * andrewbogott will say 'really?' a few more times, but in private. | 16:39 | |
| markmc | heh | 16:39 |
| andrewbogott | Does update.py look for the string 'nova' specifically? Or do I need to put some kind of $wildcard in the source? | 16:39 |
| winston-d | smoser, hi | 16:40 |
| markmc | it comes from openstack-common.conf | 16:40 |
| smoser | winston-d, hi | 16:41 |
| markmc | i.e. in openstack-common | 16:41 |
| markmc | it's 'from openstack.common import iniparse' | 16:41 |
| markmc | when it gets copied into glance | 16:41 |
| markmc | it's 'from glance.openstack.common import iniparse' | 16:41 |
| *** japage has quit IRC | 16:41 | |
| andrewbogott | markmc: got it | 16:41 |
| markmc | andrewbogott, what took you so long? | 16:42 |
| * markmc ducks :-P | 16:42 | |
| markmc | kidding | 16:42 |
| winston-d | smoser, i've encountered some problem with Nova metadata service, and I saw you checked in some patches about metadata. so I wonder maybe you know what is going on | 16:42 |
| smoser | winston-d, sure, whats up? they're regressions in the past week? or other issues. | 16:42 |
| markmc | andrewbogott, the update.py stuff sucks badly, but it's better than the copying-and-pasting stuff we were doing before | 16:42 |
| winston-d | smoser, so when i try to start metadata server via 'nova-api-meta', there's error: http://paste.openstack.org/show/18599/ | 16:44 |
| *** japage has joined #openstack-dev | 16:44 | |
| smoser | winston-d, markmc pointed out that earlier today. | 16:44 |
| smoser | your api-paste.ini needs updating | 16:45 |
| winston-d | smoser, which part? | 16:45 |
| smoser | well, likely the value in your git checkout is correct, but the file is being read from /etc | 16:45 |
| smoser | ie, you want 'sudo cp ./etc/nova/api-paste.ini /etc/nova/api-paste.ini' | 16:46 |
| winston-d | smoser, ok, let me try that. | 16:46 |
| smoser | (or fix the problem for good and: ln -sf $PWD/./etc/nova/api-paste.ini /etc/nova/api-paste.ini) | 16:46 |
| andrewbogott | markmc: OK, many more things are working now. Is there are guide someplace about proper practices when developing common? right now I just have a symlink from nova/openstack/common to /openstack-common/openstack/common which, I suspect, is not how you do it. | 16:47 |
| markmc | andrewbogott, nah, no docs ... but feel free to write something up | 16:48 |
| jaypipes | davidkranz: k, QA ML psot about negative tests sent/ | 16:48 |
| *** dhellmann has joined #openstack-dev | 16:49 | |
| markmc | andrewbogott, I mostly focus on test cases in openstack-common and then copy it across with update.py when it's mostly done | 16:49 |
| andrewbogott | markmc: That thing you just said about 'copy it across with update.py' is the part I don't know about. | 16:49 |
| andrewbogott | My tests work fine :( | 16:49 |
| *** danwent has quit IRC | 16:49 | |
| winston-d | smoser, thx. it works well now. :) | 16:50 |
| andrewbogott | markmc: OK, update.py seems to be self-documenting, so I'll tinker. | 16:50 |
| kbringard | smoser: the ipappend 2 thing works like beauty as well, you are amazing | 16:52 |
| jaypipes | smoser: you are SO cool today! | 16:52 |
| smoser | yeah. i think i'll go buy a lotto ticket, and see if i can't help kbringard out with the $million he asked about. | 16:53 |
| *** joesavak has joined #openstack-dev | 16:53 | |
| kbringard | haha | 16:53 |
| *** eglynn__ has quit IRC | 16:53 | |
| kbringard | if you win and give me a million, I'll even let you have the rest | 16:53 |
| kbringard | that's just how generous I am | 16:53 |
| kbringard | :-p | 16:53 |
| *** eglynn__ has joined #openstack-dev | 16:53 | |
| *** jsavak has quit IRC | 16:55 | |
| *** sulochan has quit IRC | 16:55 | |
| *** s0mik has joined #openstack-dev | 16:56 | |
| *** mnaser has quit IRC | 16:57 | |
| *** camm has joined #openstack-dev | 16:58 | |
| winston-d | how can i capture the unittest output of Nova since there's no run_tests.log anymore? | 16:59 |
| *** anderstj has joined #openstack-dev | 16:59 | |
| mtaylor | antonym: ping. | 17:00 |
| mtaylor | antonym: can you help us out with access to the etherpad machien? | 17:00 |
| *** jakedahn_zz is now known as jakedahn | 17:01 | |
| davidkranz | jaypipes: Great, Jay. That says it exactly right. | 17:03 |
| jaypipes | davidkranz: cheers | 17:03 |
| *** jdurgin has joined #openstack-dev | 17:04 | |
| jaypipes | winston-d: good question. perhaps mtaylor knows some tox magic? | 17:04 |
| *** utlemming has quit IRC | 17:04 | |
| winston-d | mtaylor, hi | 17:06 |
| winston-d | mtaylor, could you shed me some light on how i can capture the unittest output of Nova since there's no run_tests.log anymore? | 17:07 |
| *** danwent has joined #openstack-dev | 17:07 | |
| *** utlemming has joined #openstack-dev | 17:07 | |
| *** arosen has quit IRC | 17:09 | |
| *** davidha has joined #openstack-dev | 17:14 | |
| *** dachary has joined #openstack-dev | 17:16 | |
| *** adjohn has joined #openstack-dev | 17:16 | |
| *** epim has joined #openstack-dev | 17:19 | |
| *** harlowja has joined #openstack-dev | 17:19 | |
| jeblair | winston-d: i believe it is captured by nose, and output for failing tests. eg: https://jenkins.openstack.org/job/gate-nova-python26/763/testReport/junit/nova.tests.rpc.test_kombu/RpcKombuTestCase/test_publishing_errors_will_reconnect/ | 17:20 |
| jaypipes | winston-d: https://review.openstack.org/#/c/8680/1/glance/tests/__init__.py | 17:21 |
| jaypipes | winston-d: looks like that's how you can reenable it... | 17:21 |
| vishy | jeblair: no log for sucessful tests though ? | 17:21 |
| *** danwent has quit IRC | 17:22 | |
| jeblair | vishy: i think there's an option to nose to output everything | 17:23 |
| jeblair | vishy: -s | 17:23 |
| mtaylor | jaypipes: aroo? | 17:24 |
| jaypipes | mtaylor: regarding your https://review.openstack.org/#/c/8680/1/glance/tests/__init__.py | 17:24 |
| mtaylor | jaypipes: yes | 17:24 |
| mtaylor | jaypipes: that enables the log output to work properly with nosetests | 17:24 |
| mtaylor | jaypipes: since it already knows how to do log capturing and association with test output | 17:25 |
| vishy | jeblair: seems like we should modify run_tests to use that option then and throw the logs into a file like it used to | 17:25 |
| mtaylor | vishy: is that a useful thing? (asking just so I understand the use case) | 17:25 |
| mtaylor | so that I can be helpful and stuff | 17:26 |
| vishy | mtaylor: sure, if you're writing a test it is nice to be able to get output | 17:26 |
| mtaylor | vishy: k. but I mean it's useful to always produce a log file of all of the output on every run rather than just passing the 'gimme log output' flag? | 17:26 |
| jeblair | maybe you just want to run "nosetest -sv nova.test.newtestname" if you're building a new test...? | 17:27 |
| *** arosen has joined #openstack-dev | 17:27 | |
| vishy | mtaylor: even better might me to just make print work in the tests :) | 17:27 |
| mtaylor | vishy: well, nose will capture that - but shouldn't do if you run with -s | 17:28 |
| vishy | mtaylor: kind of hard to do that in a venv | 17:28 |
| mtaylor | vishy: 'tox -evenv nosetest -sv nova.test.newtestname' | 17:29 |
| vishy | i don't use them myself, but i think people do | 17:29 |
| vishy | does tox enable the openstack output btw? | 17:29 |
| mtaylor | it does | 17:30 |
| jgriffith | speaking of tests/output, anybody know what's up with all of the messages spewing from the nova tests? | 17:31 |
| jgriffith | Exception AttributeError: AttributeError("'_DummyThread' object has no attribute '_Thread__block'",) in <module 'threading' from '/usr/lib/python2.7/threading.pyc'> ignored | 17:31 |
| mtaylor | yeah, that's a low-level python thing, fixed in subsequent python version IIRC | 17:32 |
| vishy | jgriffith: it is a bug in python | 17:32 |
| jgriffith | Ahh, so we can't set an ignore for it or anything | 17:32 |
| vishy | jgriffith: http://bugs.python.org/issue14308 | 17:32 |
| jgriffith | In terms of the display output | 17:32 |
| jgriffith | vishy: thanks | 17:32 |
| vishy | jgriffith: the python patch is one line | 17:33 |
| vishy | if it is really annoying you | 17:33 |
| jgriffith | yeah, I'm going to pull it | 17:33 |
| *** darraghb has quit IRC | 17:33 | |
| jgriffith | Hell yeah it's annoying me :) | 17:33 |
| *** rafaduran has joined #openstack-dev | 17:34 | |
| mtaylor | jaypipes: sorry, we got distracted - did you have a question before I cut you off? | 17:35 |
| jaypipes | mtaylor: nope. | 17:35 |
| *** andrewbogott has quit IRC | 17:37 | |
| *** rkukura has joined #openstack-dev | 17:38 | |
| mtaylor | jaypipes: k. just making sure | 17:38 |
| *** torgomatic has joined #openstack-dev | 17:38 | |
| clayg | jgriffith: what was the subject on the message you sent to the ML about "run these patches to test devstack/cinder" | 17:41 |
| jgriffith | clayg: I'll have to look, or I can just give you the info again :) | 17:41 |
| *** Mandell has joined #openstack-dev | 17:41 | |
| clayg | jgriffith: even better! | 17:41 |
| jgriffith | clayg: You can either pull from j-griffith/github (devstack) | 17:41 |
| jgriffith | And it'll just pull everything from my repos which "should" be in sync for the most part and will default setting up Cinder | 17:42 |
| clayg | jgriffith: that works - i'll try that | 17:42 |
| jgriffith | Cool, lemme know | 17:42 |
| *** bencherian has joined #openstack-dev | 17:42 | |
| jgriffith | clayg: If you get a chance give exercises/euca.sh a shot... I'd like somebody else to run it | 17:43 |
| mtaylor | jaypipes: _this_ one: https://review.openstack.org/8704 is a bit more interesting | 17:45 |
| mtaylor | jaypipes: that's the attempt to allow us to collect server log files during/after test runs so that we can try to diagnose what happened during flaky runs | 17:45 |
| mtaylor | bcwaldon: ^^^ | 17:45 |
| jaypipes | ah, cool. | 17:45 |
| * jaypipes wants to see a markmc vs. mtaylor death match for client lib versioning. | 17:46 | |
| mtaylor | jaypipes: I do not want to see that | 17:46 |
| jaypipes | mtaylor: sorry, you don't get a vote. :) | 17:46 |
| mtaylor | jaypipes: neither does he | 17:47 |
| mtaylor | :) | 17:47 |
| jaypipes | lol | 17:48 |
| *** sulochan has joined #openstack-dev | 17:50 | |
| ayoung | heckj, is it possible that the openssl executable is not available on the machine that runs Jenkins? | 17:51 |
| heckj | ayoung: yes, quite. mtaylor: how could we best tell ^^ | 17:52 |
| jaypipes | which openssl? :) | 17:52 |
| *** sdague has quit IRC | 17:52 | |
| ayoung | jaypipes, the executable | 17:54 |
| jaypipes | ayoung: no, I meant `which openssl` ;) | 17:54 |
| ayoung | jaypipes, on F17 it is openssl-1.0.0j-1.fc17.x86_64 | 17:54 |
| ayoung | jaypipes, I don't have an ubuntu instance handy. Not sure what is running there. | 17:55 |
| jaypipes | jpipes@uberbox:~/repos/trystack-ops$ `which openssl` version | 17:55 |
| jaypipes | OpenSSL 1.0.0e 6 Sep 2011 | 17:55 |
| jaypipes | Xubuntu 11.10 | 17:56 |
| *** s0mik has quit IRC | 17:56 | |
| *** mestery has quit IRC | 17:57 | |
| ayoung | jaypipes, I'm guessing that will work just fine | 17:57 |
| *** GheRivero_ has joined #openstack-dev | 17:58 | |
| *** kmulvey has joined #openstack-dev | 17:59 | |
| *** dolphm has joined #openstack-dev | 18:00 | |
| *** GheRivero_ is now known as Ghe_Rivero | 18:01 | |
| Ghe_Rivero | hi people | 18:02 |
| *** mestery has joined #openstack-dev | 18:02 | |
| *** anderstj has quit IRC | 18:03 | |
| *** gyee has joined #openstack-dev | 18:04 | |
| *** Ryan_Lane has quit IRC | 18:04 | |
| *** adalbas has joined #openstack-dev | 18:05 | |
| *** apevec has quit IRC | 18:05 | |
| *** jog0 has joined #openstack-dev | 18:06 | |
| *** anotherjesse_zz is now known as anotherjesse | 18:06 | |
| *** s0mik has joined #openstack-dev | 18:07 | |
| *** andrewbogott has joined #openstack-dev | 18:07 | |
| *** anderstj has joined #openstack-dev | 18:07 | |
| * mtaylor apologizes to everyone for top posting in a mailing list | 18:07 | |
| mtaylor | heckj: openssl seems to be installed on the jenkins slaves | 18:09 |
| mtaylor | ayoung: are you hitting a problem we should look in to? | 18:09 |
| koolhead17 | hola Ghe_Rivero | 18:10 |
| winston-d | mtaylor, jeblair, jaypipes the unittest output fix you mentioned is for glance. but for nova, nova has its own logging module. | 18:11 |
| *** koolhead17 is now known as koolhead17|afk | 18:11 | |
| jaypipes | winston-d: for the tests? | 18:12 |
| mtaylor | winston-d: the equivelent nova patch should have been applied weeks ago | 18:13 |
| mtaylor | winston-d: can you explain to me more what you're trying to do that isn't working? | 18:13 |
| winston-d | jaypipes, yes, nova tests uses nova.log | 18:14 |
| winston-d | mtaylor, i didn't figure how to enable unittest output for nova. that's my problem. | 18:14 |
| mtaylor | winston-d: it should be being captured by nosetests and will associated with failed tests ... if you add -s to nosetests, it should disable tracelog output collection | 18:15 |
| mtaylor | winston-d, vishy: it seems like there might be a usecase for "I want to run the tests with all of the logging going to a file regardless of test success of failure" ... lemme see if can work up a patch that would allow you to enable that | 18:16 |
| winston-d | mtaylor, actually i'd like to debug my unittest, so i'd love to see debugging output not just failure cases. | 18:16 |
| mtaylor | winston-d: debugging tracelog output SHOULD be being collected with your test, and it SHOULD be output when the failure happens | 18:17 |
| mtaylor | winston-d: if it's not, that's a bug | 18:17 |
| mtaylor | winston-d: are you saying that the debug log output is not being output when your test case fails? | 18:17 |
| mtaylor | winston-d: (trying to understand what I should work on tracking down) | 18:18 |
| winston-d | mtaylor, no. i'd like to see the output at run time ( while the test case is running ). | 18:18 |
| vishy | winston-d: from earlier: nosetests -sv nova.tests.test_xxx.Class:method | 18:19 |
| vishy | oops | 18:19 |
| vishy | i mean xxx.:Class.method | 18:19 |
| mtaylor | winston-d: ^^ what vish just said should do what you want | 18:19 |
| ayoung | mtaylor, I'm trying to diagnose why the Signed tokens code fails the unit tests. | 18:19 |
| mtaylor | ayoung: you gotta link to a failed test? (I'll add it to my list of things to try to help out with) | 18:20 |
| *** anotherjesse is now known as anotherjesse_zz | 18:20 | |
| zykes- | jk0: ping | 18:21 |
| *** jakedahn is now known as jakedahn_zz | 18:21 | |
| ayoung | https://review.openstack.org/#/c/7754/ mtaylor | 18:21 |
| mtaylor | ayoung: thanks | 18:21 |
| *** jakedahn_zz is now known as jakedahn | 18:21 | |
| ayoung | the error message is that openssl returned nonzero error code. I don't dump the value of that output | 18:21 |
| ayoung | It is called using popen | 18:22 |
| BLZbubba | is there any sort of web interface like Horizon to manage the backend services; e.g. to see what vm's are where; migrate vm's to other hosts, etc ? | 18:22 |
| winston-d | vishy, mtaylor thx, that solved my problem. | 18:22 |
| mtaylor | winston-d: w00t! | 18:22 |
| ijw | Anyone know if people are tinkering with connection.py still? | 18:22 |
| ijw | It's much nicer than in Essex but I have a load of changes that I was going to backport that suddenly looks impractical. | 18:23 |
| *** dolphm_ has joined #openstack-dev | 18:23 | |
| ijw | (changes for my own nefarious purposes at the moment, rather than ones to be made public) | 18:24 |
| mtaylor | ayoung: yeah - all the debug logging there doesn't quite seem to be helpful does it? | 18:24 |
| *** dolphm has quit IRC | 18:24 | |
| ayoung | I guess I can try to add a bit more and resubmit the patch. Let me see... | 18:25 |
| *** sdague has joined #openstack-dev | 18:27 | |
| *** sulochan_ has joined #openstack-dev | 18:32 | |
| *** maoy has quit IRC | 18:32 | |
| *** winston-d is now known as winston-d_zzZZ | 18:34 | |
| *** sulochan has quit IRC | 18:35 | |
| *** sulochan_ is now known as sulochan | 18:35 | |
| *** Aaton_off is now known as Aaton | 18:39 | |
| BLZbubba | we already have a firewall/vlan infrastructure in place; how much pain would it be to have each tenant's vm's eth0 be on a particular vlan; but with nova-network not doing any dhcp or firewalling ? | 18:49 |
| zul | mtaylor: is run_test.sh the right way to run the tests now since everything has been converted to nose tests | 18:49 |
| *** markmc has quit IRC | 18:50 | |
| *** andyhky has quit IRC | 18:50 | |
| *** danwent_ has joined #openstack-dev | 18:53 | |
| *** pyw has quit IRC | 19:00 | |
| *** heckj has quit IRC | 19:02 | |
| *** johnpostlethwait has joined #openstack-dev | 19:03 | |
| ayoung | dolphm_, I find that hard to believe. | 19:03 |
| *** mgius has joined #openstack-dev | 19:03 | |
| *** comstud has quit IRC | 19:03 | |
| *** anotherjesse_zz is now known as anotherjesse | 19:03 | |
| ayoung | dolphm_, my last run, without the SQL patch: Ran 455 tests in 369.486s | 19:04 |
| *** andyhky has joined #openstack-dev | 19:04 | |
| *** novas0x2a|laptop has joined #openstack-dev | 19:06 | |
| dolphm_ | ayoung: just a straight up ./run_tests.sh ? | 19:11 |
| ayoung | dolphm_, yes | 19:12 |
| dolphm_ | ayoung: (approved the change btw) | 19:12 |
| dolphm_ | ayoung: /versions and /extensions | 19:12 |
| dolphm_ | ayoung: whoops... bad copy paste | 19:13 |
| ayoung | ? | 19:13 |
| dolphm_ | ayoung: http://paste.openstack.org/raw/18606/ | 19:13 |
| dolphm_ | there | 19:13 |
| ayoung | ok, something is wrong with my SQLite | 19:14 |
| dolphm_ | ayoung: want my whole output so you can compare? | 19:14 |
| ayoung | no | 19:14 |
| ayoung | I know that the problem is SQLite | 19:14 |
| ayoung | it is in the commits | 19:14 |
| bcwaldon | mtaylor: did we decide this was a good thing to do: https://review.openstack.org/#/c/8710/1 | 19:14 |
| dolphm_ | ayoung: are you testing master or your own branch? | 19:15 |
| ayoung | that one was my own branch, but the numbers are comparable | 19:15 |
| ayoung | test_add_user_to_tenant OK 2.93 | 19:16 |
| ayoung | test_authenticate OK 1.29 | 19:16 |
| ayoung | those are the first two SQL tests | 19:16 |
| ayoung | every one is > 1 second | 19:16 |
| ayoung | usually >2 seconds | 19:17 |
| dolphm_ | ayoung: ./run_tests.sh -N test_backend_sql http://paste.openstack.org/raw/18607/ | 19:19 |
| dolphm_ | ayoung: are you testing on your own hardware or in a cloud? | 19:20 |
| ayoung | My laptop | 19:20 |
| *** mnaser has joined #openstack-dev | 19:21 | |
| ayoung | I'll try it again in a few with the machine quiet | 19:21 |
| *** rkukura has quit IRC | 19:21 | |
| *** ayoung has quit IRC | 19:22 | |
| adam_g | speaking of tests, wtf: 16175 jenkins 20 0 14.6g 7.1g 8 D 1 91.1 5:30.08 nosetests | 19:23 |
| *** anderstj has quit IRC | 19:23 | |
| *** dachary has quit IRC | 19:23 | |
| *** gyee has quit IRC | 19:26 | |
| *** ayoung has joined #openstack-dev | 19:29 | |
| *** anotherjesse is now known as anotherjesse_zz | 19:30 | |
| ayoung | dolphm_, what kind of file system do you have Keystone checked out on? | 19:30 |
| dolphm_ | ayoung: mac extended / journaled | 19:31 |
| dolphm_ | ayoung: basic install of lion on an ssd | 19:31 |
| ayoung | dolphm_, I'm going to try it in a ramdisk and see what happens... | 19:32 |
| ayoung | test_authenticate OK 0.08 | 19:33 |
| dolphm_ | ayoung: nice lol | 19:33 |
| ayoung | running the whole suite. | 19:33 |
| dolphm_ | ayoung: let me know what the total time is :) | 19:33 |
| ayoung | Ran 455 tests in 26.124s | 19:33 |
| ayoung | I suspect that the culprit is the ext4 journaling | 19:34 |
| zykes- | shall I test ? :p | 19:35 |
| zykes- | got a samsung 256 gb ssd | 19:35 |
| ayoung | zykes-, please do | 19:35 |
| ayoung | you can try just one test | 19:35 |
| ayoung | ./run_tests.sh -N test_backend_sql:SqlIdentity.test_authenticate | 19:35 |
| dolphm_ | ayoung: this is a recent jenkins job: Ran 457 tests in 36.304s | 19:36 |
| zykes- | or nvm, windows box :p | 19:36 |
| dolphm_ | ayoung: "Ext4: Great overall performance, reliable, has performance issues with sqlite and some other databases." | 19:36 |
| ayoung | I wonder if I could move just the database to a ramdisk | 19:37 |
| dolphm_ | ayoung: Sqlite results in seconds> Ext4: 1690.87> Ext3: 108.77> Ext2: 50.67 | 19:38 |
| vishy | jog0: can you explain tags vs host aggregates? | 19:38 |
| jog0 | vishy: sure | 19:38 |
| jog0 | the | 19:38 |
| *** mjfork_ has joined #openstack-dev | 19:39 | |
| jog0 | the current host aggregates concept is tightly coupled with the XenAPI for example the XenAPI compute nodes directly modify the metadata, require state etc. | 19:39 |
| dolphm_ | ayoung: there's a feature in ext4 called "barriers" that you might be able to simply disable | 19:40 |
| jog0 | and after looking at what other people are trying to do (nova.conf based key, value pairs). I am exploring using tags. A tag will just be a key value pair for a computeNode that can be set via an API | 19:41 |
| ayoung | dolphm_, wait 1. I have another hack in mind...heheh | 19:41 |
| *** mjfork has quit IRC | 19:41 | |
| *** mjfork_ is now known as mjfork | 19:41 | |
| dolphm_ | ayoung: symlink the path to the test db? | 19:42 |
| ayoung | sudo mount -t tmpfs -o size=256M tmpfs /opt/stack/keystone/tests/ | 19:42 |
| ayoung | sudo chown ayoung tests/ | 19:42 |
| dolphm_ | ayoung: does it work? | 19:42 |
| ayoung | git checkout HEAD tests | 19:43 |
| vishy | jog0: The idea is to associate tags with groups of hosts though | 19:43 |
| ayoung | Ran 456 tests in 27.262s | 19:43 |
| vishy | jog0: is it really that coupled? | 19:43 |
| *** mnaser has quit IRC | 19:43 | |
| vishy | jog0: eww yuck compute hosts shouldn't be messing with the db like that | 19:44 |
| dolphm_ | ayoung: your changed merged btw | 19:44 |
| ayoung | dolphm_, thanks. | 19:44 |
| vishy | jog0: it doesn't look too bad | 19:47 |
| vishy | jog0: it seems to just be adding metadata to the aggregate | 19:47 |
| mtaylor | bcwaldon: I mean - _good_ think is a stretch - but it prolly won't hurt things | 19:47 |
| mtaylor | bcwaldon: hopefully between that and my patch to help collect server log files in jenkins, maybe we'll be able to diagnose | 19:48 |
| bcwaldon | mtaylor: honestly, increasing the timeout could go either way - maybe helpful or not | 19:48 |
| mtaylor | zul: imo, tox is the right way to run the tests - or the nosetest command if you don't want a venv | 19:48 |
| mtaylor | bcwaldon: good point | 19:48 |
| bcwaldon | mtaylor: its probably on the side of more good than bad | 19:49 |
| jog0 | vishy: Yes, that is the main one. I think I got rid of State, and most of the one aggregate per node requirment without breaking Xen too | 19:49 |
| bcwaldon | mtaylor: because if it is just a timing issue, then its probably out of our hands | 19:49 |
| bcwaldon | mtaylor: and if its not a timing issue, it will still time out | 19:49 |
| mtaylor | bcwaldon: well, mainly I'd like to see if we can determine what the issue actually is | 19:49 |
| ttx | vishy: about new rootwrap https://review.openstack.org/#/c/8530/ -- please let me know if you prefer a config file pointing to an area with filter definitions, or a config directory that includes config files and filter definitions alike | 19:49 |
| mtaylor | bcwaldon: and if it's timing, perhaps we can be cleverer about how we're starting things and stuff | 19:49 |
| ttx | vishy: if the latter, I'll rewrite tomorrow :) | 19:50 |
| *** dachary has joined #openstack-dev | 19:50 | |
| *** maoy has joined #openstack-dev | 19:50 | |
| vishy | i like the .d directory | 19:50 |
| jog0 | vishy: Modifying the current Aggregates code will also mean putting in a special case everywhere for generic hypervisor_pool support. setting an aggregate as a hypervisor_pool will only be allowed upon aggregate_creation | 19:50 |
| maoy | vishy: have you ever seen libvirt report SUSPENDED state? It's always SHUTOFF here after a VM is suspended to disk. | 19:51 |
| bcwaldon | mtaylor: I'm inclined to approve | 19:51 |
| maoy | vishy: I'm using Ubuntu 12.04 | 19:51 |
| jog0 | vishy: which I began to do | 19:52 |
| *** longhudou has quit IRC | 19:53 | |
| *** johnpur has joined #openstack-dev | 19:53 | |
| *** ChanServ sets mode: +v johnpur | 19:53 | |
| *** mgius has quit IRC | 19:54 | |
| vishy | maoy: not sure, virsh suspend gives you shutoff? | 19:55 |
| jog0 | vishy: just to clarify: You want each compute node to be allowed to have multiple aggregates, and each aggregate to have multiple sets of key,value metadata? | 19:56 |
| vishy | yup | 19:56 |
| jog0 | vishy: where the same key,value pair can be in two different aggregates? | 19:56 |
| vishy | jog0: i don't understand why you need the special case | 19:56 |
| maoy | vishy: i use the nova suspend command. I think virsh suspend actually means pause | 19:57 |
| vishy | jog0: we could just have a separate command for registering the aggregate with the hypervisor | 19:57 |
| vishy | jog0: at least that was how we discussed it during the session | 19:57 |
| vishy | maoy: yes i seem to get paused or shutoff | 20:01 |
| *** anotherjesse_zz is now known as anotherjesse | 20:01 | |
| *** rafaduran has quit IRC | 20:02 | |
| *** ncode has quit IRC | 20:02 | |
| maoy | vishy: virsh suspend -> paused, nova suspend -> shutoff ? | 20:02 |
| maoy | vishy: i'm sorting out the nasty details of sync_power_state. this came up | 20:03 |
| vishy | maoy: nova pause == virsh suspend -> paused , nova suspend == virsh managedsave -> shutoff | 20:04 |
| maoy | vishy: great. same observation here. | 20:04 |
| jog0 | vishy: There definitely can be a separate command - but it would have an overlapping code-path with the regular option, the special case is needed because you need to inform the compute-node that it is part of the hypervisor_pool, make sure each node is part of only one hypervisor_pool. | 20:04 |
| *** heckj has joined #openstack-dev | 20:04 | |
| jog0 | vishy: Most of the unusual xenAPI Host Aggregate code is in here: nova/virt/xenapi/pool.py | 20:05 |
| *** GiBa has joined #openstack-dev | 20:05 | |
| GiBa | hello | 20:05 |
| vishy | jog0: i think we can leave that to users honestly | 20:05 |
| jog0 | vishy: leave setting up resource_pools? | 20:05 |
| vishy | jog0: as in the pool_register(aggregate) can fail | 20:05 |
| vishy | jog0: so users can happily create aggregates, then they can register them as a pool | 20:06 |
| vishy | if they register two, then it just goes boom | 20:06 |
| *** joesavak has quit IRC | 20:06 | |
| jog0 | vishy: OK Sounds good to me. I will renew my decoupleXenAggregates branch | 20:06 |
| *** Ryan_Lane has joined #openstack-dev | 20:06 | |
| vishy | jog0: cool, I'm trying to avoid an explosion in metadata | 20:07 |
| *** wiliam has quit IRC | 20:07 | |
| jog0 | vishy: its almost there if we can leave the caution up to the user | 20:07 |
| vishy | if we have tags and capabilites and aggregates i think we are going to confuse the crap out of everyone | 20:07 |
| vishy | jog0: I want everything to be aggregate metadata | 20:07 |
| *** comstud has joined #openstack-dev | 20:07 | |
| maoy | vishy: I don't know what would the xenapi say about the suspend state. But I'll treat SHUTDOWN and SUSPEND both valid if vm_state is SUSPENDED | 20:08 |
| jog0 | vishy: Sounds good, I hope to get the general aggregates into F-2 but expect to be moving 'everything' to aggregate metadata in F-3 | 20:08 |
| vishy | maoy: seems reasonable for now | 20:08 |
| vishy | maoy: you could also put some logic in the vm_state in libvirt to suspended if has_managed_save() | 20:09 |
| vishy | maoy: but that might be excssive | 20:09 |
| *** GiBa has left #openstack-dev | 20:09 | |
| jog0 | vishy: thanks | 20:09 |
| *** davidha has quit IRC | 20:10 | |
| *** ohnoimdead has joined #openstack-dev | 20:10 | |
| *** dprince has quit IRC | 20:12 | |
| maoy | vishy: oh you meant in the libvirt driver get_info(). I thought you meant in libvirt itself for a second. | 20:14 |
| *** mjfork has quit IRC | 20:14 | |
| *** mjfork_ has joined #openstack-dev | 20:15 | |
| vishy | maoy: yeah :) | 20:15 |
| *** mjfork has joined #openstack-dev | 20:16 | |
| *** davidha has joined #openstack-dev | 20:19 | |
| *** mjfork_ has quit IRC | 20:19 | |
| maoy | vishy: when a VM is resized, waiting for verify, and I make a call to db.instance_get_all_by_host(), does it appear in src, dest, or both? | 20:21 |
| vishy | i believe it shows source host until the resize is complete at which point it shows dest | 20:21 |
| vishy | but it does NOT wait for verify to change the host | 20:21 |
| *** glenc has joined #openstack-dev | 20:22 | |
| *** Mandell has quit IRC | 20:23 | |
| maoy | vishy: you lost me.. | 20:23 |
| maoy | vishy: by resize is complete, what do you mean? | 20:23 |
| vishy | maoy: as in the new host has come up | 20:24 |
| *** zigo-_- has quit IRC | 20:24 | |
| vishy | * guest | 20:24 |
| maoy | vishy: vishy: before confirm/revert? | 20:24 |
| ayoung | dolphm_, my patch 3 had some issues dealing with error in CMS. I was trying to get better error reporting. | 20:24 |
| *** salgado is now known as salgado-afk | 20:25 | |
| *** glenc_ has quit IRC | 20:25 | |
| *** mjfork has quit IRC | 20:26 | |
| vishy | maoy: correct confirm is after the guest has moved | 20:26 |
| BLZbubba | does flavorid do anything other than sort order in horizon? | 20:26 |
| *** Mandell has joined #openstack-dev | 20:26 | |
| maoy | vishy: ok | 20:26 |
| vishy | maoy: although i should check. Maybe the old guest stays up | 20:26 |
| *** mjfork has joined #openstack-dev | 20:27 | |
| *** danwent_ has quit IRC | 20:29 | |
| *** danwent has joined #openstack-dev | 20:29 | |
| *** tgall__ has joined #openstack-dev | 20:36 | |
| *** Dr_Who has quit IRC | 20:36 | |
| *** zigo has joined #openstack-dev | 20:37 | |
| *** ncode has joined #openstack-dev | 20:47 | |
| jog0 | has anyone had unit test problems on OS X with "XenSMTestCase:test_create_volume"? | 20:49 |
| *** stuntmac_ has quit IRC | 20:52 | |
| andrewbogott | russellb: Do you have a moment to help me diagnose a problem with common.importutils? | 20:53 |
| andrewbogott | russelb: http://pastebin.com/BV87h8iB | 20:53 |
| *** Shrews has quit IRC | 20:54 | |
| *** zigo has quit IRC | 20:54 | |
| notmyname | bcwaldon: http://www.bioteams.com/2006/01/13/the_maximum_team.html | 20:54 |
| notmyname | bcwaldon: the first link has some good info :-) | 20:54 |
| bcwaldon | notmyname: excellent | 20:55 |
| *** japage has quit IRC | 20:55 | |
| vishy | andrewbogott: looks like there is an "import log" in that file somewhere? | 20:55 |
| russellb | well, looks like a nova.compute import failed, and that is the str() value of the original exception | 20:56 |
| maoy | vishy: does instance_get_all_by_host() return (soft) deleted instances? | 20:56 |
| russellb | so python -c 'import nova.compute' also fails? | 20:56 |
| russellb | andrewbogott: ^ | 20:57 |
| andrewbogott | russelb: Yes, that also fails. | 20:57 |
| russellb | and if so, do you get a more useful traceback? | 20:57 |
| vishy | maoy: I think so. The host should still be set until the instance is actually deleted. | 20:57 |
| andrewbogott | russellb: http://pastebin.com/fymivp60 | 20:58 |
| maoy | vishy: when vm_state is DELETED, is the host still set? | 20:58 |
| andrewbogott | I am messing with logging, so it's no surprise that I've broken something, I just can't figure out where... | 20:58 |
| maoy | vishy: there is also a deleted=True in mysql table. just want to make sure which deleted you mean... | 20:59 |
| vishy | jaypipes: does the error only occur when multiple builds are happening simultaneously? | 20:59 |
| russellb | gotcha ... like vishy said earlier, do you have a import log somewhere? | 20:59 |
| vishy | maoy: in fact i don't think the host is ever removed | 20:59 |
| vishy | maoy: so get_all_by_host filters on host and deleted=1 in the db | 20:59 |
| vishy | maoy: so it should still be returned afaik | 21:00 |
| jaypipes | vishy: by "simultaneously", what do you mean? | 21:00 |
| *** lts has quit IRC | 21:00 | |
| jaypipes | vishy: if you mean when running with --processes=X, no. | 21:00 |
| jaypipes | vishy: occurs when running tempest in a single process. | 21:00 |
| vishy | jaypipes: i mean multiple requests at once | 21:00 |
| andrewbogott | russellb: Yes, but python -c 'import nova.openstack.common.log' works fine... | 21:00 |
| vishy | jaypipes: i don't get failures running instances on devstack and neither does the gate | 21:00 |
| andrewbogott | from nova.openstack.common import log as logging in compute/api.py | 21:01 |
| *** gabrielhurley has joined #openstack-dev | 21:01 | |
| vishy | jaypipes: so I"m trying to figure out how to reproduce | 21:01 |
| jaypipes | vishy: yeah, I'm very confused, frankly | 21:01 |
| jaypipes | vishy: have you run tempest against devstack? | 21:01 |
| russellb | andrewbogott: python -c 'import nova.compute.api' fails too? | 21:01 |
| *** Ryan_Lane has quit IRC | 21:01 | |
| vishy | jaypipes: nope | 21:02 |
| *** Ryan_Lane has joined #openstack-dev | 21:02 | |
| *** steveb_ has joined #openstack-dev | 21:02 | |
| andrewbogott | russellb: Yes, same failure. Hm... | 21:02 |
| *** JStoker has quit IRC | 21:02 | |
| *** agonella has quit IRC | 21:02 | |
| russellb | andrewbogott: i can try to update importutils to give a better error ... | 21:02 |
| *** agonella has joined #openstack-dev | 21:03 | |
| andrewbogott | That might be worthwhile. I'm confused by the nested Class API cannot be found (Class API cannot be found | 21:03 |
| andrewbogott | thought maybe I was causing a recursive import or something. | 21:03 |
| *** Ghe_Rivero has quit IRC | 21:04 | |
| andrewbogott | curious, if I remove the line that imports logging from api.py, the error does not change | 21:04 |
| jaypipes | vishy: I'll keep digging. going to try running each individual failing test until I can see some pattern | 21:05 |
| *** sdake has quit IRC | 21:05 | |
| russellb | andrewbogott: http://paste.openstack.org/show/18611/ | 21:06 |
| bcwaldon | gabrielhurley: any of you nebulos interested in fixing this guy https://bugs.launchpad.net/glance/+bug/1009248 | 21:06 |
| uvirtbot | Launchpad bug 1009248 in glance "Glance HTTP Store query method fails to check response code" [Medium,Confirmed] | 21:06 |
| *** alaski has quit IRC | 21:07 | |
| andrewbogott | russellb: that helped! thank you. | 21:08 |
| russellb | cool, i'll submit that to openstack-common then | 21:10 |
| dhellmann | markwash: can I get you to have a quick look at https://review.openstack.org/#/c/8659/ for me when you have a minute? | 21:10 |
| *** rkukura has joined #openstack-dev | 21:10 | |
| markwash | sure | 21:11 |
| *** markvoelker has quit IRC | 21:11 | |
| *** mgz has quit IRC | 21:12 | |
| *** JStoker has joined #openstack-dev | 21:12 | |
| russellb | andrewbogott: https://review.openstack.org/8718 | 21:12 |
| maoy | vishy: is there existing code to handle case where the VM is marked as STOPPED/DELETED/SOFT_DELETED but power_state says RUNNING? | 21:13 |
| *** nunosantos has quit IRC | 21:14 | |
| markwash | dhellmann: looks good, thanks for preserving backwards compat | 21:14 |
| vishy | maoy: i don't think so | 21:15 |
| *** anotherjesse is now known as anotherjesse_zz | 21:15 | |
| vishy | jaypipes: cool thanks. I will look through the code and see if I can find anything. | 21:16 |
| jaypipes | vishy: would a prior call to nova-network that did not release a semaphore or lockfile properly cause an RPC timeout on the compute node during allocate_network()? | 21:16 |
| vishy | jaypipes: possibly | 21:16 |
| vishy | jaypipes: although i think you would see a log on the network host about acquiring the semaphore | 21:17 |
| jaypipes | vishy: I do. but I don't see any messages releasing it. | 21:18 |
| jaypipes | vishy: is there a way to tell if a lockfile is still around for nova-network (in a devstack env):? | 21:18 |
| vishy | jaypipes: yes look for a lockfile! | 21:19 |
| jaypipes | vishy: sorry, yes, I know that... looking for which directory. | 21:19 |
| vishy | i think they are in /opt/stack/nova | 21:19 |
| ijw | vishy: do you know if there are other changes due to land that affect connection.py? | 21:20 |
| ijw | (in libvirt) | 21:20 |
| jaypipes | vishy: I see a nova-ensure_bridge.lock in /opt/stack/nova/instances owned by the libvirt group. Would that be it? | 21:20 |
| *** kindaopsdevy has joined #openstack-dev | 21:20 | |
| vishy | yeah probably, but why is it in instances? | 21:20 |
| vishy | that is weird | 21:21 |
| jaypipes | vishy: sorry, in /opt/stack/nova, not /opt/stack/nova/instances | 21:21 |
| jaypipes | my bad. | 21:21 |
| vishy | oh ok yeah that would be ensure_bridge.lock it is still there? | 21:21 |
| jaypipes | vishy: yes, after running just one of tempest's tests, on a fresh devstack env, test failed (with that timeout error) and there is a "nova-ensure_bridge.lock" file in /opt/stack/nova | 21:22 |
| jaypipes | I don't see any other lock files. | 21:22 |
| vishy | jaypipes what commands does the test do? | 21:22 |
| jaypipes | vishy: one sec. | 21:22 |
| *** metral_ has joined #openstack-dev | 21:23 | |
| BLZbubba | to turn a snapshot into an image, do i just want to delete the extra properties e.g. "image_location" | 21:23 |
| *** sieutruc has quit IRC | 21:23 | |
| vishy | jaypipes: it seems like the likely thing is one of the commands in ensure_bridge hanging | 21:24 |
| vishy | jaypipes: so it might be nice to check the logs to see the last command that was run | 21:24 |
| jaypipes | http://paste.openstack.org/show/18612/ and then http://paste.openstack.org/show/18612/ | 21:25 |
| jaypipes | vishy: which log? net or compute? | 21:25 |
| *** metral has quit IRC | 21:25 | |
| *** metral_ is now known as metral | 21:25 | |
| jaypipes | BLZbubba: snapshot is already an image. | 21:25 |
| vishy | jaypipes: maybe both, they both call ensure bridge | 21:26 |
| jaypipes | BLZbubba: just boot it from the snapshots view in Horizon. | 21:26 |
| jaypipes | vishy: k, lemme grep for ensure_bridge | 21:26 |
| jaypipes | vishy: boom. I think we may have found our issue... | 21:27 |
| vishy | jaypipes: \o/ | 21:28 |
| *** salgado-afk is now known as salgado | 21:29 | |
| jaypipes | vishy: can you glean anything from this output? http://paste.openstack.org/show/18614/ | 21:29 |
| BLZbubba | jaypipes: ok cool, is there a reason why they are in separate areas? just curious why there are two different sections | 21:30 |
| jaypipes | vishy: if you give me some pointers, I can look deeper into the logs for something... right now, I can't seem to tell whether the lockfile is ever acquired (just says attempting to acquire) | 21:30 |
| vishy | jaypipes: can you send me the logs | 21:30 |
| jaypipes | BLZbubba: because images are "base images" and snapshots are essentially backups of an instance, saved to an image | 21:31 |
| jaypipes | vishy: sure, lemme tar em up. | 21:31 |
| vishy | jaypipes: I would be looking for the first ensure that works and looking for an executed command that is not returning | 21:31 |
| jaypipes | vishy: I will send you the api, net, cpu logs. any others? | 21:31 |
| vishy | i think that is enough | 21:32 |
| *** agonella has quit IRC | 21:32 | |
| *** agonella has joined #openstack-dev | 21:33 | |
| *** dachary has quit IRC | 21:36 | |
| *** matwood has quit IRC | 21:37 | |
| jaypipes | vishy: k, sent. thx man. | 21:38 |
| *** ncode has quit IRC | 21:39 | |
| *** andrewbogott has quit IRC | 21:41 | |
| *** andrewbogott has joined #openstack-dev | 21:46 | |
| *** andrewbogott has joined #openstack-dev | 21:46 | |
| *** nati_ueno has joined #openstack-dev | 21:48 | |
| vishy | jaypipes: as far as i can tell the bridge lock was there at the start of all the runs | 21:49 |
| dhellmann | markwash: thanks for the review! | 21:50 |
| jaypipes | vishy: hmm... ok. I will re-run after removing it by force... | 21:50 |
| jaypipes | vishy: what could leave that around? | 21:51 |
| vishy | jaypipes: well if something is killed while it has a lock open | 21:51 |
| jaypipes | k | 21:51 |
| *** Shrews has joined #openstack-dev | 21:52 | |
| vishy | it is supposed to remove them on startup but our lock removal code thinks something is still linking to it | 21:52 |
| *** tgall__ has quit IRC | 21:54 | |
| *** dolphm_ has quit IRC | 21:55 | |
| *** dhellmann has quit IRC | 21:55 | |
| adam_g | has anyone else run into issues with the memory requirements of running the volume tests in nova/tests/test_xensm.py | 21:57 |
| adam_g | ? | 21:57 |
| *** ayoung has quit IRC | 22:01 | |
| *** heckj has quit IRC | 22:04 | |
| *** matwood has joined #openstack-dev | 22:10 | |
| *** anderstj has joined #openstack-dev | 22:12 | |
| *** roge has quit IRC | 22:13 | |
| *** glenc_ has joined #openstack-dev | 22:17 | |
| lcheng | gabrielhurley: hi | 22:17 |
| *** glenc has quit IRC | 22:19 | |
| gabrielhurley | lcheng: hi, what's up? (I'm walking out the door in a couple minutes) | 22:21 |
| *** iccha has left #openstack-dev | 22:21 | |
| *** kbringard has quit IRC | 22:22 | |
| lcheng | gabrielhurley: I have some question about: https://bugs.launchpad.net/horizon/+bug/1003299 | 22:22 |
| uvirtbot | Launchpad bug 1003299 in horizon "broken instance linkage under project overview page" [High,Confirmed] | 22:23 |
| lcheng | gabrielhurley: when the user clicks on the instance link, which page should it open? | 22:23 |
| lcheng | gabrielhurley: in the code review comment. You mentioned that it should open tenant index. Is it the tenant page in the admin dashboard (/syspanel/projects/<tenant_id>/usage) ? | 22:26 |
| *** e1mer has joined #openstack-dev | 22:27 | |
| *** maoy has quit IRC | 22:29 | |
| gabrielhurley | lcheng: my recollection is that the instance ID isn't actually returned (I believe I filed a Nova bug about that), so we can't link to the instance detail page. Since this is in the Project dashboard there is not "tenant" view, so we can't link to that. If there is no logical thing we can link to then it just shouldn't be a link. | 22:31 |
| gabrielhurley | If we can get back the instance ID it should link to the instance detail page. That's really the only logical solution. It's just not available right now. | 22:31 |
| gabrielhurley | I gotta run, though, sorry | 22:31 |
| *** sulochan has quit IRC | 22:32 | |
| *** adjohn has quit IRC | 22:33 | |
| *** sulochan has joined #openstack-dev | 22:33 | |
| *** cooper has joined #openstack-dev | 22:35 | |
| *** gabrielhurley has quit IRC | 22:36 | |
| *** Gordonz has quit IRC | 22:37 | |
| *** sulochan has quit IRC | 22:37 | |
| *** sdake has joined #openstack-dev | 22:37 | |
| *** roge has joined #openstack-dev | 22:38 | |
| *** galstrom has quit IRC | 22:41 | |
| *** mattray has quit IRC | 22:45 | |
| *** metral has quit IRC | 22:48 | |
| *** AlanClark has quit IRC | 22:48 | |
| *** salgado is now known as salgado-afk | 22:49 | |
| *** metral has joined #openstack-dev | 22:49 | |
| *** littleidea has joined #openstack-dev | 22:51 | |
| *** jakedahn is now known as jakedahn_zz | 22:53 | |
| *** epim has quit IRC | 22:57 | |
| *** metral has quit IRC | 22:59 | |
| *** steveb_ has quit IRC | 23:00 | |
| *** steveb_ has joined #openstack-dev | 23:04 | |
| *** dolphm has joined #openstack-dev | 23:05 | |
| *** ewindisch has joined #openstack-dev | 23:06 | |
| *** lloydde has quit IRC | 23:07 | |
| *** dolphm has quit IRC | 23:10 | |
| *** rnirmal has quit IRC | 23:10 | |
| *** dtroyer is now known as dtroyer_zzz | 23:11 | |
| *** danwent_ has joined #openstack-dev | 23:12 | |
| *** kindaopsdevy has quit IRC | 23:12 | |
| *** danwent has quit IRC | 23:12 | |
| *** danwent_ is now known as danwent | 23:12 | |
| *** kindaopsdevy has joined #openstack-dev | 23:12 | |
| jgriffith | vishy: ping | 23:14 |
| *** andrewsmedina has quit IRC | 23:17 | |
| clayg | jgriffith: fwiw, i had a bit of problem with your cinder/devstack for not running glance-manage db_sync | 23:18 |
| *** jakedahn_zz is now known as jakedahn | 23:20 | |
| *** winston-d_zzZZ has quit IRC | 23:23 | |
| *** sleepsonthefloor is now known as sleepsonzzz | 23:24 | |
| jgriffith | clayg: Hmmm... could be that my pkgs aren't in sync with what you need? | 23:26 |
| clayg | jgriffith: oh idk... I guess I just pulled down trunk glance (this is a fresh box) | 23:27 |
| clayg | and the registry would start, but would throw 500's complaining about tables not existing | 23:27 |
| jgriffith | clayg: Hmmm... if they need any changes in nova from the past few weeks I could see that being a problem. | 23:27 |
| clayg | I could run glance-manage db_sync and get glance image-list to work, then I noticed that your cinder/devstack for didn't have that command in where I found it in upstrem devstack | 23:28 |
| jgriffith | clayg: We should have this in review and merged soon and it may solve some of this depending on what you're seeing | 23:28 |
| clayg | jgriffith: yes, that will be nice | 23:28 |
| jgriffith | Ahhhh... yes, that devstack has been floating around for quite a bit as well. Maybe trying a rebase even? | 23:28 |
| clayg | I'm really a devstack newb - I'm taking maybe it doesn't pull from master every time you start it? (is there a way to start it besides ./stack.sh) | 23:29 |
| jgriffith | clayg: Or just take it as beta right now and see what happens later this week :) | 23:29 |
| clayg | ... like if I just want to make a code change and restart all services | 23:29 |
| jgriffith | clayg: No... because my version is set to pull from my repos, otherwise you won't get the neeeded cinder changes | 23:29 |
| clayg | jgriffith: oh.. like just try and merge upstream devstack ontop of cinder/devstack? | 23:29 |
| clayg | well I'm definately getting the cinder changes, that all works good | 23:29 |
| jgriffith | clayg: yeah... | 23:30 |
| jgriffith | clayg: Then I would imagine it's a synch issue with all the pkgs | 23:30 |
| clayg | I'm just trying to understand I guess why I would run into this and you hadn't had a problem with it yet (my glance is newer?) - cause I'm assuming you haven't updated glance since you made the fork? | 23:30 |
| jgriffith | clayg: Really not worth you spending the time right now to dork with it I don't think | 23:30 |
| clayg | jgriffith: will I wouldn't... except it's sorta the only way I can really see to get cinder running atm | 23:31 |
| jgriffith | clayg: Maybe send me an outline of what you ran and what you saw and I'll see if it's something I can coorelate | 23:31 |
| jgriffith | I've been using bcwaldons vagrant setup with a modified source for devstack repo | 23:32 |
| clayg | jgriffith: i just added the glance-mange db_sync to my local checkout of your devstack and basically everything seemed to start | 23:32 |
| clayg | I was able to create a volume with cinder client... | 23:32 |
| *** pixelbeat has quit IRC | 23:32 | |
| clayg | jgriffith: oh.... that sounds kinda nice :) | 23:32 |
| jgriffith | clayg: yes!! vish and bcwaldon have shown me a few tricks like hub and vagrant that make life much easier!! | 23:33 |
| jgriffith | clayg: anyway, I'm not sure exactly what you might have run into, I'd like to wait until we're in the official repos though based on what you describe before I worry much about it. | 23:34 |
| clayg | jgriffith: hub is keyerror on my brain | 23:34 |
| jgriffith | clayg: google defunkt hub | 23:34 |
| jgriffith | clayg: let's you do like 'git clone openstack/nova' | 23:34 |
| jgriffith | clayg: shortcuts... also creating remotes on your github account etc | 23:34 |
| clayg | jgriffith: yeah boss it's fine, i was just saying your devstack fork didn't work for me out of the box - sounds like it's only an issue for me - but I think i've worked around it for now | 23:35 |
| * clayg is still waiting to see if he can get an instance booted to attach this volume to | 23:35 | |
| jgriffith | clayg: It better :) | 23:35 |
| clayg | jgriffith: the git remote add expanding to add another users fork of the current repo seems simple and handy... may be worth all the ruby trouble. | 23:37 |
| jgriffith | clayg: yeah, I definitely found it worthwhile... | 23:38 |
| jgriffith | clayg: If nothing else just the shortcuts alone are worth it | 23:38 |
| *** andrewbogott has quit IRC | 23:39 | |
| clayg | mmmhhhh looks like it can support github enterprise too... | 23:39 |
| *** edygarcia has joined #openstack-dev | 23:40 | |
| *** anderstj_ has joined #openstack-dev | 23:40 | |
| *** anderstj has quit IRC | 23:40 | |
| *** torgomatic has quit IRC | 23:40 | |
| *** Dr_Who has joined #openstack-dev | 23:40 | |
| *** Dr_Who has quit IRC | 23:40 | |
| *** Dr_Who has joined #openstack-dev | 23:40 | |
| *** torgomatic has joined #openstack-dev | 23:43 | |
| *** anderstj_ has quit IRC | 23:43 | |
| *** ayoung has joined #openstack-dev | 23:43 | |
| clayg | jgriffith: so compute blew up "no protal found" and indeed "iscsiadm -m discovery -t st -p localhost" comes up empty | 23:43 |
| *** anderstj has joined #openstack-dev | 23:43 | |
| clayg | but tgt is running and seems to be listening on 3260 | 23:43 |
| clayg | the volume says it's available... | 23:43 |
| jgriffith | clayg: hmmmm... ok, we've got something funny in your setup me suspects | 23:43 |
| clayg | jgriffith: like I said, it was a clean install... | 23:44 |
| clayg | but it is precise... are most folks still running devstack on oneiric? | 23:44 |
| jgriffith | clayg: using j-griffith/devstack right? | 23:44 |
| jgriffith | clayg: Nope, I'm using precise | 23:44 |
| clayg | origin git://github.com/j-griffith/devstack.git (fetch) | 23:44 |
| jgriffith | clayg: errr... that will merge things. I'd recommend 'rm -rf devstack;git clone j-griffith/devstack' | 23:45 |
| *** danwent has quit IRC | 23:45 | |
| jgriffith | clayg: then create your localrc and run stack.sh again | 23:46 |
| jgriffith | clayg: Just to be 'safe' even rm -rf /opt/stack and /etc/nova | 23:46 |
| jgriffith | clayg: If you wait five minutes I"ll push new udpates to my repos | 23:46 |
| clayg | sure | 23:47 |
| *** Dr_Who has quit IRC | 23:48 | |
| *** blamar has quit IRC | 23:48 | |
| *** edygarcia has quit IRC | 23:48 | |
| *** torgomatic_ has joined #openstack-dev | 23:48 | |
| *** adjohn has joined #openstack-dev | 23:49 | |
| *** anderstj has quit IRC | 23:49 | |
| *** anderstj_ has joined #openstack-dev | 23:49 | |
| *** roge has quit IRC | 23:50 | |
| *** torgomatic has quit IRC | 23:52 | |
| *** torgomatic_ is now known as torgomatic | 23:52 | |
| *** anderstj_ has quit IRC | 23:54 | |
| *** dhellmann has joined #openstack-dev | 23:57 | |
| *** matwood has quit IRC | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!