| devananda | dlaube: you need to use the same release of nova and ironic. it /should/ work with the icehouse releases... but juno rc's would make it much easier for us to help you debug things | 00:00 |
|---|---|---|
| dlaube | I completely understand | 00:01 |
| devananda | dlaube: as far as a user for ironic, it needs to be a service user, with admin privileges | 00:01 |
| devananda | but the error you're getting sounds like it's not lacking permissions, it's not passing the token somehow | 00:01 |
| dlaube | alrighty. its all very odd since ironic can auth fine to keystone otherwise… it just can seem to get a token when nova is talking to ironic | 00:01 |
| devananda | which is what makes me thinkit's a library version mismatch or such | 00:01 |
| devananda | right | 00:02 |
| dlaube | I'll go roll out nova from master and try again | 00:02 |
| devananda | dlaube: what version of python-ironicclient is installed? | 00:02 |
| dlaube | root@lab5:/etc/nova# ironic --version | 00:03 |
| dlaube | 0.3.0.14 | 00:03 |
| devananda | heh | 00:03 |
| devananda | silly packagers. I have no idea what that version is | 00:03 |
| devananda | :) | 00:03 |
| dlaube | root@lab5:/etc/nova# pip list | grep ironic | 00:03 |
| dlaube | ironic (2015.1.dev22.gce309a6) | 00:03 |
| dlaube | python-ironicclient (0.3.0.14.gaeabe6b) | 00:03 |
| devananda | dlaube: what distro? | 00:04 |
| dlaube | ubuntu 14.04.01 LTS | 00:04 |
| devananda | cool, thanks | 00:05 |
| *** eghobo has joined #openstack-ironic | 00:10 | |
| devananda | dlaube: huh. I'm wondering if ubuntu's release of that is just too old | 00:10 |
| devananda | please run: dpkg-query --list '*ironic*' | 00:11 |
| dlaube | devananda: of python-ironicclient? | 00:11 |
| devananda | ya | 00:11 |
| dlaube | ok, 1min | 00:11 |
| devananda | http://changelogs.ubuntu.com/changelogs/pool/universe/p/python-ironicclient/python-ironicclient_0.1.2-0ubuntu1/changelog | 00:11 |
| dlaube | http://pastie.org/private/zswecvja06ywovrlyx77w | 00:11 |
| dlaube | those are the old packages but I now have ironic installed from master | 00:12 |
| devananda | hm. that doesn't list the client at all | 00:12 |
| devananda | k | 00:12 |
| dlaube | I think I should backup my configs and then uninstall those old ironic packages | 00:12 |
| devananda | ++ | 00:13 |
| JayF | dlaube: I strongly suggest running any openstack services you do from pip and in a virtualenv | 00:13 |
| JayF | dlaube: but ymmv :) | 00:13 |
| devananda | dlaube: oh - if you can pull from utopic, you'll get packages of juno-rc2 | 00:14 |
| dlaube | heh, not sure how to do that | 00:15 |
| dlaube | googling | 00:16 |
| dlaube | this? https://launchpad.net/ubuntu/+source/ironic/2014.2~rc2-0ubuntu1 | 00:17 |
| devananda | dlaube: http://packages.ubuntu.com/search?suite=utopic&searchon=names&keywords=ironic | 00:17 |
| *** Kui has quit IRC | 00:17 | |
| dlaube | ahh, thanks | 00:19 |
| dlaube | I had originally tried adding these to my sources.list but it didnt work for some reason: deb http://ppa.launchpad.net/openstack-ubuntu-testing/juno/ubuntu trusty main | 00:20 |
| dlaube | deb-src http://ppa.launchpad.net/openstack-ubuntu-testing/juno/ubuntu trusty main | 00:20 |
| *** marcoemorais has quit IRC | 00:20 | |
| *** marcoemorais has joined #openstack-ironic | 00:21 | |
| jroll | devananda: I may have asked before... wdyt about a maintenance_reason field? | 00:21 |
| jroll | more of | 00:22 |
| jroll | I'm going to make this patch downstream, should I bother proposing upstream | 00:22 |
| devananda | jroll: i think we need to know the reason for things | 00:22 |
| jroll | good enough for me :) | 00:23 |
| jroll | thanks | 00:23 |
| dlaube | are people mostly running ubuntu 14.10 instead of 14.04 lts? | 00:23 |
| devananda | jroll: wdyt about splitting the node table up a bit? | 00:23 |
| jroll | we're running debian 7, fwiw | 00:23 |
| dlaube | alright thanks | 00:23 |
| jroll | devananda: I'm fine with that, more concerned about how it's split | 00:23 |
| devananda | dlaube: most folks in this channel, being developers, are running (very close to) tip of master, in venv's | 00:23 |
| jroll | dlaube: we also use virtualenvs | 00:23 |
| jroll | dlaube: and I'm speaking as a deployer, not a developer | 00:23 |
| devananda | jroll: :) | 00:24 |
| dlaube | interesting, so you would use virtualenvs even to deploy to staging/prod environments? | 00:24 |
| jroll | yes, we already do :) | 00:25 |
| dlaube | I first got up and running by following the dev quickstart guide which uses the virtualenvs | 00:25 |
| jroll | I work on rackspace OnMetal, fwiw | 00:25 |
| dlaube | but I thought it would be more prod like by installing from packages on some boxes | 00:25 |
| dlaube | instead of virtualenvs on virtualbox | 00:25 |
| *** hemna has quit IRC | 00:26 | |
| jroll | virtualbox has nothing to do with it | 00:26 |
| jroll | here's our basic flow: | 00:26 |
| jroll | git repo with downstream patches | 00:26 |
| jroll | clone repo(s) | 00:26 |
| jroll | make virtualenvs with that software inside | 00:26 |
| jroll | tarball it | 00:26 |
| *** rloo has quit IRC | 00:26 | |
| jroll | push it to control plane hosts | 00:26 |
| jroll | untar | 00:26 |
| jroll | flip a symlink | 00:26 |
| jroll | restart services | 00:26 |
| dlaube | sounds sweet | 00:27 |
| jroll | it's okay, it's much more complicated than that in reality | 00:27 |
| jroll | this is the same way rackspace deploys nova etc | 00:27 |
| jroll | shared tooling is nice | 00:27 |
| dlaube | oh neat | 00:28 |
| *** shakayumi has quit IRC | 00:45 | |
| devananda | jroll / JayF: if ya'll haven't seen the ilo doc review, http://docs-draft.openstack.org/61/127161/3/check/gate-ironic-docs/ecf98f8/doc/build/html/drivers/ilo.html | 00:47 |
| devananda | there's a section on agent-ilo-driver that I'd like to get your input on before approving | 00:48 |
| devananda | actually, nvm -- there's really not much specific to IPA in there. just the coreos command to build the image | 00:49 |
| devananda | sorry for the needless ping | 00:50 |
| *** marcoemorais has quit IRC | 00:57 | |
| *** yongli has joined #openstack-ironic | 01:04 | |
| devananda | oh also, if folks haven't voted inthe TC election, go vote :) | 01:16 |
| *** achanda has joined #openstack-ironic | 01:24 | |
| *** rwsu has quit IRC | 01:30 | |
| *** jmanko has quit IRC | 01:32 | |
| *** achanda__ has quit IRC | 01:32 | |
| *** achanda_ has joined #openstack-ironic | 01:32 | |
| *** Isotopp_ has joined #openstack-ironic | 01:32 | |
| *** stevebak` has joined #openstack-ironic | 01:32 | |
| *** rwsu has joined #openstack-ironic | 01:32 | |
| *** rushiagr_away has quit IRC | 01:33 | |
| *** Isotopp has quit IRC | 01:33 | |
| *** krtaylor has quit IRC | 01:33 | |
| *** stevebaker has quit IRC | 01:33 | |
| *** achanda has quit IRC | 01:33 | |
| *** achanda_ has quit IRC | 01:34 | |
| *** krtaylor has joined #openstack-ironic | 01:34 | |
| *** rushiagr_away has joined #openstack-ironic | 01:35 | |
| *** achanda has joined #openstack-ironic | 01:35 | |
| *** openstack has joined #openstack-ironic | 01:43 | |
| *** anteaya has joined #openstack-ironic | 01:43 | |
| *** morgabra has joined #openstack-ironic | 01:52 | |
| *** openstackgerrit has joined #openstack-ironic | 01:52 | |
| *** dvorak has joined #openstack-ironic | 01:52 | |
| *** nosnos has joined #openstack-ironic | 01:52 | |
| *** rushiagr_away has joined #openstack-ironic | 01:52 | |
| *** MattMan has joined #openstack-ironic | 01:54 | |
| *** shakamunyi has joined #openstack-ironic | 01:58 | |
| *** spandhe has quit IRC | 02:01 | |
| *** dlaube has quit IRC | 02:04 | |
| *** kongfy has joined #openstack-ironic | 02:15 | |
| *** harlowja is now known as harlowja_away | 02:21 | |
| openstackgerrit | A change was merged to openstack/ironic: Fix E131 PEP8 errors https://review.openstack.org/128397 | 02:23 |
| *** shakayumi has joined #openstack-ironic | 02:24 | |
| *** shakamunyi has quit IRC | 02:27 | |
| *** marcoemorais has joined #openstack-ironic | 02:53 | |
| *** marcoemorais1 has joined #openstack-ironic | 02:54 | |
| *** kongfy has quit IRC | 02:56 | |
| *** kongfy has joined #openstack-ironic | 02:57 | |
| *** marcoemorais has quit IRC | 02:57 | |
| *** chuckC_ has quit IRC | 03:00 | |
| *** ChuckC_ has joined #openstack-ironic | 03:01 | |
| *** ChuckC_ is now known as ChuckC | 03:01 | |
| *** chuckC_ has joined #openstack-ironic | 03:02 | |
| *** stevebak` is now known as stevebaker | 03:03 | |
| *** achanda has quit IRC | 03:14 | |
| *** achanda_ has joined #openstack-ironic | 03:17 | |
| *** kongfy has quit IRC | 03:27 | |
| *** kongfy has joined #openstack-ironic | 03:38 | |
| *** chenglch has joined #openstack-ironic | 03:39 | |
| *** kongfy has quit IRC | 03:49 | |
| *** nosnos has quit IRC | 03:50 | |
| *** nosnos has joined #openstack-ironic | 03:51 | |
| *** ramineni has joined #openstack-ironic | 03:55 | |
| *** kongfy has joined #openstack-ironic | 03:55 | |
| *** nosnos has quit IRC | 03:55 | |
| adam_g | devananda, sideways job passing at 128378 | 04:02 |
| *** vinbs has joined #openstack-ironic | 04:23 | |
| *** nosnos has joined #openstack-ironic | 04:30 | |
| *** pcrews has quit IRC | 04:33 | |
| *** pradipta_away is now known as pradipta | 04:38 | |
| *** achanda_ has quit IRC | 04:49 | |
| *** achanda_ has joined #openstack-ironic | 04:52 | |
| *** rushiagr_away is now known as rushiagr | 04:59 | |
| *** spandhe has joined #openstack-ironic | 04:59 | |
| *** chenglch has quit IRC | 05:06 | |
| *** rushiagr is now known as rushiagr_away | 05:11 | |
| *** lazy_prince has quit IRC | 05:15 | |
| *** rakesh_hs has joined #openstack-ironic | 05:19 | |
| *** yjiang5 has joined #openstack-ironic | 05:19 | |
| *** rushiagr_away is now known as rushiagr | 05:32 | |
| *** pradipta is now known as pradipta_away | 05:33 | |
| *** ifarkas has joined #openstack-ironic | 05:43 | |
| *** achanda has joined #openstack-ironic | 05:45 | |
| *** rameshg87 has joined #openstack-ironic | 05:48 | |
| *** achanda_ has quit IRC | 05:49 | |
| *** killer_prince has joined #openstack-ironic | 05:56 | |
| *** killer_prince is now known as lazy_prince | 05:57 | |
| *** pradipta_away is now known as pradipta | 05:57 | |
| *** yjiang5 has left #openstack-ironic | 06:04 | |
| *** zigo has joined #openstack-ironic | 06:08 | |
| *** vinbs has quit IRC | 06:10 | |
| *** vinbs has joined #openstack-ironic | 06:10 | |
| *** vinbs_ has joined #openstack-ironic | 06:22 | |
| *** vinbs has quit IRC | 06:23 | |
| *** vinbs_ is now known as vinbs | 06:23 | |
| *** eghobo has joined #openstack-ironic | 06:28 | |
| *** kongfy_ has joined #openstack-ironic | 06:30 | |
| *** spandhe has quit IRC | 06:32 | |
| *** kongfy has quit IRC | 06:32 | |
| *** kongfy_ is now known as kongfy | 06:32 | |
| *** eghobo has quit IRC | 06:35 | |
| *** praneshp has joined #openstack-ironic | 06:37 | |
| *** d0m0reg00dthing has joined #openstack-ironic | 06:39 | |
| *** vinbs_ has joined #openstack-ironic | 06:40 | |
| *** d0m0reg00dthing has quit IRC | 06:40 | |
| *** vinbs has quit IRC | 06:41 | |
| *** vinbs_ is now known as vinbs | 06:42 | |
| *** achanda has quit IRC | 06:45 | |
| *** foexle has joined #openstack-ironic | 06:46 | |
| *** dnikishov has joined #openstack-ironic | 06:53 | |
| *** jcoufal has joined #openstack-ironic | 06:55 | |
| *** ndipanov has joined #openstack-ironic | 07:01 | |
| *** Guest86578 is now known as d0ugal | 07:04 | |
| *** d0ugal has quit IRC | 07:05 | |
| *** d0ugal has joined #openstack-ironic | 07:05 | |
| *** bluex-pl has joined #openstack-ironic | 07:13 | |
| *** bluex-pl has quit IRC | 07:13 | |
| *** bluex-pl has joined #openstack-ironic | 07:14 | |
| GheRivero | morning @Ironic | 07:18 |
| *** dtantsur|afk is now known as dtantsur | 07:20 | |
| dtantsur | Morning Ironic! | 07:20 |
| *** chenglch|2 has joined #openstack-ironic | 07:22 | |
| yuriyz | morning GheRivero dtantsur | 07:25 |
| *** marcoemorais1 has quit IRC | 07:25 | |
| *** bluex-pl has quit IRC | 07:26 | |
| *** bluex-pl has joined #openstack-ironic | 07:26 | |
| *** andreykurilin_ has joined #openstack-ironic | 07:28 | |
| *** bluex-pl has quit IRC | 07:30 | |
| *** bluex-pl has joined #openstack-ironic | 07:30 | |
| *** pradipta is now known as pradipta_away | 07:40 | |
| *** praneshp has quit IRC | 07:41 | |
| *** achanda has joined #openstack-ironic | 07:57 | |
| *** jistr has joined #openstack-ironic | 08:02 | |
| *** achanda has quit IRC | 08:02 | |
| *** andreykurilin_ has quit IRC | 08:02 | |
| *** derekh has joined #openstack-ironic | 08:06 | |
| *** athomas has joined #openstack-ironic | 08:07 | |
| *** spandhe has joined #openstack-ironic | 08:09 | |
| *** ndipanov has quit IRC | 08:24 | |
| *** ndipanov has joined #openstack-ironic | 08:28 | |
| *** pradipta_away is now known as pradipta | 08:30 | |
| *** pelix1 has joined #openstack-ironic | 08:30 | |
| *** rwsu has quit IRC | 08:31 | |
| *** lucasagomes has joined #openstack-ironic | 08:40 | |
| *** rameshg87 has quit IRC | 08:41 | |
| *** rameshg87 has joined #openstack-ironic | 08:50 | |
| *** rwsu has joined #openstack-ironic | 08:56 | |
| *** mkerrin has quit IRC | 08:56 | |
| *** lazy_prince is now known as killer_prince | 08:57 | |
| *** ndipanov has quit IRC | 08:57 | |
| *** killer_prince has quit IRC | 09:00 | |
| *** lazy_prince has joined #openstack-ironic | 09:00 | |
| *** ndipanov has joined #openstack-ironic | 09:10 | |
| *** rushiagr is now known as rushiagr_away | 09:19 | |
| *** mkerrin has joined #openstack-ironic | 09:20 | |
| *** vdrok has quit IRC | 09:24 | |
| *** vdrok has joined #openstack-ironic | 09:26 | |
| *** mkerrin has quit IRC | 09:26 | |
| *** lsmola has quit IRC | 09:26 | |
| *** jcoufal has quit IRC | 09:29 | |
| *** jcoufal has joined #openstack-ironic | 09:29 | |
| *** igordcard has joined #openstack-ironic | 09:30 | |
| *** rushiagr_away is now known as rushiagr | 09:36 | |
| *** lazy_prince is now known as killer_prince | 09:37 | |
| openstackgerrit | Dmitry Nikishov proposed a change to openstack/ironic: Removed hardcoded IDs from "node" test resources https://review.openstack.org/127974 | 09:39 |
| *** killer_prince is now known as lazy_prince | 09:40 | |
| *** lsmola has joined #openstack-ironic | 09:41 | |
| *** lazy_prince is now known as killer_prince | 09:49 | |
| *** chenglch|2 has quit IRC | 09:59 | |
| *** mkerrin has joined #openstack-ironic | 10:00 | |
| *** vinbs_ has joined #openstack-ironic | 10:02 | |
| *** vinbs has quit IRC | 10:03 | |
| *** vinbs_ is now known as vinbs | 10:03 | |
| *** kongfy has quit IRC | 10:04 | |
| *** spandhe has quit IRC | 10:10 | |
| *** killer_prince is now known as lazy_prince | 10:16 | |
| *** yuanying has quit IRC | 10:42 | |
| *** yuanying has joined #openstack-ironic | 10:42 | |
| *** yuanying has quit IRC | 10:47 | |
| *** sambetts has joined #openstack-ironic | 10:57 | |
| *** rameshg87 has quit IRC | 11:00 | |
| *** ramineni has quit IRC | 11:02 | |
| *** Haomeng|2 has joined #openstack-ironic | 11:21 | |
| *** Haomeng has quit IRC | 11:21 | |
| *** kongfy has joined #openstack-ironic | 11:28 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: WIP: Driver VendorPassthru to support different HTTP methods https://review.openstack.org/128610 | 11:30 |
| lucasagomes | jroll ^ it's WIP still, but lemme know if something that you guys may want to have as well | 11:31 |
| *** vinbs has quit IRC | 11:36 | |
| dtantsur | lucasagomes, related note: I think we should finally move this self.vendor_routes = ... self.driver_routes = ... things to the base.VendorInterface with reasonable implementation of driver_vendor_passthru and vendor_passthru | 11:38 |
| lucasagomes | dtantsur, hmm right lemme take a look at it | 11:39 |
| lucasagomes | thanks | 11:39 |
| *** zz_jgrimm is now known as jgrimm | 11:41 | |
| *** yuanying has joined #openstack-ironic | 11:43 | |
| lucasagomes | dtantsur, :( it's so inconsistent already, one being fully async and the other being sync | 11:48 |
| dtantsur | lucasagomes, you mean vendor and just passthru? | 11:48 |
| lucasagomes | dtantsur, maybe if we, somehow make it configurable, whether u want it to be sync or async by passing a parameter | 11:49 |
| dtantsur | well, we don't have to unify them, but it would be nice to have their basic implementation in the base class | 11:49 |
| lucasagomes | dtantsur, yeah the node_vendor_passthru is fully async | 11:49 |
| lucasagomes | the driver_vendor_passthru is blocking (sync) | 11:49 |
| lucasagomes | dtantsur, right yeah | 11:49 |
| *** lazy_prince is now known as killer_prince | 11:55 | |
| *** rakesh_hs has quit IRC | 11:58 | |
| *** rakesh_hs has joined #openstack-ironic | 11:59 | |
| *** pradipta is now known as pradipta_away | 11:59 | |
| *** killer_prince is now known as lazy_prince | 12:00 | |
| dtantsur | lucasagomes, do you think we should have some driver capabilities spec? like capability of configdrive, whole-disk etc? | 12:04 |
| lucasagomes | dtantsur, hmm I haven't thought about it, but it's a good idea so we can distinguised between diff types of deployments | 12:05 |
| dtantsur | lucasagomes, my main point is to fail early (on nova driver stage), if required capability is missing from the driver | 12:05 |
| lucasagomes | cause now it's incosistent as well, with IPA deploying full disk images and not supporting ephemeral partition (afaik) and pxe deploying only partition images with support for ephemeral but not config | 12:06 |
| lucasagomes | dtantsur, yeah, it sounds good | 12:06 |
| lucasagomes | dtantsur, maybe adding the idea to the spreadsheet ? | 12:07 |
| *** lucasagomes is now known as lucas-hungry | 12:10 | |
| * lucas-hungry brb lunch | 12:10 | |
| *** lazy_prince is now known as killer_prince | 12:10 | |
| dtantsur | lucas-hungry, it's line 29, you've seen it | 12:10 |
| lucas-hungry | dtantsur, oh right, I thought it was diff things (like supporting methods on the mgmt interface in only 1 driver vs diff types of deploy) | 12:13 |
| lucas-hungry | but yeah I can see how they relate | 12:13 |
| * lucas-hungry already voted on that session :) | 12:13 | |
| *** athomas has quit IRC | 12:16 | |
| *** dprince has joined #openstack-ironic | 12:19 | |
| *** killer_prince is now known as lazy_prince | 12:23 | |
| *** athomas has joined #openstack-ironic | 12:23 | |
| *** lazy_prince is now known as killer_prince | 12:26 | |
| *** jmank has joined #openstack-ironic | 12:35 | |
| jroll | morning ironic :) | 12:47 |
| openstackgerrit | Jim Rollenhagen proposed a change to openstack/ironic: Support configdrive in agent driver https://review.openstack.org/128388 | 12:50 |
| dtantsur | jroll, morning | 12:50 |
| jroll | hey dtantsur, marked that BP there... guess it's time for a -2 :P | 12:51 |
| dtantsur | heh yes | 12:51 |
| jroll | :( | 12:51 |
| *** lucas-hungry is now known as lucasagomes | 12:58 | |
| jroll | anyone remember how to make a new migration? | 13:00 |
| *** rloo has joined #openstack-ironic | 13:04 | |
| lucasagomes | jroll, alembic revision -m "blah" AFAIR | 13:06 |
| Shrews | jroll: http://docs.openstack.org/developer/ironic/cmds/ironic-dbsync.html | 13:06 |
| Shrews | http://docs.openstack.org/developer/ironic/cmds/ironic-dbsync.html#revision, specifically | 13:07 |
| jroll | aha, thanks | 13:09 |
| jroll | lucasagomes: turns out s/alembic/ironic-dbsync | 13:12 |
| jroll | which sets up the alembic environment etc | 13:12 |
| lucasagomes | ah, nice! haven't tried using ironic-dbsync yet | 13:12 |
| lucasagomes | but good to know | 13:12 |
| *** bluex-pl has quit IRC | 13:12 | |
| *** bluex-pl has joined #openstack-ironic | 13:13 | |
| lucasagomes | jroll, what you think about driver_vendor_passthru supporting diff http methods? | 13:14 |
| lucasagomes | any opnion on it? | 13:15 |
| lucasagomes | cause I will need GET for the iPXE driver so I started looking into it | 13:15 |
| jroll | no opinion really, I could see it being useful | 13:15 |
| jroll | don't see why we shouldn't support GET | 13:15 |
| lucasagomes | jroll, right, yeah it seems useful | 13:16 |
| jroll | I like useful things | 13:16 |
| jroll | :) | 13:16 |
| lucasagomes | :) +1 | 13:16 |
| jroll | which you'll see in a moment with my next patch :P | 13:17 |
| lucasagomes | :D | 13:20 |
| lucasagomes | yay | 13:20 |
| openstackgerrit | Jim Rollenhagen proposed a change to openstack/ironic: Add Node.maintenance_reason https://review.openstack.org/128645 | 13:23 |
| openstackgerrit | Jim Rollenhagen proposed a change to openstack/ironic: Add maintenance_reason when setting maintenance mode https://review.openstack.org/128646 | 13:23 |
| jroll | whee | 13:24 |
| jroll | devananda: ^ | 13:24 |
| *** jjohnson2 has joined #openstack-ironic | 13:26 | |
| Shrews | jroll: missing migration file? | 13:27 |
| jroll | uhh | 13:27 |
| jroll | Shrews: it's there | 13:28 |
| jroll | https://review.openstack.org/#/c/128645/1/ironic/db/sqlalchemy/alembic/versions/34e43a2cb956_add_node_maintenance_reason.py | 13:28 |
| Shrews | ??? /me reloads | 13:28 |
| jroll | lol, there's only one patchset | 13:28 |
| Shrews | wow, i still don't see it | 13:29 |
| jroll | does that direct link work? | 13:29 |
| Shrews | yeah, which is really weird | 13:29 |
| jroll | it's the second one here: https://www.dropbox.com/s/xapd99mbmypb72f/Screenshot%202014-10-15%2006.29.24.png?dl=0 | 13:29 |
| jroll | maybe new view is broken? | 13:29 |
| jroll | idk | 13:29 |
| jroll | now you're going to be wondering about missing files on every patch >:) | 13:29 |
| Shrews | oh, lol. i clicked the 1st link | 13:29 |
| Shrews | doh! | 13:29 |
| jroll | heh | 13:30 |
| jroll | yeah, wo patches :P | 13:30 |
| jroll | two, even | 13:30 |
| Shrews | soooo it's obvo that i really need more coffee at this point | 13:30 |
| jroll | gogogo | 13:31 |
| Shrews | jroll: this is clearly your fault. you can't expect someone to read past the first word of the first commit line | 13:32 |
| jroll | :) | 13:33 |
| * jroll wonders who will be the first to ask for a spec | 13:34 | |
| lucasagomes | jroll, I was thinking about whether we should have a diff endpoint in our api to set the maintenance and the reason... so that when you set it to False you clean up the reason as well | 13:36 |
| lucasagomes | POST {'reason': 'blah'} v1/nodes/UUID/maintenance | 13:37 |
| lucasagomes | DELETE v1/nodes/UUID/maintenance | 13:37 |
| lucasagomes | idk if that would be useful | 13:37 |
| jroll | yeah, I was thinking about how to clean up the reason in that case | 13:37 |
| *** kongfy has quit IRC | 13:37 | |
| jroll | I was more thinking put it in the client... ironic node-set-maintenance True reason | 13:37 |
| jroll | and set False would just build the PATCH to clear it | 13:38 |
| lucasagomes | cause UPDATE node is fine, but once you have to do more actions consequently to update a node field, it kinda means that we need a diff endpoint for it | 13:38 |
| jroll | but your way might be better... I think that's a good follow-on patch | 13:38 |
| jroll | right | 13:38 |
| lucasagomes | right, yeah we can discuss it with more ppl | 13:38 |
| jroll | this is going to turn into a spec :( | 13:39 |
| lucasagomes | :/ | 13:39 |
| lucasagomes | maybe not | 13:39 |
| jroll | lol | 13:39 |
| lucasagomes | but that potentially could do more things | 13:39 |
| lucasagomes | for e.g, once it's set in maintenance we then send a signal to nova | 13:39 |
| jroll | yeah, I like the new endpoint idea | 13:39 |
| lucasagomes | saying to exclude that node from the list | 13:39 |
| lucasagomes | and when deleted the same | 13:39 |
| jroll | but we still have to make the old way work, I think | 13:39 |
| jroll | ++ | 13:39 |
| * jroll goes afk for a while | 13:41 | |
| *** rakesh_hs has quit IRC | 13:48 | |
| *** rushiagr is now known as rushiagr_away | 13:48 | |
| openstackgerrit | Dmitry Nikishov proposed a change to openstack/ironic: Removed hardcoded IDs from "node" test resources https://review.openstack.org/127974 | 14:07 |
| *** kongfy has joined #openstack-ironic | 14:12 | |
| *** shakayumi has quit IRC | 14:12 | |
| *** pelix1 is now known as pelix | 14:24 | |
| *** rainya has joined #openstack-ironic | 14:35 | |
| NobodyCa1 | good morning Ironic | 14:37 |
| lucasagomes | NobodyCa1, morning | 14:45 |
| *** pcrews has joined #openstack-ironic | 14:48 | |
| *** achanda has joined #openstack-ironic | 14:59 | |
| *** NobodyCa1 is now known as NobodyCam | 14:59 | |
| NobodyCam | morning lucasagomes | 14:59 |
| lucasagomes | NobodyCam, quick q... re VIFs allocation | 15:03 |
| lucasagomes | NobodyCam, if the node has more than 1 NIC (port) we should allocate 1 VIF for each port right? | 15:03 |
| lucasagomes | I mean, nova does it | 15:04 |
| lucasagomes | should do it* | 15:04 |
| *** bluex-pl has quit IRC | 15:07 | |
| *** achanda has quit IRC | 15:07 | |
| *** ChuckC___ has joined #openstack-ironic | 15:09 | |
| NobodyCam | yep thats what I recall | 15:12 |
| NobodyCam | I remember thinking we should be able to do better vif/nic targeting | 15:13 |
| *** ChuckC___ has quit IRC | 15:15 | |
| *** rushiagr_away is now known as rushiagr | 15:16 | |
| lucasagomes | NobodyCam, right I think it's not doing that :/ http://paste.openstack.org/show/121276/ | 15:17 |
| lucasagomes | but afair it should be doing that indeed | 15:17 |
| lucasagomes | 1 sec brb | 15:17 |
| *** lucasagomes is now known as lucas-afk | 15:17 | |
| *** kongfy has quit IRC | 15:20 | |
| *** kongfy has joined #openstack-ironic | 15:20 | |
| *** nosnos has quit IRC | 15:27 | |
| * devananda heads into the office | 15:27 | |
| *** rainya has quit IRC | 15:31 | |
| NobodyCam | morning devananda | 15:32 |
| *** lucas-afk is now known as lucasagomes | 15:32 | |
| lucasagomes | devananda, morning | 15:32 |
| *** krtaylor has quit IRC | 15:34 | |
| *** romcheg has quit IRC | 15:35 | |
| lucasagomes | devananda, ^ same q, you know if nova will try to allocate 1 VIF per Port that the node contains? | 15:36 |
| Shrews | lucasagomes: i kind of like the endpoint for setting maintenance mode idea. we could also extend that to the chassis level to set all nodes in that chassis to maint mode | 15:37 |
| lucasagomes | Shrews, right, yeah | 15:37 |
| lucasagomes | Shrews, morning :) | 15:37 |
| Shrews | morning | 15:37 |
| NobodyCam | morning Shrews | 15:38 |
| Shrews | hey hey NobodyCam | 15:38 |
| *** yongli has quit IRC | 15:40 | |
| *** spandhe_ has joined #openstack-ironic | 15:41 | |
| sambetts | Hi NobodyCam | 15:41 |
| NobodyCam | lucasagomes: how many networks where defined at boot time. | 15:41 |
| NobodyCam | hi sambetts | 15:41 |
| lucasagomes | NobodyCam, how I check it? | 15:42 |
| *** kongfy has quit IRC | 15:43 | |
| *** chuckC_ has quit IRC | 15:44 | |
| NobodyCam | lucasagomes: looks like just nova list see: http://docs.openstack.org/grizzly/openstack-compute/admin/content/using-multi-nics.html | 15:46 |
| * lucasagomes clicks | 15:47 | |
| NobodyCam | thou I'm not sure what adding after the fact will get attched | 15:47 |
| *** spandhe__ has joined #openstack-ironic | 15:47 | |
| lucasagomes | NobodyCam, so it only lists one | 15:47 |
| lucasagomes | private=10.1.0.5 | | 15:47 |
| NobodyCam | did you boot with --nic net-id= <id of first network> --nic net-id= <id of first network> options? | 15:48 |
| *** spandhe_ has quit IRC | 15:48 | |
| *** ChuckC_ has joined #openstack-ironic | 15:50 | |
| *** viktors has left #openstack-ironic | 15:51 | |
| *** ChuckC has quit IRC | 15:51 | |
| lucasagomes | NobodyCam, nop, didn't pass any --nic net-id in the boot command | 15:52 |
| NobodyCam | ahh :) | 15:52 |
| NobodyCam | need to have more than one network | 15:52 |
| jroll | morning NobodyCam :) | 15:53 |
| NobodyCam | if morning jroll :) | 15:53 |
| NobodyCam | lol | 15:53 |
| NobodyCam | morning jroll :-p | 15:53 |
| jroll | lucasagomes: from what I've seen it's 1 vif per port... or maybe one vif per network? not sure | 15:53 |
| jroll | we have 2 NICs, 2 networks :P | 15:54 |
| NobodyCam | jroll: how do you | 15:54 |
| NobodyCam | boot? | 15:54 |
| NobodyCam | with --nic option? | 15:55 |
| jroll | nova boot? | 15:55 |
| jroll | no | 15:55 |
| jroll | they're default networks, I think | 15:55 |
| jroll | they must be, virt servers get the same networks | 15:55 |
| *** ChuckC_ has quit IRC | 15:55 | |
| jroll | we don't have arbitrary networks or whatever yet | 15:55 |
| lucasagomes | jroll, thanks, yeah... I'm investigating now | 15:56 |
| *** rainya has joined #openstack-ironic | 15:56 | |
| *** dtantsur is now known as dtantsur|afk | 15:56 | |
| *** romcheg has joined #openstack-ironic | 16:02 | |
| *** romcheg has quit IRC | 16:03 | |
| *** romcheg has joined #openstack-ironic | 16:03 | |
| NobodyCam | at somepoint I i feel we should chat about chassis / node inheritance.. I think there are some use cases where we could reduce the number properites that have to be added to a node, | 16:04 |
| jroll | at the cost of an extra SELECT, no? | 16:05 |
| jroll | and really... if you're adding entire racks or whatever to ironic, you should have that scripted | 16:05 |
| NobodyCam | :) | 16:06 |
| jroll | (not that I think it's a bad idea, just throwing that out there) | 16:06 |
| NobodyCam | yep... it could so clean up our node-show | 16:07 |
| jroll | maybe? | 16:08 |
| jroll | I mean... I'd want to see the inherited properties when I do that | 16:08 |
| NobodyCam | oh ya.. if --detail was passed | 16:08 |
| NobodyCam | :-p | 16:08 |
| jroll | aha | 16:10 |
| NobodyCam | I was thinking for CI/dev where ipmi accounts and password are all the same I would love to create a chassis add the ipmi user info and not added to each node | 16:11 |
| NobodyCam | just create the node a member of chassis-x | 16:11 |
| jroll | oh yes | 16:11 |
| jroll | ++ | 16:12 |
| jroll | especially nice when rotating | 16:12 |
| NobodyCam | :) | 16:12 |
| NobodyCam | maybe be worth some chit chat at the pod area | 16:12 |
| lucasagomes | NobodyCam, +1 | 16:14 |
| lucasagomes | yeah right now we haven't been using chassis for anything apart from logically grouping nodes | 16:14 |
| *** jistr has quit IRC | 16:14 | |
| lucasagomes | NobodyCam, maybe adding to the spreedsheet? | 16:14 |
| lucasagomes | I think it's a good topic | 16:14 |
| Shrews | NobodyCam: yeah, i'd like to see that kind of attribute inheritence | 16:14 |
| NobodyCam | oh in my sickened state yesterday I came with the waco idea of a meta chassis (chassis of other chassis), but then my head hurt | 16:15 |
| NobodyCam | note i am not suggenting a meta chassis at all | 16:16 |
| Shrews | NobodyCam: careful. that's how tripleo started | 16:16 |
| NobodyCam | adding to inheritance to th Spread sheet | 16:16 |
| lucasagomes | NobodyCam, lol yeah let's make use of what we have _now_ first :P | 16:16 |
| *** marcoemorais has joined #openstack-ironic | 16:17 | |
| *** sambetts has quit IRC | 16:26 | |
| NobodyCam | added! | 16:29 |
| *** rainya has quit IRC | 16:29 | |
| *** dlaube has joined #openstack-ironic | 16:34 | |
| dlaube | so is ironic just for provisioning servers to run VMs on it, or does the other use case of just provisioning/installing an OS on a physical server to be used as a physical server still exist? | 16:36 |
| openstackgerrit | Chris Behrens proposed a change to openstack/ironic-python-agent: Refactor image writing https://review.openstack.org/128701 | 16:37 |
| *** marcoemorais has quit IRC | 16:37 | |
| jroll | dlaube: tripleo is about using ironic to provision bare metal machines to run openstack | 16:37 |
| jroll | ironic is about provisioning bare metal machines to run an operating system | 16:38 |
| *** marcoemorais has joined #openstack-ironic | 16:38 | |
| *** marcoemorais has quit IRC | 16:38 | |
| *** marcoemorais has joined #openstack-ironic | 16:39 | |
| dlaube | was just going to ask about tripleo ;) | 16:39 |
| dlaube | thank you | 16:39 |
| jroll | :) | 16:39 |
| * NobodyCam believes jroll to be psychic | 16:41 | |
| jroll | heh | 16:42 |
| JayF | dlaube: an easy way to percieve ironic as an existing cloud user | 16:42 |
| JayF | dlaube: is to think of rackspace OnMetal. tl;dr; call the API identically to how you get a virt server, but instead get a physical server | 16:42 |
| JayF | dlaube: that's utilized by other things (like TripleO) to make deploying clouds easier, but that's more of a side-effect of making bare metal provisioning painless | 16:43 |
| jroll | I like it this way: "call an API, get a computer. | 16:43 |
| jroll | " | 16:44 |
| dlaube | that's what I've thought all along but I wanted to make completely sure | 16:44 |
| lucasagomes | dlaube, just like jroll and JayF said... Ironic is generic so you can deploy whatever image containing whatever application you want in it to run on baremetal, common cases would be databases or some application that needs access to some PCI device that can't be emulated | 16:44 |
| lucasagomes | HPC | 16:45 |
| JayF | lucasagomes: I mean, honestly, we have folks who want bare-metal due to not wanting to lose perf to the hypervisor | 16:45 |
| JayF | lucasagomes: so I'd say no hypervisor is good just as a reason on it's own, who needs the other excuses ;) | 16:45 |
| lucasagomes | hah fair enuff! | 16:45 |
| jroll | or like... someone who wants to run a hypervisor and resell VMs, whether or not openstack is used for that :P | 16:46 |
| *** rainya has joined #openstack-ironic | 16:47 | |
| lucasagomes | :) indeed... I think we should go with the idea of making ironic more standalone | 16:47 |
| lucasagomes | outgrowing openstack | 16:47 |
| devananda | adam_g: what's the status of the sideways grenade job on juno/proposed? are we able to land changes there yet? | 16:52 |
| *** derekh has quit IRC | 16:52 | |
| devananda | lucasagomes: I wasn't able to land the backport of your DRAC fix, btw | 16:52 |
| lucasagomes | devananda, :( | 16:53 |
| lucasagomes | well I think it's fine | 16:53 |
| devananda | lucasagomes: might not be able to before juno/final is cut | 16:53 |
| lucasagomes | devananda, if needed we can put the patch when packaging it I suspect, it's a 1 line fix | 16:53 |
| devananda | lucasagomes: k, cool | 16:54 |
| lucasagomes | devananda, re making ironic more standalone | 16:54 |
| lucasagomes | I was thinking about it some days ago, in creating a "ironic deploy ..." command to the CLI | 16:55 |
| lucasagomes | cause right now if u wanna use ironic w/o nova, it's not very straight forward | 16:55 |
| devananda | lucasagomes: I'm working on ansible bindings for Ironic's API, fwiw | 16:55 |
| devananda | indeed | 16:55 |
| lucasagomes | needs to update the node to put things on instance_info, needs a instance_uuid (?!) | 16:55 |
| lucasagomes | hah | 16:55 |
| lucasagomes | devananda, you think that a command like that would be beneficial? | 16:55 |
| lucasagomes | devananda, oh awesome! ansible seems good | 16:55 |
| devananda | lucasagomes: sort of. what I'm finding as I write this is that our API is very painful for direct use | 16:56 |
| lucasagomes | devananda, yeah I recall the resource updates to need you to understand the json patch format and all | 16:57 |
| devananda | for example, if I know what properties (capabilities) I want a node to expose | 16:57 |
| lucasagomes | that's exacty why having a dedicated command to deploy a node would be good | 16:57 |
| devananda | and I can poll the API to get current properties | 16:58 |
| devananda | I have to actually build the JSON patch myself to converge the properties | 16:58 |
| devananda | which sucks | 16:58 |
| devananda | our python library should do that | 16:58 |
| *** rainya has quit IRC | 16:58 | |
| lucasagomes | devananda, yeah... we made the wrong choice to put this bit in the CLI, so CLI will convert the inputs to a json-patch format | 16:59 |
| lucasagomes | we need to move it to the lib | 16:59 |
| devananda | yes | 16:59 |
| devananda | but even with the CLI, I have to think about it | 16:59 |
| devananda | it should be easier | 16:59 |
| lucasagomes | devananda, right problem is the json-patch is quite complex | 16:59 |
| devananda | pass in a dict, let the library figure out what PATCH is necessary | 16:59 |
| *** marcoemorais has quit IRC | 17:00 | |
| lucasagomes | the CLI tries to simplify it by only supporting one operation per request (add, remove, ...) | 17:00 |
| *** marcoemorais has joined #openstack-ironic | 17:00 | |
| lucasagomes | but in reality a json-patch can contain multiple operations in the same request | 17:00 |
| *** ChuckC_ has joined #openstack-ironic | 17:00 | |
| devananda | indeed | 17:00 |
| devananda | I'm OK with the CLI being slightly simplistic like that | 17:01 |
| devananda | but the python lib shouldn't be | 17:01 |
| lucasagomes | devananda, yeah if we make passing a lib to the lib, and then get a delta from current resource | 17:01 |
| lucasagomes | it's def possible to generate the json patch from that | 17:01 |
| lucasagomes | I think that the json patch lib that we use | 17:01 |
| lucasagomes | actually supports doing thta | 17:02 |
| * lucasagomes checks | 17:02 | |
| devananda | oh, nice | 17:02 |
| lucasagomes | devananda, https://github.com/stefankoegl/python-json-patch/blob/master/jsonpatch.py#L149-L166 | 17:02 |
| lucasagomes | it def supports it | 17:02 |
| lucasagomes | passing a dict to the lib* | 17:03 |
| devananda | awesome | 17:03 |
| lucasagomes | :) | 17:04 |
| *** igordcard has quit IRC | 17:04 | |
| *** marcoemorais has quit IRC | 17:05 | |
| *** marcoemorais has joined #openstack-ironic | 17:05 | |
| *** marcoemorais has quit IRC | 17:05 | |
| *** marcoemorais has joined #openstack-ironic | 17:06 | |
| *** marcoemorais has quit IRC | 17:06 | |
| *** rainya has joined #openstack-ironic | 17:06 | |
| *** marcoemorais has joined #openstack-ironic | 17:06 | |
| adam_g | devananda, yeah | 17:07 |
| adam_g | devananda, https://review.openstack.org/#/c/128378/ is passing | 17:07 |
| devananda | adam_g: ooh. great | 17:08 |
| * devananda approves | 17:08 | |
| lucasagomes | right I will call it a day here | 17:08 |
| lucasagomes | have a good night everyone | 17:08 |
| NobodyCam | have a good night lucasagomes | 17:09 |
| jroll | night lucas! | 17:09 |
| *** lucasagomes is now known as lucas-dinner | 17:09 | |
| *** ChuckC_ has quit IRC | 17:10 | |
| devananda | lucas-dinner: cheers, g'night! | 17:12 |
| *** rainya has quit IRC | 17:12 | |
| *** rainya has joined #openstack-ironic | 17:13 | |
| *** praneshp has joined #openstack-ironic | 17:13 | |
| *** harlowja_away is now known as harlowja | 17:15 | |
| devananda | oh, as a reminder to folks, if you haven't already voted in the TC elections -- you should go do that today :) | 17:16 |
| adam_g | devananda, i suppose if thats voting on proposed it should also be made voting on master | 17:18 |
| devananda | adam_g: wait. huh? | 17:19 |
| devananda | adam_g: you mean the sideways migration, or did I miss a context switch? | 17:19 |
| adam_g | devananda, doh! nvm, the sideways job only runs on proposed/stable :) | 17:20 |
| devananda | heh. I mean, a grenade job on master would be good | 17:20 |
| adam_g | devananda, we should have one soon. but something i realized yesterday is that its going to take a *long* time to run that. the sideways job is 50-60 min but we only run tempest once there (after the migration). a regular grenade job runs it twice (against base, against target) | 17:22 |
| devananda | yup | 17:22 |
| devananda | tempest parallel itself is taking ~50m | 17:22 |
| *** praneshp has quit IRC | 17:23 | |
| *** rainya has quit IRC | 17:24 | |
| *** krtaylor has joined #openstack-ironic | 17:25 | |
| devananda | periodic job? | 17:25 |
| *** rushiagr is now known as rushiagr_away | 17:25 | |
| *** marcoemorais has quit IRC | 17:32 | |
| *** marcoemorais has joined #openstack-ironic | 17:32 | |
| *** praneshp has joined #openstack-ironic | 17:36 | |
| adam_g | devananda, its an option but not ideal, we'd want to have the job voting on devstack + grenade to avoid breaking changes there | 17:37 |
| *** r-daneel has joined #openstack-ironic | 17:37 | |
| devananda | adam_g: sure, but ~2hr per job is ... slower | 17:38 |
| *** rainya has joined #openstack-ironic | 17:38 | |
| adam_g | devananda, yeah.. | 17:38 |
| adam_g | i wonder how many more VMs the slaves can handle, if we can bump concurrency to 3 and see if it cuts it down | 17:39 |
| *** rainya has quit IRC | 17:40 | |
| *** athomas has quit IRC | 17:41 | |
| NobodyCam | j* Add a script to do a rolling reboot of unprovisioned nodes <- is how the LRA (Long Running Agent) would handle a public api change | 17:42 |
| JayF | you need to stick to regexp or globbing | 17:42 |
| jroll | lol | 17:42 |
| JayF | NobodyCam: You should never break existing, running agents | 17:42 |
| jroll | JayF: he's saying agent API changes, I think | 17:43 |
| JayF | NobodyCam: however if you want to upgrade agents, rebooting all unprovisioned machines is the easiest way to do that | 17:43 |
| jroll | but yeah | 17:43 |
| JayF | NobodyCam: and we basically already have and use that script regularly :P | 17:43 |
| jroll | NobodyCam: this is how we handle agent updates in onmetal, fwiw | 17:43 |
| NobodyCam | lol :) | 17:44 |
| *** achanda has joined #openstack-ironic | 17:49 | |
| NobodyCam | jroll: that [agent api changes] and also I was think about the notifications / external events items | 17:49 |
| *** athomas has joined #openstack-ironic | 17:50 | |
| *** rainya has joined #openstack-ironic | 17:52 | |
| *** rainya has quit IRC | 17:55 | |
| jroll | hmm | 17:56 |
| NobodyCam | am I correct that currently we use a hertbeat overload to send messages to the conductor? | 17:58 |
| *** ChuckC_ has joined #openstack-ironic | 17:59 | |
| *** ChuckC_ has quit IRC | 17:59 | |
| JayF | I think the heartbeat tells the conductor commands it recently finished | 17:59 |
| *** ChuckC_ has joined #openstack-ironic | 17:59 | |
| JayF | then the conductor calls /v1/commands on the agent to get the command result | 17:59 |
| jroll | ^ that's how it works today | 18:00 |
| jroll | or forever, I guess | 18:00 |
| NobodyCam | ok :) | 18:00 |
| jroll | this makes it a bit faster :) https://review.openstack.org/#/c/125464/ | 18:00 |
| jroll | and has a +2 already | 18:00 |
| jroll | </shamelessplug> | 18:00 |
| * JayF on it | 18:00 | |
| NobodyCam | lol | 18:00 |
| *** rainya has joined #openstack-ironic | 18:02 | |
| jroll | any objections to releasing "juno" IPA on pypi? | 18:02 |
| * Shrews removes his +2 to spite the shameless plugging | 18:02 | |
| jroll | going once | 18:03 |
| NobodyCam | question: how will LRA handle hash-ring changes. | 18:04 |
| NobodyCam | ie the conductor managing the node changes or dies | 18:04 |
| jroll | LRA? | 18:05 |
| JayF | long running agents | 18:05 |
| jroll | oh | 18:05 |
| jroll | it doesn't interact with a conductor | 18:05 |
| JayF | NobodyCam: only thing the conductor matters for in IPA is setting up dhcp | 18:05 |
| jroll | oh | 18:05 |
| jroll | we'll need take over things | 18:05 |
| jroll | hmm | 18:05 |
| JayF | NobodyCam: so they'll have to have that dhcp config taken over | 18:05 |
| jroll | (for pxe driver) | 18:06 |
| jroll | NobodyCam: I can call that out if you leave me a comment | 18:06 |
| JayF | but that's all the state there afaik | 18:06 |
| JayF | jroll: is LRA for pxe+IPA? | 18:06 |
| adam_g | devananda, 128378 is merged | 18:06 |
| jroll | JayF: for all the drivers that we want to support it, including that one | 18:07 |
| NobodyCam | ya | 18:08 |
| *** chuckC__ has joined #openstack-ironic | 18:12 | |
| openstackgerrit | A change was merged to openstack/ironic-python-agent: Force heartbeat immediately after async command completes https://review.openstack.org/125464 | 18:12 |
| jroll | woohoo | 18:12 |
| *** achanda has quit IRC | 18:16 | |
| NobodyCam | okay another question on the LRA spec: would the only change to ironic node-list be the power state? Ie. Provisioning State is still going to be None? | 18:18 |
| *** achanda has joined #openstack-ironic | 18:20 | |
| jroll | NobodyCam: yes, sounds good | 18:20 |
| NobodyCam | :) | 18:20 |
| *** rainya has quit IRC | 18:22 | |
| *** dprince has quit IRC | 18:25 | |
| *** rainya has joined #openstack-ironic | 18:35 | |
| *** marck has joined #openstack-ironic | 18:36 | |
| devananda | rc3 milestone created, waiting for ttx to push the button to build it | 18:37 |
| marck | Hi everyone | 18:37 |
| devananda | aaand ... no more open tags on https://bugs.launchpad.net/ironic/+bugs?field.tag=juno-rc-potential | 18:37 |
| devananda | marck: hi! | 18:37 |
| marck | I'm having trouble getting ironic to provision the first node, the log in nova-scheduler reads: INFO nova.filters [req-id...] Filter ComputeFilter returned 0 hosts | 18:39 |
| marck | I followed the instructions @ http://docs.openstack.org/developer/ironic/deploy/install-guide.html, but when restarting nova-scheduler the log said "TRACE nova ImportError: No module named ironic_host_manager" | 18:42 |
| devananda | marck: are you seeing available resources in the nova-compute log? | 18:42 |
| devananda | marck: hm. what version are you running? | 18:43 |
| *** dprince has joined #openstack-ironic | 18:43 | |
| marck | I changed scheduler_host_manager=nova.scheduler.ironic_host_manager.IronicHostManager -> scheduler_host_manager=ironic.nova.scheduler.ironic_host_manager.IronicHostManager and that went away | 18:44 |
| devananda | gotcha | 18:44 |
| devananda | what version of nova and ironic are you running? | 18:44 |
| NobodyCam | devananda: w00t :) | 18:44 |
| *** romcheg has quit IRC | 18:45 | |
| *** rainya has quit IRC | 18:45 | |
| marck | nova=2.17.0, ironic=0.1.2 | 18:46 |
| marck | ironic-api=2014.1.rc1 | 18:46 |
| devananda | marck: ok, that's the icehouse RC candidate | 18:47 |
| devananda | the guide you're following is written for tip | 18:47 |
| devananda | pulling the Juno RC's is close enough they should work as well, but icehouse configs were different | 18:47 |
| marck | ahh, my machines are Ubuntu 14.04 LTS. | 18:48 |
| marck | do I have to reconfigure the apt repos for that, or do it from source? | 18:48 |
| devananda | source would be best | 18:49 |
| marck | Is it possible to do in place or should I start from scratch? I'm just experimenting/learning right now | 18:50 |
| *** jmanko has joined #openstack-ironic | 18:52 | |
| *** rainya has joined #openstack-ironic | 18:52 | |
| *** romcheg has joined #openstack-ironic | 18:53 | |
| *** zentropy has quit IRC | 18:53 | |
| *** kbyrne has quit IRC | 18:53 | |
| *** jmank has quit IRC | 18:53 | |
| *** GheRivero has quit IRC | 18:53 | |
| *** rainya has quit IRC | 18:54 | |
| *** zentropy has joined #openstack-ironic | 18:55 | |
| *** jeblair has quit IRC | 18:55 | |
| *** rainya has joined #openstack-ironic | 18:55 | |
| *** jeblair has joined #openstack-ironic | 18:56 | |
| *** yuanying has quit IRC | 18:57 | |
| *** lynxman has quit IRC | 18:57 | |
| *** yuanying_ has joined #openstack-ironic | 18:57 | |
| *** amaurymedeiros has quit IRC | 18:57 | |
| *** lynxman has joined #openstack-ironic | 18:59 | |
| devananda | marck: either install from pip (that'll get you as close to source as possible without cloning from git directly) | 19:00 |
| devananda | marck: or if you want to consume packages, try these: http://packages.ubuntu.com/search?suite=utopic&searchon=names&keywords=ironic | 19:00 |
| *** BertieFulton has joined #openstack-ironic | 19:01 | |
| *** GheRivero has joined #openstack-ironic | 19:02 | |
| *** amaurymedeiros has joined #openstack-ironic | 19:02 | |
| *** amaurymedeiros has quit IRC | 19:03 | |
| *** amaurymedeiros has joined #openstack-ironic | 19:03 | |
| *** kbyrne has joined #openstack-ironic | 19:06 | |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Update doc headers at end of Juno https://review.openstack.org/128735 | 19:07 |
| *** achanda has quit IRC | 19:07 | |
| marck | devananda: thanks for your help, I'll give pip a try | 19:08 |
| *** marck has quit IRC | 19:09 | |
| *** tatyana has joined #openstack-ironic | 19:17 | |
| NobodyCam | ahh ramen favored food of the recovering | 19:23 |
| *** achanda has joined #openstack-ironic | 19:25 | |
| Shrews | hrm, our hash ring assumes a big-endian architecture. that's probably not good | 19:33 |
| *** ndipanov is now known as ndipanov_gone | 19:33 | |
| *** marck has joined #openstack-ironic | 19:42 | |
| *** marck has quit IRC | 19:42 | |
| *** marck has joined #openstack-ironic | 19:42 | |
| NobodyCam | humm I have not tried to run on a pure risk system | 19:45 |
| NobodyCam | s/risk/risc/ | 19:46 |
| Shrews | it's also eliminate half of the bytes in a hash, increasing the chance for collisions | 19:46 |
| Shrews | s/eliminate/eliminating/ | 19:46 |
| Shrews | and don't ask me if it works if there are collisions :/ | 19:47 |
| Shrews | i appear to have dropped into a rathole here | 19:47 |
| *** marck has quit IRC | 19:47 | |
| NobodyCam | :/ | 19:48 |
| devananda | Shrews: do we actually care about running ironic services on little-endian systems? | 19:50 |
| Shrews | *shrug* | 19:51 |
| Shrews | intel is little, no? | 19:51 |
| devananda | JayF: I haven't created the lp and pypi projects. I believe I had told you I would... | 19:51 |
| JayF | I think you started telling me how | 19:52 |
| devananda | ah | 19:52 |
| JayF | and then I saw something shiny | 19:52 |
| devananda | heh | 19:52 |
| JayF | and now it's Juno-RC3 | 19:52 |
| JayF | lol | 19:52 |
| devananda | JayF: the release is *tomorrow* | 19:52 |
| devananda | I think it's a little late to call IPA part of the integrated release -- and I think we didn't plan on doing that anyawy | 19:52 |
| JayF | I don't care what it's called | 19:53 |
| devananda | but instead releasing it on an independent cadence, like all other client libraries | 19:53 |
| JayF | I will *never* consume a build artifact of IPA officially built :P | 19:53 |
| devananda | so there's no rush to do it today | 19:53 |
| JayF | so I'm just trying to ask people who don't care how they'd want this packaged | 19:53 |
| JayF | and am getting a chorus of meh | 19:53 |
| NobodyCam | Shrews: I believe intel is little... I only know of a few pure RISC types that are not.. MIPPS / Alphas | 19:53 |
| JayF | which tends to make me not care much about releasing it either, tbh | 19:53 |
| NobodyCam | though I have not looked at any of the ARM cpu's | 19:53 |
| devananda | JayF: heh. no, i think publishing some official builds of IPA (and of DIB's ramdisk, which we need a better name for) is good | 19:53 |
| JayF | I'm willing to do what's needed to make potential deployers more happy, I just have little idea where to start (other than what I asked ttx earlier) | 19:53 |
| devananda | JayF: there are folks who want to download it and use it | 19:54 |
| JayF | So should we tag a 1.0.0 of the agent today then? | 19:54 |
| JayF | jroll: ^ | 19:54 |
| *** dprince has quit IRC | 19:54 | |
| JayF | we could tag it, then pull the latest build from tarballs.openstack.org which has the relevant ramdisk files | 19:54 |
| devananda | JayF: where do ya'll track bugs in IPA? | 19:55 |
| JayF | devananda: ironic | 19:55 |
| devananda | thought so | 19:55 |
| JayF | devananda: that was decided back in sunnyvale | 19:55 |
| JayF | ironic tag:agent | 19:55 |
| devananda | that makes it awkward to have a separate project | 19:55 |
| devananda | on LP and pypi | 19:55 |
| JayF | I mean, I think it makes sense to continue tracking bugs together | 19:56 |
| JayF | bceause there will be bugs that cross the interface | 19:56 |
| * devananda jumps on a conference call | 19:57 | |
| *** ifarkas has quit IRC | 19:57 | |
| *** romcheg has quit IRC | 19:57 | |
| JayF | Also IPA is useless without Ironic right now | 19:57 |
| JayF | although I guess that's true of python-*client | 19:58 |
| *** romcheg has joined #openstack-ironic | 19:58 | |
| *** rainya has quit IRC | 19:59 | |
| NobodyCam | brb | 20:00 |
| *** romcheg has quit IRC | 20:02 | |
| *** rainya has joined #openstack-ironic | 20:07 | |
| *** rainya has quit IRC | 20:08 | |
| *** BertieFulton has quit IRC | 20:10 | |
| *** tatyana has quit IRC | 20:12 | |
| *** rainya has joined #openstack-ironic | 20:12 | |
| *** yjiang5 has joined #openstack-ironic | 20:17 | |
| *** yjiang5 has left #openstack-ironic | 20:17 | |
| Shrews | devananda: at some point, i'd like to discuss the logic behind this line: https://github.com/openstack/ironic/blob/master/ironic/common/hash_ring.py#L94 | 20:18 |
| Shrews | something like this might work better: hashed_key = long(key_hash.hexdigest, 16) | 20:19 |
| Shrews | hexdigest()* | 20:20 |
| Shrews | or, i'm totally wrong and need flogging | 20:20 |
| *** achanda has quit IRC | 20:31 | |
| *** igordcard has joined #openstack-ironic | 20:37 | |
| *** igordcard has quit IRC | 20:37 | |
| mrda | Morning ironic | 20:38 |
| *** ChuckC___ has joined #openstack-ironic | 20:39 | |
| NobodyCam | morning mrda | 20:40 |
| *** romcheg has joined #openstack-ironic | 20:41 | |
| mrda | NobodyCam: \o | 20:42 |
| *** Kui has joined #openstack-ironic | 20:47 | |
| *** ChuckC___ has quit IRC | 20:48 | |
| *** achanda has joined #openstack-ironic | 20:49 | |
| *** jjohnson2 has quit IRC | 21:09 | |
| *** yjiang5 has joined #openstack-ironic | 21:11 | |
| *** ChuckC_ is now known as ChuckC | 21:18 | |
| *** Kui has quit IRC | 21:20 | |
| *** Kui has joined #openstack-ironic | 21:26 | |
| *** r-daneel has quit IRC | 21:28 | |
| *** yjiang5 has left #openstack-ironic | 21:28 | |
| *** yjiang5 has joined #openstack-ironic | 21:30 | |
| *** chuckC__ has quit IRC | 21:35 | |
| *** foexle has quit IRC | 21:40 | |
| *** ChuckC has quit IRC | 21:46 | |
| *** rainya has quit IRC | 21:57 | |
| *** pelix has quit IRC | 21:57 | |
| *** jcoufal has quit IRC | 22:13 | |
| NobodyCam | do we call out anywhere in our docs that pxe deploys require enough disk space to convert imsges to raw? | 22:20 |
| *** jgrimm is now known as zz_jgrimm | 22:24 | |
| *** ChuckC has joined #openstack-ironic | 22:31 | |
| *** ryanpetrello has quit IRC | 22:34 | |
| *** ryanpetrello has joined #openstack-ironic | 22:35 | |
| *** ChuckC has quit IRC | 22:35 | |
| *** ChuckC has joined #openstack-ironic | 22:35 | |
| *** enikanorov_ has quit IRC | 22:37 | |
| *** chuckC_ has joined #openstack-ironic | 22:37 | |
| *** achanda has quit IRC | 22:40 | |
| *** achanda has joined #openstack-ironic | 22:44 | |
| *** rainya has joined #openstack-ironic | 22:44 | |
| *** lucas-dinner has quit IRC | 22:47 | |
| yjiang5 | NobodyCam: Hmm, is the check_image_size() checks for raw size or qcow size? | 22:58 |
| *** marcoemorais has quit IRC | 23:05 | |
| *** marcoemorais has joined #openstack-ironic | 23:05 | |
| *** marcoemorais has quit IRC | 23:05 | |
| *** marcoemorais has joined #openstack-ironic | 23:06 | |
| *** marcoemorais has quit IRC | 23:06 | |
| *** marcoemorais has joined #openstack-ironic | 23:06 | |
| *** marcoemorais has quit IRC | 23:07 | |
| *** marcoemorais has joined #openstack-ironic | 23:07 | |
| *** romcheg1 has joined #openstack-ironic | 23:10 | |
| *** romcheg has quit IRC | 23:10 | |
| NobodyCam | yjiang5: https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/deploy_utils.py#L203-L209 | 23:13 |
| NobodyCam | that looks like the qcow size | 23:13 |
| *** romcheg1 has quit IRC | 23:13 | |
| NobodyCam | :( | 23:14 |
| *** Haomeng|2 has quit IRC | 23:14 | |
| openstackgerrit | Jim Rollenhagen proposed a change to openstack/ironic: Add maintenance_reason when setting maintenance mode https://review.openstack.org/128646 | 23:15 |
| openstackgerrit | Jim Rollenhagen proposed a change to openstack/ironic: Add Node.maintenance_reason https://review.openstack.org/128645 | 23:15 |
| *** Haomeng|2 has joined #openstack-ironic | 23:16 | |
| *** praneshp has quit IRC | 23:20 | |
| *** praneshp has joined #openstack-ironic | 23:22 | |
| *** marcoemorais has quit IRC | 23:26 | |
| *** marcoemorais has joined #openstack-ironic | 23:26 | |
| *** pcrews has quit IRC | 23:31 | |
| yjiang5 | NobodyCam: Possibly this is a generic nova issue? The nova compute api, or more precicely, the _check_requested_image() in nova/compute/api.py, should make sure this is correct? | 23:31 |
| NobodyCam | humm. I'm thinking we have a get_raw_image_size function and then drivers that need to convert the image can check the uncompressed size too | 23:33 |
| JayF | Hmm. | 23:34 |
| JayF | Is the uncompressed size stored in glance? | 23:34 |
| *** pcrews has joined #openstack-ironic | 23:34 | |
| NobodyCam | not that I know | 23:34 |
| NobodyCam | we call qemu-img convert we could easly get that data from qemu-img info | 23:36 |
| JayF | it's called from the agent | 23:36 |
| JayF | in ipa | 23:37 |
| yjiang5 | JayF: NobodyCam, I think even libvirt should check the raw size, not the qcow size? Of course, we should sanity check in ironic side, but it will be better if we can do it earlier, instead of the deploy stage? | 23:37 |
| yjiang5 | Ooops, need run away now. Will check you conclusion later. | 23:39 |
| *** yjiang5 is now known as yjiang5_away | 23:39 | |
| NobodyCam | hummm JayF the agent space is all tmp based for the image stuff/ | 23:39 |
| *** yjiang5_away has left #openstack-ironic | 23:40 | |
| JayF | NobodyCam: yeah, but you have to have enough ram on the machine | 23:40 |
| JayF | NobodyCam: to have enough space to store the image in a ramdisk | 23:40 |
| JayF | NobodyCam: so default on the IPA coreos ramdisk is ramdisk size is ram/2 | 23:40 |
| JayF | NobodyCam: so you have 32GB ram, you get 16GB temp fs, 16GB usable ram | 23:40 |
| JayF | NobodyCam: you can provide rootflags=size=20G, which, in that examle, would give you 20GB temp fs, 12GB usable ram | 23:41 |
| JayF | NobodyCam: I think you get the idea :P | 23:41 |
| openstackgerrit | OpenStack Proposal Bot proposed a change to openstack/ironic-python-agent: Updated from global requirements https://review.openstack.org/114058 | 23:41 |
| NobodyCam | oh really so agent would need 64 Gb to beploy a 32GB image | 23:41 |
| JayF | NobodyCam: possibly, yeah | 23:42 |
| JayF | NobodyCam: we need to implement streaming | 23:43 |
| NobodyCam | ++ | 23:43 |
| jroll | we're part of the way there | 23:43 |
| jroll | or getting there | 23:43 |
| * NobodyCam notes he was just looking at 32 GB deploys | 23:43 | |
| JayF | are we really? | 23:43 |
| jroll | maybe, idk | 23:44 |
| jroll | I think some of comstud's stuff will help | 23:44 |
| jroll | breaking out the image writing | 23:44 |
| JayF | that's true | 23:49 |
| *** lifeless has quit IRC | 23:52 | |
| *** lifeless has joined #openstack-ironic | 23:53 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
