| *** lucas-dinner has quit IRC | 00:06 | |
| *** lucasagomes has joined #openstack-ironic | 00:11 | |
| *** amotoki has quit IRC | 00:14 | |
| *** amotoki has joined #openstack-ironic | 00:26 | |
| *** mgoddard__ has joined #openstack-ironic | 00:26 | |
| *** mtanino_ has joined #openstack-ironic | 00:27 | |
| *** mtanino has quit IRC | 00:29 | |
| *** mgoddard_ has quit IRC | 00:29 | |
| *** achanda has joined #openstack-ironic | 00:41 | |
| *** achanda has quit IRC | 00:43 | |
| *** achanda has joined #openstack-ironic | 00:43 | |
| *** achanda has quit IRC | 00:44 | |
| *** davideagnello has quit IRC | 00:44 | |
| *** achanda has joined #openstack-ironic | 00:45 | |
| *** achanda has quit IRC | 00:47 | |
| *** Sukhdev has quit IRC | 00:48 | |
| *** achanda has joined #openstack-ironic | 00:50 | |
| *** achanda has quit IRC | 00:55 | |
| *** achanda has joined #openstack-ironic | 01:00 | |
| *** achanda has quit IRC | 01:01 | |
| *** dims has quit IRC | 01:01 | |
| *** achanda has joined #openstack-ironic | 01:05 | |
| *** vishwanathj has joined #openstack-ironic | 01:06 | |
| *** vishwanathj has quit IRC | 01:07 | |
| *** vishwanathj has joined #openstack-ironic | 01:12 | |
| *** achanda has quit IRC | 01:17 | |
| *** achanda has joined #openstack-ironic | 01:21 | |
| *** dims has joined #openstack-ironic | 01:23 | |
| *** baoli has joined #openstack-ironic | 01:24 | |
| *** achanda has quit IRC | 01:26 | |
| *** rloo has quit IRC | 01:28 | |
| *** baoli has quit IRC | 01:29 | |
| *** chenke has quit IRC | 01:30 | |
| *** chenke has joined #openstack-ironic | 01:31 | |
| *** baoli has joined #openstack-ironic | 01:38 | |
| *** baoli has quit IRC | 01:43 | |
| *** harshs has joined #openstack-ironic | 01:50 | |
| *** bandicot has joined #openstack-ironic | 01:55 | |
| *** bandicot has quit IRC | 01:56 | |
| *** kan_ has joined #openstack-ironic | 02:02 | |
| *** mtanino_ has quit IRC | 02:04 | |
| *** david-lyle has joined #openstack-ironic | 02:22 | |
| openstackgerrit | Haomeng,Wang proposed openstack/python-ironicclient: continue to delete next node if failed with previous one https://review.openstack.org/262393 | 02:33 |
|---|---|---|
| openstackgerrit | Kan proposed openstack/python-ironicclient: Add CLI to list nodes using the same driver. https://review.openstack.org/264007 | 02:36 |
| *** davideagnello has joined #openstack-ironic | 02:45 | |
| *** Haomeng has quit IRC | 02:47 | |
| *** davideagnello has quit IRC | 02:51 | |
| openstackgerrit | Haomeng,Wang proposed openstack/python-ironicclient: continue to delete next node if failed with previous one https://review.openstack.org/262393 | 02:52 |
| *** e0ne has quit IRC | 02:59 | |
| *** e0ne has joined #openstack-ironic | 03:00 | |
| *** e0ne has quit IRC | 03:01 | |
| *** thanhnt-z has joined #openstack-ironic | 03:04 | |
| openstackgerrit | Haomeng,Wang proposed openstack/python-ironicclient: continue to delete next node if failed with previous one https://review.openstack.org/262393 | 03:07 |
| *** Marga_ has joined #openstack-ironic | 03:10 | |
| *** vishwanathj has quit IRC | 03:10 | |
| *** yuanying has quit IRC | 03:15 | |
| *** dims has quit IRC | 03:20 | |
| *** achanda has joined #openstack-ironic | 03:21 | |
| *** Sukhdev has joined #openstack-ironic | 03:22 | |
| *** baoli has joined #openstack-ironic | 03:24 | |
| thanhnt-z | Hi Haomeng, are you there? | 03:26 |
| *** links has joined #openstack-ironic | 03:27 | |
| openstackgerrit | Yuiko Takada proposed openstack/ironic-inspector: Enable ironic devstack plugin in local.conf sample https://review.openstack.org/269413 | 03:44 |
| *** davideagnello has joined #openstack-ironic | 03:48 | |
| *** davideagnello has quit IRC | 03:52 | |
| *** achanda has quit IRC | 04:00 | |
| *** chenke_ has joined #openstack-ironic | 04:00 | |
| *** chenke has quit IRC | 04:03 | |
| *** yuanying has joined #openstack-ironic | 04:07 | |
| openstackgerrit | Haomeng,Wang proposed openstack/python-ironicclient: continue to delete next node if failed with previous one https://review.openstack.org/262393 | 04:16 |
| *** Nisha has joined #openstack-ironic | 04:18 | |
| *** Marga_ has quit IRC | 04:30 | |
| *** achanda has joined #openstack-ironic | 04:30 | |
| *** Marga_ has joined #openstack-ironic | 04:30 | |
| *** baoli has quit IRC | 04:31 | |
| *** achanda has quit IRC | 04:38 | |
| *** achanda has joined #openstack-ironic | 04:39 | |
| *** Marga_ has quit IRC | 04:44 | |
| *** raddaoui has joined #openstack-ironic | 04:51 | |
| *** raddaoui has quit IRC | 04:51 | |
| *** teju has joined #openstack-ironic | 05:08 | |
| openstackgerrit | Haomeng,Wang proposed openstack/python-ironicclient: continue to delete next node if failed with previous one https://review.openstack.org/262393 | 05:09 |
| *** harshs_ has joined #openstack-ironic | 05:17 | |
| *** harshs has quit IRC | 05:17 | |
| *** harshs_ is now known as harshs | 05:17 | |
| *** baoli has joined #openstack-ironic | 05:32 | |
| *** Nisha has quit IRC | 05:34 | |
| *** Nisha has joined #openstack-ironic | 05:34 | |
| *** amotoki has quit IRC | 05:39 | |
| *** harshs has quit IRC | 05:39 | |
| *** amotoki has joined #openstack-ironic | 05:44 | |
| *** smoriya_ has quit IRC | 05:57 | |
| *** jamielennox is now known as jamielennox|away | 05:59 | |
| *** yuikotakada has joined #openstack-ironic | 06:14 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ironic: Imported Translations from Zanata https://review.openstack.org/268582 | 06:14 |
| *** chenke__ has joined #openstack-ironic | 06:24 | |
| *** achanda_ has joined #openstack-ironic | 06:25 | |
| *** chenke_ has quit IRC | 06:27 | |
| *** achanda has quit IRC | 06:28 | |
| *** amotoki_ has joined #openstack-ironic | 06:32 | |
| *** amotoki has quit IRC | 06:34 | |
| *** amotoki_ has quit IRC | 06:35 | |
| *** davideagnello has joined #openstack-ironic | 06:37 | |
| *** davideagnello has quit IRC | 06:38 | |
| *** Nisha_away has joined #openstack-ironic | 06:39 | |
| *** Nisha_brb has joined #openstack-ironic | 06:41 | |
| *** ChubYann has quit IRC | 06:41 | |
| *** Nisha has quit IRC | 06:42 | |
| *** Nisha_away has quit IRC | 06:43 | |
| *** achanda_ has quit IRC | 06:43 | |
| *** amotoki has joined #openstack-ironic | 06:44 | |
| *** praneshp has joined #openstack-ironic | 06:44 | |
| openstackgerrit | Haomeng,Wang proposed openstack/python-ironicclient: continue to delete next node if failed with previous one https://review.openstack.org/262393 | 06:45 |
| *** amotoki has quit IRC | 07:00 | |
| *** Haomeng has joined #openstack-ironic | 07:01 | |
| *** jcoufal has joined #openstack-ironic | 07:03 | |
| *** ukalifon has joined #openstack-ironic | 07:09 | |
| *** Haomeng has quit IRC | 07:10 | |
| *** Haomeng has joined #openstack-ironic | 07:11 | |
| *** amotoki has joined #openstack-ironic | 07:17 | |
| *** Nisha_brb has quit IRC | 07:27 | |
| *** changzhi has joined #openstack-ironic | 07:35 | |
| *** smoriya_ has joined #openstack-ironic | 07:42 | |
| *** amotoki has quit IRC | 07:50 | |
| *** rcernin has joined #openstack-ironic | 08:01 | |
| *** ionutbalutoiu has joined #openstack-ironic | 08:14 | |
| *** praneshp has quit IRC | 08:24 | |
| *** chlong has quit IRC | 08:27 | |
| *** Nisha has joined #openstack-ironic | 08:27 | |
| *** ifarkas has joined #openstack-ironic | 08:29 | |
| *** davideagnello has joined #openstack-ironic | 08:38 | |
| *** daemontool has joined #openstack-ironic | 08:41 | |
| *** Sukhdev has quit IRC | 08:43 | |
| *** davideagnello has quit IRC | 08:44 | |
| *** ifarkas has quit IRC | 08:49 | |
| *** _degorenko|afk is now known as degorenko | 08:55 | |
| *** mbound has joined #openstack-ironic | 08:56 | |
| *** ifarkas has joined #openstack-ironic | 08:58 | |
| *** ifarkas has quit IRC | 08:59 | |
| *** ifarkas has joined #openstack-ironic | 08:59 | |
| *** phil_231 has joined #openstack-ironic | 08:59 | |
| *** daemontool has quit IRC | 09:02 | |
| *** e0ne has joined #openstack-ironic | 09:03 | |
| *** daemontool has joined #openstack-ironic | 09:03 | |
| *** sirius_ has joined #openstack-ironic | 09:04 | |
| *** amotoki has joined #openstack-ironic | 09:11 | |
| *** ndipanov has quit IRC | 09:18 | |
| *** jistr has joined #openstack-ironic | 09:20 | |
| *** ndipanov has joined #openstack-ironic | 09:21 | |
| *** alexpilotti has joined #openstack-ironic | 09:23 | |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Add portgroups to support LAG interfaces - RPC https://review.openstack.org/206243 | 09:25 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Add portgroups to support LAG interfaces - API https://review.openstack.org/206244 | 09:25 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Add portgroups to support LAG interfaces - net https://review.openstack.org/206245 | 09:25 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Update the deploy drivers with network flipping logic https://review.openstack.org/213262 | 09:25 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Allow to build user image with DIB https://review.openstack.org/256363 | 09:25 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Added operator documentation for ironic portgroups https://review.openstack.org/228496 | 09:25 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Add Link-Local-Connection info to ironic port https://review.openstack.org/256365 | 09:25 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: refactor ironic enroll-node code https://review.openstack.org/256364 | 09:25 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Add configure_provision_network function https://review.openstack.org/256367 | 09:25 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Update Ironic VM network connection https://review.openstack.org/256366 | 09:25 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Add portgroups to support LAG interfaces - objs https://review.openstack.org/206238 | 09:25 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Add Ironic/Neutron integration documentation https://review.openstack.org/258596 | 09:25 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Add network provider interface and implementations https://review.openstack.org/139687 | 09:25 |
| *** alexpilotti has quit IRC | 09:28 | |
| *** athomas has quit IRC | 09:28 | |
| *** Marga_ has joined #openstack-ironic | 09:29 | |
| *** mkovacik has quit IRC | 09:31 | |
| *** athomas has joined #openstack-ironic | 09:34 | |
| *** e0ne has quit IRC | 09:34 | |
| *** electrofelix has joined #openstack-ironic | 09:38 | |
| *** daemontool has quit IRC | 09:39 | |
| *** daemontool has joined #openstack-ironic | 09:39 | |
| *** derekh has joined #openstack-ironic | 09:43 | |
| *** boris-42 has quit IRC | 09:43 | |
| openstackgerrit | Yuriy Yekovenko proposed openstack/ironic: [WIP] Add test to verify ironic multitenancy https://review.openstack.org/269157 | 09:46 |
| *** amotoki has quit IRC | 09:48 | |
| *** e0ne has joined #openstack-ironic | 09:48 | |
| *** alexpilotti has joined #openstack-ironic | 09:50 | |
| vdrok | morning ironic! | 09:54 |
| *** alexpilo_ has joined #openstack-ironic | 09:54 | |
| *** alexpilotti has quit IRC | 09:56 | |
| zhenguo | vdrok: morning! | 09:56 |
| vdrok | morning zhenguo :) | 09:57 |
| *** e0ne has quit IRC | 09:57 | |
| openstackgerrit | Yuiko Takada proposed openstack/ironic: Migrate Tempest tests into Ironic tree https://review.openstack.org/253982 | 09:58 |
| *** alexpilo_ has quit IRC | 09:58 | |
| openstackgerrit | Sam Betts proposed openstack/ironic: [WIP] Enable tinyipa for devstack Ironic https://review.openstack.org/259089 | 10:01 |
| *** sambetts has joined #openstack-ironic | 10:01 | |
| *** alexpilotti has joined #openstack-ironic | 10:02 | |
| *** alexpilotti has quit IRC | 10:06 | |
| *** Nisha has quit IRC | 10:06 | |
| *** daemontool has quit IRC | 10:08 | |
| *** daemontool has joined #openstack-ironic | 10:08 | |
| *** daemontool has quit IRC | 10:11 | |
| *** daemontool has joined #openstack-ironic | 10:11 | |
| *** kbyrne has quit IRC | 10:12 | |
| *** daemontool has quit IRC | 10:17 | |
| mgould|afk | morning Ironic! | 10:17 |
| *** mgould|afk is now known as mgould | 10:17 | |
| *** daemontool has joined #openstack-ironic | 10:17 | |
| *** dtantsur|afk is now known as dtantsur | 10:18 | |
| mgould | morning dtantsur! | 10:18 |
| *** Marga_ has quit IRC | 10:18 | |
| dtantsur | morning Ironic, morning mgould | 10:18 |
| *** amotoki has joined #openstack-ironic | 10:18 | |
| *** daemontool has quit IRC | 10:19 | |
| *** daemontool has joined #openstack-ironic | 10:19 | |
| yuikotakada | Good morning, ironic | 10:20 |
| vdrok | good morning mgould dtantsur and yuikotakada ! | 10:20 |
| yuikotakada | morning, vdrok :) | 10:20 |
| dtantsur | morning vdrok, yuikotakada | 10:21 |
| yuikotakada | dtantsur, morning :) | 10:21 |
| sambetts | Morning all o/ | 10:21 |
| yuikotakada | sambetts, morning :) | 10:22 |
| sambetts | o/ yuikotakada | 10:22 |
| vdrok | morning sambetts | 10:23 |
| mgould | yuikotakada vdrok sambetts morning! | 10:24 |
| yuikotakada | mgould morning! | 10:25 |
| sambetts | o/ mgould | 10:25 |
| *** ionutbalutoiu has quit IRC | 10:25 | |
| *** mkovacik has joined #openstack-ironic | 10:26 | |
| *** ionutbalutoiu has joined #openstack-ironic | 10:26 | |
| *** daemontool has quit IRC | 10:38 | |
| *** davideagnello has joined #openstack-ironic | 10:40 | |
| openstackgerrit | Imre Farkas proposed openstack/ironic: DRAC: cleanup after switch to python-dracclient https://review.openstack.org/255310 | 10:42 |
| *** daemontool has joined #openstack-ironic | 10:42 | |
| *** MattMan has quit IRC | 10:44 | |
| *** MattMan has joined #openstack-ironic | 10:44 | |
| dtantsur | sambetts, morning | 10:45 |
| *** davideagnello has quit IRC | 10:46 | |
| sambetts | o/ dtantsur | 10:46 |
| *** thanhnt-z has quit IRC | 10:47 | |
| *** dims has joined #openstack-ironic | 10:48 | |
| sambetts | Man how have they not fixed that devstack swift issue yet... | 10:49 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-inspector: Pin diskimage-builder 1.7.1 to unblock the gate https://review.openstack.org/269550 | 10:50 |
| dtantsur | sambetts, yuikotakada, ifarkas, I'd like to land this to unblock inspector gate ^^ objections? | 10:50 |
| dtantsur | sambetts, yeah, fix for ironic gate fails in gate, hopefully something transient | 10:50 |
| sambetts | dtantsur: looking at the way you've done that patch, we wouldn't ever need to revert it would we? | 10:51 |
| *** kalpase has joined #openstack-ironic | 10:51 | |
| sambetts | dtantsur: it would probably be good to keep it there as a never use this version hint | 10:51 |
| dtantsur | sambetts, we should use pip_install_gr to properly use global-requirements | 10:51 |
| sambetts | ah ok, could we put at !=1.7.1 in global requirements, because its not compatible with everything? | 10:52 |
| ifarkas | dtantsur, ack, no objections | 10:52 |
| sambetts | also no objections | 10:53 |
| dtantsur | sambetts, sigh.. 1.7.1 fixes ironic gate, so if we pin it, ironic will be broken >_< | 10:53 |
| *** Marga_ has joined #openstack-ironic | 10:53 | |
| sambetts | oh... | 10:53 |
| dtantsur | ... and maybe we can't use 1.7.0, oh... | 10:53 |
| dtantsur | maybe I need to pin it to pre-1.5.0 | 10:53 |
| sambetts | oh god... can we not pin it to trunk DIB for now? | 10:54 |
| yuikotakada | dtantsur, mmm, how about make gate failing job nv? | 10:54 |
| dtantsur | yuikotakada, we need it to pass while we support the old ramdisk.. also there's no infra cores around, so we won't make it fast | 10:55 |
| * sambetts starts chanting | 10:56 | |
| sambetts | I .. P .. A ... I .. P ... A | 10:56 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-inspector: Block broken diskimage-builder versions to unblock the gate https://review.openstack.org/269550 | 11:01 |
| dtantsur | sambetts, ifarkas, yuikotakada, pinning all broken versions ^^ | 11:01 |
| dtantsur | approving this in 5 minutes, unless I get objections | 11:02 |
| sambetts | dtantsur: what happens if Ironic installs DIB too? | 11:02 |
| sambetts | does that override it ? | 11:02 |
| sambetts | or if something else also trys to install DIB after inspector? | 11:03 |
| dtantsur | after does not matter, before, however, does | 11:03 |
| dtantsur | lemme update | 11:03 |
| dtantsur | also I've changed the wrong place | 11:04 |
| yuikotakada | thanks, I'd like to try to run it in my env... | 11:04 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-inspector: Block broken diskimage-builder versions to unblock the gate https://review.openstack.org/269550 | 11:05 |
| dtantsur | this is the correct one ^^ | 11:05 |
| *** sirius_ has quit IRC | 11:05 | |
| dtantsur | I finally got people to release DIB 1.7.2, but it will take it substantial time to reach the gate, so I think we still need this pin to prepare for M2 releases | 11:09 |
| sambetts | yeah sounds like it :( | 11:14 |
| *** Marga_ has quit IRC | 11:15 | |
| dtantsur | yeah, by the way, I'd like to release inspector with the IPA patch and inspector-client with 'data save' command (to be approved) | 11:15 |
| *** changzhi has quit IRC | 11:15 | |
| lucasagomes | morning all | 11:15 |
| dtantsur | and do it all around Wed-Thu, with the remaining openstack M2 | 11:15 |
| dtantsur | lucasagomes, morning | 11:15 |
| dtantsur | yuikotakada, lemme know when you finish your testing (note that only patchset 3 is correct) | 11:16 |
| yuikotakada | dtantsur, takes 10mins, I guess | 11:17 |
| vdrok | morning lucasagomes | 11:17 |
| dtantsur | yuikotakada, no problem, take your time | 11:17 |
| yuikotakada | LGTM, otherwise | 11:17 |
| yuikotakada | lucasagomes, morning | 11:17 |
| lucasagomes | yuikotakada, morning :-) | 11:17 |
| mgould | lucasagomes, morning! | 11:18 |
| dtantsur | I suggest we make dib job non-voting for N release after mitaka is branched | 11:18 |
| yuikotakada | dtantsur, yeah, I agree for nv | 11:19 |
| sambetts | yeah, I agree, then we can start probably deprecating it | 11:20 |
| *** ionutbalutoiu has quit IRC | 11:20 | |
| *** ionutbalutoiu has joined #openstack-ironic | 11:21 | |
| lucasagomes | mgould, hi there! | 11:29 |
| *** e0ne has joined #openstack-ironic | 11:38 | |
| yuikotakada | devstack env has been not built correctly because of other issue... | 11:44 |
| dtantsur | oh... | 11:44 |
| dtantsur | I'd say we can rely on the gate to pass in this case | 11:45 |
| dtantsur | yuikotakada, what was the error btw? | 11:45 |
| yuikotakada | I cannot garantee whether it works or not, so I would like to rely on sam and ifarkas >_< | 11:45 |
| *** sergek has joined #openstack-ironic | 11:45 | |
| dtantsur | gate won't pass if it does not work, so no worries :) | 11:45 |
| yuikotakada | dtantsur, :) | 11:46 |
| yuikotakada | I've got error "losetup: could not find any free loop device" it sometimes happens and resolved after rebooting... | 11:47 |
| sambetts | dtantsur: I've tried devstacking this morning and it blows up because of the swift issue :( | 11:52 |
| phil_231 | hey everyone | 11:53 |
| phil_231 | random question, but is ironic able to handle multiple ports being assigned to a single node? | 11:55 |
| dtantsur | phil_231, ironic is, nova is not AFAIK | 11:59 |
| phil_231 | dtantsur, ok cool, just know that it had problems in the past and that makes things easier for me now :P | 12:00 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic-specs: Add pluggable credentials storage https://review.openstack.org/186056 | 12:01 |
| *** deray has joined #openstack-ironic | 12:01 | |
| *** daemontool has quit IRC | 12:05 | |
| *** daemontool has joined #openstack-ironic | 12:06 | |
| deray | hi.. I am facing an issue with the new devstack(ironic as plugin) stacking up .. Error msg is: Missing parameter(s): \n Set an authentication URL, with --os-auth-url, OS_AUTH_URL or auth.auth_url | 12:06 |
| deray | is there soemthing very basic I am missing out? | 12:07 |
| *** baoli has quit IRC | 12:07 | |
| sambetts | deray: there is currently an issue with devstack and swift causing it to fail | 12:10 |
| sambetts | deray: https://bugs.launchpad.net/ironic/+bug/1535245 | 12:11 |
| openstack | Launchpad bug 1535245 in devstack "agent_ssh gate fails on openstack object store account set --property Temp-URL-Key=secretkey" [Undecided,In progress] - Assigned to Dmitry Tantsur (divius) | 12:11 |
| sambetts | deray: just waiting on a recheck and the fix should merge soon, https://review.openstack.org/#/c/268960/1 | 12:14 |
| *** alexpilotti has joined #openstack-ironic | 12:16 | |
| deray | sambetts, oh okay .. so nimble :) | 12:23 |
| *** raildo-afk is now known as raildo | 12:27 | |
| *** davideagnello has joined #openstack-ironic | 12:42 | |
| *** smoriya_ has quit IRC | 12:44 | |
| *** links has quit IRC | 12:46 | |
| *** davideagnello has quit IRC | 12:47 | |
| jroll | morning everyone | 12:51 |
| dtantsur | morning jroll | 12:51 |
| jroll | phil_231: more clearly, the number of ports in the ironic db for a node, must be equal to the number of neutron networks attached by nova | 12:51 |
| lucasagomes | jroll, morning | 12:52 |
| jroll | hey dtantsur :) | 12:52 |
| deray | jroll, g'morning | 12:52 |
| deray | g'morning everyone | 12:52 |
| jroll | hey deray :) | 12:52 |
| deray | o/ | 12:52 |
| phil_231 | jroll, no problem thanks. and g'morning :) | 12:53 |
| openstackgerrit | Merged openstack/ironic-inspector: Block broken diskimage-builder versions to unblock the gate https://review.openstack.org/269550 | 12:55 |
| dtantsur | \o/ | 12:55 |
| sambetts | \o/ | 12:56 |
| mgould | \o/ | 12:57 |
| * dtantsur rechecked all the things | 13:03 | |
| *** dprince has joined #openstack-ironic | 13:07 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-inspector: Revert "Block broken diskimage-builder versions to unblock the gate" https://review.openstack.org/269618 | 13:08 |
| dtantsur | this ^^ should be landed when DIB is fine again | 13:08 |
| *** daemontool_ has joined #openstack-ironic | 13:10 | |
| *** daemontool has quit IRC | 13:11 | |
| *** ekarlso has quit IRC | 13:18 | |
| *** ekarlso has joined #openstack-ironic | 13:18 | |
| *** shakamunyi has joined #openstack-ironic | 13:24 | |
| NobodyCam | good morning Ironicers says the man trying to wake up | 13:31 |
| NobodyCam | morning dtantsur lucasagomes jroll sambetts vdrok jlvillal mgould and all others not listed here | 13:32 |
| dtantsur | morning NobodyCam, it's useless to wake up today | 13:32 |
| * dtantsur is still unable to | 13:32 | |
| lucasagomes | NobodyCam, hello there! | 13:32 |
| sambetts | o/ NobodyCam | 13:32 |
| jroll | hiya NobodyCam | 13:32 |
| NobodyCam | morning :) | 13:32 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic-python-agent: Extend root device hints to support device name https://review.openstack.org/269629 | 13:39 |
| *** thrash|g0ne has quit IRC | 13:39 | |
| openstackgerrit | Merged openstack/ironic-inspector: Switch to IPA as a primary ramdisk https://review.openstack.org/263730 | 13:42 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ironic: Updated from global requirements https://review.openstack.org/268447 | 13:47 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ironic-inspector: Updated from global requirements https://review.openstack.org/268448 | 13:47 |
| TheJulia | Good morning everyone | 13:48 |
| *** daemontool__ has joined #openstack-ironic | 13:49 | |
| *** daemontool_ has quit IRC | 13:50 | |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic: Extend root device hints to support device name https://review.openstack.org/269639 | 13:50 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/python-ironicclient: Updated from global requirements https://review.openstack.org/269346 | 13:52 |
| *** daemontool__ is now known as daemontool | 13:54 | |
| dtantsur | morning TheJulia | 13:55 |
| *** Marga_ has joined #openstack-ironic | 13:57 | |
| *** phil_231 has quit IRC | 13:57 | |
| *** amotoki has quit IRC | 13:58 | |
| *** kalpase has quit IRC | 14:00 | |
| *** thrash has joined #openstack-ironic | 14:01 | |
| *** thrash has quit IRC | 14:01 | |
| *** thrash has joined #openstack-ironic | 14:01 | |
| *** kalpase has joined #openstack-ironic | 14:03 | |
| *** daemontool has quit IRC | 14:06 | |
| *** daemontool has joined #openstack-ironic | 14:07 | |
| lucasagomes | TheJulia, good morning | 14:16 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic-python-agent: Extend root device hints to support device name https://review.openstack.org/269629 | 14:16 |
| *** ukalifon has quit IRC | 14:16 | |
| *** lucasagomes is now known as lucas-afk | 14:16 | |
| *** sirius_ has joined #openstack-ironic | 14:19 | |
| *** baoli has joined #openstack-ironic | 14:23 | |
| *** harshs has joined #openstack-ironic | 14:23 | |
| *** cdearborn has joined #openstack-ironic | 14:26 | |
| devananda | g'morning, all! | 14:27 |
| *** kalpase has left #openstack-ironic | 14:28 | |
| *** baoli_ has joined #openstack-ironic | 14:28 | |
| *** harshs_ has joined #openstack-ironic | 14:28 | |
| *** daemontool_ has joined #openstack-ironic | 14:30 | |
| sambetts | o/ devananda | 14:31 |
| *** baoli has quit IRC | 14:31 | |
| *** daemontool has quit IRC | 14:31 | |
| *** harshs has quit IRC | 14:31 | |
| *** harshs_ is now known as harshs | 14:31 | |
| *** rloo has joined #openstack-ironic | 14:32 | |
| *** mgould has quit IRC | 14:33 | |
| *** daemontool_ has quit IRC | 14:34 | |
| *** daemontool_ has joined #openstack-ironic | 14:34 | |
| *** daemontool_ has quit IRC | 14:37 | |
| *** daemontool_ has joined #openstack-ironic | 14:37 | |
| *** teju has quit IRC | 14:38 | |
| *** baoli_ has quit IRC | 14:39 | |
| dtantsur | morning devananda | 14:39 |
| *** sambetts_ has joined #openstack-ironic | 14:41 | |
| vdrok | morning deray jroll NobodyCam TheJulia and devananda | 14:41 |
| deray | vdrok, g'morning :) | 14:42 |
| *** davideagnello has joined #openstack-ironic | 14:43 | |
| NobodyCam | morning :) | 14:44 |
| *** sambetts has quit IRC | 14:44 | |
| NobodyCam | o/ | 14:44 |
| *** mgould has joined #openstack-ironic | 14:45 | |
| *** davideagnello has quit IRC | 14:48 | |
| *** harshs has quit IRC | 14:49 | |
| sambetts_ | jroll: No idea if I've done this right or not: https://review.openstack.org/#/c/269684/ | 14:51 |
| openstackgerrit | Debayan Ray proposed openstack/proliantutils: common, ris, ribcl changes to support firmware update https://review.openstack.org/203543 | 14:53 |
| *** ionutbalutoiu has quit IRC | 14:54 | |
| *** ionutbalutoiu has joined #openstack-ironic | 14:56 | |
| jroll | sambetts_: reviewed, just a couple nits, mostly looks good | 14:56 |
| *** deray has quit IRC | 15:02 | |
| sambetts_ | jroll: awesome, I've not done a project-config patch that complex before | 15:02 |
| *** baoli has joined #openstack-ironic | 15:03 | |
| *** baoli has quit IRC | 15:03 | |
| jroll | sambetts_: yeah, our project-config stuff is annoying | 15:03 |
| *** baoli has joined #openstack-ironic | 15:04 | |
| *** jcoufal_ has joined #openstack-ironic | 15:06 | |
| *** baoli has quit IRC | 15:07 | |
| *** sambetts_ is now known as sambetts | 15:07 | |
| *** kan_ has quit IRC | 15:08 | |
| *** baoli has joined #openstack-ironic | 15:08 | |
| *** jcoufal has quit IRC | 15:09 | |
| *** daemontool_ has quit IRC | 15:12 | |
| *** jcoufal has joined #openstack-ironic | 15:13 | |
| *** jcoufal_ has quit IRC | 15:16 | |
| *** lucas-afk is now known as lucasagomes | 15:19 | |
| lucasagomes | devananda, morning | 15:19 |
| *** ijw has joined #openstack-ironic | 15:21 | |
| *** ijw has quit IRC | 15:21 | |
| *** ijw has joined #openstack-ironic | 15:21 | |
| *** daemontool_ has joined #openstack-ironic | 15:24 | |
| *** raddaoui has joined #openstack-ironic | 15:26 | |
| ifarkas | lucasagomes, o/ could you please review https://review.openstack.org/#/c/251294/ whenever you have time? | 15:28 |
| lucasagomes | ifarkas, hi there, totally. Will do it | 15:29 |
| ifarkas | lucasagomes, thanks! | 15:29 |
| *** achanda has joined #openstack-ironic | 15:29 | |
| *** sirius_ has quit IRC | 15:30 | |
| *** sirius_ has joined #openstack-ironic | 15:32 | |
| devananda | jroll: got side tracked yesterday, finishing up https://review.openstack.org/#/c/265311/5/jenkins/jobs/devstack-gate.yaml now | 15:34 |
| jroll | devananda: cool, ty | 15:34 |
| lucasagomes | this manual clean patch looks good https://review.openstack.org/#/c/247695/ the next one in the chain is already approved as well (tho it needs a rebase) | 15:39 |
| *** achanda has quit IRC | 15:40 | |
| devananda | jroll: hrm. do you know how to do a conditional on this line? tempest-env: 'DEVSTACK_GATE_TEMPEST_REGEX=ironic' | 15:41 |
| devananda | I could add a new template, but I think that's a change in a separate project | 15:41 |
| jroll | devananda: I do not | 15:41 |
| jroll | but | 15:41 |
| jroll | the template is in the same file | 15:41 |
| jroll | devstack-virtual-ironic | 15:41 |
| devananda | ah, I meant add a new variable that puppet would expand inline | 15:43 |
| devananda | {...} | 15:43 |
| *** sirius_ has quit IRC | 15:44 | |
| openstackgerrit | Merged openstack/python-ironic-inspector-client: Updated from global requirements https://review.openstack.org/268511 | 15:46 |
| rloo | hi and morning everyone, devananda, jroll, ifarkas, lucasagomes, NobodyCam, sambetts, vdrok, dtantsur :) | 15:46 |
| lucasagomes | rloo, hello there | 15:46 |
| rloo | lucasagomes: do you have a few minutes to discuss the split-capabilities spec? | 15:46 |
| lucasagomes | rloo, I will brb in ~10min | 15:46 |
| lucasagomes | but yeah | 15:46 |
| lucasagomes | rloo, I replied to the comments in the spec this morning | 15:46 |
| rloo | lucasagomes: ok, ping me when you're avail | 15:46 |
| *** mtanino has joined #openstack-ironic | 15:46 | |
| rloo | lucasagomes: yeah, but i still need clarification. i am missing something. | 15:47 |
| jroll | devananda: oh, you got it now? | 15:47 |
| devananda | jroll: think so | 15:47 |
| jroll | devananda: just add it in the template and the jobs inheriting the template and it'll just work | 15:48 |
| jroll | cool | 15:48 |
| lucasagomes | rloo, ok, I will ping you shortly then | 15:48 |
| *** raddaoui has quit IRC | 15:50 | |
| *** hemna has joined #openstack-ironic | 15:51 | |
| NobodyCam | good morning rloo :) | 15:53 |
| NobodyCam | and others who have joined | 15:54 |
| *** Marga_ has quit IRC | 15:54 | |
| *** sirius_ has joined #openstack-ironic | 15:58 | |
| *** bradjones has joined #openstack-ironic | 15:58 | |
| *** bradjones has quit IRC | 15:58 | |
| *** bradjones has joined #openstack-ironic | 15:58 | |
| *** achanda has joined #openstack-ironic | 15:59 | |
| *** baoli has quit IRC | 16:01 | |
| *** lucasagomes is now known as lucas-brb | 16:01 | |
| *** boris-42 has joined #openstack-ironic | 16:03 | |
| *** achanda has quit IRC | 16:03 | |
| sambetts | o/ rloo | 16:06 |
| dtantsur | morning rloo | 16:06 |
| *** lucascess has joined #openstack-ironic | 16:09 | |
| *** [1]cdearborn has joined #openstack-ironic | 16:19 | |
| *** praneshp has joined #openstack-ironic | 16:20 | |
| *** degorenko is now known as _degorenko|afk | 16:22 | |
| *** Nisha has joined #openstack-ironic | 16:22 | |
| vdrok | morning rloo ! | 16:23 |
| *** lucas-brb is now known as lucasagomes | 16:28 | |
| lucasagomes | rloo, back | 16:28 |
| rloo | hi lucasagomes | 16:28 |
| lucasagomes | hi there | 16:28 |
| rloo | lucasagomes: i think there are two parts to the capabilities split. | 16:28 |
| rloo | lucasagomes: 1. moving to a table | 16:29 |
| lucasagomes | rloo, yeah, creating a new table for the capabilities | 16:29 |
| rloo | lucasagomes: 2. properties['capabilities'] vs node.capabilities | 16:29 |
| lucasagomes | yeah, which is the backward compat bits | 16:29 |
| rloo | lucasagomes: so moving to a new db table has nothing to do with microversioning, right? | 16:29 |
| rloo | lucasagomes: you'd take the value from node.properties['capabilities'] and use that to populate the db table | 16:30 |
| rloo | lucasagomes: once you populate the db table, will we still be saving node.properties['capabilities'] string in the node table? | 16:30 |
| lucasagomes | rloo, not necessarily, my idea was to have a node.capabilites field (in the API object) so it will get exposed to the API | 16:30 |
| lucasagomes | and the version will bump | 16:30 |
| rloo | lucasagomes: i think that's where i am confused. | 16:31 |
| rloo | lucasagomes: so you're saying that the db table will be done at the same time as the node.capabilities field. | 16:31 |
| lucasagomes | yeah | 16:31 |
| *** alexpilo_ has joined #openstack-ironic | 16:31 | |
| rloo | lucasagomes: but what if the user doesn't do a microversion | 16:31 |
| lucasagomes | that's what the spec is proposing | 16:31 |
| rloo | lucasagomes: sorry, i have a question before the one i just asked. to go back to what you're proposing then., | 16:32 |
| lucasagomes | rloo, so that's one reason why I want to keep both path working properties/capabilities and capabilities | 16:32 |
| rloo | lucasagomes: create db table, add a node.capabilities field. do you save node.capabilities value in the node table? | 16:32 |
| lucasagomes | the only consumer of these paths is nova | 16:32 |
| lucasagomes | and nova pins at a specific version of the API | 16:32 |
| *** cdearborn has quit IRC | 16:32 | |
| lucasagomes | rloo, yeah the node.capabilities field is a proxy to the new table | 16:33 |
| lucasagomes | it will be exposed in the API as a dict-like field | 16:33 |
| rloo | lucasagomes: so the nodes db table will not have a capabilities entry, right? | 16:33 |
| lucasagomes | maybe if I put a patch up for it, it will be cleanear about how it works? | 16:33 |
| lucasagomes | rloo, no | 16:34 |
| lucasagomes | the nodes db table won't have any changes | 16:34 |
| rloo | lucasagomes: i don't really want to read through code. i think the spec should be descriptive enough. (I think anyway) | 16:34 |
| *** alexpilotti has quit IRC | 16:34 | |
| lucasagomes | rloo, a example of how it would look like http://docs.sqlalchemy.org/en/latest/orm/extensions/associationproxy.html#proxying-to-dictionary-based-collections | 16:35 |
| rloo | lucasagomes: so you create a capabilities db table that is populated with values from node.properties['capabilities'], and at API/microversion+, you show node.capabilities which is derived from capabilities table | 16:35 |
| *** baoli has joined #openstack-ironic | 16:35 | |
| rloo | lucasagomes: and you will leave node.properties['capabilities'] value in the node DB? | 16:35 |
| lucasagomes | rloo, I think the confusion here is about migrating the data, I won't populate the table with the values from node.properties/capabilities | 16:36 |
| rloo | lucasagomes: you won't? | 16:36 |
| rloo | lucasagomes: yup, i'm confused then. | 16:37 |
| lucasagomes | rloo, no, so what I'm proposing is giving time to operators to move the data themself | 16:37 |
| lucasagomes | we can even give them a script, it will be few CLI commands to migrate the data | 16:37 |
| lucasagomes | the data they have in node.properties/capabilities will stay there after the ironic upgrade | 16:37 |
| lucasagomes | and it will continue to work as-is today | 16:38 |
| lucasagomes | nova will continue to look at node.properties/capabilities if the node.capabilities is empty (as a fallback mechanism) | 16:38 |
| lucasagomes | but it will log a wanrning to inform operators that it will be removed in the future and they should migrate it to the new field | 16:38 |
| rloo | lucasagomes: i don't like that. it means we'll always have to maintain two or more code paths wrt properties.capabilities | 16:38 |
| lucasagomes | like a configuration option | 16:38 |
| lucasagomes | rloo, just for a period of time | 16:39 |
| lucasagomes | 1 or 2 release cycles | 16:39 |
| rloo | lucasagomes: i think there should just be one source of trush. | 16:39 |
| lucasagomes | the code is minimal | 16:39 |
| lucasagomes | rloo, there's a function in nova that parses the capabilities | 16:39 |
| rloo | lucasagomes: if you give the user the choice to migrate the data whenever they want, they we don't know if they will ever migrate. | 16:39 |
| rloo | lucasagomes: i don't think we can *only* think about nova when adding this feature. | 16:39 |
| lucasagomes | rloo, nova is the only consumer of that path | 16:40 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic-specs: Add pluggable credentials storage https://review.openstack.org/186056 | 16:40 |
| jroll | how do you know? | 16:40 |
| lucasagomes | jroll, who else uses it? it's a nova filter thing | 16:40 |
| jroll | lucasagomes: any api user? | 16:40 |
| jroll | what about bifrost? | 16:40 |
| lucasagomes | jroll, that will continue to work right? | 16:41 |
| jroll | what about the scheduler I wrote myself because I don't use nova? | 16:41 |
| lucasagomes | the API will still support adding a field called capability in properties | 16:41 |
| lucasagomes | jroll, all that will continue to work | 16:41 |
| jroll | will it still support reading from that? | 16:41 |
| lucasagomes | bifrost is standalone, so it uses instance_info['capabilities'] | 16:41 |
| lucasagomes | not properties['capabilities'] | 16:41 |
| lucasagomes | jroll, sure | 16:42 |
| lucasagomes | jroll, properties is a JSON field, people can add any key/value pair there | 16:42 |
| lucasagomes | including one called capabilities | 16:42 |
| jroll | well | 16:42 |
| jroll | after the data migration, I'll no longer be able to access the data I put in node.properties[capabilities] | 16:42 |
| jroll | so it's breaking clients, right? | 16:43 |
| lucasagomes | but that's the thing, operators will migrate the data | 16:43 |
| lucasagomes | then can very well leave it in both places | 16:43 |
| lucasagomes | until they update their script/systems | 16:43 |
| jroll | hmm | 16:43 |
| rloo | whoa. that is SO not what I thought the spec was proposing. | 16:44 |
| jroll | still an API break, even if it's across microversions :/ | 16:44 |
| *** Marga_ has joined #openstack-ironic | 16:44 | |
| lucasagomes | jroll, why? | 16:44 |
| jroll | I'd almost rather not change the api for the first iteration | 16:44 |
| jroll | because the data is moving | 16:44 |
| lucasagomes | there's no data moving in the spec | 16:44 |
| jroll | in the api request/response | 16:44 |
| *** hemna has quit IRC | 16:45 | |
| lucasagomes | unless we decide to move the data, but I left that out and proposed to keep things working as-is | 16:45 |
| lucasagomes | for some time | 16:45 |
| *** mgoddard_ has joined #openstack-ironic | 16:46 | |
| jroll | oh, so the spec is just an api change, and not moving data? | 16:46 |
| *** praneshp_ has joined #openstack-ironic | 16:46 | |
| lucasagomes | jroll, yes just change the API | 16:46 |
| rloo | it can't be just an API change. you're adding a new db table. | 16:46 |
| jroll | right, so the API is breaking no? | 16:46 |
| devananda | lots of scrollback ... /me reads ... | 16:46 |
| jroll | I gotta step away for 15 | 16:46 |
| krotscheck | betherly: I has a devstack | 16:46 |
| lucasagomes | jroll, and change nova to look at the new API field (capabilities) and in case it's empty (because operator didn't move the data yet) fall back and look at properties/capabilities as-is today | 16:46 |
| krotscheck | All I need is the plugin :) | 16:47 |
| jroll | yeah, lemme think on it | 16:47 |
| lucasagomes | there's no breakage, the only thing we have to do is inform operators that the fallback mechanism in the nova driver (to look at propeties/capabilities) will be removed in the future | 16:47 |
| devananda | lucasagomes: there are definitely more consumers of the node.properties['capabilities'] resource than Nova | 16:47 |
| lucasagomes | so we are loging warning messages for it | 16:48 |
| lucasagomes | devananda, right, but we are not removing it at all | 16:48 |
| devananda | lucasagomes: this is why I asked yesterday if you can do these separately | 16:48 |
| lucasagomes | that's the thing, the properties['capabilities'] is like any other ordinary key/value pair in the JSON field | 16:48 |
| lucasagomes | devananda, I see, but the spec is not moving anything around | 16:48 |
| *** praneshp has quit IRC | 16:49 | |
| *** praneshp_ is now known as praneshp | 16:49 | |
| *** mgoddard__ has quit IRC | 16:49 | |
| lucasagomes | after the upgrade the API will look pretty much same, the only diff is that the node will have a new field called capabilities | 16:49 |
| devananda | I would also rather change the DB side first, with no API change. then do the API change. | 16:49 |
| devananda | lucasagomes: that's not "pretty much the same" :p | 16:49 |
| lucasagomes | devananda, microversion for the new field | 16:49 |
| devananda | lucasagomes: what happens to the data currently in node.properties? | 16:50 |
| rloo | devananda: ++, that's actually what i thought the spec was proposing. | 16:50 |
| lucasagomes | devananda, nothing | 16:50 |
| lucasagomes | devananda, it stays there | 16:50 |
| devananda | uuuuhh | 16:50 |
| devananda | what if an old client is still storing it there? | 16:50 |
| devananda | and a new client is trying to _also_ store the same type of data? | 16:50 |
| devananda | do they simply not see each other's data, because they're storing it in different places? | 16:50 |
| lucasagomes | devananda, the old client will continue to succeed to store anything there | 16:50 |
| lucasagomes | devananda, new client will store data there too | 16:51 |
| lucasagomes | it's a blob | 16:51 |
| lucasagomes | the migration bits I'm thinking of is like a configuration option | 16:51 |
| devananda | oh - you're making an API change to allow access to the internal JSON via a new REST resource, without any DB change? | 16:51 |
| lucasagomes | you change the name of the thing... the old name still works | 16:51 |
| lucasagomes | devananda, no the API change will just add a new field to the node's object called "capabilities" that's it | 16:52 |
| devananda | lucasagomes: so there will be two different places to store the same data | 16:52 |
| lucasagomes | then the nova driver will be updated (and pinned at that version of the API) | 16:52 |
| lucasagomes | devananda, yes | 16:52 |
| devananda | and some clients will write it at (A) and some will write it at (B) ? | 16:52 |
| devananda | that's definitely not what I had in mind | 16:52 |
| lucasagomes | devananda, but the old place won't be used by nova in the future | 16:52 |
| lucasagomes | after deprecation | 16:53 |
| lucasagomes | devananda, for some time, writing to A or B will work | 16:53 |
| devananda | "will work" | 16:53 |
| devananda | oops | 16:53 |
| lucasagomes | but we will log warning messages saying "look this will be depracated, store it in A please" | 16:53 |
| devananda | "will work" only if all clients are writing to the same location | 16:53 |
| rloo | lucasagomes: it isn't about nova only. writing to A or B won't work, you'd have to write to both to make sure they have the same values. | 16:53 |
| devananda | but we can't guarantee that all clients will do that | 16:53 |
| *** rcernin has quit IRC | 16:53 | |
| devananda | and logging warnings is useless because this isn't something the operator controls -- it's up to the user | 16:54 |
| lucasagomes | I think there's a confusion | 16:54 |
| lucasagomes | properties['capabilities'] in the standard ironic overflow is *only* used by nova | 16:54 |
| lucasagomes | right> | 16:54 |
| lucasagomes | ?* | 16:54 |
| devananda | nope | 16:54 |
| lucasagomes | who else uses it? | 16:54 |
| devananda | downstream tools | 16:55 |
| devananda | drivers | 16:55 |
| *** raddaoui has joined #openstack-ironic | 16:55 | |
| lucasagomes | what they do with it? | 16:55 |
| dtantsur | lucasagomes, tripleo does a lot with capabilities | 16:55 |
| devananda | there are in-tree drivers that key their behavior off of certain capabilities | 16:55 |
| dtantsur | just FYI | 16:55 |
| lucasagomes | devananda, that's instance_info['capabilities'] | 16:56 |
| devananda | downstream tools inject data into ironic by writing to that resource, expecting certain results | 16:56 |
| lucasagomes | the way capabilities works is | 16:56 |
| lucasagomes | nova look at propeties['capabilities'] and match it with the flavor capabiltiies | 16:56 |
| lucasagomes | say node has capabilties A,B,C and the flavor has A,B (not C) | 16:56 |
| devananda | lucasagomes: inspector writes to it, too | 16:56 |
| devananda | lucasagomes: and some tool (inspector, vendor specific, bifrost, ....) has to insert the data there | 16:57 |
| *** hemna has joined #openstack-ironic | 16:57 | |
| devananda | so even if nova were the only consumer of that resource, other clients have t interact with it | 16:57 |
| dtantsur | yes, python-tripleoclient and inspector definitely mess with capabilities | 16:57 |
| lucasagomes | devananda, right, and that will work | 16:57 |
| *** vishwanathj has joined #openstack-ironic | 16:57 | |
| lucasagomes | it will go to a deprecation process | 16:57 |
| devananda | lucasagomes: it will not work if the clients are out of sync with each other IF there are two internal locations where this data is stored | 16:58 |
| *** david-lyle has quit IRC | 16:58 | |
| *** linuxaddicts has quit IRC | 16:59 | |
| *** mgoddard_ has quit IRC | 17:00 | |
| lucasagomes | devananda, hmm yeah if one client writes it to node.capabilities to use the new model, and another to node.properties['capabilities'] (and nova is updated) yes | 17:00 |
| *** mgoddard has joined #openstack-ironic | 17:00 | |
| lucasagomes | nova will only care about node.capabilities in that spec | 17:00 |
| devananda | right | 17:01 |
| devananda | that's my concern | 17:01 |
| devananda | we need to do a DB migration to move the data into the new place and drop the old place | 17:01 |
| *** e0ne has quit IRC | 17:01 | |
| devananda | so that there is only one location to store capabilities internally | 17:01 |
| *** linuxaddicts has joined #openstack-ironic | 17:02 | |
| devananda | supporting both paths in the API is all well and good | 17:02 |
| jroll | so, let's back up a bit | 17:02 |
| jroll | what's the real goal here? | 17:02 |
| jroll | getting capabilities out of properties, right? | 17:03 |
| jroll | in the database | 17:03 |
| lucasagomes | devananda, right, it's hard cause properties is a JSON | 17:03 |
| devananda | getting all properties (including capabilities) into a db structure that can be indexed and searched | 17:03 |
| devananda | that is my goal here | 17:03 |
| lucasagomes | we can't forbid someone to create a key called "capabilities" can we? | 17:03 |
| jroll | right, so | 17:03 |
| jroll | let's not change the api. | 17:03 |
| jroll | we create the capabilities table | 17:03 |
| jroll | migrate data | 17:03 |
| jroll | at the objects layer, transform node.properties[capabilities] back and forth to the new table | 17:04 |
| lucasagomes | right, and keep using the properties/capabilities ? | 17:04 |
| jroll | the api doesn't change, the objects api doesn't change | 17:04 |
| lucasagomes | that works as well | 17:04 |
| jroll | right, for now at least | 17:04 |
| jroll | the api change is harder to get right, so I'd like to start with fixing the db | 17:04 |
| lucasagomes | ok yeah that will work | 17:05 |
| jroll | cool, yeah I was confused since I thought that's what we agreed on yesterday | 17:05 |
| lucasagomes | one thing tho, is that capabilities is a string thing... so it's hard to manage | 17:05 |
| lucasagomes | and is racy | 17:06 |
| *** david-lyle has joined #openstack-ironic | 17:06 | |
| jroll | as in, if two clients update at once? | 17:06 |
| lucasagomes | to update clients have to fetch, remove or add things in the string and write it back | 17:06 |
| lucasagomes | yeah | 17:06 |
| lucasagomes | every update affects all capabilities | 17:06 |
| jroll | yeah :( | 17:06 |
| lucasagomes | that's why I wanted to have a dict-like field for that | 17:06 |
| lucasagomes | so we support partial updates, as is today | 17:06 |
| jroll | that's already a problem today | 17:07 |
| lucasagomes | that can be done even in properties/capabilities FWIW | 17:07 |
| jroll | and we should solve it, but one thing at a time | 17:07 |
| lucasagomes | yeah | 17:07 |
| lucasagomes | jroll, but I still wonder how updates will work... today we simple dump the string in the db | 17:07 |
| lucasagomes | with the new table we have to look at any updates at propeties, check each capabilities to see if it's in the table | 17:07 |
| jroll | right, we'll need to parse it and update the table | 17:08 |
| lucasagomes | remove from the table if it was removed from the string etc... | 17:08 |
| * jroll wonders if sqlalchemy has upsert | 17:08 | |
| lucasagomes | yeah | 17:08 |
| jroll | oh, the delete is odd | 17:08 |
| jroll | FUN | 17:08 |
| lucasagomes | yes | 17:08 |
| lucasagomes | slow too | 17:08 |
| jroll | yeah :( | 17:08 |
| *** rcernin has joined #openstack-ironic | 17:09 | |
| lucasagomes | it sounds simple, but when you start thinking about it it's complex | 17:09 |
| lucasagomes | that's why I had the idea to keep both places and deprecate one | 17:09 |
| lucasagomes | with time | 17:09 |
| lucasagomes | to reduce the complexity of things | 17:09 |
| *** dtantsur is now known as dtantsur|afk | 17:09 | |
| jroll | yeah, but api deprecations are much sadface | 17:10 |
| devananda | API deprecation bad | 17:10 |
| rloo | i think it is worth doing the new db + new API with capabilities avail at the same time. due to ^slow. once data is migrated, if user is happy/fast, they should be able to use new API to get away from slowness. | 17:10 |
| lucasagomes | yeah it's not great... but in reality we can never deprecate someone from inputing in properties | 17:11 |
| devananda | we can add things no problem. but just look at ENROLL state | 17:11 |
| rloo | i don't think we want to deprecate properties['capabilities']. i think it has to be via microvesion | 17:11 |
| lucasagomes | cause it's blob... so there's no way to forbit such update | 17:11 |
| devananda | we can't break compat with older clients without *seriously* good reasons | 17:11 |
| devananda | and I do not think this is a good enough reason | 17:11 |
| jroll | +1 | 17:12 |
| lucasagomes | but think... clients add such capabilities there so nova will use it right? | 17:12 |
| devananda | even with API version headers, this is going to be challenging -- we're going to have some older clients who still access this at the current location FOR YEARS | 17:12 |
| devananda | because downstream vendor products | 17:12 |
| jroll | this is all too hard, let's go fishing | 17:12 |
| devananda | jroll: ++ | 17:12 |
| lucasagomes | heh | 17:13 |
| devananda | the pier is right by my place. come up for a drink after you catch a fish? | 17:13 |
| lucasagomes | my thinking about all this APIs stuff is that operators should be in control about what version it should use | 17:13 |
| lucasagomes | leaning curve | 17:13 |
| lucasagomes | if nova had a configuration option to set the API version as proposed before it would be easier | 17:13 |
| devananda | lucasagomes: I do not understand what you mean "operators should be in control" | 17:13 |
| devananda | in larger clouds, the team running nova != the team running ironic != the team running horizon != .... | 17:14 |
| lucasagomes | devananda, I mean the thing about pinning at a specific version of the API | 17:14 |
| devananda | I agree that nova.virt.ironic should be pinned to a specific version of Ironic's REST API. I do not think that should be configurable. | 17:15 |
| devananda | it should be a CONST in that driver. | 17:15 |
| lucasagomes | that makes upgrades difficult | 17:15 |
| devananda | not really | 17:15 |
| lucasagomes | cause you always have to upgrade ironic first | 17:15 |
| lucasagomes | if you could configure it in nova, you can update anything at any order | 17:15 |
| *** pas-ha has quit IRC | 17:15 | |
| devananda | lucasagomes: right. which is the standard thing to do -- upgrade the backend to add support for newer versions,then update the front end (client) to use those | 17:15 |
| *** ifarkas has quit IRC | 17:16 | |
| devananda | it's MORE difficult to say "do this in any order ... here are separate manuals if you do A,B or B,A or both at once" | 17:16 |
| jroll | I do think it makes upgrading ironic and nova separately in CD harder, but let's not get into that :) | 17:16 |
| devananda | jroll: bait avoided :) | 17:16 |
| devananda | back to topic of properties and capabilities | 17:17 |
| * jroll will let you know when it breaks :P | 17:17 | |
| lucasagomes | heh ok let's avoid it | 17:17 |
| devananda | my original proposal was to convert properties from JSON to key-value so we can index it | 17:17 |
| devananda | that shouldn't affect anything in the API, except for max key length (if we limit the keys to 255) | 17:17 |
| lucasagomes | right | 17:18 |
| lucasagomes | and 255 is too short for capabalities and root device hints | 17:19 |
| devananda | lucasagomes noticed that prefix indexes aren't supported in some db backends, so, eg, we couldn't define an index on (key(50), value(50)) to optimize for the case where someone decides to store a gigantic value in there and we don't want to bloat the index | 17:19 |
| devananda | that's only supported by mysql -- which, frankly, I'm fine with saying "use MySQL or suffer performance penalties" | 17:19 |
| devananda | we can store the `value` still in a longer field and index only the prefix | 17:20 |
| devananda | and thus support capabilities and root device hints | 17:20 |
| devananda | though ... I would also like to index capabilities separately, that could be done as a follow-on | 17:20 |
| lucasagomes | yes, because capabilities and root device hints even if indexable, they are not queryable... capabilities looks like "cap1:value1,cap2:value2" and root device hints in a nested json | 17:21 |
| devananda | right | 17:22 |
| lucasagomes | (that's a great summary thanks devananda) | 17:22 |
| devananda | I'm fine with that for now | 17:22 |
| devananda | lucasagomes: yvw :) | 17:22 |
| devananda | as long as we don't truncate capabilities or r_d_h and lose data ... those are not likely to be searched | 17:22 |
| devananda | and all of this internal refactoring is really meant to pave the way to enabling a search API | 17:22 |
| devananda | so I could query Ironic and ask for nodes that have 128GB RAM and 500GB of SSD in a RAID array | 17:23 |
| devananda | *that's* the goal | 17:23 |
| devananda | s/the/my/ | 17:23 |
| jroll | with the nova driver refactor, capabilities is very likely to be searched | 17:23 |
| jroll | it's part of scheduling | 17:23 |
| devananda | ahh. good point | 17:23 |
| *** derekh has quit IRC | 17:23 | |
| devananda | so yes, we will need to do that too. | 17:23 |
| jroll | right, so we do need to break capabilities out | 17:24 |
| jroll | which is why I think we should do the db migrations now | 17:24 |
| lucasagomes | devananda, so there's the idea #2 as well | 17:24 |
| jroll | but leave the api as-is | 17:24 |
| lucasagomes | which was split out capabilities and root device hints | 17:24 |
| jroll | because that's a large break | 17:24 |
| lucasagomes | then make properties values max length of 255 and make it indexable | 17:24 |
| lucasagomes | that's when that spec I put up got created | 17:25 |
| lucasagomes | so it would work in all database backends | 17:25 |
| lucasagomes | I mean, be perfomant on all backends (#1 works, but is slow when != mysql | 17:26 |
| lucasagomes | )* | 17:26 |
| * devananda is fine with not-mysql being slow | 17:26 | |
| jroll | #1 doesn't solve the problem of "capabilities needs to be indexable" | 17:26 |
| lucasagomes | yeah those are following work | 17:27 |
| jroll | so the opposition to "break capabilities into separate table but don't change api" is what exactly? that the api will be a little slow? | 17:27 |
| lucasagomes | complex | 17:27 |
| lucasagomes | and slow | 17:28 |
| jroll | the code will be complex, to be clear | 17:28 |
| jroll | and the api will be slow | 17:28 |
| jroll | but I'm not sure how slow | 17:28 |
| lucasagomes | yes | 17:28 |
| * lucasagomes doesn't have data | 17:28 | |
| *** rloo has quit IRC | 17:28 | |
| jroll | keys = _parse_capabilities().keys(); delete from capabilities where node_id=foo and key not in ($keys); for k, v in capabilities: upsert into capabilities... | 17:29 |
| jroll | shouldn't be that bad really | 17:29 |
| lucasagomes | yeah | 17:29 |
| lucasagomes | but just that any updates to properties even if do not affect capabilities needs to be split out | 17:29 |
| jroll | node.update goes over rpc anyway, if we cared about speed we'd worry about that first :) | 17:29 |
| lucasagomes | and checked to see if all keys are in the capabilities table | 17:29 |
| lucasagomes | and their values are the same | 17:30 |
| jroll | that's what upsert does | 17:30 |
| jroll | and updates to properties for keys that are not in capabilities don't need to be checked against capabilities | 17:30 |
| lucasagomes | right, re-upsert we need to check if the ORM supports it | 17:31 |
| * lucasagomes looks | 17:31 | |
| jroll | yeah it actually looks like not | 17:32 |
| jroll | although it's unclear if session.add() handles it | 17:32 |
| lucasagomes | yeah need some tests | 17:32 |
| devananda | insert ... on duplicate key update; | 17:34 |
| devananda | if sqla does not support that.... | 17:34 |
| jroll | you'd be surprised :) | 17:36 |
| devananda | have I said lately how much I dislike ORMs? | 17:36 |
| lucasagomes | devananda, I'm feeling that way at the moment too | 17:36 |
| *** davideagnello has joined #openstack-ironic | 17:37 | |
| *** Marga_ has quit IRC | 17:37 | |
| *** Marga_ has joined #openstack-ironic | 17:38 | |
| lucasagomes | [off-topic] dtantsur|afk http://logs.openstack.org/29/269629/3/check/gate-tempest-dsvm-ironic-agent_ssh-src/b0abb06/logs/devstacklog.txt.gz#_2016-01-19_16_27_29_334 | 17:38 |
| lucasagomes | [off-topic] wans't that fixed already? | 17:38 |
| * lucasagomes check the devstack patch | 17:39 | |
| jroll | it was in recheck as of this morning | 17:39 |
| lucasagomes | oh still merging | 17:39 |
| devananda | lucasagomes: even if suboptimal, sticking with the ORM, we can do this in a session | 17:39 |
| devananda | I'm waiting on that to merge before rechecking the tempestlib changes | 17:39 |
| lucasagomes | yeah that was a new patch I put up and I saw it failed | 17:40 |
| lucasagomes | but I've seem the approval from yesterday, so I thought it was merged already | 17:40 |
| lucasagomes | approval in the devstack fix I mean* | 17:40 |
| devananda | jroll: is there anything on https://review.openstack.org/#/c/253982/17 that needs to be changed? | 17:40 |
| jroll | devananda: not that I know of, but I haven't looked since last week | 17:41 |
| devananda | I see yuiko W-1'd it, but that looks like just a chicken-and-egg with the project-config change | 17:41 |
| jroll | yeah | 17:41 |
| jroll | because you did it before, I assume :P | 17:41 |
| devananda | heh | 17:41 |
| openstackgerrit | Sam Betts proposed openstack/ironic: Enable tinyipa for devstack Ironic https://review.openstack.org/259089 | 17:42 |
| * jroll +2 | 17:43 | |
| devananda | jroll: if infra lands the project-config change, it'll break our gate right now, but then we unbreak it by landing hte tempestlib stuff -- and that ensures it works when we land it | 17:43 |
| sambetts | jroll: all the patches required to get the -nv job working are up now, and just needs reviewing | 17:43 |
| JayF | jroll: I know you've been commenting about the gate and how it's been slow recently -- I've not gotten both -src jobs to pass together in IPA for a long time | 17:43 |
| jroll | devananda: that is scary | 17:44 |
| jroll | JayF: story of my life | 17:44 |
| JayF | jroll: so even if tinyipa fixes Ironic's gate, we still have effort for ipa | 17:44 |
| devananda | jroll: the other order (land tempestlib first, then infra switches us to it) results in a similar risk | 17:44 |
| jroll | JayF: I understand | 17:44 |
| jroll | devananda: sure | 17:44 |
| * sambetts heads off for the day | 17:44 | |
| JayF | jroll: I would offer to help, but have no time for $reasons as you're aware | 17:44 |
| jroll | JayF: just like everyone else :| | 17:45 |
| JayF | yep | 17:45 |
| devananda | JayF: how much of the ipa gate slowness is just timeouts? | 17:45 |
| jroll | all | 17:45 |
| devananda | thought so | 17:45 |
| devananda | then tinyipa will help | 17:45 |
| *** ChubYann has joined #openstack-ironic | 17:45 | |
| JayF | Not for the ipa-src job | 17:46 |
| JayF | will it? | 17:46 |
| jroll | IFF we change the IPA src jobs to use it | 17:46 |
| JayF | ahh. Yeah. We could do that I guess. Sorta starting to get in the "should we test >1 ramdisk" territory. | 17:46 |
| devananda | I was looking at our gate pass/fail % over the last 8mo and noticed that the old pxe jobs were A LOT more reliable | 17:46 |
| devananda | 80% pass vs. 98% pass | 17:46 |
| JayF | if we did change that ramdisk in ipa src jobs to be tinyipa, we'd have to readd the coreos builder to the gate | 17:46 |
| jroll | devananda: I also suspect some of the newish nodepool providers are slower | 17:47 |
| devananda | yah | 17:47 |
| jroll | devananda: fail rate on ovh seems high, for example | 17:47 |
| devananda | i know other things have changed, and we don't have a side-by-side | 17:47 |
| lucasagomes | JayF, yeah, I think we could add a -src job for coreos but with a limited scope | 17:48 |
| lucasagomes | JayF, just build the ramdisk and deploy (to ensure it boots) | 17:48 |
| lucasagomes | we don't need to run cleaning (so it won;t boot again) nor rebuild or antyhing like that | 17:48 |
| JayF | We already have a job that just does the build, that since we added tempest for IPA only runs in post | 17:48 |
| *** raddaoui has quit IRC | 17:48 | |
| *** jistr has quit IRC | 17:48 | |
| lucasagomes | JayF, cool, yeah would be good to see if it actually boots on the machine and IPA starts | 17:49 |
| lucasagomes | to ensure the service files are correct etc | 17:49 |
| jroll | JayF: might be ^ | 17:49 |
| jroll | yeah what lucas said | 17:49 |
| JayF | yep I agree | 17:49 |
| * devananda afks for food | 17:49 | |
| jroll | ditto, starving | 17:50 |
| sambetts | night all o/ | 17:52 |
| devananda | g'night, sambetts ! | 17:52 |
| *** sambetts has quit IRC | 17:52 | |
| NobodyCam | night sambetts | 17:55 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Add proxy related parameters to agent driver https://review.openstack.org/254296 | 18:00 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Add documentation for proxies usage with IPA https://review.openstack.org/250878 | 18:00 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Add ability to cache swift temporary URLs https://review.openstack.org/254295 | 18:00 |
| *** mkovacik has quit IRC | 18:08 | |
| openstackgerrit | Merged openstack/bifrost: Add inventory_dns feature to bifrost https://review.openstack.org/265630 | 18:09 |
| openstackgerrit | Merged openstack/python-ironic-inspector-client: Implement 'introspection data save' command https://review.openstack.org/266409 | 18:10 |
| NobodyCam | wow https://review.openstack.org/#/c/251995 was approved on 1/15 and is still in waiting to land | 18:16 |
| lucasagomes | NobodyCam, it has a dependent patch on it | 18:16 |
| lucasagomes | I mean it depends on a patch* | 18:16 |
| *** e0ne has joined #openstack-ironic | 18:16 | |
| lucasagomes | this one https://review.openstack.org/#/c/247695 | 18:16 |
| * lucasagomes +2'd today | 18:16 | |
| *** praneshp has quit IRC | 18:16 | |
| NobodyCam | gah..TY lucasagomes | 18:17 |
| NobodyCam | I still am not use to the new layout | 18:17 |
| lucasagomes | it lgtm btw, we should review approve it because manual cleaning is priority | 18:17 |
| lucasagomes | NobodyCam, yeah I hate it | 18:17 |
| lucasagomes | I mean, parts of it | 18:17 |
| lucasagomes | the search annoys me every single time | 18:17 |
| NobodyCam | :) +++ | 18:17 |
| lucasagomes | the CTRL+F bind for that slow-ass search is terrible | 18:17 |
| lucasagomes | find another bind... | 18:17 |
| NobodyCam | they seem to have lost the whole Depends on chain | 18:18 |
| *** awiddersheim has quit IRC | 18:18 | |
| lucasagomes | ouch | 18:18 |
| *** awiddersheim has joined #openstack-ironic | 18:19 | |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Add more unit tests for NO_PROXY validation https://review.openstack.org/269800 | 18:21 |
| *** piet has joined #openstack-ironic | 18:21 | |
| *** rloo has joined #openstack-ironic | 18:25 | |
| *** spandhe has joined #openstack-ironic | 18:25 | |
| mjturek1 | hey I have a quick question. When the migration to tempest plugins lands (https://review.openstack.org/#/c/253982/) will config options still live in tempest.conf? | 18:28 |
| *** baoli has quit IRC | 18:29 | |
| devananda | mjturek1: I do not think so. see http://docs.openstack.org/developer/tempest/plugin.html#plugin-structure | 18:29 |
| *** trown is now known as trown|lunch | 18:30 | |
| *** mgould has quit IRC | 18:32 | |
| mjturek1 | devanada might be missing something, but I'm talking about the actual key value pairs (ie active_timeout=500 in tempest.conf), not the definitions of each option (config.py) | 18:32 |
| *** Nisha has quit IRC | 18:33 | |
| mjturek1 | it does say below "an external plugin can rely on using any configuration option coming from Tempest, there will be at least a full deprecation cycle for any option before it’s removed" is that what ironic is doing? | 18:34 |
| *** kozhukalov has quit IRC | 18:34 | |
| devananda | mjturek1: "When adding configuration options the register_opts method gets passed the CONF object from tempest." | 18:34 |
| *** romcheg has quit IRC | 18:34 | |
| *** Ng has quit IRC | 18:34 | |
| *** kozhukalov has joined #openstack-ironic | 18:34 | |
| devananda | mjturek1: so I believe the config file is global, as there is one CONF object, which is passed to each plugin | 18:34 |
| *** Ng_ has joined #openstack-ironic | 18:34 | |
| mjturek1 | ahhh okay | 18:35 |
| *** sergek has quit IRC | 18:35 | |
| devananda | mjturek1: so I misspoke initially. I believe the config will continue to live in tempest.conf | 18:35 |
| mjturek1 | devananda: very cool, thanks! | 18:35 |
| *** romcheg has joined #openstack-ironic | 18:35 | |
| *** sergek has joined #openstack-ironic | 18:37 | |
| *** mgoddard_ has joined #openstack-ironic | 18:39 | |
| *** mgoddard has quit IRC | 18:42 | |
| *** vishwanathj has quit IRC | 18:47 | |
| *** boris-42 has quit IRC | 18:53 | |
| *** mkovacik has joined #openstack-ironic | 19:00 | |
| openstackgerrit | Merged openstack/ironic-inspector: Updated from global requirements https://review.openstack.org/268448 | 19:00 |
| *** alex_xu has quit IRC | 19:04 | |
| *** alex_xu has joined #openstack-ironic | 19:07 | |
| *** ionutbalutoiu has quit IRC | 19:13 | |
| *** ionutbalutoiu has joined #openstack-ironic | 19:13 | |
| *** alexpilotti has joined #openstack-ironic | 19:15 | |
| *** raddaoui has joined #openstack-ironic | 19:15 | |
| *** alexpilo_ has quit IRC | 19:17 | |
| *** jcoufal has quit IRC | 19:19 | |
| *** alexpilotti has quit IRC | 19:21 | |
| *** absubram has joined #openstack-ironic | 19:31 | |
| *** trown|lunch is now known as trown | 19:31 | |
| *** Ng_ is now known as Ng | 19:32 | |
| *** piet has quit IRC | 19:33 | |
| *** athomas has quit IRC | 19:34 | |
| *** harshs has joined #openstack-ironic | 19:36 | |
| *** athomas has joined #openstack-ironic | 19:45 | |
| *** mbound has quit IRC | 19:46 | |
| *** krtaylor has quit IRC | 19:49 | |
| *** piet has joined #openstack-ironic | 19:52 | |
| *** sirius_ has quit IRC | 19:58 | |
| lucasagomes | folks I'm going call it a day | 20:01 |
| lucasagomes | devananda, jroll I will be away from tomorrow until next week (I'm going back to Dublin tomorrow), so we can continue the split next week | 20:02 |
| lucasagomes | the split of capablities/idexable fields* | 20:02 |
| NobodyCam | have a good night lucasagomes :) | 20:02 |
| devananda | lucasagomes: ack. travel safe o/ | 20:02 |
| lucasagomes | thanks | 20:02 |
| lucasagomes | see you guys around, have a great night/weekend | 20:02 |
| NobodyCam | have a good flight lucasagomes | 20:03 |
| lucasagomes | NobodyCam, thanks! Hope to see you in dub soon :-) | 20:03 |
| *** lucasagomes is now known as lucas-pto | 20:03 | |
| *** krtaylor has joined #openstack-ironic | 20:03 | |
| *** ashaw_ has quit IRC | 20:04 | |
| jroll | thanks lucas-pto, enjoy :) | 20:04 |
| *** e0ne has quit IRC | 20:04 | |
| *** dprince has quit IRC | 20:09 | |
| *** Sukhdev has joined #openstack-ironic | 20:18 | |
| *** ashaw_ has joined #openstack-ironic | 20:18 | |
| *** ashaw_ has quit IRC | 20:18 | |
| *** e0ne has joined #openstack-ironic | 20:22 | |
| *** e0ne has quit IRC | 20:22 | |
| *** lucascess has left #openstack-ironic | 20:23 | |
| NobodyCam | lucas-pto: end of march-ish :) | 20:24 |
| *** electrofelix has quit IRC | 20:35 | |
| *** boris-42 has joined #openstack-ironic | 20:42 | |
| *** praneshp has joined #openstack-ironic | 20:43 | |
| mrda | Morning Ironic | 20:47 |
| NobodyCam | morning mrda | 20:50 |
| NobodyCam | :) | 20:50 |
| *** harshs has quit IRC | 20:51 | |
| jroll | devananda: any reason we don't try/except here? https://github.rackspace.com/O3Eng/ironic/blob/ironic.173/ironic/conductor/base_manager.py#L94 | 20:52 |
| jroll | no explosions if that happens, maybe due to refactor | 20:52 |
| jroll | oh wrong link heh | 20:52 |
| jroll | https://github.com/openstack/ironic/blob/master/ironic/conductor/base_manager.py#L94-L96 | 20:52 |
| devananda | jroll: I think we want it to blow up (stop the process) if it fails there | 20:54 |
| jroll | devananda: it doesn't though | 20:54 |
| jroll | that seems to run in a thread | 20:54 |
| devananda | yea | 20:54 |
| devananda | I saw that last week | 20:54 |
| devananda | I don't know when it changed | 20:54 |
| jroll | I just went to fix it and realized I have no idea how to make main thread blow up heh | 20:54 |
| devananda | urgh. my draft bug didn't get filed E_DISTRACTION | 20:55 |
| * devananda files it now | 20:55 | |
| mrda | hey NobodyCam | 20:55 |
| jroll | sweet, ty deva | 20:55 |
| *** harshs has joined #openstack-ironic | 20:55 | |
| jroll | I also don't see how that's in a thread, hrm | 20:56 |
| jroll | oh, bah | 20:56 |
| devananda | oh, right. I saw this while reviewing a patch to change BaseConductorManager's base class | 20:56 |
| devananda | *parent class | 20:57 |
| devananda | wait, did that land? | 20:57 |
| jroll | yeah, oslo.service starts it in a thread now | 20:58 |
| jroll | idk | 20:58 |
| *** daemontool_ has quit IRC | 20:59 | |
| devananda | now i have no idea what patch that was | 21:01 |
| devananda | found it - https://review.openstack.org/#/c/264720 | 21:03 |
| * devananda files a bug, since this isn't just a result of that patc | 21:04 | |
| devananda | patch | 21:04 |
| jroll | thank you | 21:05 |
| devananda | https://bugs.launchpad.net/ironic/+bug/1535908 | 21:11 |
| openstack | Launchpad bug 1535908 in Ironic "ironic-conductor no longer stops if exception raised during init_host()" [High,Triaged] | 21:11 |
| jroll | devananda: left a comment with a way worse scenario | 21:14 |
| jroll | no output whatsoever if driver fails to load | 21:14 |
| *** raddaoui has quit IRC | 21:15 | |
| devananda | oh - nice | 21:15 |
| devananda | and yea, the conductor does not register with the db, because init_host doesn't finish | 21:15 |
| *** ijw_ has joined #openstack-ironic | 21:21 | |
| *** ijw has quit IRC | 21:21 | |
| *** ijw has joined #openstack-ironic | 21:23 | |
| *** e0ne has joined #openstack-ironic | 21:23 | |
| *** raddaoui has joined #openstack-ironic | 21:24 | |
| *** Sukhdev has quit IRC | 21:25 | |
| *** ijw_ has quit IRC | 21:26 | |
| *** mbrennan has joined #openstack-ironic | 21:27 | |
| *** e0ne has quit IRC | 21:33 | |
| openstackgerrit | Merged openstack/ironic: Add more unit tests for NO_PROXY validation https://review.openstack.org/269800 | 21:37 |
| *** baoli has joined #openstack-ironic | 21:38 | |
| *** baoli_ has joined #openstack-ironic | 21:39 | |
| openstackgerrit | John L. Villalovos proposed openstack/ironic: Clean up 'no_proxy' unit tests https://review.openstack.org/269875 | 21:41 |
| *** penick has joined #openstack-ironic | 21:42 | |
| *** baoli has quit IRC | 21:43 | |
| jlvillal | JayF: Any thoughts on https://review.openstack.org/#/c/267219/ The proxy support for Docker inside IPA. I don't have any better ideas at this time. | 21:46 |
| JayF | jlvillal: lots of thoughts. Most of them conflicted. | 21:47 |
| jlvillal | :) | 21:47 |
| rloo | hey, is our gate working now? Something merged! | 21:48 |
| jlvillal | rloo: Unittest merged. | 21:48 |
| jlvillal | rloo: So a limited set of tests pass for only unittest changes. | 21:49 |
| jlvillal | pep8, python27, python34 and maybe something else. | 21:49 |
| rloo | jlvillal: oh. sigh. | 21:49 |
| *** baoli_ has quit IRC | 21:49 | |
| jlvillal | No devstack stuff. | 21:49 |
| jlvillal | Yeah :( | 21:49 |
| JayF | jlvillal: I'm not landing anything until IPA gate is behaving better, pretty much | 21:49 |
| *** baoli has joined #openstack-ironic | 21:49 | |
| *** harshs has quit IRC | 21:49 | |
| rloo | jlvillal: thx. I guess I should be happy... | 21:49 |
| *** baoli has quit IRC | 21:50 | |
| jlvillal | JayF: No hurry. | 21:50 |
| JayF | jlvillal: no problem with landing that really ... well, I have some problems, but I'm not going to hold you to account for docker's shitty decisions | 21:50 |
| JayF | lol | 21:50 |
| rloo | JayF: gate is broken. probably for ipa too but not sure. | 21:50 |
| jlvillal | I appreciate that! :) | 21:50 |
| JayF | I mean, our gate has been busted in IPA for a week or more | 21:50 |
| JayF | I just can't spare the time to dig in deeply :( | 21:50 |
| * jlvillal goes off to find soda... | 21:50 | |
| *** baoli has joined #openstack-ironic | 21:50 | |
| jlvillal | rloo: My patch submission above is only unit test changes ;) | 21:51 |
| rloo | jlvillal: what, is that a hint that you'd like me to review it? :) | 21:52 |
| * NobodyCam will bbiab, need to run and get some vape juice | 21:56 | |
| *** baoli has quit IRC | 22:01 | |
| *** baoli has joined #openstack-ironic | 22:01 | |
| *** alexpilotti has joined #openstack-ironic | 22:02 | |
| jlvillal | rloo: :) For sure it is a low priority patch! | 22:03 |
| rloo | jlvillal: :) I decided to pick a patch that hasn't been reviewed for awhile. | 22:04 |
| jlvillal | That's a good plan. | 22:04 |
| *** alexpilo_ has joined #openstack-ironic | 22:04 | |
| jlvillal | rloo: I don't expect our gate to be fixed until tomorrow. devstack patches take about 18 hours to make it through the gate. Assuming no error. | 22:06 |
| rloo | jlvillal: yeah, i noticed that. i wonder why it takes so long but anyway, not much we can do about it. | 22:07 |
| jlvillal | Yeah. | 22:07 |
| rloo | jlvillal: it is times like this where i think it is good to have some manual intervention to override these computers and just get it merged! | 22:07 |
| *** alexpilotti has quit IRC | 22:08 | |
| jlvillal | +1. They need a prioritization system. So patches fixing a project's gate breakage could move to the top of the queue. | 22:08 |
| rloo | jlvillal: yup | 22:09 |
| *** Sukhdev has joined #openstack-ironic | 22:11 | |
| *** raddaoui has quit IRC | 22:13 | |
| *** baoli has quit IRC | 22:13 | |
| *** baoli has joined #openstack-ironic | 22:14 | |
| *** harshs has joined #openstack-ironic | 22:14 | |
| openstackgerrit | John L. Villalovos proposed openstack/ironic: Clean up 'no_proxy' unit tests https://review.openstack.org/269875 | 22:20 |
| *** rcernin has quit IRC | 22:21 | |
| jlvillal | TheJulia: Looking at http://docs.openstack.org/developer/bifrost/readme.html#json-file-format gmmaha and I were wondering where the UUID came from? I was thinking when a node got created with 'ironic node-create' that is when the UUID was generated. | 22:21 |
| *** Sukhdev has quit IRC | 22:22 | |
| jlvillal | So didn't quite understand why we are supposed to provide a UUID. | 22:23 |
| *** Sukhdev has joined #openstack-ironic | 22:23 | |
| *** trown is now known as trown|outttypeww | 22:25 | |
| *** baoli has quit IRC | 22:25 | |
| *** baoli has joined #openstack-ironic | 22:26 | |
| * jlvillal wonders if cinerama knows? ^^^ | 22:26 | |
| cinerama | hi jlvillal let me look | 22:27 |
| jlvillal | Thanks! | 22:27 |
| *** harshs has quit IRC | 22:31 | |
| *** baoli has quit IRC | 22:31 | |
| *** mbrennan has quit IRC | 22:31 | |
| gmmaha | thanks cinerama | 22:33 |
| * jlvillal sees firefox at 10.3GB of memory usage. Decides it is time to restart it... | 22:34 | |
| gmmaha | jlvillal: i thought its only chrome that was that hungry.. | 22:34 |
| jlvillal | gmmaha: I'm not sure how many weeks Firefox has been running. Plus I probably have 200+ tabs across 10 windows... | 22:35 |
| *** jamielennox|away is now known as jamielennox | 22:35 | |
| gmmaha | jlvillal: :D | 22:35 |
| jlvillal | gmmaha: I was noticing a slow down for sure | 22:35 |
| *** baoli has joined #openstack-ironic | 22:36 | |
| *** harshs has joined #openstack-ironic | 22:36 | |
| *** harshs has quit IRC | 22:36 | |
| *** davidlenwell has quit IRC | 22:37 | |
| *** harshs has joined #openstack-ironic | 22:38 | |
| * jlvillal does not like that new Gerrit causes Edit->Find to not work | 22:38 | |
| *** davidlenwell has joined #openstack-ironic | 22:38 | |
| * devananda watches https://review.openstack.org/#/c/268960/ and continues waiting for it to land | 22:39 | |
| *** ionutbalutoiu has quit IRC | 22:40 | |
| * jlvillal thinks devananda will be waiting a lot longer... | 22:42 | |
| jlvillal | :( | 22:43 |
| rloo | devananda: wrt https://review.openstack.org/#/c/206244, should i review it after https://review.openstack.org/#/c/206245/, or should i do the non-api related patches first? | 22:45 |
| devananda | :( | 22:46 |
| *** baoli_ has joined #openstack-ironic | 22:46 | |
| devananda | rloo: I've broken it up into pre-API change and post-API change | 22:47 |
| rloo | devananda: so you want the pre-API changes done so we can merge those, before the post-API changes? | 22:48 |
| devananda | yah | 22:48 |
| devananda | everything up to https://review.openstack.org/#/c/206244/48 | 22:48 |
| *** baoli_ has quit IRC | 22:48 | |
| devananda | and also, pls review that one in depth -- since it's the point where all this becomes harder to unwind | 22:49 |
| rloo | devananda: oh. so just the first 4 then. | 22:49 |
| *** baoli_ has joined #openstack-ironic | 22:49 | |
| devananda | the stuff after that is actually pretty easy to review, fwiw | 22:49 |
| *** baoli has quit IRC | 22:49 | |
| devananda | smaller patches to devstack and docs, mostly | 22:49 |
| rloo | devananda: not sure our orders are the same orders. This one is after 206244 but it doesn't look simple and is more than API: https://review.openstack.org/#/c/139687/ | 22:51 |
| devananda | rloo: ah. you're right. I thought that one was before the API change, but it's not | 22:53 |
| devananda | vsaienko: is it possible to rebase https://review.openstack.org/#/c/139687/47 to come *before* https://review.openstack.org/#/c/206244/48 ? | 22:53 |
| devananda | vsaienko: actually, never mind that. I need to read both patches | 22:54 |
| rloo | devananda: i have to take off soon, will try to resume tomorrow. now i know what ordering you'd like :) | 22:54 |
| devananda | rloo: so both these patches are making significant API changes ... I need to dig into the second one more | 22:54 |
| rloo | devananda: 'significant' is not good. sigh. maybe i'll do something else instead :) | 22:55 |
| devananda | rloo: cool. I need to run too. ciao! | 22:55 |
| devananda | haha | 22:56 |
| devananda | well, the first 4 patches don't touch the API | 22:56 |
| * devananda really runs | 22:56 | |
| *** baoli_ has quit IRC | 23:00 | |
| *** baoli has joined #openstack-ironic | 23:01 | |
| zer0c00l | is ironic inspector part of liberty or kilo? | 23:03 |
| openstackgerrit | John L. Villalovos proposed openstack/ironic: Clean up 'no_proxy' unit tests https://review.openstack.org/269875 | 23:11 |
| *** baoli has quit IRC | 23:12 | |
| *** baoli has joined #openstack-ironic | 23:15 | |
| *** baoli has quit IRC | 23:25 | |
| *** absubram has quit IRC | 23:25 | |
| openstackgerrit | John L. Villalovos proposed openstack/ironic: DO_NOT_MERGE_OR_REVIEW: Test patch for Grenade testing. https://review.openstack.org/269924 | 23:29 |
| *** alex_xu has quit IRC | 23:30 | |
| *** alex_xu has joined #openstack-ironic | 23:31 | |
| [1]cdearborn | hey guys - i'm trying to write a little code that uses ironic client & running into a problem that i haven't been able to debug | 23:32 |
| [1]cdearborn | Here is the code: http://paste.openstack.org/show/484353/ | 23:34 |
| [1]cdearborn | here is the error: http://paste.openstack.org/show/484354/ | 23:35 |
| [1]cdearborn | at the time of failure, the variables have these values: http://paste.openstack.org/show/484355/ | 23:36 |
| [1]cdearborn | the line that fails is: val_fromjson = fromjson(attrdef.datatype, value[attrdef.name]) | 23:36 |
| [1]cdearborn | so it's trying to index into the string "path" using a string | 23:37 |
| [1]cdearborn | i have no idea why it's doing this | 23:37 |
| [1]cdearborn | i took this code directly from a unit test in the current code. i'm probably using a slightly older version of the client | 23:39 |
| [1]cdearborn | so was wondering if there might have been a bug fix that corrected this problem | 23:39 |
| jroll | [1]cdearborn: I believe patch needs to be a list of dicts | 23:42 |
| jlvillal | [1]cdearborn: Darn jroll beat me to it. But yes I think a list also | 23:42 |
| jlvillal | @wsme.validate(types.uuid, [NodePatchType]) | 23:42 |
| jlvillal | @expose.expose(Node, types.uuid_or_name, body=[NodePatchType]) | 23:42 |
| jlvillal | def patch(self, node_ident, patch): | 23:42 |
| *** harshs has quit IRC | 23:43 | |
| [1]cdearborn | thx very much! wondering why the unit test passes, and yet does not have the patch as a list: https://github.com/openstack/python-ironicclient/blob/master/ironicclient/tests/unit/v1/test_node.py#L621 | 23:47 |
| jlvillal | [1]cdearborn: I don't think the unit test actually calls via wsme into Ironic. But updating the unit test would be good, to match how it should be. | 23:48 |
| *** piet has quit IRC | 23:49 | |
| *** smoriya_ has joined #openstack-ironic | 23:51 | |
| jlvillal | [1]cdearborn: Maybe a patch for https://github.com/openstack/python-ironicclient/blob/master/ironicclient/v1/node.py#L191-L193 to check and make sure it is a list of dicts? And corresponding unit tests. If you like and would have time. | 23:52 |
| [1]cdearborn | changed to a list of dicts, and now getting: http://paste.openstack.org/show/484356/ | 23:53 |
| jlvillal | [1]cdearborn: What is 'new-driver'? | 23:54 |
| [1]cdearborn | here is the current code: http://paste.openstack.org/show/484357/ | 23:54 |
| jlvillal | [1]cdearborn: Right. It specifies 'new-driver'. | 23:55 |
| jlvillal | Is that registered with the conductor? | 23:55 |
| jlvillal | That is the error | 23:55 |
| jlvillal | $ git grep 'registered which supports' | 23:55 |
| jlvillal | conductor/rpcapi.py: reason = (_('No conductor service registered which supports ' | 23:55 |
| *** piet has joined #openstack-ironic | 23:55 | |
| [1]cdearborn | oh - didn't realize that it would actually check to see if that driver was available - i assumed if the unit test passed then it was valid - let me switch to the code i was really trying to write... | 23:56 |
| jlvillal | [1]cdearborn: The unit tests for python-ironicclient never actually call ironic. | 23:56 |
| [1]cdearborn | gotcha - that makes it tough to learn how to use the library though with limited docs and unit tests that aren't really valid examples... | 23:57 |
| jlvillal | Yep. | 23:58 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!