| JayF | How do you intend on doing your deployments? | 00:00 |
|---|---|---|
| JayF | Localboot or pxe ? | 00:00 |
| klindgren | Localboot | 00:00 |
| JayF | meaning in the end state, will your instances boot themselves locally or pxe boot | 00:00 |
| JayF | cool | 00:00 |
| JayF | yeah, I don't know how to solve that problem beyond different images for different hardware | 00:00 |
| *** piet has quit IRC | 00:01 | |
| klindgren | yea :-/ which sucks because end users wont/shouldn't know what image to use for x hardware | 00:02 |
| JayF | If you're putting it behind nova, nova supports doing that | 00:02 |
| JayF | or... hmm | 00:02 |
| JayF | that might be a rackspace patch downstream :/ | 00:02 |
| * JayF gives up on the summit schedule website after it breaks firefox again | 00:02 | |
| JayF | TheJulia: I think I'm going to try on windows chrome tonight and see if it's less terrible | 00:03 |
| * JayF getting offline for the night o/ | 00:03 | |
| TheJulia | goodnight JayF | 00:03 |
| klindgren | thats def a down stream patch. You can make sure that you have properties on an image that only work on x hardware. but no way of presenting say a cent7 image and on the backend saying this cent7 image is for x hardware and this cent7 image is for y hardware. | 00:03 |
| *** garthb__ has quit IRC | 00:04 | |
| *** piet has joined #openstack-ironic | 00:04 | |
| *** shu-mutou has joined #openstack-ironic | 00:05 | |
| *** baoli has quit IRC | 00:13 | |
| *** piet has quit IRC | 00:24 | |
| *** alex_xu has quit IRC | 00:26 | |
| *** alex_xu has joined #openstack-ironic | 00:27 | |
| *** ppiela has quit IRC | 00:29 | |
| *** jaypipes has quit IRC | 00:30 | |
| *** suro-patz has quit IRC | 00:31 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ironic: Updated from global requirements https://review.openstack.org/306942 | 00:37 |
| *** rloo has quit IRC | 00:41 | |
| *** hoangcx has joined #openstack-ironic | 00:42 | |
| *** alex_xu has quit IRC | 00:47 | |
| *** alex_xu has joined #openstack-ironic | 00:49 | |
| *** mtanino_ has joined #openstack-ironic | 00:54 | |
| *** mtanino has quit IRC | 00:55 | |
| *** mtanino_ has quit IRC | 00:59 | |
| *** ayoung has joined #openstack-ironic | 01:09 | |
| *** Sukhdev has quit IRC | 01:11 | |
| *** ChrisAusten has joined #openstack-ironic | 01:15 | |
| *** Sukhdev has joined #openstack-ironic | 01:23 | |
| devananda | krotscheck: I am recreating my ironic-webclient test env - but when I acces it in firefox, all I get is "loading" | 01:26 |
| TheJulia | devananda: Does the ironic-webclient-npm-publish-draft link load on the ironic-webclient reviews? Works just fine for me on chrome. | 01:33 |
| devananda | TheJulia: yea, the publish-draft link works fine | 01:44 |
| devananda | and it's not a firewall issue locally (tried disabling it) | 01:44 |
| Haomeng | devananda: can you help me review this spec if you have time - https://review.openstack.org/#/c/266688/, thanks, and good night:) | 01:49 |
| Haomeng | devananda: no rush, tomorrow is ok:) | 01:50 |
| *** keedya has quit IRC | 01:55 | |
| krotscheck | devananda: Are you getting anything from the console? | 02:03 |
| *** fragatin_ has quit IRC | 02:26 | |
| *** Sukhdev has quit IRC | 02:26 | |
| *** absubram has joined #openstack-ironic | 02:28 | |
| *** ijw has quit IRC | 02:28 | |
| *** keedya has joined #openstack-ironic | 02:30 | |
| *** ijw has joined #openstack-ironic | 02:30 | |
| *** deray has joined #openstack-ironic | 02:35 | |
| *** ijw has quit IRC | 02:39 | |
| *** jaybeale has joined #openstack-ironic | 02:39 | |
| *** sekrit is now known as CIA | 02:44 | |
| *** amotoki has joined #openstack-ironic | 02:44 | |
| *** fragatina has joined #openstack-ironic | 02:46 | |
| *** jaybeale has quit IRC | 02:49 | |
| *** fragatina has quit IRC | 02:51 | |
| *** yuanying has quit IRC | 02:56 | |
| *** fragatina has joined #openstack-ironic | 03:07 | |
| *** fragatina has quit IRC | 03:12 | |
| *** phuongnh has joined #openstack-ironic | 03:13 | |
| *** links has joined #openstack-ironic | 03:21 | |
| *** ayoung has quit IRC | 03:40 | |
| *** yuanying has joined #openstack-ironic | 03:48 | |
| *** jaybeale has joined #openstack-ironic | 03:52 | |
| *** piet has joined #openstack-ironic | 04:07 | |
| *** fragatina has joined #openstack-ironic | 04:09 | |
| *** ekarlso has quit IRC | 04:11 | |
| *** fragatina has quit IRC | 04:13 | |
| *** smoriya has quit IRC | 04:16 | |
| *** smoriya_afk is now known as smoriya | 04:16 | |
| *** smoriya_afk has joined #openstack-ironic | 04:17 | |
| *** piet has quit IRC | 04:19 | |
| *** piet has joined #openstack-ironic | 04:20 | |
| *** ChrisAusten has quit IRC | 04:21 | |
| *** ChrisAusten has joined #openstack-ironic | 04:22 | |
| *** jaybeale has quit IRC | 04:24 | |
| *** ekarlso has joined #openstack-ironic | 04:25 | |
| *** fragatina has joined #openstack-ironic | 04:26 | |
| *** fragatina has quit IRC | 04:31 | |
| *** piet has quit IRC | 04:34 | |
| *** piet has joined #openstack-ironic | 04:54 | |
| *** Sukhdev has joined #openstack-ironic | 05:00 | |
| *** piet has quit IRC | 05:10 | |
| *** yonglihe has quit IRC | 05:22 | |
| *** sivaramakrishna has joined #openstack-ironic | 05:25 | |
| *** vmud213 has joined #openstack-ironic | 05:33 | |
| *** itamarl has joined #openstack-ironic | 05:47 | |
| *** chip_ has joined #openstack-ironic | 05:49 | |
| *** chip_ has quit IRC | 05:52 | |
| *** chip_ has joined #openstack-ironic | 05:53 | |
| *** ChubYann has quit IRC | 05:56 | |
| *** yuanying_ has joined #openstack-ironic | 05:57 | |
| *** stendulker has joined #openstack-ironic | 05:57 | |
| *** yuanying has quit IRC | 05:58 | |
| *** harlowja_at_home has joined #openstack-ironic | 06:00 | |
| *** fragatina has joined #openstack-ironic | 06:02 | |
| openstackgerrit | Merged openstack/bifrost: curl is required for diskimage-builder https://review.openstack.org/305659 | 06:04 |
| *** fragatina has quit IRC | 06:07 | |
| *** thiagop has quit IRC | 06:12 | |
| openstackgerrit | Shivanand Tendulker proposed openstack/ironic-specs: Group management of Ironic nodes https://review.openstack.org/235344 | 06:13 |
| *** xavierr has quit IRC | 06:14 | |
| *** suro-patz has joined #openstack-ironic | 06:14 | |
| *** chip_ has quit IRC | 06:16 | |
| *** rcernin has joined #openstack-ironic | 06:21 | |
| *** raildo is now known as raildo-afk | 06:25 | |
| *** e0ne has joined #openstack-ironic | 06:38 | |
| *** e0ne has quit IRC | 06:43 | |
| *** e0ne has joined #openstack-ironic | 06:44 | |
| *** e0ne_ has joined #openstack-ironic | 06:45 | |
| *** yuanying_ has quit IRC | 06:47 | |
| *** tesseract has joined #openstack-ironic | 06:47 | |
| *** yuanying has joined #openstack-ironic | 06:48 | |
| *** e0ne has quit IRC | 06:48 | |
| *** tesseract is now known as Guest67082 | 06:48 | |
| *** e0ne_ has quit IRC | 06:50 | |
| *** harlowja_at_home has quit IRC | 06:59 | |
| *** Sukhdev has quit IRC | 07:04 | |
| *** suro-patz has quit IRC | 07:07 | |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Allow to set Neutron port setup delay from config https://review.openstack.org/293876 | 07:14 |
| *** jmccrory has quit IRC | 07:22 | |
| *** jlvillal has quit IRC | 07:22 | |
| *** jlvillal has joined #openstack-ironic | 07:22 | |
| *** ifarkas has joined #openstack-ironic | 07:24 | |
| *** ohamada has joined #openstack-ironic | 07:24 | |
| *** jmccrory has joined #openstack-ironic | 07:24 | |
| *** athomas has joined #openstack-ironic | 07:28 | |
| *** aarefiev22 has joined #openstack-ironic | 07:36 | |
| *** itamarl_ has joined #openstack-ironic | 07:37 | |
| *** itamarl_ has quit IRC | 07:37 | |
| *** itamarl_ has joined #openstack-ironic | 07:37 | |
| *** itamarl has quit IRC | 07:40 | |
| *** itamarl_ is now known as itamarl | 07:40 | |
| *** moshele has joined #openstack-ironic | 07:40 | |
| *** yuanying has quit IRC | 07:49 | |
| *** yuanying has joined #openstack-ironic | 07:52 | |
| *** ndipanov has joined #openstack-ironic | 07:54 | |
| *** lucas-dinner is now known as lucasagomes | 07:55 | |
| lucasagomes | morning all | 07:55 |
| lucasagomes | vdrok, https://review.openstack.org/#/c/307705/ the patch in tempest seems to work :-) | 07:56 |
| *** pece has joined #openstack-ironic | 07:57 | |
| *** aarefiev_ has joined #openstack-ironic | 07:59 | |
| *** zzzeek has quit IRC | 08:00 | |
| *** zzzeek has joined #openstack-ironic | 08:04 | |
| *** yuanying has quit IRC | 08:05 | |
| *** mbound has joined #openstack-ironic | 08:07 | |
| *** karimb has joined #openstack-ironic | 08:08 | |
| *** kromanenko has joined #openstack-ironic | 08:14 | |
| Haomeng | lucasagomes: morning, can you help to review my bp - https://review.openstack.org/#/c/266688/ if you have time? Thank you, nice day:) | 08:16 |
| *** jistr has joined #openstack-ironic | 08:18 | |
| *** MattMan has quit IRC | 08:18 | |
| *** MattMan has joined #openstack-ironic | 08:18 | |
| vdrok | morning ironic, lucasagomes and Haomeng ! | 08:20 |
| vdrok | yep, seen that, great :) | 08:20 |
| Haomeng | vdrok: morning:) | 08:20 |
| *** pece has quit IRC | 08:21 | |
| alineb | morning all | 08:21 |
| lucasagomes | Haomeng, sure thing, just updating the tempest patch and will look at it | 08:21 |
| lucasagomes | alineb, morning | 08:21 |
| Haomeng | lucasagomes: sure, no rush, thanks:) | 08:21 |
| vmud213 | morning Haomeng,vdrok,lucasagomes,alineb and rest all | 08:25 |
| Haomeng | vmud213: :) | 08:25 |
| vdrok | morning alineb and vmud213 | 08:26 |
| *** yonglihe has joined #openstack-ironic | 08:27 | |
| *** keedya has quit IRC | 08:28 | |
| lucasagomes | vmud213, morning | 08:28 |
| *** derekh has joined #openstack-ironic | 08:37 | |
| openstackgerrit | Merged openstack/ironic-inspector: use openstack cli instead of keystone cli https://review.openstack.org/307523 | 08:39 |
| openstackgerrit | Merged openstack/ironic-specs: Add node name regexp and wildcard filter to API https://review.openstack.org/266688 | 08:42 |
| *** karimb has quit IRC | 08:42 | |
| *** mgould has joined #openstack-ironic | 08:47 | |
| openstackgerrit | Merged openstack/ironic-inspector: Updated from global requirements https://review.openstack.org/307752 | 08:48 |
| *** fhubik has joined #openstack-ironic | 08:52 | |
| *** karimb has joined #openstack-ironic | 08:53 | |
| *** electrofelix has joined #openstack-ironic | 08:54 | |
| lucasagomes | vdrok, if you have some time later today, mind taking a look at https://review.openstack.org/#/c/293912/ and https://review.openstack.org/#/c/294688/ ? I want to amke a new release for staging drivers containing such changes | 08:55 |
| lucasagomes | this week | 08:55 |
| *** david-lyle has quit IRC | 08:59 | |
| *** david-lyle has joined #openstack-ironic | 09:00 | |
| lucasagomes | jroll, this needs ur ack: https://review.openstack.org/#/c/307914/ | 09:04 |
| openstackgerrit | Yuriy Zveryanskyy proposed openstack/ironic-specs: Ansible deploy driver https://review.openstack.org/241946 | 09:05 |
| vdrok | lucasagomes: sure, will do | 09:08 |
| openstackgerrit | lokesh s proposed openstack/ironic: Add support for the audit middleware https://review.openstack.org/272658 | 09:10 |
| sambetts|afk | Morning all | 09:14 |
| *** sambetts|afk is now known as sambetts | 09:14 | |
| *** dtantsur|afk is now known as dtantsur | 09:18 | |
| dtantsur | Morning Ironic | 09:18 |
| sambetts | o/ dtantsur | 09:18 |
| dtantsur | lucasagomes, mind reviewing a short spec https://review.openstack.org/305864 please? | 09:24 |
| lucasagomes | dtantsur, not at all! | 09:26 |
| *** ndipanov has quit IRC | 09:41 | |
| *** ndipanov has joined #openstack-ironic | 09:43 | |
| sambetts | dtantsur, lucasagomes: https://bugs.launchpad.net/nova/+bug/1572472 | 09:48 |
| openstack | Launchpad bug 1572472 in OpenStack Compute (nova) "Ironic Nova host manager Unsupported Operand" [Undecided,New] | 09:48 |
| mgould | morning Ironic | 09:51 |
| *** athomas has quit IRC | 09:54 | |
| vmud213 | dtantsur: Can you review https://review.openstack.org/#/c/303503 when you get a chance. | 09:56 |
| dtantsur | sure, though I don't promise the chance will be soon :) | 09:57 |
| lucasagomes | dtantsur, reviwed, I left some questions in-line. It may be the case of my lacking of understading about how it works, but it seems that we need to have more information in place to the switch network to happen | 09:57 |
| vmud213 | thanks :) | 09:57 |
| lucasagomes | sambetts, looking | 09:57 |
| dtantsur | lucasagomes, thanks, valid questions. I'll clarify as soon as I finish rewriting the driver composition spec :) | 09:58 |
| lucasagomes | cool thank you | 09:58 |
| *** mbound has quit IRC | 09:59 | |
| *** athomas has joined #openstack-ironic | 10:00 | |
| *** mbound has joined #openstack-ironic | 10:02 | |
| *** mbound has quit IRC | 10:04 | |
| *** mbound has joined #openstack-ironic | 10:05 | |
| sambetts | lucasagomes: I think I've worked out the possible path thats causing that bug... and its a weird one... | 10:13 |
| lucasagomes | sambetts, :-/ right, it's funny that free_disk_gb is set to None being an integer column in the database | 10:14 |
| lucasagomes | well maybe it's nullable | 10:14 |
| sambetts | lucasagomes: according to the model its a nullable value | 10:14 |
| lucasagomes | yeah | 10:14 |
| sambetts | but our driver shouldn | 10:15 |
| sambetts | shouldn't let it be null | 10:15 |
| sambetts | the logic in our driver is If node has instance id then local_gb_used = local_gb else local_gb_used = 0 | 10:16 |
| openstackgerrit | Tan Lin proposed openstack/ironic-specs: Support rolling upgrades https://review.openstack.org/299245 | 10:16 |
| lucasagomes | hmm I don't think we do have control over free_disk_size right? that's the resource tracker bits | 10:16 |
| lucasagomes | sambetts, oh right | 10:16 |
| sambetts | so it seems like a node that isn't valid to scheduale onto is being given the instance id and therfore is blowing up | 10:17 |
| lucasagomes | sambetts, and local_gb will always be populated right? Cause it's mandatory for ironic | 10:17 |
| sambetts | it should be for a node thats been scheduled onto, the api tests add a bunch of fakedriver nodes that don't have all the info populaters | 10:18 |
| *** stendulker has quit IRC | 10:18 | |
| sambetts | populated | 10:18 |
| sambetts | and it appears to be one of those blowing it up | 10:19 |
| *** mbound has quit IRC | 10:19 | |
| lucasagomes | oh :-/ yeah fake drivers may be a problem | 10:20 |
| *** mbound has joined #openstack-ironic | 10:20 | |
| lucasagomes | but in any case, maybe the nova ironic driver should do a propert.get('local_gb', 0) | 10:20 |
| lucasagomes | playing the safe card there | 10:20 |
| sambetts | yeah probably :/ | 10:21 |
| lucasagomes | because there's some arithmetics that will use that field which can blow up if it's non-int | 10:21 |
| lucasagomes | such as 'disk_available_least': local_gb - local_gb_used | 10:21 |
| sambetts | yeah... see I don't know why I don't see an error from that line | 10:22 |
| sambetts | unless None - None is valid in python | 10:22 |
| *** mbound has quit IRC | 10:23 | |
| lucasagomes | I don't think so | 10:23 |
| lucasagomes | yeah it's not | 10:24 |
| lucasagomes | off-topic: the tempest tests are failing in the -kilo jobs due a version conflict :-( http://logs.openstack.org/52/307952/4/check/gate-tempest-dsvm-full-kilo/fa4c1ab/logs/devstacklog.txt.gz#_2016-04-20_08_54_21_194 | 10:24 |
| *** mbound has joined #openstack-ironic | 10:25 | |
| sambetts | lucasagomes: this was the pbr version enforced by python_openstackclient above in that log pbr!=0.7,<1.0,>=0.6 | 10:27 |
| sambetts | lucasagomes: it actually does use a .get to pull that information out of the properties with 0 being the default :/ | 10:32 |
| lucasagomes | oh what? How it gets to None then? resource tracker maybe? | 10:32 |
| sambetts | I'm starting to wonder if this is caused by a race between the resource tracker and the ComputeNode updater thing or something :/ | 10:32 |
| *** karimb has quit IRC | 10:33 | |
| lucasagomes | yeah I don't know off the top of my head :-/ gotta check. Can you easily reproduce that problem? | 10:36 |
| *** karimb has joined #openstack-ironic | 10:37 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-specs: [WIP] Driver composition reform spec https://review.openstack.org/188370 | 10:37 |
| dtantsur | boom! big update ^^^ sambetts, devananda | 10:37 |
| *** nathan-h has joined #openstack-ironic | 10:38 | |
| sambetts | dtantsur: sweet! | 10:38 |
| *** hoangcx has quit IRC | 10:45 | |
| *** sivaramakrishna has quit IRC | 10:47 | |
| *** thrash|g0ne is now known as thrash | 10:49 | |
| *** smoriya has quit IRC | 10:54 | |
| *** smoriya_afk is now known as smoriya | 10:54 | |
| *** smoriya has quit IRC | 10:55 | |
| *** cfarquhar has quit IRC | 10:59 | |
| *** cfarquhar has joined #openstack-ironic | 11:00 | |
| *** cfarquhar has quit IRC | 11:00 | |
| *** cfarquhar has joined #openstack-ironic | 11:00 | |
| *** smoriya_afk has joined #openstack-ironic | 11:04 | |
| *** smoriya_afk is now known as smoriya | 11:04 | |
| vdrok | morning dtantsur sambetts and mgould ! | 11:04 |
| *** smoriya has quit IRC | 11:05 | |
| *** smoriya_afk has joined #openstack-ironic | 11:07 | |
| *** smoriya_afk is now known as smoriya | 11:07 | |
| mgould | vdrok, morning! | 11:07 |
| jroll | morning everyone | 11:07 |
| jroll | lucasagomes: ack'd | 11:08 |
| jroll | sambetts: someone else filed a similar bug that mriedem and I couldn't figure out at all | 11:08 |
| lucasagomes | jroll, awesome, thanks | 11:08 |
| lucasagomes | jroll, did you guys know how to reproduce the problem? | 11:09 |
| jroll | lucasagomes: nope :/ | 11:09 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic: Document how to run the tempest tests https://review.openstack.org/307695 | 11:09 |
| * lucasagomes added debug information ^ | 11:09 | |
| jroll | and we only pass ints up from our driver, so | 11:09 |
| jroll | ¯\_(ツ)_/¯ | 11:09 |
| lucasagomes | jroll, :-/ yeah that's hard to fix then | 11:09 |
| lucasagomes | jroll, yeah I wonder if it's the resource tracker messing up with something | 11:10 |
| jroll | lemme see if I can find it | 11:10 |
| sambetts | yeah thats what I'm just working through now | 11:10 |
| sambetts | looking that the resource tracker code etc | 11:10 |
| dtantsur | morning vdrok, jroll | 11:10 |
| vdrok | morning jroll | 11:11 |
| jroll | heya vdrok, dtantsur | 11:11 |
| jroll | sambetts: lucasagomes: https://bugs.launchpad.net/nova/+bug/1567434 | 11:12 |
| openstack | Launchpad bug 1567434 in OpenStack Compute (nova) "new hypervisor should appear with available resources set to 0" [Undecided,Incomplete] | 11:12 |
| lucasagomes | ty will take a look | 11:13 |
| jroll | going to move sam's info over and mark 1572472 as a dupe | 11:13 |
| lucasagomes | it's good to see a different error other than "no valid hosts" in the nova show output :-) | 11:15 |
| *** deray has quit IRC | 11:15 | |
| sambetts | I'd really like to track this down, its blemishing an otherwise stable gate record on my CI ;) | 11:15 |
| jroll | lucasagomes: lol | 11:16 |
| lucasagomes | not good, but somewhat relieving | 11:16 |
| sambetts | could this be related to some of our gate timeout issues? | 11:16 |
| sambetts | I guess we should probably make our hostmanager safe against those sort of values being None, because they are nullable according to the model :/ | 11:18 |
| ifarkas | lucasagomes, dtantsur, re jbod volumes, I registered a bug launchpad: https://bugs.launchpad.net/ironic/+bug/1572511 | 11:20 |
| openstack | Launchpad bug 1572511 in Ironic "RAID interface doesn't support JBOD volumes" [Undecided,New] | 11:20 |
| ifarkas | is it ok as a bug or should I make it an RFE? wdyt? | 11:20 |
| lucasagomes | jroll, sambetts maybe stating something obvious here, but would worth checking the value of compute.updated_at prior to try to update the informations int he host_manager ? | 11:22 |
| lucasagomes | see https://github.com/openstack/nova/blob/master/nova/scheduler/ironic_host_manager.py#L64 | 11:22 |
| lucasagomes | maybe checking "if compute.updated_at is not None" prior to trying to do any arithmetics in that method | 11:23 |
| jroll | hrm | 11:23 |
| lucasagomes | perhaps just return if that's not set | 11:23 |
| jroll | I wonder if we just need to be sure we force ints here? | 11:23 |
| jroll | I totally forgot about this thing | 11:23 |
| lucasagomes | see https://github.com/openstack/nova/blob/master/nova/scheduler/host_manager.py#L184-L186 | 11:23 |
| lucasagomes | there's some checks in the normal host_manager which the ironic one lacks of | 11:24 |
| jroll | oooo | 11:24 |
| sambetts | hmmm, good point, that might help prevent a race there | 11:24 |
| lucasagomes | it may be a normal thing to pass an unbounded/non-updated compute object to host managers by the scheduler | 11:24 |
| jroll | why do we override that, anyway? | 11:24 |
| lucasagomes | which assumes that the host manager code will detect | 11:24 |
| lucasagomes | jroll, no idea heh | 11:24 |
| lucasagomes | probably to work with resource tracker nicelly in our clusterized-way of managing nodes | 11:25 |
| lucasagomes | ifarkas, awesome! Thanks | 11:25 |
| *** trown|outtypewww is now known as trown | 11:25 | |
| jroll | sure, just don't see what that is :P | 11:25 |
| lucasagomes | yeah no idea off the top of my head | 11:26 |
| jroll | hm | 11:26 |
| jroll | brb | 11:27 |
| lucasagomes | I assume it's something to do with the get 100% of the resources when the node is picked for scheduling | 11:27 |
| lucasagomes | since we can't partially allocate resources with baremetal machines (like we do for hypervisors) | 11:28 |
| lucasagomes | anyway I will go eat something, I will add a comment to the bug later | 11:28 |
| * lucasagomes brb's | 11:28 | |
| *** lucasagomes is now known as lucas-hungry | 11:28 | |
| sambetts | hmmm this is really interesting!!! I'm looking at the 2 nova logs n-cpu.log and n-sch.log, and in n-cpu.log I can see it going through the resource tracker, and in the schedular I can see the host_manager error, the host_manager error happens time wise half way through the resource tracker logic | 11:28 |
| sambetts | jroll, lucas-hungry ^ | 11:29 |
| jroll | aha | 11:29 |
| *** electrofelix has quit IRC | 11:29 | |
| jroll | lucas-hungry: I put up https://review.openstack.org/308279 out of curiousity | 11:29 |
| lucas-hungry | jroll, hah cool aight | 11:29 |
| sambetts | it happens between the log that says, Final Resource view .... and Compute_service record updated | 11:29 |
| jroll | really brb now | 11:30 |
| lucas-hungry | sambetts, can you get the value of compute.updated_at when the code hits the host manager? | 11:30 |
| lucas-hungry | sambetts, just print it in a log or something | 11:30 |
| sambetts | lucas-hungry: updated at is None | 11:31 |
| *** electrofelix has joined #openstack-ironic | 11:31 | |
| jroll | sambetts: please keep https://bugs.launchpad.net/nova/+bug/1567434 up to date with your findings, so nova people and everyone else can see | 11:31 |
| openstack | Launchpad bug 1567434 in OpenStack Compute (nova) "new hypervisor should appear with available resources set to 0" [Undecided,Incomplete] | 11:31 |
| lucas-hungry | sambetts, ha, so, maybe that's why the fields are unbonded ? | 11:31 |
| lucas-hungry | unbounded* | 11:31 |
| lucas-hungry | the compute object does not have the information yet | 11:31 |
| * lucas-hungry brb for lunch | 11:32 | |
| sambetts | lucas-hungry: hmmm looking at the code though, when it first creates the ComputeNode object in the db it should be setting all that information on it before it hits create() | 11:32 |
| dtantsur | ifarkas, I | 11:32 |
| dtantsur | ifarkas, I'd call it an RFE | 11:32 |
| *** karimb has quit IRC | 11:33 | |
| ifarkas | dtantsur, ok, thanks | 11:33 |
| ifarkas | will update the bug | 11:34 |
| *** electrofelix has quit IRC | 11:37 | |
| *** electrofelix has joined #openstack-ironic | 11:37 | |
| *** electrofelix has quit IRC | 11:38 | |
| *** electrofelix has joined #openstack-ironic | 11:38 | |
| dtantsur | jroll, I wonder if we should build a list of standard node capabilities for inspect interfaces to discover... | 11:42 |
| dtantsur | I know that iLO folks have something in their spec, but not everything seems actually implemented | 11:43 |
| dtantsur | (spec: http://specs.openstack.org/openstack/ironic-specs/specs/kilo-implemented/ilo-properties-capabilities-discovery.html#proposed-change_ | 11:43 |
| *** electrofelix has quit IRC | 11:44 | |
| *** phuongnh has quit IRC | 11:44 | |
| *** electrofelix has joined #openstack-ironic | 11:44 | |
| sambetts | lucas-hungry, jroll I'm pretty sure I've worked it out now... | 11:47 |
| *** baoli has joined #openstack-ironic | 11:47 | |
| *** baoli has quit IRC | 11:49 | |
| *** baoli has joined #openstack-ironic | 11:49 | |
| *** baoli has quit IRC | 11:49 | |
| *** daemontool has joined #openstack-ironic | 11:50 | |
| *** stacker has joined #openstack-ironic | 11:56 | |
| *** dprince has joined #openstack-ironic | 11:57 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-python-agent: [WIP] Support capabilities discovery https://review.openstack.org/308292 | 11:59 |
| jroll | dtantsur: yes, I'm thinking we should | 12:00 |
| dtantsur | jroll, do you feel like a spec or just a documentation update? | 12:01 |
| jroll | dtantsur: idk, I feel like it should be in code or something | 12:02 |
| jroll | s/or something/somehow | 12:03 |
| *** baoli has joined #openstack-ironic | 12:04 | |
| dtantsur | jroll, wdym by "in code"? | 12:05 |
| *** baoli_ has joined #openstack-ironic | 12:06 | |
| sambetts | jroll, lucas-hungry: yup definatly got it... its a race that the if updated_at is not None would prevent | 12:06 |
| jroll | dtantsur: well, what's the goal here, just to document it? or to enforce something? | 12:06 |
| jroll | sambetts: fun! | 12:06 |
| jroll | nice find | 12:06 |
| dtantsur | jroll, that's the question :) I would start with documenting. I just don't want to end up with something different from what iLO does (or plans to do) | 12:07 |
| dtantsur | as iLO has their capabilities defined in a spec, we might want to have an informative spec too... | 12:07 |
| dtantsur | but maybe a documentation update is just fine | 12:07 |
| jroll | dtantsur: yeah, I think docs are fine in that case | 12:08 |
| *** baoli has quit IRC | 12:09 | |
| dtantsur | jroll, I wonder if iLO folks would hate me for asking to update http://specs.openstack.org/openstack/ironic-specs/specs/kilo-implemented/ilo-properties-capabilities-discovery.html#proposed-change to be a bit more consistent... | 12:09 |
| dtantsur | e.g. they have supported_boot_mode which is actually just boot_mode, and then support_secure_boot which is probably secure_boot_supported | 12:10 |
| jroll | dtantsur: idk why they would hate you :) | 12:10 |
| jroll | makes sense to me | 12:10 |
| *** raildo-afk is now known as raildo | 12:10 | |
| sambetts | jroll: I'll put together a patch into nova to sort it out, although I don't know whether to just add the check to the host manager or to try to solve the real cause of the race :/ | 12:12 |
| dtantsur | actually what they have is quite different: https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/ilo/inspect.py#L33 | 12:12 |
| dtantsur | ok, I have to document it :) | 12:12 |
| jroll | sambetts: probably the former, honestly, there's a reason why the standard host manager has this check :) | 12:13 |
| jroll | BootMode? really? | 12:13 |
| sambetts | jroll: yeah, I'm going to git blame that first to see if it was added to fix a similar issue or not ... | 12:13 |
| jroll | fair enough | 12:14 |
| *** xavierr has joined #openstack-ironic | 12:17 | |
| *** piet has joined #openstack-ironic | 12:18 | |
| xavierr | good morning Ironic :) | 12:19 |
| mgould | random question: do we ever plan to support ARM-based nodes? I mentioned I was working on baremetal provisioning to a former colleague and now he's telling me horror stories about the Aarch64 boot process | 12:19 |
| mgould | xavierr, morning! | 12:19 |
| dtantsur | mgould, I would not -2 someone's effort to support ARM.. | 12:19 |
| sambetts | I'm sure there is a rasppi usecase somewhere ;) | 12:20 |
| mgould | heh, yes | 12:20 |
| TheJulia | good morning everyone | 12:20 |
| sambetts | o/ TheJulia | 12:20 |
| xavierr | morning mgould sambetts TheJulia | 12:20 |
| mgould | morning TheJulia | 12:21 |
| mgould | sambetts, they do have an RPi model intended for rack-mounting or embedded use: https://www.raspberrypi.org/products/compute-module/ | 12:21 |
| mgould | anyway, apparently you have to tell the MMU to treat all memory as non-device or you're in for hours of hilarious debugging | 12:22 |
| sambetts | I think one probelm I know of with the rasppi is that it doesn't support any builtin remote control, not even WOL if I remember correctly :( I really wanted to built a pi cluster | 12:24 |
| mgould | oh yeah, that would be a big problem for us | 12:24 |
| sambetts | we'd probably end up with something like the lego power interface again | 12:24 |
| mgould | network -> RPi -> robot arm -> power connection on RPi host nodes | 12:25 |
| sambetts | exactly | 12:25 |
| mgould | frankly, I'm astonished we don't have a lego-mindstorm-robot-arm driver already | 12:25 |
| sambetts | we do, just not in tree | 12:25 |
| mgould | hahaha, of course | 12:25 |
| mgould | :-) | 12:25 |
| sambetts | we also had the "MeatPowerDriver" at one point too, which all it did was send an email to tell someone to turn the machine on | 12:26 |
| mgould | ahahahaha | 12:27 |
| * TheJulia thinks we could use a coffee driver and that it could be CI tested :) | 12:27 | |
| mgould | HTCPCP backend? | 12:27 |
| sambetts | haha | 12:28 |
| TheJulia | possibly :) | 12:28 |
| * sambetts is sure he's not the only one who googled that just no w | 12:28 | |
| TheJulia | we might have to write an upgrade to that in order to ensure only one cup of coffee is produced each time | 12:28 |
| mgould | only one cup of coffee? | 12:29 |
| mgould | who are you, and what have you done with the real TheJulia? | 12:29 |
| TheJulia | mgould: gues, that or a very tiny coffee pot | 12:29 |
| mgould | :-) | 12:29 |
| TheJulia | mgould: I'm just julia with a really bad migraine since around 4:30PM yesterday local time | 12:29 |
| mgould | :-( | 12:30 |
| mgould | get well soon! | 12:30 |
| sambetts | :( | 12:32 |
| * xavierr thinking the CI running the tests without mocks for coffee driver | 12:32 | |
| *** xavierr is now known as xavierr_coffee | 12:34 | |
| *** xavierr_coffee has quit IRC | 12:35 | |
| jroll | mgould: sambetts: unfortunately raspis don't do pxe booting, unless you load a pxe rom onto the sd card, in which case you have nothing to write to :/ | 12:35 |
| sambetts | yeah :( unless you've got external storage, but then IPA would need to understand writing to a USB mounted disk and also understand how not to blow up the SD car | 12:36 |
| sambetts | card | 12:37 |
| *** keedya has joined #openstack-ironic | 12:37 | |
| openstackgerrit | Imre Farkas proposed openstack/ironic: RAID interface to support JBOD volumes https://review.openstack.org/308306 | 12:40 |
| mat128 | morning ironicers | 12:41 |
| mat128 | oh interesting discussion, raspi! | 12:41 |
| mgould | jroll, bummer :-( | 12:41 |
| mat128 | while understanding and reversing BMC for test purposes, I came to the conclusion that the raspi would make a beautiful, reliable BMC | 12:41 |
| mat128 | except for the video capture part :) | 12:41 |
| mgould | mat128, what do you need to do video capture? | 12:42 |
| sambetts | mat128: I've been thinking about exactly that | 12:42 |
| mat128 | mgould: regular BMCs use a video chip and digitalize VGA data (yup...) | 12:42 |
| mat128 | they are just electrically inline with the VGA output lol | 12:42 |
| sambetts | I've even got some idea about how to allow VirtualMedia boot support too | 12:43 |
| mgould | jroll, my former colleague tells me that ARM are targeting the server market, so we may have to care about them eventually | 12:43 |
| mat128 | yes, usb host and stuff | 12:43 |
| jroll | mgould: oh, we should totally do ARM. just not (the current version) raspberry pis :) | 12:43 |
| mgould | mat128, makes sense | 12:43 |
| mat128 | for an ultra basic BMC, you could wire the arm's GPIO to the power / reset switches on your main machine | 12:44 |
| *** karimb has joined #openstack-ironic | 12:44 | |
| TheJulia | so a bolt on bmc for raspberry pi modules? | 12:44 |
| mgould | TheJulia, using an RPi as a BMC for an x86 box, I think | 12:45 |
| mat128 | exactly ^ | 12:45 |
| sambetts | mat128: Yeah I was thinking to just hook it up to the buttons in the PC case, short press == soft shut down, long press == hard shut and reset = reboot etc | 12:45 |
| TheJulia | interesting | 12:45 |
| mat128 | sambetts: thats it :) | 12:45 |
| jroll | neat | 12:45 |
| mat128 | thats what they do anyway | 12:45 |
| mat128 | and for more elaborate stuff, the bmc behaves as a usb device and talks to the main machine | 12:46 |
| *** ipukha has joined #openstack-ironic | 12:46 | |
| mat128 | so that would require cooperation with the main board | 12:46 |
| mgould | TheJulia, a BMC for RPi would be cool too, but I don't know how you'd do it | 12:46 |
| *** links has quit IRC | 12:46 | |
| mat128 | I havent looked into racking RPis, but you could easy have something control the power to them | 12:46 |
| mat128 | kind of a PDU for RPis | 12:47 |
| sambetts | yeah that would be pretty straight forward, they just don't have a nice soft shutdown | 12:47 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic: Update the inspection documentation https://review.openstack.org/308310 | 12:47 |
| dtantsur | jroll, wdyt ^^^? | 12:47 |
| mat128 | sambetts: you could always emulate a keyboard and send the equivalent of shiny new power buttons that are support to cooperate with the OS to shut it down | 12:48 |
| mat128 | sambetts: http://hackaday.com/2016/01/25/raspberry-pi-zero-cluster-packs-a-punch/ | 12:48 |
| sambetts | mat128: for video, i didn't know if it would be possible to get something like a USB capture card and then stream it | 12:49 |
| mat128 | all you need is power control | 12:49 |
| sambetts | now that is awesome! | 12:49 |
| mat128 | sambetts: I wonder how much computing that requires | 12:49 |
| mat128 | sambetts: but for serial capture, this is super easy :) | 12:49 |
| mat128 | an alternative idea to capturing analog data if to present the board as a video card | 12:50 |
| mat128 | they now do USB cards, or maybe even using a PI on a PCI card and exposing a basic video card would work better than doing capture | 12:50 |
| mat128 | now the BIOS has to use your device instead of the onboard one | 12:50 |
| *** lucas-hungry is now known as lucasagomes | 12:51 | |
| lucasagomes | sambetts, alright cool! :-) | 12:51 |
| lucasagomes | so just adding that check at the beggining seems good? | 12:51 |
| sambetts | yeah :/ hmmm, mat128 you coming to summit? | 12:51 |
| mat128 | unfortunately no :( | 12:51 |
| TheJulia | :( | 12:51 |
| TheJulia | this is where a teleporter or star trek like transporter would be useful | 12:51 |
| TheJulia | just beam people to the conference | 12:52 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-python-agent: [WIP] Support capabilities discovery https://review.openstack.org/308292 | 12:52 |
| jroll | TheJulia: and be able to sleep in your own bed during confs \o/ | 12:52 |
| TheJulia | jroll: oh that would be nice | 12:52 |
| sambetts | lucasagomes: yeah, so the problem stems from the fact the resource manager does ComputeNode create with the basic required resource info, then do a bunch of process of the resources, then update the ComputeNode with info like free_disk_gb, so if the schedular happens to process that node between the two when updated_at == None then free_disk_gb won't be set | 12:53 |
| lucasagomes | mgould, about lego mindstorm someone has the driver for it | 12:55 |
| lucasagomes | mgould, we saw it in a summit, I think was Paris | 12:55 |
| mgould | awesome | 12:55 |
| * mgould will look for the video :-) | 12:55 | |
| *** mgould is now known as mgould|lunch | 12:55 | |
| jroll | dtantsur: +2 | 12:55 |
| jroll | mgould|lunch: https://cloudbase.it/rebot/ | 12:56 |
| sambetts | damn... jroll beat me | 12:56 |
| mgould|lunch | jroll, thanks! | 12:56 |
| *** piet has quit IRC | 12:56 | |
| jroll | so I don't think we'll be able to release ironic this week, because gate | 12:57 |
| * jroll moves that todo for week after summit | 12:57 | |
| sambetts | :( | 12:57 |
| lucasagomes | mgould|lunch, https://goo.gl/photos/nJg6ocES25XyEEPo7 | 12:57 |
| lucasagomes | sambetts, cool | 12:57 |
| dtantsur | jroll, if it makes you feel better, I was planning an inspector release after the summit as well | 12:58 |
| *** piet has joined #openstack-ironic | 12:58 | |
| jroll | dtantsur: cool :) | 12:58 |
| ChrisAusten | mat128: I was at the OpenCompute conference in San Jose last month. A company there did exactely that for a multi server config. He had a raspberypi just driving gpios to the servers for simple stuff. If you want to get fancier a real BMC can typically get you a bunch of information about the system without having access to the OS. | 12:59 |
| *** karimb has quit IRC | 13:00 | |
| *** trown is now known as trown|brb | 13:00 | |
| *** alexpilotti has joined #openstack-ironic | 13:00 | |
| sambetts | ChrisAusten: Oh sweet! | 13:01 |
| * sambetts is starting to think no idea is unique any more... | 13:01 | |
| ChrisAusten | I think it is pretty limited in functionality but a fun project nevertheless | 13:02 |
| ChrisAusten | I code for BMCs so we build servers with fully wired arm soc, meaning I get access to thermal, cooling, vpd, pci, etc away from the host os. | 13:03 |
| *** karimb has joined #openstack-ironic | 13:05 | |
| *** [1]cdearborn has joined #openstack-ironic | 13:06 | |
| *** Goneri has joined #openstack-ironic | 13:06 | |
| openstackgerrit | xiexs proposed openstack/python-ironicclient: Introduce FakeBaremetal class https://review.openstack.org/302037 | 13:06 |
| vdrok | seems like one more bug in resource tracking - https://bugs.launchpad.net/nova/+bug/1572555 | 13:07 |
| openstack | Launchpad bug 1572555 in OpenStack Compute (nova) "Nova reports memory_mb=0 for available Ironic node" [Undecided,New] | 13:07 |
| dtantsur | jroll, interesting concern on the agent API spec. Do we want to keep lookup tied to a deploy driver? or maybe we request all implementations to provide the same inventory and make the API generic? | 13:08 |
| jroll | dtantsur: that is interesting | 13:09 |
| jroll | I'm not sure right now | 13:09 |
| dtantsur | in one case we end up with /v1/drivers/<DRIVER>/lookup, in the other - with /v1/nodes/lookup | 13:10 |
| dtantsur | the former is more close to what we have now ofc | 13:10 |
| sambetts | I think we should standardise | 13:10 |
| *** xavierr has joined #openstack-ironic | 13:10 | |
| jroll | ah, I do much perfer the latter | 13:10 |
| jroll | interesting that this comes up while I'm working on the node search spec... we could almost use that api | 13:11 |
| jroll | and then for config, /nodes/uuid/agent_config | 13:11 |
| dtantsur | jroll, ... except for this one is not authenticated :( | 13:11 |
| jroll | ah, right | 13:11 |
| *** rloo has joined #openstack-ironic | 13:11 | |
| jroll | I wish we could do client certs here but surely someone would throw a fit | 13:11 |
| jroll | because ssl is too hard | 13:12 |
| dtantsur | jroll, speaking of simplification: we only support lookup by MAC's.. do we really need to send the whole inventory? | 13:12 |
| dtantsur | it might be safer and simpler to only send MAC's... | 13:12 |
| jroll | dtantsur: today, no we don't... I'd like to support matching on things like serial number or something | 13:12 |
| jroll | or even storing the inventory | 13:12 |
| dtantsur | which gets us at the border with inspector :) | 13:13 |
| jroll | yeah :) | 13:13 |
| jroll | I'd almost like to start talking about unifying the two | 13:13 |
| jroll | but idk | 13:13 |
| lucasagomes | dtantsur, jroll I think that was the idea for lookup and heartbeat | 13:13 |
| dtantsur | which? :) | 13:13 |
| lucasagomes | jroll, dtantsur once the bash ramdisk goes away we should promote those apis | 13:13 |
| *** vmud213 has quit IRC | 13:13 | |
| dtantsur | that's what I'm working on :) | 13:13 |
| lucasagomes | yup | 13:14 |
| *** karimb_ has joined #openstack-ironic | 13:14 | |
| dtantsur | the question here is whether we should continue sending the whole inventory on lookup. 1. we only use MAC's now, 2. the less we send over non-authenticated channel - the better, 3. not all ramdisks have the same inventory format or want to replicate it | 13:14 |
| dtantsur | jroll, lucasagomes ^^^ | 13:14 |
| jroll | dtantsur: sure, I'm fine with only sending macs, we should bump the inventory version if we do so in case people are using out of tree subclasses that depend on more inventory being sent | 13:15 |
| jroll | though I doubt we'd avoid breaking those anyway | 13:15 |
| dtantsur | jroll, we will still send the whole inventory to the passthru's on fallback | 13:16 |
| jroll | right | 13:16 |
| jroll | fair enough | 13:16 |
| *** karimb has quit IRC | 13:16 | |
| dtantsur | jroll, also, with API versioning in place, do we even need a version of inventory? | 13:16 |
| jroll | dtantsur: probably not | 13:16 |
| dtantsur | isn't it contained in the API version itself? | 13:16 |
| jroll | if the api is versioned :) | 13:16 |
| *** ChrisAusten has quit IRC | 13:16 | |
| jroll | vendor passthru is not | 13:16 |
| jroll | vendor passthru body* | 13:17 |
| *** mkovacik_ has joined #openstack-ironic | 13:17 | |
| dtantsur | I'm not touching vendor passthru at all, if we fallback to it - we will send the same payload | 13:17 |
| jroll | right | 13:17 |
| *** Haomeng has quit IRC | 13:17 | |
| jroll | so now that it is a real api, we do not need a version | 13:17 |
| jroll | is what I'm saying | 13:17 |
| *** dansmith has quit IRC | 13:17 | |
| dtantsur | got it :) | 13:17 |
| *** piet has quit IRC | 13:18 | |
| *** KennethWilke has quit IRC | 13:18 | |
| lucasagomes | +1 for versioning the proper api and +1 for the fallback, cause since it's driver specific we may have drivers out of tree relying on more items from inventory | 13:18 |
| *** marlinc has quit IRC | 13:19 | |
| *** trown|brb is now known as trown | 13:19 | |
| *** mkovacik has quit IRC | 13:19 | |
| *** marlinc has joined #openstack-ironic | 13:19 | |
| *** dansmith has joined #openstack-ironic | 13:20 | |
| *** dansmith is now known as Guest64767 | 13:20 | |
| NobodyCam | good morning Ironicers | 13:21 |
| TheJulia | Good morning | 13:22 |
| *** piet has joined #openstack-ironic | 13:23 | |
| NobodyCam | :) | 13:23 |
| *** KennethWilke has joined #openstack-ironic | 13:25 | |
| rloo | morning NobodyCam, TheJulia, jroll, dtantsur, lucasagomes, etc | 13:25 |
| dtantsur | morning rloo, NobodyCam | 13:26 |
| jroll | good morning rloo and NobodyCam :) | 13:26 |
| lucasagomes | rloo, NobodyCam good morning | 13:26 |
| NobodyCam | good morning rloo, dtantsur, and lucasagomes | 13:26 |
| NobodyCam | :) | 13:26 |
| TheJulia | good morning rloo | 13:26 |
| NobodyCam | and the TheJulia :) | 13:26 |
| rloo | how is Sir Gate today? | 13:26 |
| lucasagomes | rloo, broken still, at least we know that the patch in tempest fix it | 13:27 |
| jroll | a fix is up, failing kilo CI (a fix is up for that) | 13:27 |
| rloo | great. progress! | 13:27 |
| rloo | jroll: you still aiming for a release this week? | 13:27 |
| jroll | rloo: not with the gate like this :( | 13:27 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic-specs: Node search API spec https://review.openstack.org/306092 | 13:28 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-specs: Promote agent vendor passthru to core API https://review.openstack.org/306418 | 13:28 |
| dtantsur | lucasagomes, jroll ^^^ | 13:28 |
| jroll | does anyone mind a quick initial review on that? dtantsur lucasagomes ^ | 13:28 |
| jroll | I feel like these operators will get bike-sheddy | 13:28 |
| * lucasagomes looks | 13:28 | |
| *** thiagop has joined #openstack-ironic | 13:28 | |
| thiagop | Good morning, Ironic | 13:28 |
| * rloo doesn't look. wants others to bikeshed first. | 13:29 | |
| rloo | hi thiagop | 13:29 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic-specs: Node search API spec https://review.openstack.org/306092 | 13:29 |
| jroll | added a quick sentence | 13:29 |
| dtantsur | jroll, it's fine, except that we probably want to accept just a number as a synonym to "==number" | 13:30 |
| jroll | dtantsur: yeah, makes sense | 13:30 |
| jroll | I was just thinking the same :) | 13:30 |
| thiagop | rloo: o/ | 13:31 |
| lucasagomes | dtantsur, +1 for the alias | 13:33 |
| NobodyCam | morning thiagop | 13:33 |
| lucasagomes | jroll, also, we don't want to have some operators for string? At least != and == ? Say give me all nodes != <arch> | 13:33 |
| *** piet has quit IRC | 13:33 | |
| thiagop | hiya NobodyCam o/ | 13:33 |
| lucasagomes | maybe not for the inital filtering | 13:33 |
| jroll | lucasagomes: hm | 13:34 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic-specs: Node search API spec https://review.openstack.org/306092 | 13:34 |
| jroll | added the alias | 13:34 |
| jroll | lucasagomes: maybe if someone has a use case, we can add it later? I don't really see the need for that, iirc nova only does exact match on arch | 13:34 |
| lucasagomes | yeah it's fine, and easy to extend if needed | 13:35 |
| jroll | lucasagomes: this all means the integer things will need to be ints in the db, btw, you good with that on the "make properties ~json" thing? | 13:36 |
| jroll | dtantsur: just one thing on your spec | 13:37 |
| dtantsur | jroll, the optional bit was by yuriyz' request, I think I can remove it though after it's no longer driver-specific | 13:38 |
| jroll | oh, right | 13:38 |
| jroll | +1 | 13:38 |
| lucasagomes | jroll, yes seems fine. Better keep the data model as-is because when we brainstormed chainging it to indexable all the options were really ugly | 13:38 |
| rloo | jroll, when you have a minute. the spec was approved so you can remove your -2: https://review.openstack.org/#/c/250112/ | 13:39 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-specs: Promote agent vendor passthru to core API https://review.openstack.org/306418 | 13:39 |
| dtantsur | jroll, ^^^ | 13:39 |
| jroll | lucasagomes: well, if we don't make it indexable, it's going to be super slow | 13:40 |
| jroll | rloo: done | 13:40 |
| jroll | lucasagomes: I'll think on it | 13:40 |
| rloo | thx jroll | 13:40 |
| jroll | dtantsur: +2 | 13:40 |
| jroll | brb | 13:40 |
| lucasagomes | jroll, cool yeah | 13:41 |
| *** sturivnyi_ has quit IRC | 13:42 | |
| *** ayoung has joined #openstack-ironic | 13:50 | |
| *** links has joined #openstack-ironic | 13:52 | |
| *** absubram has quit IRC | 13:56 | |
| *** vishwanathj has joined #openstack-ironic | 13:58 | |
| dtantsur | jroll, are you fine with s/macs/addresses/ per lucas' comment? | 14:02 |
| jroll | dtantsur: yep | 14:02 |
| *** ChrisAusten has joined #openstack-ironic | 14:02 | |
| * dtantsur updates | 14:02 | |
| sambetts | ++ for supprt for infiniband etc | 14:02 |
| sambetts | jroll, lucasagomes: https://review.openstack.org/#/c/308348/ | 14:03 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-specs: Promote agent vendor passthru to core API https://review.openstack.org/306418 | 14:03 |
| dtantsur | done ^^^ | 14:03 |
| *** mgould|lunch is now known as mgould | 14:04 | |
| lucasagomes | sambetts, thanks | 14:04 |
| jroll | thanks sambetts, dtantsur | 14:05 |
| jroll | [/b 52 | 14:05 |
| jroll | oops | 14:05 |
| dtantsur | B-52, not so bad :D | 14:05 |
| jroll | heh, that's the nova channel | 14:05 |
| jroll | seems fitting | 14:06 |
| rloo | you thinking of http://theb52s.com/ or https://en.wikipedia.org/wiki/B-52_(cocktail) or http://www.boeing.com/defense/b-52-bomber/ ? | 14:07 |
| * lucasagomes wonders as well | 14:07 | |
| *** [1]cdearborn has quit IRC | 14:07 | |
| jroll | the last :) | 14:08 |
| rloo | :D | 14:08 |
| *** baoli_ has quit IRC | 14:08 | |
| lucasagomes | heh I mean no offense, but a russian and an american talking about bombers sounds scary :-) | 14:09 |
| jroll | well, it depends, currently I would like to drop a bomb on the ComputeCapabilitiesFilter which seems not too scary | 14:10 |
| lucasagomes | lol I would that too | 14:10 |
| *** cdearborn has joined #openstack-ironic | 14:10 | |
| lucasagomes | the b-52 cocktail looks pretty good tho | 14:10 |
| lucasagomes | it has coffee | 14:11 |
| *** ametts has joined #openstack-ironic | 14:11 | |
| jroll | kahlua isn't quite coffee, just roughly tastes like it :P | 14:11 |
| lucasagomes | oh :-( | 14:13 |
| thiagop | Kahlua remembers me of "black russians" | 14:13 |
| thiagop | the nicest I had | 14:14 |
| jroll | yesss | 14:14 |
| * dtantsur actually meant a cocktail, but is fine with other variants | 14:14 | |
| NobodyCam | lol drinking this early in the morning | 14:14 |
| thiagop | black russians? | 14:15 |
| *** cdearborn has quit IRC | 14:15 | |
| jroll | NobodyCam: it's okay if it tastes like coffee, right? | 14:15 |
| thiagop | never, that's a drink that I take when I want to get crazy | 14:15 |
| *** jaypipes has joined #openstack-ironic | 14:16 | |
| NobodyCam | lol | 14:19 |
| *** mbound has quit IRC | 14:20 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic-specs: Node search API spec https://review.openstack.org/306092 | 14:21 |
| jroll | dtantsur: lucasagomes: adds capabilities ^^ | 14:21 |
| * lucasagomes looks | 14:22 | |
| jroll | idk if I should put a blurb in for moving this to a library: https://github.com/openstack/nova/blob/master/nova/scheduler/filters/extra_specs_ops.py | 14:22 |
| lucasagomes | jroll, that actually wouldn't be a bad idea, e.g I could reuse it for the root device hints in IPA | 14:25 |
| jroll | yep | 14:25 |
| lucasagomes | jroll, maybe we should just reuse the same operators for properties and capabilities? | 14:28 |
| jroll | lucasagomes: it is, we just aren't using the string operators | 14:28 |
| lucasagomes | well there are things like "=" which seems to be an alias for ">=" | 14:29 |
| lucasagomes | I don't like it, but at least it will be consistent on both fields | 14:29 |
| jroll | yeah, I'm indifferent I guess | 14:29 |
| jroll | my plan either way was to make match() take an optional list like op_methods | 14:29 |
| *** Sukhdev has joined #openstack-ironic | 14:32 | |
| rloo | dtantsur: let me know if you want to make any of these changes, otherwise I'll +A: https://review.openstack.org/#/c/308310 | 14:32 |
| *** absubram has joined #openstack-ironic | 14:32 | |
| dtantsur | rloo, I can follow-up with these | 14:33 |
| dtantsur | thanks for the review! | 14:33 |
| *** dprince has quit IRC | 14:33 | |
| rloo | dtantsur: ok, +A'd. | 14:33 |
| dtantsur | thnx | 14:33 |
| rloo | dtantsur: thank you for doing the work! | 14:34 |
| *** moshele has quit IRC | 14:34 | |
| openstackgerrit | Kyrylo Romanenko proposed openstack/python-ironicclient: Use name randomizer from tempest https://review.openstack.org/307904 | 14:35 |
| sambetts | jroll, lucasagomes ... just started working to add the test for the nova change, and I've just realised that their check actually only checks if compute.updated_at is not None if the host managers self.updated is not None :/ | 14:37 |
| *** electrofelix has quit IRC | 14:39 | |
| openstackgerrit | Merged openstack/ironic: Update the inspection documentation https://review.openstack.org/308310 | 14:42 |
| *** electrofelix has joined #openstack-ironic | 14:42 | |
| lucasagomes | sambetts, yeah, in our case we are setting updated right? So it's expected to be None | 14:46 |
| openstackgerrit | Andre keedy proposed openstack/ironic-inspector: Replace node data via uuid/data/unprocessed API https://review.openstack.org/308379 | 14:46 |
| lucasagomes | unless something else will set it | 14:46 |
| sambetts | lucasagomes: in the host manager init it does self.updated = None | 14:47 |
| sambetts | and it also does the self.updates = compute.updated_at | 14:47 |
| *** athomas has quit IRC | 14:47 | |
| *** watanabe_isao has joined #openstack-ironic | 14:48 | |
| lucasagomes | sambetts, right, so we only need to check "compute.updated_at is not None and compute.updated_at > self.updated" ? | 14:48 |
| * sambetts is wondered whether I need to go and fix the root of the problem, ie. that they populate the ComputeNode info in 2 steps | 14:48 | |
| sambetts | lucasagomes: yeah, but I'm wondering why the nova one doesn't have to do it | 14:49 |
| sambetts | lucasagomes: hmmm I think I know why... | 14:50 |
| *** athomas has joined #openstack-ironic | 14:52 | |
| lucasagomes | sambetts, why is that? | 14:52 |
| * lucasagomes looks at the code | 14:52 | |
| sambetts | lucasagomes: if you look at the nova host manager logic, it switches out free_disk_gb with disk_available_least if disk_available_least is not None | 14:52 |
| sambetts | lucasagomes: disk_available_least gets set on the ComputeNode create | 14:53 |
| sambetts | so would exist at the time the schedular is picking up the object | 14:53 |
| *** Sukhdev has quit IRC | 14:54 | |
| sambetts | hmmm... although the min(least_gb, free_gb) call returns None when either options is None... | 14:56 |
| sambetts | so that should still faill... | 14:56 |
| sambetts | gaaaaah... | 14:57 |
| *** garthb has joined #openstack-ironic | 14:58 | |
| lucasagomes | :-/ | 14:59 |
| *** dprince has joined #openstack-ironic | 14:59 | |
| *** Guest64767 is now known as dansmith | 15:00 | |
| *** amit213 has quit IRC | 15:06 | |
| *** baoli has joined #openstack-ironic | 15:08 | |
| *** amit213 has joined #openstack-ironic | 15:08 | |
| sambetts | lucasagomes: I think that nova itself might actually suffer from this same issue, however its only triggered when you first add a computenode into the cluster and its a very narrow failure window, so because Ironic adds a compute node for every ironic node we're more likely to see it that someone testing with 1 or maybe 2 nova computes | 15:09 |
| sambetts | or even someone who has way more nova computes than that, becuase I doubt they are ever adding so many in one go as we are | 15:10 |
| devananda | morning, all | 15:11 |
| *** suro-patz has joined #openstack-ironic | 15:11 | |
| sambetts | Hey devananda | 15:11 |
| openstackgerrit | Jarrod Johnson proposed openstack/pyghmi: Add system X firmware information https://review.openstack.org/307393 | 15:11 |
| jroll | morning devananda :) | 15:11 |
| *** rajinir has joined #openstack-ironic | 15:12 | |
| lucasagomes | sambetts, oh that's interesting | 15:12 |
| lucasagomes | yeah every node is actually a computenode | 15:12 |
| lucasagomes | devananda, morning | 15:12 |
| openstackgerrit | Zhenguo Niu proposed openstack/ironic: Add nodes tagging support - objs https://review.openstack.org/233357 | 15:13 |
| openstackgerrit | Zhenguo Niu proposed openstack/ironic: Add nodes tagging support - API https://review.openstack.org/250478 | 15:13 |
| openstackgerrit | Zhenguo Niu proposed openstack/ironic: Joined 'tags' column while getting node https://review.openstack.org/253065 | 15:13 |
| openstackgerrit | Zhenguo Niu proposed openstack/ironic: Add DB layer support for node tags filter https://review.openstack.org/272624 | 15:13 |
| devananda | wow, much backscroll | 15:13 |
| devananda | mgould: we already support ARM (or did at one point) | 15:13 |
| *** jaypipes has quit IRC | 15:13 | |
| dtantsur | morning devananda | 15:15 |
| *** suro-patz has quit IRC | 15:15 | |
| sambetts | lucasagomes: and actually going back and reading the git blames for the reason they added that if statement in the nova host manager, it was because of simultainous schedules onto one ComputeNode which isn't possible in Ironic | 15:16 |
| *** e0ne has joined #openstack-ironic | 15:18 | |
| jroll | lucasagomes: https://review.openstack.org/#/q/topic:extra-specs-ops-library | 15:19 |
| lucasagomes | jroll, nice! seems pretty straight forward | 15:20 |
| jroll | lucasagomes: yep, easy peasy | 15:20 |
| jroll | I'll add this to the spec | 15:21 |
| *** alexpilotti has quit IRC | 15:24 | |
| NobodyCam | morning devananda | 15:24 |
| lucasagomes | jroll, thanks, one behavior that I find strange in the match() is that, if a operator that doesn't exist is passed it only returns False | 15:24 |
| lucasagomes | it doesn't raise any exception or things like | 15:25 |
| lucasagomes | making it hard to validate the supported operators | 15:25 |
| lucasagomes | jroll, do you think that, as part of the patch to allow passing custom operators, we should change that behavior? | 15:25 |
| lucasagomes | tho this may break nova if we do, idk... just leave it I guess | 15:26 |
| * sambetts would like stevedored operators | 15:28 | |
| *** moshele has joined #openstack-ironic | 15:29 | |
| JayF | Anyone else want to take a look at https://review.openstack.org/#/c/306418 (promote agent vendor passthru to core api) before I land it? | 15:29 |
| NobodyCam | I'm ona call but would love to | 15:30 |
| sambetts | JayF: I'd like to take a quick read through | 15:30 |
| JayF | then I'll wait :) it's a good simple spec | 15:31 |
| *** baoli has quit IRC | 15:31 | |
| dtantsur | 3x +2 on https://review.openstack.org/#/c/306418/, how many more do we need? ;) | 15:32 |
| dtantsur | also morning JayF | 15:32 |
| JayF | lol | 15:32 |
| JayF | good morning :) | 15:32 |
| *** links has quit IRC | 15:32 | |
| openstackgerrit | Alexey Galkin proposed openstack/python-ironicclient: Fix for tox 'testenv:cover' command https://review.openstack.org/307325 | 15:34 |
| openstackgerrit | Alexey Galkin proposed openstack/ironic: Fix tox cover command https://review.openstack.org/306854 | 15:36 |
| * devananda reads the promote agent vendor passthru spec | 15:36 | |
| *** Sukhdev has joined #openstack-ironic | 15:37 | |
| dtantsur | good chance to -1 it for a grammar nit \o/ | 15:37 |
| dtantsur | :D | 15:37 |
| sambetts | dtantsur, JayF looks good to me :) | 15:38 |
| devananda | um, so I disagree with it already | 15:38 |
| dtantsur | even better :) | 15:38 |
| devananda | pls give me a bit more time to read before approving | 15:38 |
| *** alexpilotti has joined #openstack-ironic | 15:41 | |
| *** thiagop has quit IRC | 15:44 | |
| *** jaybeale has joined #openstack-ironic | 15:44 | |
| *** rama_y has joined #openstack-ironic | 15:45 | |
| *** xavierr has quit IRC | 15:46 | |
| *** itamarl has quit IRC | 15:46 | |
| *** alexpilotti has quit IRC | 15:46 | |
| openstackgerrit | Debayan Ray proposed openstack/ironic: [WIP] Boot from iSCSI volume https://review.openstack.org/308425 | 15:47 |
| dtantsur | devananda, any suggestions on better endpoint URL's? | 15:47 |
| dtantsur | devananda, please see other answers inline | 15:49 |
| *** jaybeale has quit IRC | 15:50 | |
| *** baoli has joined #openstack-ironic | 15:51 | |
| *** akrish9 has joined #openstack-ironic | 15:53 | |
| *** raildo is now known as raildo-afk | 15:56 | |
| lucasagomes | devananda, very interesting review on the lookup name collision | 15:57 |
| *** akrish9 has quit IRC | 15:57 | |
| dtantsur | yeah, name collision is the only critical thing there IMO | 15:57 |
| dtantsur | I wonder if it should be /v1/lookup or what?... | 15:58 |
| *** ijw has joined #openstack-ironic | 15:59 | |
| lucasagomes | dtantsur, devananda and funny thing, we have a v1/node/detail endpoint | 16:00 |
| dtantsur | OUCH | 16:00 |
| lucasagomes | I just named a node "detail" and when I do a node-show | 16:00 |
| lucasagomes | check it out | 16:00 |
| dtantsur | so yeah, I'm not the 1st to screw it :D and at least I was caught | 16:00 |
| lucasagomes | http://paste.openstack.org/show/494844/ | 16:00 |
| lucasagomes | dtantsur, ^ | 16:00 |
| lucasagomes | yeah... I will file a bug about it | 16:01 |
| dtantsur | that's pretty bad. I've no idea what to do about it | 16:01 |
| lucasagomes | yeah | 16:01 |
| dtantsur | we will probably have to forbid name "detail", which is super-weird, but still... | 16:02 |
| lucasagomes | yeah | 16:02 |
| lucasagomes | and validate as well | 16:02 |
| dtantsur | now I need ideas about the new endpoints :) | 16:02 |
| lucasagomes | we also have a v1/node/validate | 16:02 |
| * lucasagomes names it validate to see | 16:02 | |
| dtantsur | speaking of resetfulness - neither of these 2 are restful, so I tend to think about restfulness as a nit | 16:02 |
| lucasagomes | yeah it's very abstract | 16:04 |
| *** moshele has quit IRC | 16:04 | |
| sambetts | things like requesting detail should have been query options on the api e.g. GET /v1/node/<uuid>?detail=True | 16:05 |
| dtantsur | yep | 16:05 |
| dtantsur | ok, leaving for a day, thanks all for reviews! | 16:05 |
| sambetts | Cya dtantsur!! | 16:07 |
| lucasagomes | dtantsur, see ya | 16:08 |
| mgould | dtantsur, o/ | 16:08 |
| *** dtantsur is now known as dtantsur|afk | 16:08 | |
| lucasagomes | all this name collision reminds me that I can't name a label TODO in gmail | 16:08 |
| * lucasagomes tried that before | 16:08 | |
| sambetts | huh thats interesting | 16:09 |
| lucasagomes | devananda, jroll funny one https://bugs.launchpad.net/ironic/+bug/1572651 | 16:11 |
| openstack | Launchpad bug 1572651 in Ironic "Naming a node "detail" or "validate" confuses Ironic" [Medium,Confirmed] - Assigned to Lucas Alvares Gomes (lucasagomes) | 16:11 |
| *** kromanenko has quit IRC | 16:11 | |
| lucasagomes | apparently "validate" is not under nodes/, that's just a documentation bug that is not showing the node parameter in it | 16:15 |
| sambetts | nice! we should totally move that logic into query URL things | 16:15 |
| *** alexpilotti has joined #openstack-ironic | 16:15 | |
| *** ChubYann has joined #openstack-ironic | 16:17 | |
| *** ohamada has quit IRC | 16:18 | |
| *** watanabe_isao has quit IRC | 16:21 | |
| *** jistr has quit IRC | 16:22 | |
| jlvillal | lucasagomes: facepalm | 16:22 |
| lucasagomes | jlvillal, :D | 16:22 |
| *** karimb_ has quit IRC | 16:23 | |
| * jlvillal has a feeling we made a mistake in the API design there. | 16:23 | |
| *** karimb has joined #openstack-ironic | 16:23 | |
| lucasagomes | jlvillal, perhaps we should have enforced nodes/?name=<blah> but well | 16:24 |
| lucasagomes | happens | 16:24 |
| jlvillal | Yep | 16:25 |
| *** sacharya has joined #openstack-ironic | 16:26 | |
| *** bnemec has quit IRC | 16:26 | |
| devananda | lucasagomes: hah, awesome, yea, that sucks :( | 16:28 |
| *** karimb has quit IRC | 16:29 | |
| lucasagomes | update: the patch capping testresources (which prevents the tempest patch fixing the gate to pass) has been approved https://review.openstack.org/#/c/307858/ | 16:30 |
| openstackgerrit | Andre keedy proposed openstack/ironic-inspector: Replace node data via uuid/data/unprocessed API https://review.openstack.org/308379 | 16:34 |
| *** rbudden has joined #openstack-ironic | 16:35 | |
| *** bnemec has joined #openstack-ironic | 16:35 | |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic: Document how to run the tempest tests https://review.openstack.org/307695 | 16:35 |
| *** penick has joined #openstack-ironic | 16:37 | |
| lucasagomes | calling it a day | 16:37 |
| lucasagomes | have a great evening all, see you all tomorrow! | 16:37 |
| NobodyCam | night lucasagomes :) | 16:37 |
| NobodyCam | see you monday | 16:37 |
| lucasagomes | NobodyCam, yup, IRL! | 16:37 |
| lucasagomes | looking fwd to it | 16:37 |
| NobodyCam | yes | 16:38 |
| jlvillal | lucasagomes: Any idea on the tempest error? | 16:38 |
| jlvillal | https://review.openstack.org/307952 | 16:38 |
| lucasagomes | jlvillal, yup, https://review.openstack.org/#/c/307952/ | 16:38 |
| jlvillal | Is it just a matter of recheck? | 16:39 |
| lucasagomes | jlvillal, that fixes it | 16:39 |
| lucasagomes | jlvillal, yeah see the Ironic test patch https://review.openstack.org/#/c/307705/ | 16:39 |
| lucasagomes | depending on it | 16:39 |
| jlvillal | I mean, since that patch has Jenkins -1 | 16:39 |
| jlvillal | I was wondering if just keep doing recheck on it. | 16:39 |
| lucasagomes | jlvillal, oh that's because we need https://review.openstack.org/#/c/307858/ | 16:39 |
| jlvillal | Or is something else broken. | 16:39 |
| jlvillal | Ah, thanks. | 16:39 |
| jlvillal | Thanks lucasagomes. Have a great night! | 16:40 |
| lucasagomes | jlvillal, the -kilo jobs are broken due that patch ^ we need to cap the version of that library | 16:40 |
| lucasagomes | it's being approved, so after it merges we can recheck again | 16:40 |
| jlvillal | Thank you | 16:40 |
| lucasagomes | np | 16:40 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: DO NOT REVIEW https://review.openstack.org/296432 | 16:40 |
| lucasagomes | thank you for keeping an eye on it | 16:40 |
| jlvillal | vsaienko: ^^^ | 16:40 |
| *** suro-patz has joined #openstack-ironic | 16:40 | |
| lucasagomes | see y'all later! | 16:40 |
| *** lucasagomes is now known as lucas-dinner | 16:40 | |
| *** raildo-afk is now known as raildo | 16:41 | |
| *** ppiela has joined #openstack-ironic | 16:44 | |
| *** derekh has quit IRC | 16:44 | |
| *** Guest67082 has quit IRC | 16:46 | |
| devananda | dtantsur|afk: further comments posted - need to jump into a meeting now | 16:50 |
| devananda | lucas-dinner: g'night! | 16:50 |
| *** watanabe_isao has joined #openstack-ironic | 16:50 | |
| *** ijw has quit IRC | 16:51 | |
| *** Sukhdev has quit IRC | 16:55 | |
| *** ijw has joined #openstack-ironic | 16:55 | |
| *** alexpilotti has quit IRC | 16:56 | |
| *** amotoki has quit IRC | 16:59 | |
| *** nathan-h has quit IRC | 17:00 | |
| *** rpioso has joined #openstack-ironic | 17:03 | |
| *** electrofelix has quit IRC | 17:05 | |
| *** mbound has joined #openstack-ironic | 17:06 | |
| jroll | lucas-dinner: actually if an operator doesn't exist, it falls back to using == | 17:07 |
| jroll | also, I told you all node names were a bad idea :D | 17:07 |
| *** fragatina has joined #openstack-ironic | 17:08 | |
| *** athomas has quit IRC | 17:13 | |
| *** mtanino has joined #openstack-ironic | 17:14 | |
| openstackgerrit | Michael Krotscheck proposed openstack/ironic-webclient: Modified IronicNode mock to use single source of data https://review.openstack.org/307927 | 17:19 |
| openstackgerrit | Michael Krotscheck proposed openstack/ironic-webclient: Changed action from 'deploy' to 'active' https://review.openstack.org/307926 | 17:19 |
| openstackgerrit | Michael Krotscheck proposed openstack/ironic-webclient: IronicNodeProvision resource https://review.openstack.org/282457 | 17:19 |
| openstackgerrit | Michael Krotscheck proposed openstack/ironic-webclient: Added provision state manipulation https://review.openstack.org/307928 | 17:19 |
| openstackgerrit | Jay Faulkner proposed openstack/ironic-python-agent: Force container builds to use http.us.debian.org https://review.openstack.org/307474 | 17:20 |
| *** ipukha has quit IRC | 17:21 | |
| *** penick has quit IRC | 17:22 | |
| *** penick has joined #openstack-ironic | 17:22 | |
| watanabe_isao | sambetts, may I ask a stupid question? | 17:23 |
| sambetts | hey, sure :D | 17:23 |
| watanabe_isao | What do you mean by make UCSM CI voting? Any configure on gerrit or zuul,do you mean, please? | 17:23 |
| watanabe_isao | ^^^ sambetts, sorry for got to add your name. | 17:24 |
| devananda | jroll: we have about half a dozen node names, that I now know of, which are effectively not allowed | 17:24 |
| devananda | so yea, fun. | 17:24 |
| *** mbound has quit IRC | 17:26 | |
| sambetts | watanabe_isao: so, right now we 2 different CI jobs one for the CIMC driver, and one for the UCSM driver, and Success or Failure of the Cisco Ironic CI is determined by the success of all the jobs in our CI, right now the UCSM job in our zuul config is voting = false so that it doesn't contribute the final sucess/failure vote, but now its stable we plan to make it vote towards the Success or | 17:27 |
| sambetts | failure of the Cisco CI | 17:27 |
| *** xavierr has joined #openstack-ironic | 17:27 | |
| *** trown is now known as trown|lunch | 17:27 | |
| *** penick has quit IRC | 17:29 | |
| sambetts | watanabe_isao: you can see this more clearly by looking at the messages our CI posts, here are some examples http://paste.openstack.org/show/494861/ | 17:30 |
| watanabe_isao | sambetts, I got it. Thank you. BTW in my understand, we need to provide CI start with non-voting, do you know if there is period to keep it, then change it to voting? | 17:31 |
| sambetts | watanabe_isao: as I understand it none of the third party CIs leave a blocking vote on a patch, so its down to us to decide when to make it voting or not | 17:32 |
| sambetts | krtaylor: ^ can you confirm | 17:32 |
| * mgould -> home; good night! | 17:32 | |
| *** mgould is now known as mgould|afk | 17:32 | |
| sambetts | cya mgould|afk | 17:32 |
| *** yarkot has joined #openstack-ironic | 17:32 | |
| *** alexpilotti has joined #openstack-ironic | 17:32 | |
| jlvillal | Woo hoo, the patch to tempest stable/kilo has merged. So now the patch to fix our gate should hopefully pass Jenkins. | 17:33 |
| *** akrish9 has joined #openstack-ironic | 17:33 | |
| jlvillal | Now all we need are some reviews of the +2 variety on that tempest patch :) | 17:33 |
| sambetts | w00p, although /me still isn't a fan of fixing something without knowing why it broke :'( | 17:33 |
| jlvillal | sambetts: Yeah, does seem odd that it just started happening. | 17:34 |
| devananda | sambetts: you are correct - third-party CI can vote as a normal user, but can not vote in a way that would block the gate | 17:34 |
| watanabe_isao | sambetts, thank you. Maybe I have a mis-understand about voting non-voting. I'll check for that. | 17:37 |
| sambetts | watanabe_isao, devananda, krtaylor: I'm not sure if there might be a difference between having a job voting and having a CI voting, e.g. I'm talking about jobs voting towards a CIs vote, but the Cisco Ironic CI doesn't currently place a +1/-1 in the verified column | 17:38 |
| sambetts | not sure if thats down to us to do or down to a upstream setting | 17:39 |
| * krtaylor reads scrollback | 17:39 | |
| *** fragatina has quit IRC | 17:42 | |
| *** fragatina has joined #openstack-ironic | 17:42 | |
| krtaylor | the right way is to start with commenting only, pushing comments that meet the infra requirements, then when that is stable for a few weeks, ask the team if it is ok to start voting, but a third party CI system will never be able to stop a merge (gate) | 17:43 |
| krtaylor | watanabe_isao, ^^^ does that help? | 17:44 |
| watanabe_isao | krtaylor, if sambetts's CI wants to put +1 in Verified. Do we need to configure anything in gerrit? Or just configure zuul of CI? Do you know that, please? | 17:45 |
| * sambetts has got to head out for dinner | 17:46 | |
| sambetts | please continue this convo and I'll pick up the scrollback when I'm back :) | 17:46 |
| *** sambetts is now known as sambetts|afk | 17:46 | |
| krtaylor | watanabe_isao, sambetts|afk, re: verify voting, it's all here: http://docs.openstack.org/infra/system-config/third_party.html#posting-result-to-gerrit | 17:49 |
| krtaylor | I strongly recommend reading every word on that page | 17:49 |
| *** ifarkas has quit IRC | 17:51 | |
| watanabe_isao | krtaylor, thanks for the information. And I think I have figured out what I want. | 17:51 |
| krtaylor | additional good reading: https://specs.openstack.org/openstack/ironic-specs/specs/not-implemented/third-party-ci.html | 17:51 |
| krtaylor | watanabe_isao, ^^^ | 17:51 |
| *** jaypipes has joined #openstack-ironic | 17:52 | |
| watanabe_isao | krtaylor, sambetts|afk : please let me explain. | 17:53 |
| *** daemontool has quit IRC | 17:53 | |
| *** ChrisAusten has quit IRC | 17:54 | |
| watanabe_isao | CIs can give its result to a project, but will not be able to Verified +1/-1, unless it is add to gerrit Verified group of that project. | 17:55 |
| watanabe_isao | e.x. In cinder you may find many Cis are working, but only VMware NSX CI can give +1/-1 Verified. | 17:56 |
| watanabe_isao | The reason is:https://review.openstack.org/#/admin/groups/508,members | 17:56 |
| watanabe_isao | ^^^ and the above Group cinder-ci | 17:57 |
| watanabe_isao | is configured as Label Verified in :https://review.openstack.org/#/admin/projects/openstack/cinder,access | 17:57 |
| watanabe_isao | Currently we don't have a label like that in ironic. | 17:58 |
| watanabe_isao | ^^^https://review.openstack.org/#/admin/projects/openstack/ironic,access | 17:58 |
| jlvillal | watanabe_isao: I would imagine it would be a long time before we (Ironic) would allow a 3rd Party CI to block the gate. | 17:59 |
| jlvillal | There would have to be a large amount of trust built up that we can believe in this CI and that there are people available 24/7 to fix issues if the CI broke. | 18:00 |
| *** ChrisAusten has joined #openstack-ironic | 18:00 | |
| jlvillal | I don't know of any plans to allow any 3rd Party CI to vote in the Ironic CI at this time. | 18:00 |
| watanabe_isao | jlvillal, do you mean if a CI Verified a -1 would block the gate? | 18:00 |
| jlvillal | I mean, to vote as to block the gate. | 18:00 |
| jlvillal | Voting is fine, as long as it doesn't block the gate. For example if I vote -1, it doesn't block the gate from merging if I am over-ruled. | 18:01 |
| watanabe_isao | jlvillal, yes. And I'm expecting CI vote -1 is the same as me. | 18:02 |
| jlvillal | Agreed. It is informational only. No power to block the gate. | 18:03 |
| watanabe_isao | jlvillal, yes. And if Ironic is not going to add a +1/-1 Verified label for 3rd party CI (like in cinder), CIs would not even vote +1/-1. And it is just fine. | 18:05 |
| *** ndipanov has quit IRC | 18:06 | |
| jlvillal | I think we want them to vote. And if we trust them then core reviewers would then look at -1 votes to see if it is an issue. | 18:06 |
| jlvillal | krtaylor likely knows more about this than me though. As he is in charge of coordinating our 3rd Party CI efforts | 18:06 |
| *** alexpilotti has quit IRC | 18:07 | |
| *** ChubYann has quit IRC | 18:07 | |
| watanabe_isao | jlvillal, I am good with the current situation. I just want to explain of what I have found of, if we want trust-able 3rd party CIs to vote +1/-1 like Jenkins do, what should we do. And about "do we need to do it", I think it would be another long discussion at other place.:) | 18:11 |
| watanabe_isao | krtaylor, sambetts|afk : hope my explain ^^^ helps you. | 18:13 |
| *** watanabe_isao has quit IRC | 18:15 | |
| krtaylor | watanabe_isao, I understand, but verify is done like it's documented, I don't believe that anything has been set up to prevent verification votes in ironic | 18:15 |
| krtaylor | and jlvillal, we wouldn't want third-party to block gate | 18:16 |
| krtaylor | no good reason to, verify +1/-1 should be all that is allowed | 18:17 |
| *** garthb has quit IRC | 18:19 | |
| jlvillal | krtaylor: +1 on not blocking gate. Unless we really really really trust them and we think it is very important to our project to make sure they don't break. | 18:20 |
| devananda | actually we can not allow third-party CI to block the gate | 18:20 |
| *** ChubYann has joined #openstack-ironic | 18:20 | |
| devananda | it is expressly not allowed by openstack-infra | 18:20 |
| jlvillal | devananda: Oh, in the conversation before here it was saying that Cinder has a 3rd Party CI. VMware NSX CI. I didn't verify that though. | 18:20 |
| devananda | the question is when/whether we choose to allow a given third-party CI system to vote +1/-1 -- or merely to post an advisory result. | 18:21 |
| devananda | jlvillal: yes, those systems "vote" in that they place a +1/-1 PASS/FAIL | 18:21 |
| devananda | they are not able to prevent patches from merging, however | 18:21 |
| jlvillal | But they don't block the gate. | 18:21 |
| devananda | right | 18:21 |
| jlvillal | Ah okay. Good. | 18:21 |
| jlvillal | devananda: So they can do the "Verified" field: https://review.openstack.org/#/admin/projects/openstack/cinder,access | 18:22 |
| jlvillal | If I'm reading that right. | 18:23 |
| devananda | http://docs.openstack.org/infra/system-config/third_party.html#requirements | 18:23 |
| jlvillal | But maybe Jenkins only cares about "Jenkins" vote in the verified field | 18:23 |
| *** Sukhdev has joined #openstack-ironic | 18:23 | |
| devananda | yes | 18:23 |
| *** alexpilotti has joined #openstack-ironic | 18:23 | |
| jlvillal | Okay. Good. and thanks. | 18:23 |
| *** Sukhdev has quit IRC | 18:24 | |
| devananda | jlvillal: note that only has +1/-1 rights | 18:24 |
| devananda | compare to the "Continuous Integration Tools" group at https://review.openstack.org/#/admin/projects/All-Projects,access | 18:24 |
| devananda | which has +2/-2 rights | 18:24 |
| devananda | that is, IIUC, Jenkins | 18:24 |
| *** alexpilotti has quit IRC | 18:25 | |
| *** ayoung has quit IRC | 18:25 | |
| *** alexpilotti has joined #openstack-ironic | 18:25 | |
| jlvillal | Okay. So we can make the 3rd Party CI do +1/-1. But no +2/-2. | 18:25 |
| jlvillal | And a patch only gets merged with +2 Verified, +2 Code Review, and +1 Workflow | 18:26 |
| *** yolanda has quit IRC | 18:26 | |
| jlvillal | If I'm understanding it correctly. | 18:26 |
| devananda | yep | 18:27 |
| *** Sukhdev has joined #openstack-ironic | 18:28 | |
| *** yolanda has joined #openstack-ironic | 18:29 | |
| *** alexpilotti has quit IRC | 18:29 | |
| *** penick has joined #openstack-ironic | 18:35 | |
| *** daemontool has joined #openstack-ironic | 18:35 | |
| *** aarefiev_ has quit IRC | 18:45 | |
| *** trown|lunch is now known as trown | 18:46 | |
| *** vishwanathj has quit IRC | 18:50 | |
| *** vishwanathj has joined #openstack-ironic | 18:50 | |
| *** garthb has joined #openstack-ironic | 18:51 | |
| *** e0ne has quit IRC | 18:55 | |
| *** yolanda has quit IRC | 18:58 | |
| *** rcernin has quit IRC | 19:05 | |
| openstackgerrit | Jarrod Johnson proposed openstack/pyghmi: Add system X firmware information https://review.openstack.org/307393 | 19:14 |
| *** jayc_ has joined #openstack-ironic | 19:17 | |
| *** alexpilotti has joined #openstack-ironic | 19:21 | |
| *** yolanda has joined #openstack-ironic | 19:23 | |
| *** alexpilotti has quit IRC | 19:25 | |
| rloo | does anyone know if we have a list of supported APIs? We discussed in the last summit I think. Trying to remember if methods in drivers/modules/deploy_utils.py or dhcp/neutron.py can be removed. | 19:25 |
| JayF | dhcp/neutron.py isn't the api def, it's somewhere else | 19:26 |
| JayF | deploy_utils isn't a part of any api aiui | 19:26 |
| JayF | wouldn't https://github.com/openstack/ironic/blob/master/ironic/dhcp/base.py be the thing that can't be changed w/r/t dhcp | 19:26 |
| JayF | not the neutron implementation of it specifically/ | 19:27 |
| rloo | JayF: dunno. you may be right. | 19:30 |
| rloo | JayF: I'm going to assume that you are right :) | 19:30 |
| JayF | I'm probably the worst person to make that assumption about wrt api design :P | 19:30 |
| *** ijw has quit IRC | 19:31 | |
| rloo | JayF: ha ha, it is just that the code patch follows what you just said, so that makes it easier for me not to deal with it. | 19:31 |
| *** rcernin has joined #openstack-ironic | 19:32 | |
| *** e0ne has joined #openstack-ironic | 19:33 | |
| *** ametts has quit IRC | 19:33 | |
| *** Sukhdev has quit IRC | 19:46 | |
| *** alexpilotti has joined #openstack-ironic | 19:57 | |
| openstackgerrit | Milan Kováčik proposed openstack/ironic-inspector-specs: High Availability for Ironic Inspector https://review.openstack.org/253675 | 20:06 |
| *** ChrisAusten has quit IRC | 20:08 | |
| openstackgerrit | Jarrod Johnson proposed openstack/pyghmi: Add system X firmware information https://review.openstack.org/307393 | 20:08 |
| jlvillal | jroll: FYI: https://review.openstack.org/#/c/307952/ got a +1 from Jenkins. I pinged in QA channel about it to get reviews. | 20:09 |
| jlvillal | To fix our gate we need that patch. | 20:09 |
| *** comstud has quit IRC | 20:14 | |
| rloo | JayF: is this fixed? https://bugs.launchpad.net/ironic/+bug/1531217 | 20:16 |
| openstack | Launchpad bug 1531217 in Ironic "[RFE] Nodes should not always be powered off on failure" [Wishlist,In progress] - Assigned to Jay Faulkner (jason-oldos) | 20:16 |
| JayF | I guess so; the other piece is bigger than that and can be a diff bug | 20:16 |
| rloo | JayF: OK. I changed the status to Fix Committed. (I hope that was right). | 20:19 |
| JayF | fix released is where it goes now | 20:19 |
| JayF | I'll do that | 20:19 |
| rloo | JayF: was it released? | 20:19 |
| jroll | release team always marks things fix released now | 20:19 |
| jroll | they skip committed | 20:19 |
| JayF | what jroll said | 20:19 |
| NobodyCam | W00f h00 on the road to Austin | 20:20 |
| rloo | OH. ok then. we should delete 'fix committed' as a choice then! | 20:20 |
| rloo | NobodyCam: with the dogs? | 20:20 |
| NobodyCam | Not this time. Got sitter at home for them | 20:20 |
| jroll | rloo: dunno if we have the controls in LP to do that | 20:21 |
| rloo | jroll: probably not or someone would have done it already. I think. | 20:21 |
| jroll | you never know :) | 20:21 |
| rloo | NobodyCam: all roads lead to Austin. See you there! Safe trip! | 20:21 |
| jroll | NobodyCam: safe travels \o/ | 20:22 |
| *** ijw has joined #openstack-ironic | 20:25 | |
| TheJulia | 307952 got +A'ed about 5 minutes ago, headed for gate | 20:29 |
| jroll | woot. | 20:30 |
| jroll | I assume that's the tempest thing? | 20:30 |
| TheJulia | yup | 20:30 |
| jlvillal | Woo hoo: https://review.openstack.org/#/c/307952/ has got +2 and +A :) | 20:30 |
| jlvillal | Will fix our gate soon :) | 20:30 |
| TheJulia | yup | 20:31 |
| jlvillal | And TheJulia already said all that :) | 20:31 |
| JayF | if TheJulia and jroll wanna hit https://review.openstack.org/#/c/307474/ with their +2s again, and +A it, that'll fix IPA gate when the other merges (I set depends-on) | 20:31 |
| JayF | then maybe tomorrow morning we can actually merge changes \o/ | 20:32 |
| TheJulia | maybe! :) | 20:33 |
| jroll | bam | 20:33 |
| *** Sukhdev has joined #openstack-ironic | 20:33 | |
| jroll | yeah dude, that's like 16 hours from now, who knows what could happen | 20:33 |
| TheJulia | The queue is not bad at all at the moment | 20:33 |
| TheJulia | Well, post queue is bad | 20:33 |
| jlvillal | Yeah, gate queue isn't too bad. | 20:34 |
| jlvillal | And check queue isn't too bad at 4 hours. Since yesterday I saw one at 16 hours | 20:34 |
| TheJulia | jlvillal: I'd almost say surprisingly... good to be honest | 20:34 |
| jlvillal | :) | 20:35 |
| rloo | anyone know why the queues are so long? That doesn't seem to bode well, given my assumption that there isn't much activity/work just prior to the summit | 20:37 |
| TheJulia | rloo: one of the nodepool accounts disappeared earlier in the week from what I understand | 20:38 |
| TheJulia | so at a high, I think it was something like 39 hours | 20:38 |
| *** stacker has quit IRC | 20:40 | |
| jroll | there was also an image bug for a day that killed everything right before that | 20:40 |
| *** lucas-dinner has quit IRC | 20:40 | |
| *** rajinir has quit IRC | 20:45 | |
| *** lucasagomes has joined #openstack-ironic | 20:47 | |
| rloo | I suppose we should be thankful that this didn't happen the week of a release... | 20:47 |
| *** xavierr has quit IRC | 20:51 | |
| NobodyCam | Thank you thank you:-) see y'all there :-) | 20:53 |
| *** trown is now known as trown|outtypewww | 20:53 | |
| *** daemontool has quit IRC | 20:56 | |
| *** dprince has quit IRC | 20:57 | |
| jroll | devananda: hey, do you want to moderate the "managing hardware pools" session? :) | 20:58 |
| devananda | ohhai | 20:58 |
| devananda | jroll: so, I do not have any new information on that since the discussion at the last summit | 20:59 |
| jroll | devananda: at the midcycle we agreed to not move forward on a bunch of specs, with the promise we'd figure out an architectural opinion on it during the summit | 21:00 |
| jroll | and I was thinking you have more insight into these things than I do | 21:00 |
| jroll | as far as knowledge about the HW | 21:00 |
| devananda | fair 'nuf | 21:01 |
| *** baoli has quit IRC | 21:01 | |
| jroll | cool, thanks | 21:01 |
| devananda | I can moderate it, but I don't have a lot to add yet, as far as a proposed architecture | 21:01 |
| jroll | right | 21:01 |
| jroll | I asked for ideas int he pad but none have shown up yet :/ | 21:02 |
| devananda | compared to the driver composition one, where I've got a proposed architecture laid out already in the 'pad | 21:02 |
| devananda | I wonder if this is one we should punt on again :( | 21:02 |
| jroll | if nothing else, I can bring some drinks and we can just hang out | 21:02 |
| jroll | but really, we really need to stop kicking this down the road | 21:02 |
| devananda | I :( because punting makes me sad -- i know folks want it, but if no one's coming forward with a proposal, we have a lot of other things that do | 21:03 |
| jroll | right | 21:03 |
| openstackgerrit | Jarrod Johnson proposed openstack/pyghmi: Add system X firmware information https://review.openstack.org/307393 | 21:04 |
| jroll | so we'll call it a "brainstorming session" | 21:04 |
| rloo | jroll, devananda: maybe send email and ask? although there isn't much time til the summit. | 21:04 |
| jroll | or maybe enough vendors will show up and we can get some requirements? | 21:04 |
| jroll | rloo: yeah, we could | 21:04 |
| jroll | I still need to send an email about semver | 21:04 |
| jroll | maybe I'll do that after summit though | 21:04 |
| jroll | since we aren't releasing this week anyway | 21:05 |
| devananda | I feel like I have a grasp of the requirements based on the last two summits' discussions | 21:05 |
| rloo | jroll: i just looked and dtantsur|afk, sinval, gabriel-bezerra, thiagop, krtaylor and jroll said they were interested in HW pools. | 21:05 |
| devananda | but I do not have the desire to work on it myself, nor that type of hardware to test on | 21:05 |
| rloo | devananda: so a good start is to write down the requirements and make sure there is agreement on that at least | 21:05 |
| jroll | rloo: well, I'm interested in unblocking folks | 21:05 |
| jroll | rloo: ++ | 21:05 |
| jroll | devananda: could you do that? I'm not sure I have a good grasp on it | 21:06 |
| devananda | ack | 21:06 |
| jroll | thanks | 21:06 |
| jroll | https://etherpad.openstack.org/p/ironic-newton-summit-hardware-pools | 21:06 |
| jroll | for reference | 21:06 |
| devananda | anyone else wants to brainstorm with me, you're welcome to | 21:11 |
| *** Goneri has quit IRC | 21:13 | |
| devananda | looking back, I don't see notes from Liberty, but there's this from Mitaka: https://etherpad.openstack.org/p/summit-mitaka-ironic-group-management | 21:13 |
| *** jayc_ has quit IRC | 21:15 | |
| *** Sukhdev has quit IRC | 21:15 | |
| *** xek has quit IRC | 21:16 | |
| *** e0ne has quit IRC | 21:20 | |
| krtaylor | devananda, I can add some bits there too, have to finish up a couple of other things first | 21:23 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic-specs: Node search API spec https://review.openstack.org/306092 | 21:23 |
| *** jayc_ has joined #openstack-ironic | 21:27 | |
| devananda | krtaylor: so, i think what you just added -- molteniron -- is a completely different topic | 21:27 |
| krtaylor | sorry devananda, I dumped that in the middle of your SOC, I'll add some more in a bit, meeting... | 21:27 |
| krtaylor | I was going to ask, so this is groups not pools as in nodepool | 21:28 |
| devananda | krtaylor: this session is about ironic interacting with certain types of hardware that is grouped | 21:28 |
| devananda | eg, blade systems | 21:28 |
| devananda | or 4-node SOCs where you can not power off each system independently | 21:28 |
| krtaylor | devananda, gotcha, my bad, removing | 21:29 |
| devananda | each one has an IP address and PXE boots separately -- but they share firmware and power state | 21:29 |
| devananda | things like that, which we can't model today | 21:29 |
| *** xek has joined #openstack-ironic | 21:29 | |
| devananda | krtaylor: I'll clarify the wording at the top of the pad :) | 21:29 |
| krtaylor | devananda, thanks, there has been discussion on managing pools of HW for testing (nodepool++), I was confused | 21:30 |
| devananda | at least ,that's whta I think it's about .... jroll ? | 21:30 |
| jroll | devananda: correct | 21:30 |
| devananda | k k | 21:30 |
| * devananda continues | 21:30 | |
| krtaylor | I'll bring up managing physical systems for testing in the CI section of gate/qa | 21:31 |
| krtaylor | I was wondering why we needed a fishbowl for that :) | 21:31 |
| devananda | krtaylor: yah, that is a better forum for it | 21:33 |
| devananda | infra team knows they (will) need it | 21:33 |
| devananda | gabriel-bezerra: in reading this oneview spec: http://specs.openstack.org/openstack/ironic-specs/specs/not-implemented/oneview-drivers-dynamic-allocation.html | 21:34 |
| devananda | gabriel-bezerra: I do not see how that relates to the topic of ironic managing groups or pools of servers | 21:35 |
| devananda | but it was somehow added to the etherpad for that session | 21:35 |
| devananda | any ideas why? | 21:35 |
| devananda | jroll: ^ ? | 21:35 |
| devananda | krtaylor: actually, that spec is somewhat closer to the resource pool management you're talking about | 21:36 |
| devananda | tldr; making ironic better cooperate with an external pool manager (OneView) | 21:36 |
| krtaylor | devananda, it does | 21:37 |
| jroll | devananda: not sure, honestly | 21:38 |
| *** Sukhdev has joined #openstack-ironic | 21:38 | |
| rloo | devananda, jroll: I just remembered this spec, is it related to the hw pool? https://review.openstack.org/#/c/235344/ (We discussed this last summit i think) | 21:39 |
| jroll | rloo: yes | 21:39 |
| rloo | jroll: oh, then you just need to +2 it! (maybe not) | 21:40 |
| devananda | rloo: yes, and I've already added it to the pad | 21:40 |
| devananda | I'm going through all the spec proposals right now, in fact, looking for more unapproved ones about this topic | 21:40 |
| devananda | like https://review.openstack.org/134866 | 21:40 |
| rloo | devananda: thx | 21:40 |
| * jroll has to bounce for now, catch y'all later | 21:40 | |
| *** thrash is now known as thrash|g0ne | 21:41 | |
| *** e0ne has joined #openstack-ironic | 21:41 | |
| rloo | oh, chassis. sigh. I wonder if we'll get anywhere but yeah, (the royal) we probably need to do something. | 21:42 |
| *** Sukhdev has quit IRC | 21:43 | |
| devananda | rloo: heh | 21:44 |
| *** Sukhdev has joined #openstack-ironic | 21:47 | |
| jlvillal | Unleash the dogs of war! The gate should work now :) | 21:49 |
| *** e0ne has quit IRC | 21:51 | |
| rloo | jlvillal: YAY! | 21:53 |
| *** piet has joined #openstack-ironic | 21:57 | |
| mrda | Morning Ironic | 21:57 |
| mrda | And jlvillal, yay! | 21:58 |
| jlvillal | Hi mrda. Good morning :) | 21:58 |
| jlvillal | mrda when does your 20+ hour journey start? | 21:59 |
| * jlvillal assumes it is 20+ hours... | 21:59 | |
| *** Sukhdev has quit IRC | 21:59 | |
| mrda | jlvillal: I arrive in AUS 15 hours before conference opening :) | 22:00 |
| jlvillal | mrda you get to do like that Huey Lewis & The News song and "Go Back In Time" :) | 22:00 |
| mrda | jlvillal: so that means I leave Sunday morning 4:30am | 22:00 |
| jlvillal | Ouch. | 22:01 |
| jlvillal | Though I think I leave on Sunday at 7am | 22:01 |
| mrda | jlvillal: it's more like Cher's, "If I could turn back time" | 22:01 |
| jlvillal | :) | 22:01 |
| mrda | :) | 22:02 |
| *** Sukhdev has joined #openstack-ironic | 22:05 | |
| TheJulia | That moment you realize that you had the BMC for the power hog of a machine in the other room to resume previous power state... and that the BMC does not do that until it has booted up 1-2 minutes later... after walking out of the room. Doh. | 22:07 |
| devananda | mrda: ouch, no time to recover from jetlag :( | 22:11 |
| devananda | after reviewing open specs, I don't see any others | 22:11 |
| *** jaybeale has joined #openstack-ironic | 22:13 | |
| *** jaybeale has quit IRC | 22:16 | |
| *** jaybeale has joined #openstack-ironic | 22:16 | |
| *** rpioso has quit IRC | 22:21 | |
| *** suro-patz has quit IRC | 22:23 | |
| *** suro-patz has joined #openstack-ironic | 22:23 | |
| *** jayc_ has quit IRC | 22:25 | |
| openstackgerrit | Julia Kreger proposed openstack/ironic: Active Node Creation via adopt state https://review.openstack.org/275766 | 22:25 |
| openstackgerrit | Julia Kreger proposed openstack/ironic: Active Node Creation via adopt state https://review.openstack.org/275766 | 22:26 |
| mrda | devananda: unfortunately no :( | 22:29 |
| *** ayoung has joined #openstack-ironic | 22:33 | |
| *** jaybeale has quit IRC | 22:36 | |
| *** UForgotten has quit IRC | 22:48 | |
| *** UForgotten has joined #openstack-ironic | 22:48 | |
| devananda | ok - think I'm done summarizing things on https://etherpad.openstack.org/p/ironic-newton-summit-hardware-pools | 22:50 |
| *** krotscheck is now known as krotscheck_dcm | 22:57 | |
| *** rbudden has quit IRC | 22:59 | |
| *** piet has quit IRC | 23:03 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ironic: Updated from global requirements https://review.openstack.org/306942 | 23:13 |
| openstackgerrit | Merged openstack/ironic-python-agent: Force container builds to use http.us.debian.org https://review.openstack.org/307474 | 23:15 |
| TheJulia | hmmmm :) | 23:16 |
| *** ozialien has joined #openstack-ironic | 23:19 | |
| *** Sukhdev has quit IRC | 23:22 | |
| *** jaypipes has quit IRC | 23:29 | |
| *** mat128 has quit IRC | 23:29 | |
| *** mat128 has joined #openstack-ironic | 23:30 | |
| * TheJulia very cautiously suspects the gate might be happier :) | 23:34 | |
| *** Sukhdev has joined #openstack-ironic | 23:41 | |
| *** penick has quit IRC | 23:45 | |
| devananda | JayF: question on agent lookup() method - I thought this could only be called during DEPLOYWAIT | 23:47 |
| devananda | JayF: but in reading the code, I am not getting confirmation of my belief. Am I wrong? | 23:48 |
| *** Sukhdev has quit IRC | 23:48 | |
| jlvillal | TheJulia: Gate does appear to be happier. Looking at a job in zuul. | 23:51 |
| TheJulia | yup, two so far have passed | 23:51 |
| TheJulia | should be 3 shortly | 23:51 |
| jlvillal | That was a relatively long breakage on the gate. | 23:51 |
| TheJulia | Yeah :( | 23:52 |
| jlvillal | TheJulia: Do you want to update the etherpad or should I? | 23:52 |
| TheJulia | If you would that would be awesome, it is dinner time here | 23:52 |
| jlvillal | TheJulia: will do :) | 23:52 |
| jlvillal | Bon appetit | 23:52 |
| TheJulia | jlvillal: thank you, and have a good evening | 23:52 |
| * TheJulia goes off to nom pizza | 23:52 | |
| *** rloo has quit IRC | 23:53 | |
| *** ChrisAusten has joined #openstack-ironic | 23:55 | |
| openstackgerrit | Merged openstack/ironic: Make sure Cisco drivers are documented on IRONIC_DEPLOY_DRIVER https://review.openstack.org/305868 | 23:56 |
| jlvillal | FYI: Etherpad updated that gate is currently golden | 23:57 |
| *** yuikotakada has quit IRC | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!