| adam_g | wrong review heh | 00:00 |
|---|---|---|
| adam_g | https://review.openstack.org/#/c/81958/ | 00:00 |
| devananda | um | 00:00 |
| devananda | so tempest already does taht i thought | 00:00 |
| adam_g | devananda, does what? | 00:00 |
| adam_g | the API tests in tempest do basic CRUD testing, with or withou the virt driver | 00:00 |
| devananda | right | 00:01 |
| adam_g | devananda, i'd like to have everything that scenario test running and passing with the rest of whats green in the results pasted above | 00:01 |
| devananda | ahh | 00:01 |
| devananda | second review | 00:01 |
| devananda | :) | 00:01 |
| adam_g | or at least 1) that scenario test 2) the tempest.api.baremetal.* tests | 00:02 |
| adam_g | we can target just those easily with a regexp in devstack-gate-vm.sh | 00:02 |
| devananda | adam_g: side note, it'd be great if you tag me in the reviews so I know about them :) | 00:02 |
| devananda | i don't follow all of devstack's review queue | 00:02 |
| adam_g | devananda, im happy to. didn't know of such feature | 00:02 |
| devananda | ahh | 00:03 |
| adam_g | oh there it | 00:03 |
| * devananda points to "Add reviewer" button | 00:03 | |
| adam_g | "Add Reviewer" | 00:03 |
| adam_g | :) | 00:03 |
| devananda | :) | 00:03 |
| devananda | thanks! | 00:03 |
| adam_g | devananda, anyway, if that scenario test gets merged that gives us basic functional testing of a nova-driven provision | 00:04 |
| adam_g | we can add an _advanced_ops if we want to test some of the things like ma addr update, etc | 00:04 |
| adam_g | *mac addr | 00:04 |
| devananda | awesome | 00:04 |
| devananda | so | 00:04 |
| devananda | there's still a problem | 00:05 |
| devananda | AIUI, if tempest is enabled -at-all- it will test any service it discovers | 00:05 |
| adam_g | the issue with the other tests that are failing now are problems like parallel testing (can be turned off), incorrect tenants (random tenants created for each test. we need to spawn on tenant /w the same network as nodes) | 00:05 |
| devananda | it sounds like we need to prevent tempest running any of the nova tests while nova is using the virt.ironic driver | 00:06 |
| devananda | which we can't do | 00:06 |
| adam_g | devananda, well, long term it would be great to be able to define feature requirements for individual tests--thats another conversation | 00:06 |
| adam_g | in the meantime | 00:06 |
| adam_g | http://git.openstack.org/cgit/openstack-infra/devstack-gate/tree/devstack-vm-gate.sh#n330 | 00:06 |
| adam_g | should let us easily target specific tests, or blacklist others | 00:07 |
| devananda | interesting | 00:07 |
| devananda | i didnt know about that | 00:07 |
| adam_g | devananda, another possible reason why we dont have more green is the issues around deleting an instance that hasn't been fully provisioned on the ironic side | 00:10 |
| adam_g | its possible an early test boots something, it fails during provision, test deletes it and moves on.. meanwhile, ironic node is never freed and all subsequent tests can't boot instances | 00:11 |
| devananda | adam_g: right | 00:12 |
| devananda | adam_g: and the bit of reading in that log that I did suggests the conductor never got the do_node_deploy message at all | 00:13 |
| adam_g | devananda, it could be the random tenant spawning the instance does not have appropriate config in neutron and the driver is dying during vif setup? | 00:14 |
| devananda | adam_g: ahh. could be | 00:15 |
| devananda | so an instance gets created and associated in ironic | 00:15 |
| devananda | but the virt driver never makes it to deploy | 00:15 |
| devananda | taht seems likely | 00:15 |
| devananda | then it tries to clean up, but ironic rejects taht request | 00:15 |
| devananda | adam_g: looks like DEVSTACK_GATE_TEMPEST_REGEX is the way to go for now | 00:16 |
| devananda | adam_g: do you know how to trigger the virtual-ironic job in tempest's queue? | 00:17 |
| devananda | it should be available in tempest/devstack/devstack-gate, just not running by default yet | 00:17 |
| adam_g | devananda, oh looking closer at those logs, its using the wrong flavor | 00:17 |
| adam_g | http://logs.openstack.org/31/83231/1/check/check-tempest-dsvm-virtual-ironic/ff61a33/logs/screen-n-cpu.txt.gz?level=TRACE | 00:17 |
| adam_g | https://review.openstack.org/#/c/82703/ | 00:18 |
| *** matsuhashi has joined #openstack-ironic | 00:19 | |
| devananda | yep | 00:20 |
| * adam_g needs to run out for 20 mins back soon | 00:20 | |
| NobodyCam | :) run fast | 00:20 |
| NobodyCam | hehehe | 00:20 |
| devananda | i will be wrapping up soon too | 00:21 |
| NobodyCam | :) | 00:21 |
| NobodyCam | I too will be doing the same shortly | 00:22 |
| NobodyCam | trying to see why my compute_manager setting is not getting set. | 00:23 |
| *** lucas-afk has quit IRC | 00:25 | |
| *** epim has quit IRC | 00:26 | |
| *** rloo has quit IRC | 00:36 | |
| *** rloo has joined #openstack-ironic | 00:37 | |
| *** rloo has quit IRC | 00:45 | |
| *** rloo has joined #openstack-ironic | 00:45 | |
| *** blamar has joined #openstack-ironic | 00:57 | |
| *** eghobo has quit IRC | 01:02 | |
| *** matsuhashi has quit IRC | 01:12 | |
| *** harlowja has quit IRC | 01:12 | |
| *** zigo has quit IRC | 01:12 | |
| *** zul has quit IRC | 01:12 | |
| *** jrist has quit IRC | 01:12 | |
| *** john-n-seattle has quit IRC | 01:12 | |
| *** Haomeng has quit IRC | 01:12 | |
| *** harlowja has joined #openstack-ironic | 01:17 | |
| *** rloo has quit IRC | 01:34 | |
| *** eghobo has joined #openstack-ironic | 01:41 | |
| *** nosnos has joined #openstack-ironic | 01:45 | |
| *** blamar has quit IRC | 01:45 | |
| *** eghobo has quit IRC | 01:45 | |
| *** matsuhashi has joined #openstack-ironic | 01:49 | |
| *** zigo has joined #openstack-ironic | 01:49 | |
| *** zul has joined #openstack-ironic | 01:49 | |
| *** jrist has joined #openstack-ironic | 01:49 | |
| *** john-n-seattle has joined #openstack-ironic | 01:49 | |
| *** hemna_ has quit IRC | 01:53 | |
| openstackgerrit | A change was merged to openstack/ironic: Do not use __builtin__ in python3 https://review.openstack.org/80144 | 01:58 |
| *** harlowja is now known as harlowja_away | 02:19 | |
| *** vkozhukalov has joined #openstack-ironic | 02:30 | |
| *** lifeless has joined #openstack-ironic | 02:30 | |
| lifeless | NobodyCam: I enabled tlh for Ubuntu years ago | 02:34 |
| lifeless | NobodyCam: you should totally be able to do it | 02:35 |
| *** lifeless has quit IRC | 02:35 | |
| *** harlowja_away is now known as harlowja | 02:37 | |
| *** matsuhashi has quit IRC | 02:37 | |
| *** matsuhashi has joined #openstack-ironic | 02:38 | |
| *** matsuhashi has quit IRC | 02:47 | |
| *** killer_prince has quit IRC | 02:50 | |
| *** Haomeng has joined #openstack-ironic | 02:51 | |
| *** killer_prince has joined #openstack-ironic | 02:51 | |
| *** lifeless has joined #openstack-ironic | 02:53 | |
| *** nosnos has quit IRC | 03:04 | |
| *** linggao has quit IRC | 03:05 | |
| *** linggao has joined #openstack-ironic | 03:30 | |
| *** blamar has joined #openstack-ironic | 03:35 | |
| *** blamar has quit IRC | 03:36 | |
| *** harlowja is now known as harlowja_away | 03:47 | |
| *** nosnos has joined #openstack-ironic | 03:48 | |
| *** matsuhashi has joined #openstack-ironic | 03:48 | |
| *** eghobo has joined #openstack-ironic | 04:20 | |
| *** saju_m has joined #openstack-ironic | 04:38 | |
| *** rameshg87 has joined #openstack-ironic | 04:40 | |
| *** matsuhashi has quit IRC | 04:45 | |
| *** matsuhas_ has joined #openstack-ironic | 04:49 | |
| *** matsuhas_ has quit IRC | 05:12 | |
| *** pradipta_away is now known as pradipta | 05:19 | |
| *** loki_ has joined #openstack-ironic | 05:25 | |
| *** matsuhashi has joined #openstack-ironic | 05:29 | |
| *** sabah has joined #openstack-ironic | 05:57 | |
| openstackgerrit | Jenkins proposed a change to openstack/ironic: Imported Translations from Transifex https://review.openstack.org/82993 | 06:07 |
| *** saju_m has quit IRC | 06:07 | |
| *** killer_prince is now known as lazy_prince | 06:07 | |
| *** dwalleck has joined #openstack-ironic | 06:08 | |
| *** dwalleck has quit IRC | 06:08 | |
| *** vkozhukalov has quit IRC | 06:10 | |
| *** mrda is now known as mrda_away | 06:18 | |
| *** russellb has quit IRC | 06:32 | |
| *** killer_prince has joined #openstack-ironic | 06:36 | |
| *** saju_m has joined #openstack-ironic | 06:50 | |
| *** max_lobur has joined #openstack-ironic | 07:22 | |
| *** ndipanov has joined #openstack-ironic | 07:24 | |
| *** sabah has quit IRC | 07:26 | |
| *** max_lobur has quit IRC | 07:26 | |
| *** sirushti is now known as shortstop | 07:27 | |
| *** eghobo has quit IRC | 07:28 | |
| *** max_lobur has joined #openstack-ironic | 07:30 | |
| *** subah has joined #openstack-ironic | 07:30 | |
| *** subah has quit IRC | 07:32 | |
| *** sabah has joined #openstack-ironic | 07:34 | |
| *** vkozhukalov has joined #openstack-ironic | 07:38 | |
| openstackgerrit | Sergey Lukjanov proposed a change to openstack/ironic: Start using oslosphinx theme for docs https://review.openstack.org/83043 | 07:41 |
| *** jistr has joined #openstack-ironic | 07:47 | |
| *** russellb has joined #openstack-ironic | 07:47 | |
| *** greghaynes has quit IRC | 08:02 | |
| *** athomas has joined #openstack-ironic | 08:02 | |
| *** athomas has quit IRC | 08:05 | |
| *** greghaynes has joined #openstack-ironic | 08:07 | |
| *** ifarkas has joined #openstack-ironic | 08:15 | |
| *** sabah has quit IRC | 08:29 | |
| *** matsuhashi has quit IRC | 08:33 | |
| *** max_lobur has quit IRC | 08:34 | |
| *** saju_m has quit IRC | 08:43 | |
| *** nosnos has quit IRC | 08:43 | |
| *** mdurnosvistov_ has joined #openstack-ironic | 08:44 | |
| dtantsur | Morning Ironic | 08:46 |
| mdurnosvistov_ | dtantsur: morning :) | 08:50 |
| mdurnosvistov_ | Morning all! ;) | 08:50 |
| Haomeng | morning all:) | 08:50 |
| *** saju_m has joined #openstack-ironic | 08:55 | |
| *** saju_m has quit IRC | 09:16 | |
| *** max_lobur has joined #openstack-ironic | 09:17 | |
| *** max_lobur1 has joined #openstack-ironic | 09:17 | |
| *** sabah has joined #openstack-ironic | 09:20 | |
| openstackgerrit | Andrey Kurilin proposed a change to openstack/python-ironicclient: Sync cliutils from oslo https://review.openstack.org/72417 | 09:21 |
| *** max_lobur has quit IRC | 09:21 | |
| *** derekh has joined #openstack-ironic | 09:24 | |
| openstackgerrit | Andrey Kurilin proposed a change to openstack/python-ironicclient: Sync cliutils from oslo https://review.openstack.org/72417 | 09:24 |
| *** saju_m has joined #openstack-ironic | 09:27 | |
| lifeless | adam_g: devananda: https://bugs.launchpad.net/ironic/+bug/1298274 may interest | 09:36 |
| *** jistr has quit IRC | 09:38 | |
| *** matsuhashi has joined #openstack-ironic | 09:45 | |
| openstackgerrit | A change was merged to openstack/ironic: Fix strings post landing for c63e1d9f6 https://review.openstack.org/83044 | 09:46 |
| *** jistr has joined #openstack-ironic | 10:04 | |
| *** martyntaylor has joined #openstack-ironic | 10:04 | |
| *** saju_m has quit IRC | 10:24 | |
| *** EmilienM has quit IRC | 10:28 | |
| *** EmilienM has joined #openstack-ironic | 10:28 | |
| *** martyntaylor has quit IRC | 10:31 | |
| *** lazy_prince has quit IRC | 10:34 | |
| *** Manishanker has joined #openstack-ironic | 10:44 | |
| *** nosnos has joined #openstack-ironic | 10:44 | |
| *** martyntaylor has joined #openstack-ironic | 10:51 | |
| *** romcheg has joined #openstack-ironic | 11:04 | |
| *** matsuhashi has quit IRC | 11:07 | |
| openstackgerrit | Sandhya Balakrishnan proposed a change to openstack/ironic: Ironic User Guide https://review.openstack.org/83343 | 11:11 |
| *** saju_m has joined #openstack-ironic | 11:14 | |
| *** ifarkas has quit IRC | 11:18 | |
| *** sabah has quit IRC | 11:29 | |
| *** matsuhashi has joined #openstack-ironic | 11:32 | |
| *** saju_m has quit IRC | 11:33 | |
| *** saju_m has joined #openstack-ironic | 11:35 | |
| *** ifarkas has joined #openstack-ironic | 11:45 | |
| *** matsuhashi has quit IRC | 11:45 | |
| *** lynxman has quit IRC | 11:48 | |
| *** lynxman has joined #openstack-ironic | 11:48 | |
| *** athomas has joined #openstack-ironic | 12:00 | |
| openstackgerrit | Aleksandr Gordeev proposed a change to openstack/ironic-python-agent: Get rid of modes. Introduce pluggable extensions https://review.openstack.org/83087 | 12:02 |
| agordeev | good morning Ironic! | 12:05 |
| *** matsuhashi has joined #openstack-ironic | 12:05 | |
| Shrews | agordeev: morning | 12:29 |
| * Shrews is slow in responding until at least his 2nd cup of coffee | 12:30 | |
| *** jdob has joined #openstack-ironic | 12:30 | |
| agordeev | Shrews: morning :) | 12:31 |
| jroll | morning agordeev :) | 12:32 |
| romcheg | Morning follks! | 12:32 |
| mdurnosvistov_ | agordeev: morning/afternoon =) | 12:32 |
| mdurnosvistov_ | romcheg: morning! | 12:32 |
| romcheg | Guys, I'm not familiar with building dos in OpenStack so I have a question about this patch https://review.openstack.org/#/c/83043/2/doc/source/conf.py | 12:33 |
| jroll | morning all :) | 12:33 |
| romcheg | I would appreciate if some one explains me where the oslosphinx theme is set to be used? | 12:33 |
| *** martyntaylor has quit IRC | 12:33 | |
| romcheg | I can only see that theme settings were commented on L64-L66 | 12:34 |
| jroll | agordeev: looking over this patch right now btw | 12:35 |
| Shrews | romcheg: i believe that's all handled in pbr | 12:37 |
| Shrews | romcheg: http://docs.openstack.org/developer/pbr/#sphinx-autodoc | 12:37 |
| romcheg | Shrews: Thanks I'll check that out | 12:37 |
| *** rloo has joined #openstack-ironic | 12:42 | |
| *** matsuhashi has quit IRC | 12:45 | |
| *** matsuhashi has joined #openstack-ironic | 12:45 | |
| *** loki_ has quit IRC | 12:48 | |
| *** martyntaylor has joined #openstack-ironic | 12:49 | |
| *** killer_prince has quit IRC | 12:49 | |
| *** matsuhashi has quit IRC | 12:50 | |
| agordeev | jroll: morning! | 12:55 |
| agordeev | romcheg: mdurnosvistov_ morning, guys! | 12:55 |
| romcheg | Morning agordeev | 12:55 |
| jroll | agordeev: hey :) | 12:56 |
| jroll | agordeev: I've got one thing on your patch, trying to find the best way to explain it | 13:00 |
| jroll | the tests happened to match the bug as well | 13:00 |
| * jroll needs more coffee | 13:00 | |
| *** pradipta is now known as pradipta_away | 13:01 | |
| *** ifarkas has quit IRC | 13:02 | |
| agordeev | jroll: what thing? let me know | 13:06 |
| jroll | agordeev: just put up a review | 13:07 |
| jroll | agordeev: I'm planning on working on some integration things today, so if we can get that merged this morning, I can (if everything else goes okay) get some real usage on it :) | 13:08 |
| *** matty_dubs|gone is now known as matty_dubs | 13:15 | |
| jroll | agordeev: by the wat, let me know if that doesn't make sense | 13:15 |
| *** matsuhashi has joined #openstack-ironic | 13:18 | |
| *** matsuhashi has quit IRC | 13:25 | |
| *** matsuhashi has joined #openstack-ironic | 13:26 | |
| *** jistr has quit IRC | 13:26 | |
| *** matsuhashi has quit IRC | 13:26 | |
| *** matsuhashi has joined #openstack-ironic | 13:26 | |
| *** max_lobur1 has quit IRC | 13:28 | |
| *** mdurnosvistov has quit IRC | 13:28 | |
| openstackgerrit | Aleksandr Gordeev proposed a change to openstack/ironic-python-agent: Get rid of modes. Introduce pluggable extensions https://review.openstack.org/83087 | 13:29 |
| agordeev | jroll: yup, you are right. It was a bug :) That tells the reason why i can't get access to ext.obj before | 13:30 |
| jroll | heh, yeah :) | 13:30 |
| jroll | checking it out now | 13:31 |
| jroll | lgtm, thanks agordeev! | 13:31 |
| *** mdurnosvistov has joined #openstack-ironic | 13:31 | |
| jroll | as soon as jenkins gets done I'll +2 it | 13:32 |
| *** blamar has joined #openstack-ironic | 13:32 | |
| *** jbjohnso_ has joined #openstack-ironic | 13:33 | |
| agordeev | jroll: np, thanks for lightning fast reviewing :) | 13:34 |
| jroll | I do my best :) | 13:35 |
| *** krtaylor has quit IRC | 13:35 | |
| jroll | does anyone know how to watch a project for new reviews? | 13:36 |
| openstackgerrit | A change was merged to openstack/ironic: Start using oslosphinx theme for docs https://review.openstack.org/83043 | 13:40 |
| *** linggao has joined #openstack-ironic | 13:42 | |
| *** jistr has joined #openstack-ironic | 13:44 | |
| rloo | jroll. I think you can set it in your 'Settings' (at https://review.openstack.org/#/settings/projects) | 13:45 |
| *** lucasagomes has joined #openstack-ironic | 13:46 | |
| *** Manishanker has left #openstack-ironic | 13:46 | |
| devananda | morning, all | 13:59 |
| *** jdob has quit IRC | 14:01 | |
| rloo | hi devananda. | 14:01 |
| linggao | Good morning devananda rloo | 14:01 |
| *** zigo has quit IRC | 14:01 | |
| *** jdob has joined #openstack-ironic | 14:02 | |
| agordeev | devananda: morning! | 14:02 |
| agordeev | rloo linggao good morning! | 14:02 |
| linggao | agordeev, good morning. | 14:02 |
| rloo | hello linggao & agordeev :-) | 14:02 |
| *** zigo has joined #openstack-ironic | 14:03 | |
| *** mtaylor has joined #openstack-ironic | 14:05 | |
| *** rameshg87 has left #openstack-ironic | 14:06 | |
| *** krtaylor has joined #openstack-ironic | 14:06 | |
| *** blamar has quit IRC | 14:06 | |
| *** blamar has joined #openstack-ironic | 14:06 | |
| *** mordred has quit IRC | 14:08 | |
| *** mtaylor has quit IRC | 14:08 | |
| *** mtaylor has joined #openstack-ironic | 14:08 | |
| *** mtaylor is now known as mordred | 14:08 | |
| NobodyCam | good morning Ironic says the man making coffee.. | 14:10 |
| agordeev | NobodyCam: morning! | 14:10 |
| NobodyCam | morning agordee, rloo, linggao and devananda :) | 14:11 |
| rloo | hiya NobodyCam! | 14:11 |
| lucasagomes | morning NobodyCam rloo agordeev | 14:11 |
| linggao | Hi NobodyCam. | 14:12 |
| lucasagomes | morning devananda linggao :) | 14:12 |
| NobodyCam | ahh didn't see ya there lucasagomes :) | 14:12 |
| lucasagomes | no worries | 14:12 |
| NobodyCam | :) | 14:12 |
| linggao | Good morning lucasagomes | 14:12 |
| agordeev | lucasagomes: morning! | 14:13 |
| linggao | lucasagomes, do you know if this doc is up to late or not? http://docs.openstack.org/developer/ironic/webapi/v1.html | 14:13 |
| linggao | late->date | 14:13 |
| lucasagomes | linggao, I think it's cause it's auto-generated | 14:14 |
| linggao | lucasagomes, that's great. someone asked me and I was not sure. | 14:15 |
| *** blamar has quit IRC | 14:17 | |
| *** blamar has joined #openstack-ironic | 14:17 | |
| *** max_lobur1 has joined #openstack-ironic | 14:19 | |
| *** blamar has quit IRC | 14:21 | |
| lucasagomes | yup :D | 14:21 |
| *** blamar has joined #openstack-ironic | 14:21 | |
| NobodyCam | :) coffee started, kids quickly walked, /me dragged behind :-/ | 14:27 |
| rloo | NobodyCam: you have to show them who's boss! | 14:28 |
| NobodyCam | I do...THay are :-p | 14:29 |
| NobodyCam | hehehe | 14:29 |
| rloo | :-) | 14:31 |
| *** killer_prince has joined #openstack-ironic | 14:32 | |
| *** thehybridtech has joined #openstack-ironic | 14:33 | |
| NobodyCam | off the workers here at the park carries "cookies" (dob bones) they ran up to his golf cart and I wasn;t ready for it.. lol | 14:33 |
| NobodyCam | s/off/ one of/ | 14:33 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Add DiskPartitioner https://review.openstack.org/83396 | 14:34 |
| rloo | ha ha | 14:34 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Use DiskPartitioner https://review.openstack.org/83399 | 14:36 |
| *** jgrimm has joined #openstack-ironic | 14:36 | |
| NobodyCam | lucasagomes: what do you think about making a common dir under test for testing file in common? woth it or over kill? | 14:37 |
| jroll | rloo: thanks! | 14:37 |
| rloo | jroll: yw | 14:38 |
| lucasagomes | NobodyCam, I think it's needed heh | 14:38 |
| * NobodyCam goes bace to coffee.. as typing is not working well yet | 14:38 | |
| NobodyCam | gah | 14:38 |
| NobodyCam | :) | 14:38 |
| lucasagomes | so the test dir structure looks like the ironic dir structure | 14:38 |
| lucasagomes | today's a tad confusing | 14:38 |
| lucasagomes | (not a big deal tho) | 14:38 |
| NobodyCam | ya.. I was looking for a test of one of the common files the over day and it ended up a bug hunt | 14:40 |
| lucasagomes | lol | 14:49 |
| lucasagomes | yeah | 14:49 |
| openstackgerrit | A change was merged to openstack/ironic-python-agent: Get rid of modes. Introduce pluggable extensions https://review.openstack.org/83087 | 15:02 |
| NobodyCam | brb | 15:10 |
| *** matsuhashi has quit IRC | 15:18 | |
| *** nosnos has quit IRC | 15:19 | |
| *** datajerk has quit IRC | 15:20 | |
| *** dhellmann_ is now known as dhellmann | 15:21 | |
| *** matty_dubs is now known as matty_dubs|czech | 15:26 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Use DiskPartitioner https://review.openstack.org/83399 | 15:31 |
| rloo | lucasagomes: so we don't have any docn for the ironic CLI, right? something like http://docs.openstack.org/developer/python-keystoneclient/? | 15:35 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Use DiskPartitioner https://review.openstack.org/83399 | 15:35 |
| lucasagomes | rloo, no :( | 15:36 |
| devananda | lucasagomes: have you gotten API translations to work at all? | 15:36 |
| devananda | I've been trying this morning with no success | 15:36 |
| lucasagomes | devananda, to be honest I never tried | 15:36 |
| devananda | based on https://review.openstack.org/#/c/74626/ | 15:37 |
| devananda | but still doesn't work | 15:37 |
| devananda | but ... i may be doing it wrong | 15:37 |
| lucasagomes | hmm lemme try, finishing something before and then I will give it a go | 15:37 |
| lucasagomes | devananda, there's anyone working on translation? | 15:38 |
| devananda | it might be worth trying on a system which already has default locale != en | 15:38 |
| lucasagomes | maybe the strings are simply not translated to they fallback to english | 15:38 |
| lucasagomes | so they* | 15:38 |
| devananda | lucasagomes: yes - -https://www.transifex.com/projects/p/ironic/ | 15:38 |
| lucasagomes | devananda, ah wow! brazilian ptgz 46%! | 15:38 |
| NobodyCam | lucasagomes: also: #openstack-translation | 15:39 |
| Shrews | how do we not have a Tagalog translation yet????? | 15:40 |
| devananda | ahhh | 15:40 |
| devananda | so | 15:40 |
| devananda | i think something's broken with our transifex integration | 15:40 |
| devananda | the translations exist in transifex but NOT in master ( | 15:41 |
| lucasagomes | ohh | 15:43 |
| lucasagomes | brazilian ptgz is now 47% :) did my 1% contribution heh | 15:43 |
| lucasagomes | yay | 15:43 |
| *** saju_m has quit IRC | 15:44 | |
| devananda | hehe | 15:44 |
| devananda | Shrews: klingon. we need a klingon translation. | 15:44 |
| lucasagomes | Haomeng, ping re lazy translation, how did u test it? | 15:44 |
| lucasagomes | devananda, +1! | 15:44 |
| lucasagomes | irish is not even listed there :( | 15:45 |
| *** killer_prince is now known as lazy_prince | 15:45 | |
| NobodyCam | that the tlh I keep wanting | 15:47 |
| *** ewindisch has quit IRC | 15:47 | |
| Shrews | devananda: want an update on https://review.openstack.org/83105? or after coffee/other stuff? | 15:47 |
| *** hemna_ has joined #openstack-ironic | 15:47 | |
| devananda | mmm, yes. also need moar coffee | 15:48 |
| Shrews | devananda: So, what I've put up is working like a champ under manual testing with devstack. get_available_nodes() is called by n-cpu every minute via a periodic task... | 15:49 |
| Shrews | devananda: I can kill the API server and watch it retry, and bring it up in between tries and it works again. | 15:49 |
| Shrews | So functionally, I'm feeling pretty confident. | 15:50 |
| *** ewindisch has joined #openstack-ironic | 15:51 | |
| Shrews | I've limited the scope of changes in that review just for testing. | 15:51 |
| devananda | sounds good | 15:53 |
| devananda | Shrews: feel like adding a second patch, based on that, to do retry for things like delete? | 15:54 |
| devananda | Shrews: also, for reuse, I'd put IronicClientWrapper in its own file | 15:55 |
| Shrews | devananda: well, that brings up my next question: fully implementing this would change LOTS of areas of code. Should we implement this piecemeal, or as one large patch? | 15:55 |
| Shrews | s/LOTS of areas/LOTS of methods/ | 15:56 |
| devananda | oh, and retry_excs should have HTTPConflict too, I /think/ | 15:56 |
| Shrews | Since we don't really have unit tests, I'm not real comfortable with changing everything | 15:57 |
| devananda | Shrews: piecemeal is almost always easier on reviewers | 15:57 |
| devananda | yea | 15:57 |
| devananda | and we dont have functional tests yet of this code either | 15:57 |
| Shrews | devananda: ++ on conflict | 15:57 |
| devananda | though adam_g seems to be getting pretty close | 15:57 |
| devananda | Shrews: conflict should address teh "NodeLocked" exceptions | 15:57 |
| devananda | which will be a real issue in production where periodic-tasks in ir-cond are firing off | 15:57 |
| devananda | oh, also | 15:58 |
| devananda | Shrews: take a look at my comment on https://review.openstack.org/#/c/82948/ | 15:58 |
| devananda | i think i found why the client isn't handling some exceptions | 15:58 |
| devananda | lucasagomes: ^ | 15:59 |
| Shrews | devananda: so, I'm just going to pick a handful of methods to change, test those through devstack, and open the review up | 15:59 |
| devananda | Shrews: ++ | 15:59 |
| Shrews | devananda: i tried running the unit tests directly under nova using your tests from https://review.openstack.org/51328, but that didn't work and i jumped into using devstack before looking into that too much | 16:01 |
| Shrews | just fyi | 16:01 |
| devananda | didn't work how? | 16:01 |
| Shrews | testr threw up all over my laptop | 16:01 |
| NobodyCam | icky | 16:02 |
| Shrews | some sort of long, undecipherable message that made me hide in a corner | 16:02 |
| devananda | hah | 16:02 |
| devananda | that's usually a parse error | 16:02 |
| rloo | jroll: do you remember the diagram you guys had at the midcycle meet-up? | 16:02 |
| *** eghobo has joined #openstack-ironic | 16:03 | |
| lucasagomes | devananda, I think we shouldn't address Conflict | 16:03 |
| lucasagomes | NodeLocked is TemporaryFailure | 16:03 |
| devananda | lucasagomes: https://github.com/openstack/python-ironicclient/blob/master/ironicclient/exc.py | 16:03 |
| devananda | is incomplete | 16:03 |
| jroll | rloo: we had lots of diagrams :P I think the one you're looking for should be somewhere here: https://wiki.openstack.org/wiki/Ironic-python-agent#Architecture | 16:03 |
| lucasagomes | devananda, ah | 16:04 |
| lucasagomes | right | 16:04 |
| devananda | woops, bad link. should have had L150 in there | 16:04 |
| devananda | https://github.com/openstack/python-ironicclient/blob/master/ironicclient/exc.py#L151 | 16:04 |
| lucasagomes | devananda, I think there was a patch before to use exceptions from oslo to the clients | 16:04 |
| devananda | the mapping is at the end of that file | 16:04 |
| devananda | yea, it wasn't based properly // never landed | 16:04 |
| lucasagomes | but we hold it because it would affect the driver and consenquently the graduation | 16:04 |
| lucasagomes | but now it might be good to land it | 16:04 |
| lucasagomes | right | 16:04 |
| devananda | lucasagomes: how would it help this situation? | 16:05 |
| rloo | jroll: thx. I don't actually know which diagram, it was something you guys had written on the board. I didn't look :-( I'll take what I can get though ;) | 16:06 |
| jroll | rloo: what sort of thing are you looking for? | 16:06 |
| lucasagomes | well idk how complete is the oslo exceptions for the clients | 16:06 |
| rloo | jroll: *any*, I'm desperate. Just joking. Trying to put together docn for ironic. | 16:07 |
| jroll | heh | 16:07 |
| jroll | rloo: I'm happy to draw things up if those don't suffice :) | 16:07 |
| JayF | Just let any of us know if there are questions or you want more :) | 16:08 |
| rloo | jroll: deva had mentioned the diagram you guys drew. i think it might have had sequence/flow, so the IPA-Startup-Flow diagram might be useful, w/o the teeth part perhaps. | 16:08 |
| jroll | it says teeth? :/ | 16:08 |
| devananda | rloo: sounds like you want a "what we have today" vs "what we're planning" diagram | 16:08 |
| jroll | oh, we changed that name | 16:09 |
| * jroll edits | 16:09 | |
| NobodyCam | rloo: hope i wasn't overly critical in my review | 16:09 |
| rloo | devananda: yeah, well, I want a 'what we will have when RC goes out' which should be similar to what we have today. | 16:09 |
| devananda | right | 16:09 |
| devananda | and IPA isn't part of that | 16:09 |
| JayF | None of the agent stuff will be in icehouse at all | 16:09 |
| *** epim has joined #openstack-ironic | 16:09 | |
| lucasagomes | rloo, https://github.com/ifarkas/TuskarSummit2014March-Ironic/tree/master/images | 16:09 |
| rloo | thx lucasagomes. | 16:10 |
| rloo | anyone want to try updating this: http://docs.openstack.org/admin-guide-cloud/content/conceptual-architecture.html? | 16:11 |
| devananda | https://github.com/ifarkas/TuskarSummit2014March-Ironic/blob/master/images/PXEandIPMIFlow.png is slightly misleading, IMHO | 16:11 |
| rloo | or how about adding ironic to this: http://docs.openstack.org/admin-guide-cloud/content/logical-architecture.html | 16:11 |
| devananda | lucasagomes: first two lines should be reflexive | 16:11 |
| *** dshulyak has quit IRC | 16:11 | |
| *** dshulyak has joined #openstack-ironic | 16:11 | |
| lucasagomes | devananda, +1 | 16:12 |
| lucasagomes | well I've the google docs files I can share it and we fix it there | 16:12 |
| lucasagomes | and export to png again | 16:12 |
| lucasagomes | or use another tool to recreate then | 16:12 |
| devananda | there's also http://git.openstack.org/cgit/openstack/tripleo-incubator/tree/images | 16:13 |
| lucasagomes | https://www.draw.io/ looks quite good | 16:13 |
| devananda | which may or may not be out of date now | 16:13 |
| rloo | NobodyCam: nope, not overly critical. thx for the feedback. | 16:13 |
| jroll | rloo: updated https://dl.dropboxusercontent.com/u/363486/IPA-Startup-Flow.png | 16:13 |
| NobodyCam | :) | 16:14 |
| *** epim_ has joined #openstack-ironic | 16:14 | |
| devananda | rloo: right, so tripleo-incubator/tree/images still references nova baremetal ... | 16:15 |
| devananda | very out of date | 16:15 |
| lucasagomes | devananda, rloo NobodyCam I was looking at barbican docs, they are quite good | 16:15 |
| lucasagomes | look | 16:15 |
| lucasagomes | https://github.com/cloudkeep/barbican/wiki/Architecture | 16:15 |
| NobodyCam | lucasagomes: ya I did too | 16:15 |
| devananda | nice | 16:16 |
| NobodyCam | looks like just what we were looking of from my initial read over | 16:16 |
| *** epim has quit IRC | 16:16 | |
| *** epim_ is now known as epim | 16:16 | |
| rloo | so do I have a volunteer to write up a user guide/architecture docn for ironic? | 16:17 |
| * rloo can only hope | 16:17 | |
| lucasagomes | maybe everyone should help a bit with ^ | 16:17 |
| rloo | yeah ;) Guess I'll put together a first stab at something, and I'll leave lots of room for everyone else to contribute ;) | 16:19 |
| NobodyCam | :) rloo that would be awesome. poke me and I'll throw in my 2 cents | 16:21 |
| rloo | NobodyCam: thx! | 16:21 |
| romcheg | Hi lucasagomes, NobodyCam, rloo, jroll, devananda, Shrews… there are so many of you :) | 16:22 |
| lucasagomes | romcheg, morning :) | 16:22 |
| JayF | Good morning. | 16:22 |
| dtantsur | lucasagomes, I'm getting InstanceDeployFailure: Failed to provision instance d90bf752-2cab-4414-9cc6-334ebf86078e: PXE driver failed to continue deployment. | 16:22 |
| dtantsur | in nova-scheduler journald logs | 16:22 |
| lucasagomes | rloo, cheers, I can try to give u a hand | 16:22 |
| rloo | thx lucasagomes! | 16:23 |
| lucasagomes | dtantsur, right, anything on the ir-cond logs? | 16:23 |
| romcheg | lucasagomes: I'm looking at https://review.openstack.org/#/c/83396/1/ironic/common/disk_partitioner.py this file right now and I have a question about initialisind start and end on L88 and L90 | 16:23 |
| NobodyCam | hey hey romcheg :) nice to see you :) | 16:23 |
| jroll | morning romcheg | 16:23 |
| lucasagomes | romcheg, sure what's the question? | 16:23 |
| lucasagomes | romcheg, end is the offset to the end of the last partition | 16:24 |
| lucasagomes | romcheg, which starts with 0 (disk empty), on the next interaction it will be 0 + part_size | 16:25 |
| romcheg | It looks odd to me. Why not just start=0; end=start+part['size'] | 16:25 |
| dtantsur | lucasagomes, SSHException: Error reading SSH protocol bannerSecond simultaneous read on fileno 7 detected. Unless you really know what you're doing, make sure that | 16:25 |
| dtantsur | only one greenthread can read any particular socket. Consider using a pools.Pool. If you do know what you're doing and want to disable this error, call eventlet.debug.hub_prevent_multiple_readers(False) | 16:25 |
| lucasagomes | so the start of the next partition is at the end of the previous | 16:25 |
| dtantsur | the only error I found | 16:25 |
| romcheg | Maybe I'm missing some logic there? | 16:25 |
| lucasagomes | romcheg, that logic works only for the first partition | 16:25 |
| lucasagomes | romcheg, the second one needs to start at the end of the first | 16:25 |
| lucasagomes | third at the end of the second etc... | 16:25 |
| romcheg | Ah, right! | 16:26 |
| lucasagomes | dtantsur, oh eww | 16:26 |
| lucasagomes | dtantsur, right haven't seem it, maybe my env is too old | 16:27 |
| lucasagomes | I will try to recreate it | 16:27 |
| romcheg | lucasagomes: sorry for the dumb question | 16:27 |
| lucasagomes | romcheg, it's not dumb :) | 16:27 |
| Shrews | rloo: someone put up a first-pass user guide today | 16:27 |
| * Shrews digs up review # | 16:27 | |
| Shrews | https://review.openstack.org/83343 | 16:28 |
| rloo | Shrews: cool. | 16:28 |
| * rloo takes a look | 16:29 | |
| * NobodyCam looks too | 16:29 | |
| lucasagomes | dtantsur, ur on f20? | 16:29 |
| *** athomas has quit IRC | 16:30 | |
| dtantsur | lucasagomes, yes | 16:30 |
| lucasagomes | ack | 16:30 |
| openstackgerrit | Jenkins proposed a change to openstack/ironic: Updated from global requirements https://review.openstack.org/83471 | 16:33 |
| NobodyCam | bbt.... brb | 16:33 |
| *** epim has quit IRC | 16:34 | |
| NobodyCam | post bbt walkies.. bbiafm... | 16:40 |
| *** zigo has quit IRC | 16:41 | |
| *** akrivoka has joined #openstack-ironic | 16:42 | |
| SpamapS | DriverNotFound: FakeSeaMicroDriver | 17:01 |
| SpamapS | DriverNotFound: PXEAndSeaMicroDriver | 17:01 |
| SpamapS | Any ideas why that might happen? | 17:02 |
| SpamapS | http://logs.openstack.org/59/81959/22/check-tripleo/check-tripleo-ironic-seed-precise/fb05cfd/ <-- test run showing this fail in seed_logs.tgz | 17:02 |
| jroll | SpamapS: I believe those are just warnings, but that happens when the seamicro deps are not installed | 17:02 |
| JayF | Is that the broken test? | 17:03 |
| * JayF checks | 17:03 | |
| JayF | ah, so the one that's been failing for me is check-tripleo-ironic-undercloud-precise | 17:04 |
| JayF | idt that the seed one has been failing, but icbw | 17:04 |
| SpamapS | jroll: is there a bug open for that? | 17:04 |
| SpamapS | jroll: tracebacks in logs != warnings | 17:04 |
| SpamapS | (stevedore.extension): 2014-03-27 06:06:40,516 ERROR Could not load 'fake_seamicro': FakeSeaMicroDriver | 17:05 |
| SpamapS | ERROR | 17:05 |
| SpamapS | devananda: over here. ;) | 17:05 |
| SpamapS | devananda: so error tracebacks in logs == High priority bug IMO. | 17:06 |
| devananda | lol | 17:06 |
| devananda | hi | 17:06 |
| SpamapS | as an admin that says to me "We have no idea what the program is doing now." | 17:06 |
| SpamapS | https://review.openstack.org/#/c/81959/22 anyway.. I am having a hard time figuring out what these two failures are | 17:07 |
| *** matty_dubs|czech is now known as matty_dubs|lunch | 17:07 | |
| jroll | SpamapS: I'm not sure if there's a bug filed | 17:07 |
| devananda | SpamapS: "these two failures"? do you mean anything aside from not loading teh SeaMicro drivers? | 17:08 |
| *** harlowja_away is now known as harlowja | 17:08 | |
| devananda | SpamapS: so the root cause here is, this is how stevedore handles missing plugins | 17:09 |
| SpamapS | devananda: the sea micro failures are just distracting | 17:09 |
| devananda | SpamapS: if I don't raise an exception from within the plugin, stevedore is going to load it. and I need stevedore to *not* load it when a dependency is missing | 17:09 |
| devananda | SpamapS: ok.w hat's the real problem? | 17:09 |
| SpamapS | devananda: so you can't catch that exception? Anywhere? | 17:09 |
| SpamapS | devananda: the real problem is what I can't figure out ;) | 17:10 |
| devananda | no, i cant | 17:10 |
| SpamapS | devananda: bug in stevedore? | 17:10 |
| devananda | notice what's logging it -- stevedore.extension | 17:10 |
| devananda | not ironic | 17:10 |
| SpamapS | yeah I see that, so.. bug in stevedore? | 17:10 |
| SpamapS | Thats a broken API IMO | 17:10 |
| SpamapS | plugins are nasty business.. you have to be able to handle failures gracefully. | 17:10 |
| devananda | the exception raised at File "/opt/stack/new/ironic/ironic/drivers/pxe.py", line 102 is how ironic is supposed to signal stevedore | 17:10 |
| devananda | that the plugin should not be loaded | 17:10 |
| devananda | dhellmann: ^ | 17:10 |
| * dhellmann reads scrollback | 17:11 | |
| devananda | I may be wrong :) | 17:11 |
| dhellmann | which extension manager is being used? | 17:11 |
| SpamapS | if that is how you say "don't load me" .. stevedore should probably not log it at ERROR | 17:11 |
| devananda | dhellmann: NameDispatch | 17:11 |
| devananda | https://github.com/openstack/ironic/blob/master/ironic/common/driver_factory.py#L49 | 17:12 |
| devananda | nd | 17:12 |
| devananda | https://github.com/openstack/ironic/blob/master/ironic/drivers/pxe.py#L102 | 17:12 |
| *** datajerk has joined #openstack-ironic | 17:13 | |
| dhellmann | ok | 17:13 |
| dhellmann | why don't you want the plugin loaded? | 17:14 |
| devananda | it's a third-party driver | 17:14 |
| devananda | the required libraries for taht driver may not be present on the system | 17:14 |
| devananda | SpamapS: one work around is to install said third-party libs, even though you guys won't need them | 17:15 |
| dhellmann | the philosophy with stevedore is if the manager tries to load the plugin, it should load -- whether or not you can use it is an issue that can be evaluated later | 17:15 |
| dhellmann | we are relying on that, for example, to get configuration settings from drivers | 17:15 |
| devananda | dhellmann: how would that be done when, eg. that module is missing a dependent library? | 17:16 |
| SpamapS | pretty normal for a plugin to be present but decline to load. | 17:16 |
| dhellmann | the class registered as the plugin can report an exception when something tries to use it | 17:16 |
| *** vkozhukalov has quit IRC | 17:16 | |
| devananda | dhellmann: um... so that's going to break some other assumptions in ironic | 17:16 |
| *** romcheg has quit IRC | 17:16 | |
| dhellmann | SpamapS: it would work the way you want if you were using the DriverManager, but you're using a "load everything" manager | 17:17 |
| *** romcheg has joined #openstack-ironic | 17:17 | |
| devananda | if the class loads, ironic-conductor assumes it's usable, and will advertise support for it | 17:17 |
| devananda | afaik, we couldn't UNload it if it failed | 17:17 |
| SpamapS | For instance, Asterisk has plugins for proprietary codecs, but you have to have a license for the proprietary codecs or they don't load. | 17:17 |
| dhellmann | devananda: you can, after loading the plugins, ask them "are you ready to be used?" by invoking a method | 17:17 |
| *** romcheg has quit IRC | 17:17 | |
| SpamapS | devananda: why aren't we using DriverManager ? | 17:17 |
| dhellmann | SpamapS: I'm aware of other usage patterns. I'm describing what is implemented. | 17:18 |
| devananda | SpamapS: dunno. dhellmann, should we be using DriverManager? | 17:18 |
| dhellmann | devananda: it sounds like you're using the manager you want -- maybe we should give you a way to adjust that log message level, or only log it if there is no error handler callback provided | 17:19 |
| devananda | dhellmann: oh, if I provide an error handler, then I could choose to squelch that particular exception? | 17:19 |
| dhellmann | devananda: not now, but we could make it work that way | 17:19 |
| devananda | that would work fine, since it's an ironicException I want to squelch anyway | 17:19 |
| devananda | ahh heh | 17:19 |
| dhellmann | devananda: simple change, move the log message to an else clause on the existing if statement | 17:20 |
| JayF | That sounds awesome, and I believe would help with IPA as well | 17:20 |
| dhellmann | devananda: give me a sec | 17:20 |
| SpamapS | dhellmann: sorry for that. :) It seems that we are just in need of a merging of understanding, not learning new things in particular. ;) | 17:20 |
| dhellmann | SpamapS: agreed :-) | 17:20 |
| devananda | dhellmann: thanks much. I should apologize that I keep using stevedore in unexpected ways, but it's working great for them ;) | 17:21 |
| *** romcheg has joined #openstack-ironic | 17:21 | |
| dhellmann | devananda: well, I expected to need to make changes, so don't apologize for pushing boundaries :-) | 17:22 |
| lucasagomes | romcheg, thanks for the reviews I will address them :) | 17:22 |
| romcheg | lucasagomes: My pleasure to review some good code. And to -1 it then :-P | 17:23 |
| lucasagomes | romcheg, lol! \o/ | 17:23 |
| dhellmann | devananda, SpamapS : https://review.openstack.org/83509 | 17:24 |
| JayF | That's great, thanks! | 17:24 |
| jroll | thanks for that, dhellmann :) | 17:27 |
| dhellmann | do you all have a bug associated with this? If so, I'd like to add it to oslo for tracking. | 17:29 |
| devananda | SpamapS: ? | 17:30 |
| devananda | if not, i'll file one | 17:30 |
| dhellmann | thanks | 17:30 |
| SpamapS | I did not file one yet no | 17:32 |
| devananda | k | 17:33 |
| *** martyntaylor has quit IRC | 17:36 | |
| devananda | dhellmann: think there's any likelyhood of that making it in soon, given FF / RC's? | 17:38 |
| devananda | dhellmann: *and into a release of stevedore | 17:38 |
| dhellmann | yeah, I'll try to get it to land today | 17:38 |
| devananda | awesome | 17:38 |
| devananda | https://bugs.launchpad.net/ironic/+bug/1298529 | 17:38 |
| dhellmann | and do a release tomorrow if that happens | 17:38 |
| *** zigo has joined #openstack-ironic | 17:39 | |
| NobodyCam | humm seed is booting but undercloud vm appears to not be getting and dhcp at all | 17:41 |
| *** romcheg has quit IRC | 17:42 | |
| NobodyCam | SpamapS: are you testing with the latest devtest patches? are you seeing that on the seed or undercloud? | 17:42 |
| SpamapS | NobodyCam: https://review.openstack.org/#/c/81959/22 | 17:44 |
| NobodyCam | yep thats what I'm using | 17:45 |
| NobodyCam | SpamapS: is your undercloud booting? | 17:45 |
| SpamapS | NobodyCam: see the logs there | 17:47 |
| SpamapS | NobodyCam: I am not testing this | 17:47 |
| SpamapS | CI is | 17:47 |
| NobodyCam | ack :) | 17:47 |
| *** romcheg has joined #openstack-ironic | 17:54 | |
| lucasagomes | aight I came to the office today so I'm heading back home now | 18:00 |
| lucasagomes | have a g'night everyone | 18:00 |
| NobodyCam | have a good night lucasagomes :) | 18:00 |
| lucasagomes | NobodyCam, devananda romcheg rloo, etc... night :) | 18:00 |
| rloo | g'night lucasagomes! | 18:00 |
| romcheg | Bye lucasagomes | 18:00 |
| romcheg | Have a good evening | 18:00 |
| *** lucasagomes has quit IRC | 18:01 | |
| *** matty_dubs|lunch is now known as matty_dubs | 18:03 | |
| NobodyCam | SpamapS: strange error in the console log. looks like it couldn't find the pm_type: http://logs.openstack.org/59/81959/22/check-tripleo/check-tripleo-ironic-seed-precise/fb05cfd/console.html#_2014-03-27_10_28_29_602 | 18:04 |
| SpamapS | NobodyCam: agreed | 18:06 |
| *** jistr has quit IRC | 18:07 | |
| hemna_ | devananda, ping | 18:10 |
| devananda | hemna_: sorta here. what's up | 18:16 |
| *** blamar has quit IRC | 18:17 | |
| *** blamar has joined #openstack-ironic | 18:18 | |
| Shrews | anyone know what's up with nothing being scheduled for the Friday of the summit? http://openstacksummitmay2014atlanta.sched.org | 18:19 |
| *** derekh has quit IRC | 18:19 | |
| hemna_ | devananda, I proposed as session for ironic boot from cinder for atlanta. trying to coordinate with Wan-yen on it. | 18:19 |
| devananda | Shrews: it's staggered? | 18:20 |
| devananda | Shrews: mon-thu presentations. tues-fri design. I think. | 18:20 |
| devananda | hemna_: thanks! | 18:21 |
| hemna_ | devananda, np. I thought it would be useful to chat about it. | 18:21 |
| devananda | hemna_: fwiw, we only get 4 sessions, so they will probably have to be focused on the project's central needs during Juno | 18:21 |
| hemna_ | devananda, ok | 18:21 |
| hemna_ | maybe we can meet offline there then? | 18:21 |
| devananda | hemna_: one of them will probably be to address vendor support broadly | 18:22 |
| hemna_ | 4 sessions isn't very much :( | 18:22 |
| devananda | hemna_: BUT there's a lot of unconference space for more detailed discussions | 18:22 |
| devananda | hemna_: yea... incubated projects don't get much :( | 18:22 |
| devananda | and it's all on tuesday | 18:22 |
| devananda | hemna_: actuallym, you should propose one to the cinder track | 18:22 |
| hemna_ | dang ok. when I get there I'll try and fill out an unconference room for it then | 18:23 |
| hemna_ | devananda, from what I can tell, there isn't much that needs to be done in cinder for this to work. | 18:23 |
| devananda | hemna_: it's as much asbout cinder support as ironic support, and probably fits better in that project | 18:23 |
| devananda | hemna_: last time i talked with the cinder PTL, there was a lot missing in terms of a way to describe volumes | 18:23 |
| devananda | hemna_: genericaly enough that they could be used for ironic | 18:23 |
| devananda | but that was at the last summit. maybe it's been done already :) | 18:24 |
| hemna_ | ok I'll ping john about his thoughts on it then. I think it can work w/o changes to Cinder, but he knows the internals better than I do. | 18:24 |
| hemna_ | Ironic still has some major things to work on to config the node for boot from san. | 18:25 |
| devananda | hemna_: indeed it does | 18:25 |
| devananda | hemna_: there's two possible routes,a nd we need to define an API that supports both | 18:25 |
| hemna_ | yah, at this point defining that API is probably a great start. | 18:25 |
| devananda | hemna_: configure BIOS via local channels in the ironic-python-agent || remote configuration via BMC | 18:26 |
| devananda | gotta step out for a few... bbs | 18:35 |
| *** akrivoka has quit IRC | 18:41 | |
| *** epim has joined #openstack-ironic | 18:43 | |
| *** epim has quit IRC | 18:44 | |
| *** epim has joined #openstack-ironic | 18:44 | |
| *** ndipanov has quit IRC | 19:00 | |
| *** mdurnosvistov_ has quit IRC | 19:04 | |
| *** epim has quit IRC | 19:06 | |
| * devananda is back | 19:08 | |
| NobodyCam | wb devananda | 19:08 |
| devananda | having difficulty figuring out why transifex is throwing up | 19:09 |
| *** jdob has quit IRC | 19:09 | |
| *** jdob has joined #openstack-ironic | 19:09 | |
| NobodyCam | devananda: reading over lucas's ML thread on user creds, should we perpose a summit session on the topic? | 19:10 |
| NobodyCam | we only have 4 sessions thou? | 19:10 |
| devananda | NobodyCam: yes, only 4 40-minute sessions | 19:11 |
| NobodyCam | :( | 19:11 |
| devananda | so they need to be very deliberate | 19:12 |
| devananda | what's critical to the project that can only be unblocked / addressed by the critical mass of folks there? | 19:12 |
| devananda | there are *tons* of useful discussions that we should have | 19:12 |
| NobodyCam | ya ! | 19:12 |
| devananda | see Summit Ideas @ https://etherpad.openstack.org/p/IronicWhiteBoard | 19:13 |
| devananda | please add to it | 19:13 |
| devananda | or propose a talk | 19:13 |
| devananda | s/talk/session/ | 19:14 |
| devananda | and I'll condense things as needed | 19:14 |
| devananda | if we end up with 20 sessions proposed, it's not a bad thing :) | 19:14 |
| JayF | devananda: we have a proposal up to talk about the agent, that seems like it might be worthy of a slot? | 19:14 |
| NobodyCam | :) | 19:14 |
| devananda | JayF: yes | 19:14 |
| JayF | ah cool Jim must've already added that to the etherpad | 19:14 |
| devananda | JayF: I'd prefer as much info ahead of time as possible in the BP/etherpad/session proposal | 19:15 |
| devananda | JayF: i'll probably into the topic, then you guys should present the code layout / arch / etc, then open the remaining ~30 min for discussion | 19:16 |
| devananda | JayF: also we should have clear goals to come out of the session | 19:16 |
| devananda | so that the discussion stays on topic and results in action | 19:16 |
| JayF | Hopefully that will be more than preseting code layout, I hope we can have a live prototype demo | 19:16 |
| devananda | I think that was a shortcoming of my organization last time. we had 3 sessions of talking and then attempted to wrap it all together in the last session for "action items" | 19:17 |
| jroll | JayF: I proposed that session a few weeks back on summit.openstack.org | 19:17 |
| devananda | rather than having each session start with "here are the goals" | 19:17 |
| jroll | hence the link :P | 19:17 |
| devananda | JayF: so. demo. It'll be great to have it working. But. | 19:18 |
| devananda | JayF: the format of the design sessions is not to present / demo soemthing | 19:18 |
| JayF | yeah I didn't know it had been linked into the mentioned etherpad though | 19:18 |
| JayF | but we should polish it up, come up with clear goals like deva said, before going to the summit | 19:18 |
| JayF | and hopefully we can talk about how it works while showing it working :) | 19:18 |
| devananda | JayF: as long as the demo/presentation part is kept under, say, 10 minutes, it's probably OK | 19:19 |
| JayF | I understand, demo is secondary, but it's important to us as a proof of concept, that what we're trying to do works /then/ while we're talking about it rather than talking about what might be possible :) | 19:19 |
| devananda | JayF: absolutely! | 19:19 |
| JayF | I like things that work more than I like ideas that might work :P | 19:19 |
| devananda | JayF: it should totally be working AND you should totally demo it | 19:19 |
| linggao | devananda, NobodyCam, talking about session, have proposed 2 sessions. | 19:21 |
| devananda | JayF: my point is about format of the summit sessions. considering that we, as a project, only have 4 x 40m slots in which to resolve discussions that require the whole group, a lengthy demo is not efficient use of that time | 19:21 |
| JayF | I get that. It's actually a bit shocking that such little time is able to be able to dedicate to the project | 19:21 |
| JayF | Like I guess I was expecting there to be more limitations on 'we need everyone' time vs 'ironic' time, but having not been to a summit that seems to be very wrong | 19:22 |
| devananda | JayF: one of the challenges is that our 4 slots overlap with the cross-project day | 19:23 |
| devananda | so actually | 19:23 |
| openstackgerrit | Adam Gandelman proposed a change to openstack/ironic: Raise unexpected exceptions during destroy() https://review.openstack.org/82948 | 19:23 |
| devananda | some of our core team will NEED to be in those sessions | 19:23 |
| devananda | and not participate in our own design | 19:23 |
| devananda | I've already done what I can about this, but been overruled | 19:23 |
| NobodyCam | :( | 19:23 |
| NobodyCam | brb | 19:24 |
| devananda | linggao: hi! I'll be reviewing the sessions as we get closer to the summit. I expect that folks will continue to propose more | 19:24 |
| linggao | devananda, NobodyCam, one is for chassis level functions and the other one is more hardware control functions. | 19:25 |
| openstackgerrit | Jim Rollenhagen proposed a change to openstack/ironic-python-agent: Fix URL for node lookup https://review.openstack.org/83539 | 19:25 |
| devananda | linggao: if you've followed this discussion, you will hopefully understand why I will need to condense many sessions together, or push them to unconference space | 19:25 |
| *** max_lobur1 has quit IRC | 19:25 | |
| linggao | devananda, sure. I just wonder should I start writing blueprint for them or should wait after the summit. | 19:26 |
| devananda | linggao: please file BPs ahead of the sumit and link them to the proposals, if that is appropriate | 19:27 |
| devananda | linggao: or reference an already existing BP if one exists | 19:27 |
| openstackgerrit | Adam Gandelman proposed a change to openstack/ironic: Remove some obsolete settings from DevStack doc https://review.openstack.org/83540 | 19:28 |
| linggao | devananda, sure. will do. | 19:28 |
| *** GheRivero has quit IRC | 19:28 | |
| devananda | Haomeng: ping | 19:29 |
| *** vkozhukalov has joined #openstack-ironic | 19:31 | |
| linggao | He may be asleep :). It is 3 am for him. | 19:32 |
| *** mtrisoline has joined #openstack-ironic | 19:32 | |
| openstackgerrit | Jim Rollenhagen proposed a change to openstack/ironic-python-agent: Fix URL for node lookup https://review.openstack.org/83539 | 19:35 |
| NobodyCam | linggao: I hope he is a 3 am :-p | 19:36 |
| *** mtrisoline has left #openstack-ironic | 19:37 | |
| *** mriedem has joined #openstack-ironic | 19:39 | |
| *** mriedem has left #openstack-ironic | 19:39 | |
| linggao | NobodyCam, :) | 19:41 |
| devananda | anyone looking for things to delve into? | 19:41 |
| *** epim has joined #openstack-ironic | 19:44 | |
| Shrews | why is check_uptodate.sh borking on the sample config from master? | 19:44 |
| Shrews | i mean, how does that happen if it causes pep8 to fail? | 19:45 |
| *** mtrisoline has joined #openstack-ironic | 19:46 | |
| *** GheRivero has joined #openstack-ironic | 19:46 | |
| mtrisoline | Is it possible someone may be able to help me with updated docs? This seems out of date, I am trying to implement nova-baremetal as part of a requirment to use openstack. #openstack-nova pointed me here. https://wiki.openstack.org/wiki/Baremetal | 19:47 |
| linggao | devananda, I am looking for something to do. | 19:47 |
| NobodyCam | Shrews: what is the error you are getting | 19:47 |
| NobodyCam | hi mtrisoline | 19:48 |
| Shrews | NobodyCam: just that the sample is out of date. specifically: check_uptodate.sh: ironic.conf.sample is not up to date. | 19:48 |
| NobodyCam | mtrisoline: you need nova-baremetal? not Ironic? we are working on replacing nova baremetal | 19:49 |
| NobodyCam | Shrews: is there a diff? | 19:49 |
| NobodyCam | have you pulled any patches, there was a hostname issue but I thought that was fixed awhile ago | 19:50 |
| mtrisoline | NobodyCam: I know, but I have to work off of Havana atm. Unless there is a way I can patch in Ironic. | 19:50 |
| Shrews | NobodyCam: indeed. http://paste.openstack.org/show/74483/ | 19:50 |
| NobodyCam | mtrisoline: I would have to know a little more, Ironic does at this point have a working (deff not bug free, but working) nova driver | 19:51 |
| mtrisoline | O, I was not aware of that., | 19:52 |
| NobodyCam | Shrews: are your on master branch? | 19:53 |
| NobodyCam | Shrews: I'll need to test that, going to be a minute | 19:53 |
| NobodyCam | mtrisoline: do you use devtst or devstack at all? | 19:54 |
| mtrisoline | NobodyCam: Basically we want to replace our vmware with openstack and offer up IaaS internally to the company, but for production we run all our databases and hadoop clusters on bare metal. We were looking into the functionality to replace our kickstart setup with the baremetal functionality of openstack. Does that help? | 19:54 |
| Shrews | NobodyCam: i am on master | 19:55 |
| NobodyCam | mtrisoline: so at this point you have no openstack infrastructure | 19:56 |
| Shrews | NobodyCam: HEAD is f1e702 | 19:57 |
| devananda | linggao: ok! which would you prefer -- writing new scheduler filters? debugging our jenkins' translation job | 19:57 |
| mtrisoline | Nope just one baremetal node acting as a single controller and everything else virtualized w/in an old hypervisor, but I have 4 servers on the way on top of a few more that are being pulled from production to throw into a full dev environment for this. | 19:57 |
| NobodyCam | I would suggest looking at the TripleO stuff. https://github.com/openstack/tripleo-incubator | 19:57 |
| openstackgerrit | A change was merged to openstack/ironic-python-agent: Fix URL for node lookup https://review.openstack.org/83539 | 19:57 |
| devananda | mtrisoline: I don't know how much the internal nova virt driver API has changed since Havana | 19:58 |
| devananda | mtrisoline: I suspect that will be the biggest area of potential conflict | 19:58 |
| linggao | devananda, I prefer the first one. | 19:58 |
| mtrisoline | Right now in my dev environment I have everything else 100% working and functional. | 19:58 |
| devananda | https://bugs.launchpad.net/tripleo/+bug/1291396 | 19:59 |
| mtrisoline | Our existing setup works extremely well for the baremetal stuff, so I know I can push this off. I was just really trying to get this into my proof of concept. I definitely appreciate the info. I will look into the ironic driver in Havana. | 19:59 |
| devananda | and see my comments on https://bugs.launchpad.net/ironic/+bug/1291420 | 19:59 |
| devananda | linggao: ^ | 19:59 |
| rloo | Shrews, NobodyCam: wondering if this change has anything to do with it: https://review.openstack.org/#/c/83471/1/requirements.txt | 20:00 |
| devananda | Rohan (k4n0) said a week ago he was working on it, but i don't see any patches up | 20:00 |
| devananda | linggao: description of what I think needs to be done -- https://bugs.launchpad.net/ironic/+bug/1291420/comments/3 | 20:00 |
| linggao | davananda, ok. Let me read it. | 20:01 |
| Shrews | rloo, NobodyCam: I mean, it has to be something with my environment, right? A change that doesn't pass the pep8 test would never get merged into master. This is new magic I do not yet comprehend. | 20:01 |
| Shrews | perhaps i should rebuild the tox venv | 20:02 |
| adam_g | Shrews, i was scratching my head similarly the other day. my branch's sample was out of date, but so was masters? | 20:02 |
| NobodyCam | Shrews: unless they changed the pep8 rules in mid landing of a patch | 20:02 |
| NobodyCam | Shrews: ya :) | 20:02 |
| *** harlowja is now known as harlowja_away | 20:03 | |
| NobodyCam | mtrisoline: are you getting a error (I may have missed seeing it) | 20:04 |
| * Shrews rebuilds venv | 20:04 | |
| devananda | eek | 20:10 |
| * devananda realizes we broke a unit test for seamicro | 20:10 | |
| *** mdurnosvistov_ has joined #openstack-ironic | 20:10 | |
| devananda | right. so this is interesting | 20:10 |
| devananda | allow third-party drivers to load plugins not in requirements.txt (which we should do, cause, vendor stuff) | 20:10 |
| devananda | and more sane handling of errors while loading said drivers | 20:11 |
| devananda | means that we DONT actually test them. EVEN THE UNIT TESTS | 20:11 |
| devananda | mordred: thoughts on ^ ? | 20:12 |
| NobodyCam | brb | 20:13 |
| JayF | devananda: should we have a separate test-requirements.txt which installs those 'unapproved dependencies' for testing purposes? Maybe even make a separate check job that checks with the vendor libs vs without? | 20:15 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Have stevedore convert DriverNotFound to a warning https://review.openstack.org/83572 | 20:16 |
| devananda | JayF: probably. however. I don't know if -infra approves of such things, which is why I pinged mordred | 20:17 |
| mrda_away | Morning all | 20:17 |
| JayF | devananda: wdyt about lifeless's comments here? https://review.openstack.org/#/c/83231/ I gtg but would love to be able to see your comments on the review when I get back this evening, so I can work on it again. | 20:17 |
| devananda | JayF: there's a larger discussion around such things for several projects | 20:17 |
| NobodyCam | morning mrda_away | 20:17 |
| JayF | It's just strange that these ongoing conversations happen somewhere other than the ML tbh | 20:17 |
| devananda | JayF: ah, right | 20:17 |
| lifeless | I didn't know there was a larger discussion | 20:18 |
| devananda | by that i mean all the third-party testing support | 20:18 |
| JayF | lifeless: sorry, there are like two conversations happening simultaneously :) we were talking about seamicro not your comments. | 20:18 |
| devananda | which has been on the ML, IRC, and elsewhere | 20:18 |
| lifeless | ohright | 20:18 |
| JayF | ah, gotcha | 20:18 |
| *** mrda_away is now known as mrda_ | 20:19 | |
| devananda | JayF: w.r.t. lifeless' comments, I stand by my bug report (low-level messages shouldn't go into node.last_error) | 20:19 |
| *** packet has joined #openstack-ironic | 20:19 | |
| devananda | JayF: and you need unit tests :) | 20:19 |
| NobodyCam | morning lifeless | 20:20 |
| * JayF may need to learn how to write unit tests | 20:20 | |
| *** mrda_ has quit IRC | 20:20 | |
| *** jbjohnso_ has quit IRC | 20:20 | |
| JayF | but that's OK, I knew it was coming :) No time like the present | 20:20 |
| devananda | JayF: also, i think you're missing a bunch of other parts of conductor.manager | 20:20 |
| lifeless | devananda: its your call, but I'd really love it if you got fully involved in tripleo-cd ops before making it | 20:20 |
| devananda | JayF: so this should have a a Partial-bug, not a Closes-bug, tag | 20:20 |
| lifeless | devananda: because, wearing my ops hat, you're wrong :) | 20:21 |
| devananda | lifeless: I love it when I'm wrong :) | 20:21 |
| JayF | devananda: I'll try to get more coverage of those error types, and get some tests in. | 20:21 |
| JayF | bbl | 20:21 |
| devananda | JayF: thanks! | 20:21 |
| devananda | lifeless: if we assume the only consumer of ironic is a trusted operational team, then you'd be correct -- we should expose *all* the plubming | 20:22 |
| devananda | lifeless: but that assumption is incorrect | 20:22 |
| openstackgerrit | Matt Wagner proposed a change to openstack/ironic: Add documentation for Ironic config file https://review.openstack.org/83577 | 20:23 |
| lifeless | devananda: I didn't argue for exposing all the plumbing | 20:23 |
| lifeless | devananda: so thats a straw man vs my actual position | 20:23 |
| NobodyCam | lifeless: quick sidetrack question.. have you seen any "Falied to resize filesystem" errors on the undercloud when deploying? | 20:23 |
| devananda | lifeless: ok. more precisely, exposing the low-level exception traces from drivers | 20:23 |
| devananda | lifeless: which could include eg. output from executils, ssh, sfdisk, or other nitty gritty bits of drivers that we don't control | 20:24 |
| lifeless | NobodyCam: yes, there's a bug open in diskimage-builder and ubuntu | 20:24 |
| NobodyCam | :) TY | 20:24 |
| *** mrda_ has joined #openstack-ironic | 20:24 | |
| lifeless | NobodyCam: and you can often work around by tweaking the 0.X multiplier in disk-image-create | 20:24 |
| lifeless | devananda: yes but thats a driver concern | 20:25 |
| *** mrda_ has left #openstack-ironic | 20:26 | |
| *** mrda_ has joined #openstack-ironic | 20:26 | |
| devananda | lifeless: ahh | 20:27 |
| * devananda looks for rloo's patch | 20:27 | |
| rloo | what patch? | 20:28 |
| lifeless | devananda: and e.g. output from sfdisk is *useful* and not a security concern | 20:28 |
| devananda | rloo: iirc you were cleaning up the driver exception messages at one point | 20:28 |
| devananda | this one | 20:29 |
| devananda | https://bugs.launchpad.net/ironic/+bug/1267693 | 20:29 |
| rloo | devananda: I think I was cleaning up the exceptions so they would all be IronicExceptions | 20:29 |
| devananda | yep | 20:29 |
| rloo | yeah, that one. Not necessarily the messages? | 20:29 |
| * rloo scrolls back to understand context | 20:29 | |
| devananda | lifeless: if drivers always caught exceptions from the tools they invoked, and then raised IronicExceptions with meaningful messages | 20:31 |
| devananda | lifeless: that would be something I think should be exposed in the API | 20:32 |
| devananda | lifeless: the issue is, right now, driver's aren't well policed about that. | 20:32 |
| devananda | lifeless: and I can understand how hiding too much is worse right now than showing too much, so if you agree with that ^ direction, I think we can punt on that until Juno. | 20:34 |
| *** harlowja_away is now known as harlowja | 20:35 | |
| rloo | devananda, if I understand, seems like we (or me I guess) should just clean up the exception msgs from the drivers, and leave the code in manager.py as is. | 20:35 |
| lifeless | devananda: I think saying 'drivers raising other than IronicException is a bug regardless of whether the operation failed or not' | 20:35 |
| lifeless | devananda: is valid - if thats a reasonable rephrase of your viewpoint I think we have consensus | 20:35 |
| devananda | lifeless: ++ | 20:35 |
| *** jbjohnso_ has joined #openstack-ironic | 20:35 | |
| devananda | or more precisely | 20:36 |
| *** mrda_ is now known as mrda | 20:36 | |
| devananda | nvm. good enough | 20:36 |
| * devananda updates bug desc | 20:36 | |
| devananda | and closes the other | 20:37 |
| devananda | rloo: yep. so that's two things: make sure drivers are raising suitable user-visible strings, and make sure drivers are really careful not to allow exceptions (from the things they call) to propagate up | 20:38 |
| rloo | devananda: do you see those as two separate bugs, and/or do you want me to do them both in the same bug: 1267693? | 20:39 |
| rloo | I was so close too. Only had pxe.py to do, but now who knows what has changed. ha ha. ;) | 20:39 |
| lifeless | devananda: point is to me that we should be able to get e.g. sfdisk output wrapped in IronicException - that would be ok right ? | 20:40 |
| devananda | lifeless: yes. if eg. the PXE driver captures the excutils exception,"sanitizes" it, and raises an IronicException (or subclass), that's fine | 20:41 |
| *** EmilienM has quit IRC | 20:42 | |
| devananda | lifeless: my objection and the reason for that bug was seeing full dumps from excutils in node.last_error | 20:42 |
| linggao | devananda, I digged a bit on this defect (exact match filter), so the need is to write new filters under nova/schduler/filters, | 20:42 |
| devananda | lifeless: but I agree, this is a better solution | 20:42 |
| linggao | devananda, the filters will make sure the host has nodes that match the flavor exactly. | 20:42 |
| *** EmilienM has joined #openstack-ironic | 20:43 | |
| devananda | rloo: I could go either way. (a) improve the try: handling in drivers, so nothing "slips through" (b) sanitize the low-level errors and use good, informative messages | 20:43 |
| linggao | davananda, then overrides scheduler_default_filters CONF setting in ironic_host_manager. | 20:43 |
| devananda | rloo: if you feel that's worth two bugs, please open another | 20:43 |
| linggao | davananda, am I on the right track? | 20:43 |
| devananda | linggao: yes, except this seems off: "make sure the host has nodes" | 20:44 |
| rloo | devananda. I don't really care if it is two bugs or not ;) I'll just add a comment to the existing bug to remember to check/do that. | 20:44 |
| devananda | linggao: the nova-scheduler is choosing a specific node, represented in nova as (host, hyperisor_hostname) or (host, node) | 20:44 |
| devananda | rloo: k k | 20:44 |
| linggao | devananda, I am lost here. Is the scheduler choosing a compute host or choosing a baremetal node in our case? | 20:46 |
| lifeless | devananda: so can we reword that bug ? | 20:47 |
| devananda | linggao: both | 20:48 |
| devananda | linggao: but for ironic, it does not matter which compute host the scheduler picks -- it only matters which node | 20:48 |
| devananda | linggao: all compute hosts see the same pool of nodes | 20:49 |
| linggao | devananda, so in the exact-match-filters, what items should be exactly match? | 20:51 |
| *** mrda has quit IRC | 20:51 | |
| devananda | linggao: cpu, ram, disk | 20:51 |
| devananda | lifeless: i closed https://bugs.launchpad.net/ironic/+bug/1265587. how do you propose updating https://bugs.launchpad.net/ironic/+bug/1267693 | 20:52 |
| devananda | lifeless: i think the wording is already fairly close to your earlir suggestion | 20:52 |
| linggao | cpu of one of the baremetal nodes or cpu of the compute host? | 20:52 |
| devananda | linggao: ironic node. the compute host is functionally not much more than a proxy | 20:53 |
| *** mrda has joined #openstack-ironic | 20:54 | |
| linggao | devanand, I think I got it. I'll study more and work on it tomorrow. | 20:54 |
| linggao | davananda, thanks. | 20:55 |
| *** jdob has quit IRC | 20:55 | |
| *** romcheg has quit IRC | 20:55 | |
| NobodyCam | humm .. /me thinks he needs some food stuffs | 20:57 |
| NobodyCam | brb | 20:58 |
| *** vkozhukalov has quit IRC | 20:59 | |
| lifeless | devananda: looking | 21:01 |
| openstackgerrit | Ruby Loo proposed a change to openstack/ironic: Install guide for Ironic https://review.openstack.org/81645 | 21:02 |
| lifeless | devananda: looks fine to me | 21:02 |
| devananda | lifeless: :) | 21:02 |
| *** derekh has joined #openstack-ironic | 21:03 | |
| *** romcheg has joined #openstack-ironic | 21:04 | |
| *** linggao has quit IRC | 21:04 | |
| *** linggao has joined #openstack-ironic | 21:05 | |
| *** linggao has quit IRC | 21:06 | |
| adam_g | where/how was the nova driver being unit tested prior to it being merged into the ironic tree? this predates me and im having trouble finding traces of it in the nova or ironic trees | 21:08 |
| lifeless | tested? | 21:09 |
| lifeless | ironic/nova/tests/ I think you'll find | 21:09 |
| devananda | yes ^ | 21:09 |
| adam_g | doh | 21:10 |
| devananda | urgh! LP won't let me update this one bug! https://bugs.launchpad.net/ironic/+bug/1229324/+editstatus | 21:10 |
| devananda | other bugs are fine... | 21:10 |
| devananda | that one's been timing out since yesterday | 21:10 |
| *** jbjohnso_ has quit IRC | 21:11 | |
| comstud | devananda: the edit screen loads for me | 21:11 |
| comstud | i dunno if you get that far or not | 21:11 |
| devananda | comstud: edit loads for me, too | 21:12 |
| comstud | ok | 21:12 |
| devananda | comstud: "save" times out | 21:12 |
| comstud | What do you want changed? I can attempt, although I dunno if I'll even have perms | 21:12 |
| devananda | comstud: thanks. status -> fix commit. importance -> wishilist | 21:12 |
| NobodyCam | devananda: timesout for me | 21:12 |
| devananda | heh | 21:12 |
| comstud | Sorry, something just went wrong in Launchpad. | 21:13 |
| comstud | We’ve recorded what happened, and we’ll fix it as soon as possible. Apologies for the inconvenience. | 21:13 |
| comstud | Trying again in a couple of minutes might work. | 21:13 |
| devananda | yea | 21:13 |
| comstud | heh | 21:13 |
| devananda | so, that bug is broken | 21:13 |
| *** florentflament has joined #openstack-ironic | 21:13 | |
| devananda | yay! | 21:13 |
| comstud | file a bug about the bug | 21:13 |
| devananda | hahaha | 21:13 |
| NobodyCam | lol my bug is not working :-p | 21:14 |
| NobodyCam | sounds like a bug to me | 21:14 |
| NobodyCam | :-/ | 21:14 |
| lifeless | devananda: can I describe a problem for you | 21:16 |
| lifeless | devananda: when configuring Ironic, if there are no conductors running, we can't add nodes | 21:16 |
| lifeless | devananda: but there's no API to see if conductors are running | 21:16 |
| lifeless | devananda: analogous to nova service-list or neutron agent-list | 21:16 |
| devananda | lifeless: sure there is. ironic driver-lsit | 21:17 |
| devananda | *list | 21:17 |
| lifeless | ahha | 21:17 |
| lifeless | that was entirely unobvious | 21:17 |
| lifeless | thank you | 21:17 |
| devananda | :) | 21:17 |
| devananda | hmmm | 21:18 |
| devananda | the CLI help text is not very helpful | 21:18 |
| devananda | "List drivers." | 21:18 |
| lifeless | see ;) | 21:18 |
| devananda | lifeless: thoughts on how to "hide" the fake* drivers in a production install? do you see a reason not to do that? | 21:21 |
| lifeless | require a config option to enable them | 21:22 |
| lifeless | set that in devstack | 21:22 |
| devananda | hm | 21:22 |
| devananda | would work | 21:22 |
| *** mdurnosvistov_ has quit IRC | 21:26 | |
| *** derekh has quit IRC | 21:30 | |
| rloo | devananda: why isn't https://bugs.launchpad.net/bugs/1298628 a python-ironicclient bug? | 21:35 |
| devananda | rloo: because I am opening lots of bugs and tagged it wrong :p | 21:37 |
| rloo | devananda: :-( | 21:37 |
| devananda | actually | 21:37 |
| lifeless | oh I had a new failure for you | 21:37 |
| devananda | rloo: there, i moved it | 21:38 |
| lifeless | ah you saw it | 21:38 |
| rloo | devananda. the person who takes that bug thanks you ;) | 21:38 |
| rloo | devananda. wrt docn, do you want a quick status? | 21:40 |
| rloo | devananda. forget it, I think it might be better to put the status in the etherpad. will do that later ;) | 21:40 |
| devananda | rloo: i think i'm fairly up on it, but ... | 21:40 |
| devananda | rloo: that's even better | 21:40 |
| devananda | oh | 21:41 |
| devananda | as a general reminder to folks (including myself) | 21:41 |
| devananda | if you start an etherpad with interesting things in it | 21:41 |
| devananda | please reference it here: https://etherpad.openstack.org/p/IronicWhiteBoard | 21:41 |
| rloo | devananda. thx, good to know (or maybe I forgot...) | 21:42 |
| *** tatyana has joined #openstack-ironic | 21:43 | |
| devananda | lifeless: is there further action required based on the content of https://etherpad.openstack.org/p/ironic-nova-friction in the immediate future? | 21:43 |
| lifeless | devananda: I suspect not | 21:50 |
| NobodyCam | brb | 21:50 |
| lifeless | devananda: separately though we'll need a thing to stop ironic deploying to itself | 21:51 |
| lifeless | devananda: this affects the hash ring | 21:51 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Fix seamicro validate() method definition https://review.openstack.org/83598 | 21:51 |
| devananda | lifeless: hah | 21:51 |
| devananda | lifeless: or at least deploying itself into the sme control plane | 21:51 |
| *** epim has quit IRC | 21:52 | |
| lifeless | devananda: specifically the conductor shouldn't try to deploy an image to a node which it is running on | 21:52 |
| devananda | lifeless: if ironic and ironic' use different DB instances (and none of the enroled hw is duplicated) it should be OK | 21:52 |
| devananda | wait | 21:52 |
| comstud | devananda: question for you re: object code in ironic when you have a minute | 21:52 |
| devananda | ironic won't deploy an image to a node that has an instance on it | 21:53 |
| devananda | lifeless: so i dont think taht's a problem. you need to delete the current instance first | 21:53 |
| *** matty_dubs is now known as matty_dubs|gone | 21:53 | |
| devananda | comstud: shoot | 21:53 |
| lifeless | devananda: nova rebuild | 21:53 |
| comstud | devananda: There's a slight bug so I was going to put up a fix. It's related to the @base.remotable decorators... for example, on Node.save | 21:54 |
| comstud | devananda: Do you ever intend to allow those to be proxied to anything? I don't think they are needed in Ironic | 21:54 |
| comstud | in nova... they are used to force a save() to proxy over to nova-conductor to execute there. | 21:54 |
| devananda | lifeless: oh. you're thinking, "what if novarebuild of a ironic instance that ironic deployed" | 21:54 |
| devananda | lifeless: yea... that's fun. braincandy for another day :) | 21:55 |
| lifeless | devananda: scenario: take three machines, deploy ironic on all of them using ironic from a fourth machine. uninstall ironic from the fourth machine. Now do 'nova rebuild' of one of those nodes. | 21:55 |
| lifeless | devananda: this is the undercloud upgrade process :) | 21:55 |
| devananda | comstud: well. hypothetically, rpcapi.update_node could be wrapped in @base.remotable | 21:56 |
| lifeless | devananda: but equally, 'nova delete' on one, may turn itself off hard without cmopleting the delete. | 21:56 |
| comstud | devananda: What service would you proxy it back to? | 21:56 |
| devananda | comstud: i dont see any particular benefit to that. do you? | 21:56 |
| devananda | comstud: conductor | 21:56 |
| comstud | from API? | 21:56 |
| comstud | I'm not sure I do | 21:56 |
| devananda | comstud: yes. that's what we do now, we just didn't use @remotable | 21:57 |
| comstud | ah | 21:57 |
| comstud | Gotcha | 21:57 |
| devananda | when API needs to udpate a node, it calls rpcapi.update_node | 21:57 |
| devananda | all the work happens in conductor | 21:57 |
| comstud | under a lock I assume | 21:57 |
| devananda | exactly | 21:57 |
| devananda | and with some other handling | 21:57 |
| comstud | If you wanted to use the proxy stuff built into objects... | 21:57 |
| comstud | that would mean moving the locking into save() | 21:57 |
| devananda | comstud: i would only want that to effect save() on the API side | 21:58 |
| comstud | right | 21:58 |
| devananda | i dunno if objects could do that? | 21:58 |
| comstud | So, you could... a bit hackily maybe.. | 21:58 |
| comstud | How the indirection stuff works.. | 21:58 |
| *** eguz has joined #openstack-ironic | 21:58 | |
| comstud | in conductor you'd have a object_action method | 21:58 |
| devananda | yea, i didn't dig into the indirection enough to grok it. seems needlessly complex for ironic's use case | 21:59 |
| comstud | it gets as args the object method to call with its args | 21:59 |
| comstud | you could do it in there ahead of calling object.save() | 21:59 |
| comstud | yeah, that's what I'm thinking here | 21:59 |
| comstud | so I'm considering just removing the decorators | 22:00 |
| *** tatyana has left #openstack-ironic | 22:00 | |
| comstud | I'll give you a quick background on the issue I want to fix.. | 22:00 |
| comstud | objects are supposed to cache the context used on get() | 22:00 |
| comstud | such that you only need to do object.save() | 22:00 |
| *** tatyana has joined #openstack-ironic | 22:00 | |
| devananda | lifeless: right. i get the scenarios. need to map out the implications and interactions, and dont have the bandwidth right now | 22:00 |
| comstud | however, save() has a context arg... because unfrotunately RPC requires it... | 22:00 |
| comstud | but it's only required if you have the decorator | 22:00 |
| devananda | comstud: ahhh. i think this is a thing that's been annoying me for a while | 22:01 |
| comstud | the bug right now is... the Node object doesn't store the context on get_by_uuid() and it should | 22:01 |
| devananda | yep | 22:01 |
| devananda | comstud: please fix that :-D | 22:01 |
| comstud | I pointed jroll to a couple of diffs and am going to fix it | 22:01 |
| comstud | yeah | 22:01 |
| devananda | awesome! | 22:01 |
| lifeless | devananda: I think its fairly easy | 22:01 |
| lifeless | devananda: monday perhaps ? | 22:01 |
| comstud | as a part of it.. I could remove the decorators to remove confusion :) | 22:01 |
| comstud | or I could just leave them. | 22:01 |
| devananda | lifeless: when the list of RC1 bugs reaches 0 :) | 22:02 |
| lifeless | devananda: mmmm | 22:02 |
| * devananda points out that we almost reached 0 today, and now it's up to 5 | 22:02 | |
| devananda | https://launchpad.net/ironic/+milestone/icehouse-rc1 | 22:02 |
| *** eguz has quit IRC | 22:02 | |
| comstud | hah | 22:03 |
| *** eghobo has quit IRC | 22:03 | |
| devananda | lifeless: anything else you feel is reeeeealy important for icehouse, please file it and poke me to tag it | 22:03 |
| *** eghobo has joined #openstack-ironic | 22:03 | |
| devananda | lifeless: ya'll having tripleo start using ironic now is awesome but oh how I wish it had happened just one week earlier :) | 22:03 |
| lifeless | devananda: so do I but people bandwidth is what it is | 22:07 |
| NobodyCam | so close | 22:07 |
| devananda | lifeless: indeed it is | 22:08 |
| *** derekh has joined #openstack-ironic | 22:08 | |
| *** thehybridtech has quit IRC | 22:13 | |
| openstackgerrit | Josh Gachnang proposed a change to openstack/ironic-python-agent: Adding DynamicLoopingCall around lookup https://review.openstack.org/83604 | 22:19 |
| *** tatyana has left #openstack-ironic | 22:25 | |
| devananda | JoshNang: take a look at https://review.openstack.org/#/c/83105/ | 22:28 |
| devananda | JoshNang: and https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/ipmitool.py#L154 | 22:29 |
| devananda | JoshNang: maybe im totally wrong, but it might be neat to reuse some of this code eventually | 22:29 |
| devananda | JoshNang: that's three places that are basically doing the same thing | 22:29 |
| JoshNang | devananda: oh that's handy | 22:30 |
| JoshNang | not sure how i missed the ipmitool one when grepping around to see how other people did this | 22:30 |
| devananda | JoshNang: it landed this week ;) | 22:31 |
| JoshNang | ha that'll do it | 22:31 |
| devananda | speaking of... Shrews -- you might want to incorporate some of that into the retry stuff you're working on, too (eventually) | 22:32 |
| devananda | Shrews: exponential back-off of retries, i mean | 22:32 |
| JoshNang | it seems like the main issue is not having a timeout on DynamicLoopingCall | 22:33 |
| JoshNang | or having a sub/replacement class that does ExponentialBackoffLoopingCall. seems like a very handy thing to have | 22:34 |
| devananda | ExponentialBackoffWithTimeout | 22:34 |
| devananda | or such | 22:34 |
| devananda | seems like a reasonable thing to add to oslo | 22:35 |
| openstackgerrit | Alex Gaynor proposed a change to openstack/ironic-python-agent: Added PyPy env to tox.ini https://review.openstack.org/83605 | 22:35 |
| JoshNang | definitely. i'll add it to my to do list | 22:37 |
| openstackgerrit | Josh Gachnang proposed a change to openstack/ironic-python-agent: Adding DynamicLoopingCall around lookup https://review.openstack.org/83604 | 22:45 |
| openstackgerrit | Jim Rollenhagen proposed a change to openstack/ironic-python-agent: Change argument name in service file https://review.openstack.org/83608 | 22:46 |
| devananda | SpamapS: i believe this should fix the issue you ran into this morning, if you also pull the latest stevedore from trunk -- https://review.openstack.org/#/c/83572/ | 22:46 |
| SpamapS | devananda: you mean stevedore 0.15 ? ;) | 22:47 |
| SpamapS | (read the ML :) | 22:47 |
| devananda | ahh | 22:47 |
| devananda | yep | 22:47 |
| devananda | and sure enough, it works | 22:47 |
| devananda | http://paste.openstack.org/show/74507/ | 22:48 |
| devananda | :) | 22:48 |
| devananda | dhellmann: ^ tanks! | 22:48 |
| devananda | lifeless: NobodyCam: have a moment to review https://review.openstack.org/#/c/83572/ ? | 22:49 |
| * NobodyCam clicks | 22:49 | |
| NobodyCam | devananda: +2'd I did not test as i'm in the middle of a devtest run.. so I used your paste :) | 23:02 |
| devananda | ack | 23:02 |
| NobodyCam | afternoon walkies bbaifm | 23:04 |
| devananda | lifeless: comment on https://review.openstack.org/#/c/66621/8 -- are you concerned about the performance cost of compiling a regex on every API request? | 23:05 |
| jroll | devananda: I can't speak for lifeless, but it's a fairly simple regex and only two URLs it's compiling for | 23:09 |
| jroll | I was thinking about the same thing, but I think it's fairly negligible | 23:10 |
| devananda | humm | 23:13 |
| devananda | jroll: so I just added a LOG line to auth_token to see when it's actually compiling the regex | 23:14 |
| devananda | perhaps i've been staring at this screen for too long now | 23:14 |
| NobodyCam | why? | 23:14 |
| devananda | but it's not logging | 23:14 |
| *** eghobo has quit IRC | 23:15 | |
| jroll | devananda: is that middleware even installed? | 23:16 |
| devananda | that's waht i'm wondering | 23:16 |
| jroll | I don't think it is | 23:16 |
| jroll | heh | 23:16 |
| devananda | heh | 23:17 |
| devananda | so what's this patch for? :) | 23:17 |
| jroll | for fun, clearly | 23:17 |
| jroll | *should* we be using that middleware? | 23:17 |
| *** romcheg has quit IRC | 23:18 | |
| jroll | oh wait | 23:18 |
| jroll | the acl hook uses the AuthTokenMiddleware | 23:18 |
| devananda | right | 23:18 |
| jroll | yeah, so it should be used | 23:20 |
| devananda | the module is loaded | 23:20 |
| devananda | but the class isn't isntantiated | 23:20 |
| jroll | it should be, with the default pecan config | 23:20 |
| jroll | enable_acl is True | 23:20 |
| jroll | acl.install() returns the wsgi app wrapped with AuthTokoenMiddleware | 23:21 |
| jroll | s/Tokoen/Token | 23:21 |
| devananda | oh, sec | 23:22 |
| openstackgerrit | A change was merged to openstack/ironic: Nova Ironic driver to set pxe_swap_mb in Ironic https://review.openstack.org/83131 | 23:22 |
| devananda | yea, it's late ... | 23:22 |
| devananda | heh | 23:22 |
| devananda | wrong test env | 23:22 |
| jroll | lol | 23:22 |
| devananda | i had auth disabled in /taht/ env | 23:22 |
| jroll | right | 23:23 |
| devananda | testing again | 23:23 |
| jroll | oh yeah, this shouldn't be compiling on every request | 23:23 |
| jroll | just at setup time | 23:23 |
| * jroll has also been staring too long | 23:23 | |
| jroll | only the match() is executed at request time | 23:24 |
| devananda | yea | 23:24 |
| devananda | and it is | 23:24 |
| devananda | +A'd | 23:24 |
| jroll | nice | 23:24 |
| devananda | ok... with that, i'm done for a while | 23:25 |
| * NobodyCam thinks devananda needs more tea (or coffee) | 23:25 | |
| NobodyCam | lol | 23:25 |
| devananda | NobodyCam: ++ !!! | 23:25 |
| NobodyCam | :) | 23:25 |
| NobodyCam | have a good night devananda | 23:25 |
| jroll | night deva | 23:25 |
| devananda | NobodyCam: one last one for you -- since I dont want to push my own patch through | 23:26 |
| devananda | https://review.openstack.org/#/c/83598/ | 23:26 |
| NobodyCam | can we write a test that tests that our tests have run? | 23:28 |
| lifeless | devananda: it compiles once | 23:28 |
| NobodyCam | I really tried to get another test into that one | 23:28 |
| lifeless | devananda: which is why I was fine with it | 23:28 |
| *** jgrimm has quit IRC | 23:28 | |
| *** packet has quit IRC | 23:29 | |
| openstackgerrit | A change was merged to openstack/ironic: Raise unexpected exceptions during destroy() https://review.openstack.org/82948 | 23:36 |
| NobodyCam | lifeless: question: can ironic reffer to itself as "an OpenStack project" or do we have to use "an Incubated OpenStack project"?? do you know? | 23:39 |
| NobodyCam | in say our user doc's | 23:40 |
| NobodyCam | mordred: ^^^ would know the answer? | 23:40 |
| mordred | NobodyCam: I believe it's well within scope in your user docs to call yourself an OpenStack project | 23:43 |
| mordred | NobodyCam: you receive ATC credit for working on it - you are contributing to openstack in an approved program | 23:43 |
| mordred | Ironic is not part of the Integrated Release | 23:43 |
| NobodyCam | :) | 23:43 |
| mordred | but that does not make it a part of the OpenStack effort | 23:43 |
| NobodyCam | very good... Thak you | 23:43 |
| NobodyCam | thank even :) | 23:44 |
| jroll | looking at the nova driver | 23:48 |
| jroll | seems pretty hardcoded to the pxe driver | 23:48 |
| jroll | are we going to need to write another nova driver for the agent, or just abstract things out a bunch? | 23:49 |
| *** early has quit IRC | 23:49 | |
| NobodyCam | jroll: https://github.com/openstack/ironic/blob/master/ironic/nova/virt/ironic/ironic_driver_fields.py was attempting to deal with just that issue | 23:51 |
| NobodyCam | where are you seeing conflict | 23:51 |
| jroll | yeah | 23:51 |
| jroll | I saw pxe all over driver.py | 23:51 |
| jroll | but didn't read close enough until now to see "if 'pxe' in node.driver" | 23:51 |
| jroll | so we can just add to that, I suppose | 23:52 |
| NobodyCam | we tried not to pigon hole our self's into just pxe.. we may not have be secussful everywhere | 23:53 |
| *** early has joined #openstack-ironic | 23:53 | |
| jroll | heh | 23:53 |
| NobodyCam | successful even | 23:53 |
| jroll | don't worry, I'll make sure of that very soon :) | 23:53 |
| NobodyCam | driver fields file | 23:54 |
| jroll | yeah | 23:54 |
| jroll | I'm going to implement the agent stuff as much as possible over the next few days / maybe a week | 23:54 |
| *** stevehuang has quit IRC | 23:55 | |
| jroll | NobodyCam: do you know, should cpu_arch be 'amd64' for 64-bit machines? | 23:55 |
| jroll | or something else | 23:55 |
| NobodyCam | yes | 23:56 |
| jroll | thanks | 23:56 |
| *** Madasi has quit IRC | 23:57 | |
| NobodyCam | jroll: i should it can be something else but the standard toolng uses amd64 | 23:57 |
| *** derekh has quit IRC | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
