| *** Sukhdev has quit IRC | 00:09 | |
| *** naohirot has joined #openstack-ironic | 00:11 | |
| *** chlong has joined #openstack-ironic | 00:12 | |
| *** garthb has quit IRC | 00:20 | |
| *** shadower has quit IRC | 00:23 | |
| *** shadower has joined #openstack-ironic | 00:23 | |
| *** spandhe has joined #openstack-ironic | 00:27 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Fix RPC API compat for 1.29 upgrade https://review.openstack.org/234450 | 00:29 |
|---|---|---|
| jroll | ^ that should be good to go, I think | 00:29 |
| jroll | would like to backport that to liberty as well | 00:30 |
| *** sdake has joined #openstack-ironic | 00:43 | |
| *** baoli has joined #openstack-ironic | 00:43 | |
| *** baoli_ has joined #openstack-ironic | 00:46 | |
| *** baoli has quit IRC | 00:49 | |
| *** baoli_ has quit IRC | 00:52 | |
| *** links has joined #openstack-ironic | 00:53 | |
| *** lazy_prince has joined #openstack-ironic | 00:58 | |
| *** sdake has quit IRC | 00:58 | |
| *** baoli has joined #openstack-ironic | 01:00 | |
| *** Sukhdev_ has quit IRC | 01:00 | |
| *** jcoufal_ has quit IRC | 01:01 | |
| *** sdake has joined #openstack-ironic | 01:03 | |
| *** zhenguo has joined #openstack-ironic | 01:15 | |
| *** baoli has quit IRC | 01:15 | |
| *** praneshp has quit IRC | 01:15 | |
| *** baoli has joined #openstack-ironic | 01:18 | |
| rloo | jroll: awesome about the ironic versioned documentation. is it automatic now (ie it just happens, i don't have to know how) | 01:24 |
| rloo | jroll: wrt the rpc fix, i'll try to look tonight or tomorrow morning | 01:24 |
| jroll | rloo: yes, it's automatic, always has been apparently | 01:25 |
| jroll | and thanks. | 01:25 |
| *** Nisha has joined #openstack-ironic | 01:26 | |
| *** mmedvede has quit IRC | 01:26 | |
| rloo | jroll: cool. now to figure out how to document where the documentation is :D | 01:26 |
| jroll | hahaha | 01:26 |
| *** Nisha has quit IRC | 01:30 | |
| *** lazy_prince has quit IRC | 01:31 | |
| *** lazy_prince has joined #openstack-ironic | 01:32 | |
| *** chenglch has joined #openstack-ironic | 01:32 | |
| *** stendulker has joined #openstack-ironic | 01:32 | |
| *** killer_prince has joined #openstack-ironic | 01:33 | |
| *** dims has quit IRC | 01:33 | |
| *** dims has joined #openstack-ironic | 01:34 | |
| *** baoli has quit IRC | 01:34 | |
| *** baoli has joined #openstack-ironic | 01:35 | |
| *** lazy_prince has quit IRC | 01:37 | |
| rloo | jroll: it seems to me that you were going to update the documentation for liberty, but i don't recall what/why. was it the release notes or upgrade stuff? | 01:40 |
| jroll | rloo: both? :) | 01:40 |
| jroll | rloo: 4.2.1 release notes, and some stuff about the out-of-tree driver breakage | 01:41 |
| rloo | jroll: I mean, after the stable/liberty branch was cut. i have such a short memory. | 01:41 |
| rloo | jroll: good, you remember! | 01:41 |
| *** ijw has quit IRC | 01:41 | |
| jroll | rloo: well, that's what I remember, I'm not sure that's accurate :) | 01:41 |
| rloo | jroll: i was thinking that now i know there's a doc ../ironic/4.2.1 url, we could go to town on our install guide etc, but NO, not going to do that. | 01:42 |
| jroll | rloo: hah | 01:44 |
| *** baoli has quit IRC | 01:49 | |
| *** baoli has joined #openstack-ironic | 01:50 | |
| *** baoli has quit IRC | 02:00 | |
| *** baoli has joined #openstack-ironic | 02:00 | |
| *** alex_xu has joined #openstack-ironic | 02:02 | |
| *** spandhe has quit IRC | 02:02 | |
| *** baoli has quit IRC | 02:05 | |
| *** sdake has quit IRC | 02:06 | |
| *** chlong has quit IRC | 02:06 | |
| *** stendulker has quit IRC | 02:09 | |
| *** harshs has quit IRC | 02:18 | |
| *** mtanino has quit IRC | 02:20 | |
| openstackgerrit | Naohiro Tamura proposed openstack/ironic-python-agent: Add param docstring into the normalize func https://review.openstack.org/234519 | 02:22 |
| *** chlong has joined #openstack-ironic | 02:23 | |
| *** boris-42 has quit IRC | 02:28 | |
| *** baoli has joined #openstack-ironic | 02:31 | |
| *** killer_prince has quit IRC | 02:40 | |
| *** killer_prince has joined #openstack-ironic | 02:40 | |
| *** Haomeng|2 has joined #openstack-ironic | 02:45 | |
| *** killer_prince has quit IRC | 02:48 | |
| *** lazy_prince has joined #openstack-ironic | 02:48 | |
| *** Haomeng has quit IRC | 02:48 | |
| *** penick has joined #openstack-ironic | 02:50 | |
| *** lazy_prince has quit IRC | 02:53 | |
| *** lazy_prince has joined #openstack-ironic | 02:54 | |
| *** achanda has joined #openstack-ironic | 02:57 | |
| *** lazy_prince has quit IRC | 03:05 | |
| *** amotoki has joined #openstack-ironic | 03:14 | |
| *** getvasanth_ has joined #openstack-ironic | 03:19 | |
| *** penick has quit IRC | 03:19 | |
| *** getvasanth has quit IRC | 03:22 | |
| *** chlong has quit IRC | 03:29 | |
| *** amotoki has quit IRC | 03:30 | |
| openstackgerrit | Merged openstack/ironic: Move FakeOneViewDriver to the fake.py module https://review.openstack.org/233753 | 03:32 |
| *** rameshg87 has joined #openstack-ironic | 03:35 | |
| *** penick has joined #openstack-ironic | 03:35 | |
| *** Nisha has joined #openstack-ironic | 03:40 | |
| *** amotoki has joined #openstack-ironic | 03:41 | |
| *** purp_too is now known as purp | 03:45 | |
| *** chlong has joined #openstack-ironic | 03:45 | |
| *** baoli has quit IRC | 03:46 | |
| *** amotoki has quit IRC | 03:50 | |
| *** harshs has joined #openstack-ironic | 03:54 | |
| *** stendulker has joined #openstack-ironic | 03:54 | |
| *** amotoki has joined #openstack-ironic | 03:55 | |
| *** david-lyle has quit IRC | 04:01 | |
| *** stendulker_ has joined #openstack-ironic | 04:02 | |
| *** stendulker has quit IRC | 04:02 | |
| *** rloo has quit IRC | 04:03 | |
| *** MattMan has quit IRC | 04:10 | |
| *** MattMan has joined #openstack-ironic | 04:10 | |
| *** stendulker has joined #openstack-ironic | 04:13 | |
| *** stendulker_ has quit IRC | 04:13 | |
| *** praneshp has joined #openstack-ironic | 04:14 | |
| *** amotoki has quit IRC | 04:18 | |
| *** garthb has joined #openstack-ironic | 04:25 | |
| *** spandhe has joined #openstack-ironic | 04:25 | |
| *** amotoki has joined #openstack-ironic | 04:26 | |
| *** garthb has quit IRC | 04:28 | |
| *** Sukhdev has joined #openstack-ironic | 04:28 | |
| *** garthb has joined #openstack-ironic | 04:28 | |
| *** ijw has joined #openstack-ironic | 04:40 | |
| *** praneshp_ has joined #openstack-ironic | 04:42 | |
| *** praneshp has quit IRC | 04:43 | |
| *** praneshp_ is now known as praneshp | 04:43 | |
| *** harshs has quit IRC | 04:44 | |
| *** puranamr has joined #openstack-ironic | 04:46 | |
| *** chlong has quit IRC | 04:49 | |
| *** penick has quit IRC | 04:57 | |
| *** pradipta has joined #openstack-ironic | 05:04 | |
| *** kozhukalov_ has joined #openstack-ironic | 05:13 | |
| *** Sukhdev has quit IRC | 05:20 | |
| *** lazy_prince has joined #openstack-ironic | 05:22 | |
| *** boris-42 has joined #openstack-ironic | 05:22 | |
| *** VikasC has joined #openstack-ironic | 05:25 | |
| *** mrmoje has joined #openstack-ironic | 05:28 | |
| *** lintan_ has joined #openstack-ironic | 05:30 | |
| *** BharatK has joined #openstack-ironic | 05:34 | |
| *** ijw has quit IRC | 05:45 | |
| *** puranamr has quit IRC | 05:53 | |
| *** kozhukalov_ has quit IRC | 05:56 | |
| *** smoriya has quit IRC | 05:59 | |
| *** VikasC_ has joined #openstack-ironic | 06:02 | |
| *** VikasC has quit IRC | 06:05 | |
| *** spandhe has quit IRC | 06:10 | |
| *** chlong has joined #openstack-ironic | 06:20 | |
| *** meghal has joined #openstack-ironic | 06:22 | |
| meghal | hello…had a quick question on hash ring implementation in ironic-conductor…currently, at various points the mapping function is called to check if a node belongs to a particular conductor…how frequently is this mapping updated ? is it only during sync_local_states when the hash ring is reset ? | 06:25 |
| meghal | jroll: devananda ^ ? | 06:25 |
| *** ukalifon1 has joined #openstack-ironic | 06:27 | |
| *** chlong has quit IRC | 06:27 | |
| *** harshs has joined #openstack-ironic | 06:27 | |
| *** meghal has quit IRC | 06:29 | |
| *** meghal has joined #openstack-ironic | 06:30 | |
| *** Marga_ has quit IRC | 06:42 | |
| *** harshs has quit IRC | 06:47 | |
| *** yog_ has joined #openstack-ironic | 06:48 | |
| *** garthb has quit IRC | 06:55 | |
| *** romainh has joined #openstack-ironic | 07:00 | |
| *** yog_ has quit IRC | 07:00 | |
| Haomeng|2 | meghal: I think so, need other to confirm:) | 07:01 |
| *** openstackgerrit has quit IRC | 07:01 | |
| *** openstackgerrit has joined #openstack-ironic | 07:02 | |
| *** Nisha has quit IRC | 07:02 | |
| *** Nisha has joined #openstack-ironic | 07:08 | |
| meghal | Haomeng|2: ok…will wait for others to confirm | 07:11 |
| *** Marga_ has joined #openstack-ironic | 07:11 | |
| Haomeng|2 | meghal: k, welcome:) | 07:11 |
| meghal | was wondering if everybody runs sync_local_state periodic job, because if I am not interested in preparing the pxe config env on take over (update of conductor affinity) then sync_local_state is not needed | 07:12 |
| meghal | but I do not want to affect the hash ring calculation | 07:12 |
| *** yog_ has joined #openstack-ironic | 07:13 | |
| *** Marga_ has quit IRC | 07:14 | |
| *** Marga_ has joined #openstack-ironic | 07:14 | |
| Haomeng|2 | meghal: I understand the periodic job is auto-run, so no need to manually run it | 07:14 |
| meghal | Haomeng|2: I meant that if there was an option to disable the auto-run | 07:16 |
| Haomeng|2 | meghal: let me check code, looks like there is no such option | 07:16 |
| meghal | because if a particular use case does not need it, then I can avoid the run of those periodic workers | 07:16 |
| meghal | Haomeng|2: such an option does not exist | 07:16 |
| Haomeng|2 | meghal: yes | 07:16 |
| Haomeng|2 | meghal: so you mean that that should be useful if there it is | 07:17 |
| meghal | what I want to understand is the implications of not running sync_local_state on hash ring | 07:17 |
| meghal | and if there are none, then I would like to have an option to disable the auto run | 07:17 |
| meghal | and we can discuss more on that | 07:17 |
| meghal | but first lets try to get an answer on this | 07:17 |
| *** dtantsur|afk is now known as dtantsur | 07:20 | |
| dtantsur | Morning Ironic | 07:20 |
| *** mrmoje has quit IRC | 07:21 | |
| *** mrmoje has joined #openstack-ironic | 07:24 | |
| Haomeng|2 | dtantsur: morning:) | 07:24 |
| Haomeng|2 | meghal: ok | 07:24 |
| *** Haomeng|2 is now known as Haomeng | 07:24 | |
| *** achanda has quit IRC | 07:26 | |
| *** achanda has joined #openstack-ironic | 07:27 | |
| openstackgerrit | chenglch proposed openstack/ironic: Add iscsi_opc and agent_opc to manage the OpenPOWER servers https://review.openstack.org/209824 | 07:29 |
| *** ifarkas has joined #openstack-ironic | 07:32 | |
| *** achanda has quit IRC | 07:33 | |
| *** mkovacik has joined #openstack-ironic | 07:33 | |
| *** ijw has joined #openstack-ironic | 07:37 | |
| *** getvasanth has joined #openstack-ironic | 07:38 | |
| *** getvasanth_ has quit IRC | 07:39 | |
| *** meghal1 has joined #openstack-ironic | 07:40 | |
| *** Nisha has quit IRC | 07:40 | |
| *** mkovacik has quit IRC | 07:40 | |
| *** meghal has quit IRC | 07:41 | |
| dtantsur | jroll, something we might want to get fixed soon (maybe also in L): https://bugs.launchpad.net/ironic-python-agent/+bug/1505923 | 07:50 |
| openstack | Launchpad bug 1505923 in ironic-python-agent "Does not works with the latest tgt default socket name" [High,Triaged] | 07:50 |
| *** marzif has quit IRC | 07:52 | |
| *** david-lyle has joined #openstack-ironic | 07:57 | |
| *** jcoufal has joined #openstack-ironic | 07:59 | |
| *** Marga_ has quit IRC | 08:01 | |
| *** ijw has quit IRC | 08:01 | |
| *** ijw has joined #openstack-ironic | 08:03 | |
| *** marzif has joined #openstack-ironic | 08:06 | |
| *** ijw has quit IRC | 08:07 | |
| *** athomas has joined #openstack-ironic | 08:08 | |
| *** mgoddard has joined #openstack-ironic | 08:10 | |
| *** mbound has joined #openstack-ironic | 08:15 | |
| *** pradipta has quit IRC | 08:17 | |
| *** jistr has joined #openstack-ironic | 08:18 | |
| *** bapalm has quit IRC | 08:23 | |
| *** Cameron_C_ has quit IRC | 08:25 | |
| *** bapalm has joined #openstack-ironic | 08:26 | |
| *** ndipanov has joined #openstack-ironic | 08:27 | |
| getvasanth | Why there is only one ironic-conductor running on a 4 core box, where has the heat-engine runs 5 engines (1 Parent, 4 forks), do we have any config like num_engine_worker(HEAT) in Ironic too? | 08:28 |
| *** marzif has quit IRC | 08:38 | |
| *** marzif has joined #openstack-ironic | 08:39 | |
| Haomeng | getvasanth: no, conductor will not run per-core I think | 08:39 |
| *** meghal1 has quit IRC | 08:39 | |
| Haomeng | getvasanth: we have no such workers configuration for conductor | 08:40 |
| *** jamielennox is now known as jamielennox|away | 08:40 | |
| *** lucasagomes has joined #openstack-ironic | 08:41 | |
| *** mkovacik has joined #openstack-ironic | 08:41 | |
| *** lucasagomes_ has joined #openstack-ironic | 08:42 | |
| *** Nisha has joined #openstack-ironic | 08:42 | |
| *** lucasagomes has quit IRC | 08:46 | |
| *** derekh has joined #openstack-ironic | 08:50 | |
| *** mrmoje has quit IRC | 08:53 | |
| *** subscope has joined #openstack-ironic | 08:54 | |
| *** lucasagomes has joined #openstack-ironic | 08:58 | |
| *** subscope has quit IRC | 09:01 | |
| *** subscope has joined #openstack-ironic | 09:01 | |
| *** ijw has joined #openstack-ironic | 09:03 | |
| *** athomas has quit IRC | 09:05 | |
| *** lucasagomes has quit IRC | 09:07 | |
| *** lucasagomes_ is now known as lucasagomes | 09:08 | |
| lucasagomes | o/ streamed raw images in IPA works | 09:08 |
| * lucasagomes writes tests now | 09:08 | |
| *** ijw has quit IRC | 09:09 | |
| *** athomas has joined #openstack-ironic | 09:14 | |
| *** pelix has joined #openstack-ironic | 09:14 | |
| *** lintan has joined #openstack-ironic | 09:15 | |
| *** lintan_ has quit IRC | 09:16 | |
| *** dims_ has joined #openstack-ironic | 09:17 | |
| *** marzif has quit IRC | 09:19 | |
| *** marzif has joined #openstack-ironic | 09:19 | |
| *** dims has quit IRC | 09:20 | |
| dtantsur | lucasagomes, awesome! | 09:22 |
| *** links has quit IRC | 09:22 | |
| *** dtantsur is now known as dtantsur|brb | 09:24 | |
| zhenguo | hi ironic, does somebody knows why ListType return set() here https://github.com/openstack/ironic/blob/master/ironic/api/controllers/v1/types.py#L204 | 09:28 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ironic: Updated from global requirements https://review.openstack.org/234664 | 09:29 |
| zhenguo | I find the comments said set() to remove duplicated items, but why not use list(set()) to convert back to list? | 09:29 |
| *** mmnelemane has joined #openstack-ironic | 09:29 | |
| openstackgerrit | Merged openstack/ironic-python-agent: Add param docstring into the normalize func https://review.openstack.org/234519 | 09:36 |
| lucasagomes | zhenguo, yes I think it's just to remove the duplicates | 09:38 |
| lucasagomes | zhenguo, good point, it could do that too | 09:39 |
| *** subscope has quit IRC | 09:39 | |
| zhenguo | lucasagomes: so it's ok that I push a patch to update that | 09:39 |
| lucasagomes | zhenguo, right it's fine yes | 09:39 |
| lucasagomes | zhenguo, tho, what are the gains with it? (Perhaps it's more intuite to be a list... not sure) | 09:39 |
| zhenguo | lucasagomes: as I need to use Listtype as a response, but set() is not json serializable | 09:40 |
| lucasagomes | zhenguo, alright yeah ++ | 09:41 |
| zhenguo | lucasagomes: ok, thanks | 09:41 |
| sambetts | o/ Morning all | 09:45 |
| *** sambetts has quit IRC | 09:56 | |
| *** sambetts has joined #openstack-ironic | 09:57 | |
| openstackgerrit | Zhenguo Niu proposed openstack/ironic: Convert set() to list in ListType https://review.openstack.org/234697 | 09:57 |
| *** naohirot has quit IRC | 09:57 | |
| zhenguo | lucasagomes, dtantsur: can you help to approve https://review.openstack.org/#/c/192576/ , already got two +2s. | 10:02 |
| lucasagomes | zhenguo, comment on the list type, just update the test to check for the type please | 10:03 |
| lucasagomes | will take a look | 10:03 |
| zhenguo | lucasagomes: thanks, will update soon :) | 10:03 |
| lucasagomes | cool thank YOU for finding the problem | 10:04 |
| *** ijw has joined #openstack-ironic | 10:04 | |
| *** athomas has quit IRC | 10:05 | |
| *** chenglch has quit IRC | 10:08 | |
| *** ijw has quit IRC | 10:09 | |
| *** ionutbalutoiu has joined #openstack-ironic | 10:11 | |
| *** romainh has quit IRC | 10:11 | |
| openstackgerrit | Zhenguo Niu proposed openstack/ironic: Convert set() to list in ListType https://review.openstack.org/234697 | 10:13 |
| * zhenguo goes home | 10:13 | |
| *** zhenguo has quit IRC | 10:13 | |
| *** chlong has joined #openstack-ironic | 10:14 | |
| *** kozhukalov_ has joined #openstack-ironic | 10:14 | |
| *** marzif has quit IRC | 10:20 | |
| *** marzif has joined #openstack-ironic | 10:20 | |
| *** subscope has joined #openstack-ironic | 10:22 | |
| *** ionutbalutoiu has quit IRC | 10:24 | |
| *** ionutbalutoiu has joined #openstack-ironic | 10:27 | |
| *** ionutbalutoiu has quit IRC | 10:27 | |
| *** ionutbalutoiu has joined #openstack-ironic | 10:27 | |
| *** kozhukalov_ has quit IRC | 10:32 | |
| *** dtantsur|brb is now known as dtantsur | 10:34 | |
| dtantsur | morning sambetts | 10:35 |
| sambetts | Morning dtantsur | 10:41 |
| *** kozhukalov_ has joined #openstack-ironic | 10:44 | |
| * sambetts needs coffee | 10:47 | |
| *** mrmoje has joined #openstack-ironic | 10:49 | |
| *** marzif has quit IRC | 10:50 | |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic: Documentation for Swift-less iLO Virtual Media deploy https://review.openstack.org/228020 | 10:52 |
| *** Marga_ has joined #openstack-ironic | 10:52 | |
| *** marzif has joined #openstack-ironic | 10:53 | |
| *** e0ne has joined #openstack-ironic | 10:56 | |
| *** mrmoje has quit IRC | 10:57 | |
| *** ijw has joined #openstack-ironic | 11:05 | |
| * rameshg87 goes home | 11:06 | |
| *** rameshg87 has quit IRC | 11:06 | |
| *** Nisha has quit IRC | 11:09 | |
| ifarkas | dtantsur, hi, could you please review https://review.openstack.org/#/c/214160/ and https://review.openstack.org/#/c/230445/ ? | 11:10 |
| *** ijw has quit IRC | 11:11 | |
| *** marzif has quit IRC | 11:12 | |
| *** ionutbalutoiu has quit IRC | 11:13 | |
| *** marzif has joined #openstack-ironic | 11:13 | |
| *** ionutbalutoiu has joined #openstack-ironic | 11:16 | |
| dtantsur | ifarkas, sure | 11:18 |
| ifarkas | dtantsur, thanks! | 11:21 |
| *** e0ne has quit IRC | 11:23 | |
| *** stendulker has quit IRC | 11:27 | |
| dtantsur | ifarkas, reviewer, hope you don't regret making me a core already :D | 11:30 |
| dtantsur | * reviewed | 11:30 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic-python-agent: Add support for streaming raw images directly onto the disk https://review.openstack.org/234339 | 11:30 |
| ifarkas | dtantsur, lol, no worries, thank for the reviews :D | 11:31 |
| ifarkas | I will fix them | 11:31 |
| lucasagomes | ionutbalutoiu, ^ you may want to look at it to stream ur windows image | 11:31 |
| *** bmpbike has joined #openstack-ironic | 11:32 | |
| ionutbalutoiu | lucasagomes, nice work :) | 11:33 |
| lucasagomes | ionutbalutoiu, :-) lemme know if that works for you, you will need to patch IPA and regenerate the ramdisk | 11:33 |
| *** BharatK has quit IRC | 11:35 | |
| ionutbalutoiu | Sure, not sure if I can give it a shot today, but I'll let you know as soon as I tested it. Thanks for the patch! | 11:36 |
| *** mrmoje has joined #openstack-ironic | 11:41 | |
| *** baoli has joined #openstack-ironic | 11:48 | |
| *** BharatK has joined #openstack-ironic | 11:48 | |
| *** baoli_ has joined #openstack-ironic | 11:49 | |
| *** ionutbalutoiu has quit IRC | 11:50 | |
| *** ionutbalutoiu has joined #openstack-ironic | 11:51 | |
| *** baoli has quit IRC | 11:52 | |
| *** subscope has quit IRC | 11:54 | |
| *** subscope has joined #openstack-ironic | 11:57 | |
| lucasagomes | ionutbalutoiu, no bothers :-) thanks | 11:59 |
| *** baoli_ has quit IRC | 12:00 | |
| *** persia has quit IRC | 12:02 | |
| *** bizarrochristy has joined #openstack-ironic | 12:03 | |
| *** cemason has quit IRC | 12:03 | |
| *** ijw has joined #openstack-ironic | 12:03 | |
| *** persia has joined #openstack-ironic | 12:03 | |
| *** lucasagomes is now known as lucas-hungry | 12:03 | |
| *** ijw has quit IRC | 12:07 | |
| *** subscope has quit IRC | 12:07 | |
| getvasanth | Haomeng: thanks for the info, but on a 4 core box i see only one conductor running | 12:08 |
| *** bizarrochristy has quit IRC | 12:12 | |
| *** thrash|g0ne is now known as thrash | 12:13 | |
| *** achanda has joined #openstack-ironic | 12:16 | |
| *** achanda has quit IRC | 12:17 | |
| *** achanda has joined #openstack-ironic | 12:18 | |
| *** achanda has quit IRC | 12:22 | |
| *** cemason has joined #openstack-ironic | 12:25 | |
| *** priteau has joined #openstack-ironic | 12:27 | |
| *** cemason has quit IRC | 12:30 | |
| *** marzif has quit IRC | 12:31 | |
| *** marzif has joined #openstack-ironic | 12:31 | |
| openstackgerrit | Julia Kreger proposed openstack/bifrost: Initial support for installation of ironic-inspector https://review.openstack.org/229930 | 12:31 |
| *** cemason has joined #openstack-ironic | 12:33 | |
| *** dprince has joined #openstack-ironic | 12:34 | |
| *** athomas has joined #openstack-ironic | 12:34 | |
| TheJulia | NobodyCam: cinerama: please review https://review.openstack.org/#/c/216372/6 when you have a chance | 12:34 |
| *** trown|outttypeww is now known as trown | 12:35 | |
| TheJulia | yay, just get started and a hard power outage | 12:35 |
| *** cemason has quit IRC | 12:37 | |
| *** baoli has joined #openstack-ironic | 12:37 | |
| *** subscope has joined #openstack-ironic | 12:43 | |
| *** sinval has quit IRC | 12:44 | |
| *** puranamr has joined #openstack-ironic | 12:45 | |
| *** ionutbalutoiu has quit IRC | 12:49 | |
| jroll | morning all | 12:51 |
| jroll | busy busy day | 12:51 |
| dtantsur | morning jroll, TheJulia | 12:51 |
| jroll | dtantsur: do we support building IPA with anything this latest tgt is packaged with? https://bugs.launchpad.net/ironic-python-agent/+bug/1505923 | 12:52 |
| openstack | Launchpad bug 1505923 in ironic-python-agent "Does not works with the latest tgt default socket name" [High,Triaged] | 12:52 |
| dtantsur | jroll, probably not now, at least I personally haven't hit it | 12:53 |
| jroll | dtantsur: yeah, not sure we need to backport it | 12:53 |
| *** puranamr has quit IRC | 12:54 | |
| *** BharatK has quit IRC | 12:54 | |
| dtantsur | jroll, I'd ask afazekas how he figured it out | 12:54 |
| *** puranamr has joined #openstack-ironic | 12:54 | |
| * dtantsur asks on #openstack-qa | 12:56 | |
| *** rloo has joined #openstack-ironic | 12:56 | |
| TheJulia | jroll: there is the ironic-agent element that someone use | 12:57 |
| dtantsur | seems like all supported distributions have old enough tgt | 12:58 |
| dtantsur | jroll, I'm much more worried about https://bugs.launchpad.net/ironic-python-agent/+bug/1504562 | 12:59 |
| openstack | Launchpad bug 1504562 in ironic-python-agent "Please support Linux-IO in extensions/iscsi in addition to tgtd" [Medium,Triaged] | 12:59 |
| sambetts | I hit this the other day when working with tinyipa | 12:59 |
| jroll | yeah, that's a thing that needs to happen | 13:00 |
| sambetts | debian packages v1.0.51 of tgt and I initally installed the most recent version of tgt and it blew up | 13:00 |
| jroll | so yeah, I don't think we need to backport either of those. let's get them fixed and make a release. upstream builders work fine right now. | 13:01 |
| *** amotoki has quit IRC | 13:02 | |
| dtantsur | jroll, I'm trying to understand this Linux-IO thing... so few docs and examples | 13:02 |
| jroll | dtantsur: yeah, I'm ignoring that until we're done with liberty :) | 13:03 |
| TheJulia | sambetts: the patch mentioned in th ebug is not listed in the changes for the debian package, fwiw | 13:03 |
| *** mgoddard has quit IRC | 13:03 | |
| sambetts | TheJulia: That what I meant the copy in the debian repo is the older version before this change | 13:05 |
| TheJulia | ahh | 13:05 |
| sambetts | The tinycore version was newer and so blew up, sorry it made sense in my head haha | 13:06 |
| jroll | xek: if we fix this RPC version bug thing, can you truly do a rolling upgrade? | 13:06 |
| jroll | xek: or is it still broken even if we fix this (I hope not, but I think it might be) | 13:06 |
| TheJulia | sambetts: Stable, but unstable has it :) | 13:06 |
| openstackgerrit | Julia Kreger proposed openstack/bifrost: Initial support for installation of ironic-inspector https://review.openstack.org/229930 | 13:07 |
| rloo | jroll: i would think it is still broken. I thought our versioned objects stuff was broken and we fixed things in liberty so we could do rolling upgrades from liberty->future. | 13:07 |
| sambetts | TheJulia: does testing? I think thats the version we pull for IPA | 13:07 |
| jroll | rloo: that's what I'm thinking, but our objects might be "close enough" to deal with it... | 13:07 |
| xek | rloo, you are right | 13:08 |
| jroll | rloo: if we drop this patch we can release 4.2.1 today, the rest of the backports are coming down | 13:08 |
| rloo | jroll: my understanding from what dan mentioned, is that it isn't. | 13:08 |
| jroll | ugh, so what's the use in backporting this | 13:08 |
| * jroll sighs and abandons his patch | 13:08 | |
| rloo | jroll: that's what/why i asked you yesterday. | 13:08 |
| openstackgerrit | Julia Kreger proposed openstack/bifrost: Initial support for installation of ironic-inspector https://review.openstack.org/229930 | 13:08 |
| xek | jroll, the commit I cited was just an example, we need to have some test to address similar bugs in the future | 13:08 |
| jroll | rloo: yeah, I'm a hopeful person | 13:08 |
| rloo | jroll: sorry, maybe i should have been clearer. | 13:08 |
| jroll | xek: agree | 13:08 |
| TheJulia | sambetts: testing has it | 13:08 |
| rloo | jroll: hope is good. and if the patch was 'good enough' but ... | 13:08 |
| jroll | xek: needs moar grenade-partial | 13:09 |
| xek | jroll, yep | 13:09 |
| rloo | jroll: sorry to ask, but do we want to get this doc into liberty: https://review.openstack.org/228020 | 13:09 |
| sambetts | TheJulia: :/ hmmm, not sure then | 13:10 |
| dtantsur | rloo, morning. last time I checked we didn't publish docs for releases | 13:10 |
| rloo | jroll: and did we decide to punt on the patch for AgentDeploy/boot interface | 13:10 |
| jroll | rloo: I'm not concerned about it. as long as it's out there for people to use, I'm happy | 13:10 |
| jroll | dtantsur: we found out yesterday we do \o/ | 13:10 |
| rloo | dtantsur: turns out we have been. jroll mentioned it yesterday! | 13:10 |
| dtantsur | ORLY?? Oo | 13:10 |
| dtantsur | any links at hand? | 13:10 |
| jroll | rloo: yep, both have breakage, without the 'fix' the break is easier to understand and fix downstream | 13:10 |
| jroll | dtantsur: sec | 13:10 |
| rloo | dtantsur: http://docs.openstack.org/developer/ironic/4.2.0/ | 13:10 |
| jroll | dtantsur: http://docs.openstack.org/developer/ironic/2015.1.2/ | 13:11 |
| jroll | every tag gets a subtree | 13:11 |
| *** subscope has quit IRC | 13:11 | |
| dtantsur | dear god, why didn't I know about it?? | 13:11 |
| rloo | dtantsur: why didn't any of us know about it?!! | 13:11 |
| *** lucas-hungry is now known as lucasagomes | 13:12 | |
| jroll | ++ | 13:13 |
| dtantsur | such surprise very docs wow publish | 13:13 |
| *** BharatK has joined #openstack-ironic | 13:13 | |
| dtantsur | wow wow | 13:13 |
| rloo | jroll: so i will assume unless you ping us that you don't need anything reviewed for stable/liberty (except doc/release notes updates) | 13:13 |
| jroll | hehe | 13:13 |
| *** marzif has quit IRC | 13:13 | |
| jroll | rloo: yeah, we'll need release notes and that should be it | 13:13 |
| *** romainh has joined #openstack-ironic | 13:14 | |
| *** amotoki has joined #openstack-ironic | 13:15 | |
| dtantsur | rloo, if you're interesting in helping with futurist change, please review https://review.openstack.org/#/c/232517 (a preliminary refactor) | 13:15 |
| rloo | dtantsur: ok. | 13:16 |
| *** marzif has joined #openstack-ironic | 13:16 | |
| rloo | dtantsur: but i was thinking i should look at the ironic patch that uses ironic-lib. is it ready to be reviewed? | 13:16 |
| dtantsur | rloo, IIRC not yet; and we still don't have ironic-lib bumped in g-r.. | 13:17 |
| rloo | dtantsur: I *hope* this ironic-lib will prove to be useful... | 13:17 |
| rloo | dtantsur: sigh. ok, will look at your patch then :) | 13:17 |
| jroll | I've filed https://blueprints.launchpad.net/ironic/+spec/online-upgrade-support | 13:18 |
| jroll | we need to do this stuff this cycle. | 13:18 |
| dtantsur | ++ | 13:18 |
| rloo | jroll: do we have a priorities list started anywhere? | 13:18 |
| jroll | we can review the ironic-lib changeover without the g-r bump :P | 13:18 |
| jroll | rloo: nothing updated in some time, no | 13:18 |
| jroll | I've kinda started sketching one out locally but nothing solid | 13:19 |
| * dtantsur dives into Linux-IO and iSCSI, which him the best of luck | 13:19 | |
| *** puranamr has quit IRC | 13:19 | |
| dtantsur | s/which/wish/ | 13:19 |
| rloo | jroll: as long as after the summit (or during) we have a good idea of the priorities | 13:19 |
| rloo | dtantsur: good luck | 13:19 |
| jroll | rloo +1 | 13:19 |
| *** subscope has joined #openstack-ironic | 13:19 | |
| sambetts | dtantsur: if you want any help ping me :) | 13:20 |
| dtantsur | cool, thanks :) | 13:20 |
| dtantsur | rtslib is AGPLv3, can't decide if wow or not wow | 13:24 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic: Fix an annoying detail in the developer quick-start https://review.openstack.org/234783 | 13:25 |
| lucasagomes | morning jroll rloo TheJulia | 13:26 |
| rloo | dtantsur: qq wrt 232517. Are you splitting it up cuz you want separate classes, or splitting it up cuz you don't want it all in one file? | 13:26 |
| rloo | hi lucasagomes (and everyone else!) | 13:27 |
| jroll | heya lucasagomes :) | 13:27 |
| dtantsur | rloo, cuz I hate editing this huge file and this huge class :) | 13:27 |
| rloo | dtantsur: so both then. | 13:27 |
| dtantsur | yep | 13:27 |
| *** subscope has quit IRC | 13:28 | |
| jroll | we fixed a whole one bug on 4.2.1 \o/ :| | 13:31 |
| jroll | https://launchpad.net/ironic/+milestone/4.2.1 | 13:31 |
| *** amotoki has quit IRC | 13:32 | |
| rloo | jroll: that should really have been two bugs :) | 13:33 |
| *** cdearborn has joined #openstack-ironic | 13:33 | |
| jroll | I know. | 13:33 |
| * jroll not happy about that but realized it too late | 13:33 | |
| jroll | I still don't know how someone successfully tested the first bug fix | 13:33 |
| rloo | jroll: if i had known you'd care, i would have said something. | 13:34 |
| jroll | let's just run our CI in every locale linux supports so we don't have to do this again | 13:34 |
| jroll | rloo: it's fine, it is what it is | 13:34 |
| *** caiobo has joined #openstack-ironic | 13:35 | |
| rloo | jroll: my guess/thought was that it only became an issue if the command failed and spewed stuff in stderr. | 13:35 |
| jroll | rloo: oh, that's a good point | 13:36 |
| rloo | jroll: i think that's why we already have some commands coded using standard locale; they are fixed when someone encounters the problem. | 13:37 |
| *** marzif has quit IRC | 13:37 | |
| jroll | yeah. | 13:37 |
| jlvillal | Good morning Ironic | 13:38 |
| jroll | hey jlvillal | 13:38 |
| *** subscope has joined #openstack-ironic | 13:39 | |
| jlvillal | jroll, Hey there. | 13:39 |
| *** amotoki has joined #openstack-ironic | 13:40 | |
| *** subscope has quit IRC | 13:44 | |
| *** subscope has joined #openstack-ironic | 13:44 | |
| *** amotoki has quit IRC | 13:47 | |
| *** puranamr has joined #openstack-ironic | 13:47 | |
| *** kkoski has joined #openstack-ironic | 13:49 | |
| lucasagomes | jroll, :-/ | 13:49 |
| lucasagomes | probably the bugs are just not tagged correctly? /me not sure | 13:49 |
| lucasagomes | jlvillal, morning! | 13:50 |
| jroll | lucasagomes: no, the bug was reused for the second bug fix | 13:51 |
| jroll | I looked through everything | 13:51 |
| lucasagomes | oh, alright cool | 13:51 |
| openstackgerrit | Julia Kreger proposed openstack/bifrost: Initial support for installation of ironic-inspector https://review.openstack.org/229930 | 13:54 |
| *** jistr is now known as jistr|mtg | 13:56 | |
| *** puranamr has quit IRC | 13:58 | |
| *** dims_ has quit IRC | 14:01 | |
| TheJulia | dtantsur: w/r/t that database error on friday, it is because utf8 and the engine default on CI nodes is likely no innodb where as my laptop is. https://dev.mysql.com/doc/refman/5.0/en/innodb-foreign-key-constraints.html | 14:01 |
| *** dims has joined #openstack-ironic | 14:01 | |
| *** amotoki has joined #openstack-ironic | 14:02 | |
| krtaylor | morning ironic | 14:05 |
| *** marzif has joined #openstack-ironic | 14:09 | |
| jlvillal | lucasagomes, Good morning :) | 14:09 |
| jlvillal | krtaylor, Good morning | 14:09 |
| TheJulia | dtantsur: yup, set my engine to MyISAM and kaboom :( | 14:10 |
| *** [1]cdearborn has joined #openstack-ironic | 14:16 | |
| rloo | dtantsur: maybe this is a dumb question, but wrt 232517, if it is wrt futurist, did you think of moving the periodic tasks into a separate file/class? | 14:19 |
| NobodyCam | good morning Ironicers :) | 14:22 |
| NobodyCam | happy hump day! | 14:22 |
| rloo | morning NobodyCam, jlvillal, krtaylor | 14:22 |
| NobodyCam | morning jroll rloo dtantsur devananda jlvillal sambetts TheJulia lucasagomes krtaylor vdrok | 14:23 |
| NobodyCam | morning rloo :) | 14:23 |
| lucasagomes | NobodyCam, morning | 14:23 |
| NobodyCam | :) | 14:23 |
| *** cemason has joined #openstack-ironic | 14:26 | |
| *** cdearborn has quit IRC | 14:27 | |
| openstackgerrit | Stephanie Miller proposed openstack/bifrost: ironic-inspector install doc polish https://review.openstack.org/234810 | 14:29 |
| TheJulia | good morning NobodyCam | 14:29 |
| *** kozhukalov_ has quit IRC | 14:30 | |
| openstackgerrit | Julia Kreger proposed openstack/bifrost: Initial support for installation of ironic-inspector https://review.openstack.org/229930 | 14:33 |
| jroll | morning NobodyCam :) | 14:33 |
| NobodyCam | :) morning :) | 14:34 |
| openstackgerrit | Merged openstack/bifrost: Generate network_info json using a module https://review.openstack.org/216372 | 14:34 |
| openstackgerrit | Merged openstack/bifrost: Support setting vlan in config drive https://review.openstack.org/216431 | 14:34 |
| jlvillal | Good morning NobodyCam rloo | 14:39 |
| *** cemason has quit IRC | 14:45 | |
| *** kozhukalov_ has joined #openstack-ironic | 14:45 | |
| sambetts | o/ jlvillal | 14:46 |
| sambetts | o/ NobodyCam | 14:46 |
| openstackgerrit | Zhenguo Niu proposed openstack/ironic: Convert set() to list in ListType https://review.openstack.org/234697 | 14:46 |
| jlvillal | sambetts, \o | 14:47 |
| *** caiobo has quit IRC | 14:47 | |
| *** baoli has quit IRC | 14:48 | |
| *** baoli_ has joined #openstack-ironic | 14:48 | |
| *** cemason has joined #openstack-ironic | 14:50 | |
| *** thrash is now known as thrash|biab | 14:51 | |
| *** puranamr has joined #openstack-ironic | 14:59 | |
| NobodyCam | :) | 15:00 |
| rloo | jroll, devananda: are we abandoning this AgentDeploy fix? https://review.openstack.org/#/c/231215/ | 15:01 |
| dtantsur | sorry, I'm in the middle of meetings | 15:01 |
| *** cemason has quit IRC | 15:01 | |
| dtantsur | rloo, I hoped that manager.py will contain everything futurist-related from now on | 15:02 |
| *** jistr|mtg is now known as jistr | 15:02 | |
| dtantsur | TheJulia, oh... do we need to support MyISAM at all? | 15:02 |
| *** wshao has joined #openstack-ironic | 15:02 | |
| *** baoli_ has quit IRC | 15:02 | |
| rloo | dtantsur: it just seemed to me that the periodic tasks could be subclassed instead of the other stuff :-) Anyway, I commented in your patch. | 15:02 |
| *** cemason has joined #openstack-ironic | 15:03 | |
| *** puranamr has quit IRC | 15:03 | |
| *** garthb has joined #openstack-ironic | 15:04 | |
| TheJulia | dtantsur: don't know, I just find it weird that the bare bones trusty nodes bifrost uses for CI have a different db config. I've updated the mode setting to set the engine to innodb in my config, waiting on CI run | 15:05 |
| *** cemason has quit IRC | 15:07 | |
| *** VikasC_ has quit IRC | 15:07 | |
| jroll | rloo: yeah, we are | 15:08 |
| lucasagomes | rloo, jroll oh we don't want that fix anymore? | 15:09 |
| rloo | jroll: ok, will you abandon it then. I don't feel right doing that :) | 15:09 |
| jroll | rloo: I'll let devananda do it when he sends the email / release notes patch about it | 15:09 |
| jroll | lucasagomes: no, we decided either way it breaks something, and without the patch the breakage is easy to see and easy to fix | 15:10 |
| dtantsur | rloo, not sure... it was annoying enough to do this change :D | 15:10 |
| lucasagomes | jroll, right, yeah it's set it to the driver definition | 15:10 |
| rloo | jroll: but I don't want folks spending time on that patch. | 15:10 |
| lucasagomes | ok seems good | 15:10 |
| rloo | jroll: just in case someone was going to spend time on it ;) | 15:11 |
| jroll | rloo: I -2'd | 15:11 |
| rloo | dtantsur: yeah, i can imagine. Just that it seems easier/clearer to have a periodic tasks file/class, than have to deal with how to separate/name the other classes :) | 15:11 |
| rloo | jroll: thx, that'll work too | 15:11 |
| dtantsur | rloo, if we don't split API bits, it's still going to be huge, so dunno... let's wait for what jroll says | 15:12 |
| jroll | says about what | 15:12 |
| dtantsur | jroll, about conductor split. rloo suggests to split it in a different way, but I'm so lazy to redo it :) | 15:12 |
| rloo | dtantsur: i thought you'd only then need to edit the periodic tasks file and leave manager.py alone :) | 15:12 |
| rloo | dtantsur: we should do what is 'right', whatever that is :) | 15:13 |
| jroll | dtantsur: I'm not opinionated, y'all can work it out | 15:13 |
| rloo | jroll: and that wasn't an opinion of yours | 15:13 |
| dtantsur | ok, I'll give it one more thought tomorrows with fresh mind | 15:13 |
| dtantsur | (fresh mind in the morning, lol) | 15:14 |
| rloo | dtantsur: no worries. it isn't that urgent. | 15:14 |
| *** jcoufal has quit IRC | 15:19 | |
| *** david-lyle has quit IRC | 15:25 | |
| openstackgerrit | Lilia Sampaio proposed openstack/ironic: Add OneView driver documentation https://review.openstack.org/234396 | 15:26 |
| *** mgoddard has joined #openstack-ironic | 15:26 | |
| *** david-lyle has joined #openstack-ironic | 15:27 | |
| *** e0ne has joined #openstack-ironic | 15:27 | |
| *** mbound has quit IRC | 15:27 | |
| *** Nisha has joined #openstack-ironic | 15:29 | |
| *** cemason has joined #openstack-ironic | 15:31 | |
| *** getvasanth has quit IRC | 15:31 | |
| *** thiagop has joined #openstack-ironic | 15:33 | |
| thiagop | Good morning, Ironic | 15:33 |
| *** kkoski has quit IRC | 15:33 | |
| *** kozhukalov_ has quit IRC | 15:35 | |
| *** bizarrochristy has joined #openstack-ironic | 15:36 | |
| *** cemason has quit IRC | 15:39 | |
| *** cemason_ has joined #openstack-ironic | 15:39 | |
| *** marzif has quit IRC | 15:39 | |
| jlvillal | thiagop, Good morning | 15:41 |
| thiagop | hey jlvillal, have you taken a look at the docs? | 15:41 |
| *** BharatK has quit IRC | 15:42 | |
| thiagop | (or maybe "did you take a look"...) | 15:42 |
| jlvillal | thiagop, A quick glance. But I've been mostly head down on functional testing. Looking at multiple projects to see how they do it. And figuring out what I want to copy to use for us. | 15:43 |
| TheJulia | dtantsur: http://lists.openstack.org/pipermail/openstack-dev/2014-July/040773.html | 15:43 |
| *** bizarrochristy has quit IRC | 15:43 | |
| *** kkoski has joined #openstack-ironic | 15:43 | |
| thiagop | jlvillal: np. Just askin' since you talked about yesterday | 15:43 |
| jlvillal | thiagop, thanks | 15:44 |
| dtantsur | TheJulia, oh wut? so what do they expect people to do? | 15:44 |
| * devananda has the plague today :( | 15:44 | |
| *** mgoddard_ has joined #openstack-ironic | 15:45 | |
| jlvillal | devananda, Sorry to hear that. Get better! | 15:45 |
| jroll | devananda: :( | 15:45 |
| devananda | making tea then going back to bed to write the release notes update | 15:46 |
| *** david-ly_ has joined #openstack-ironic | 15:46 | |
| TheJulia | dtantsur: explicitly set their engine of choice it looks like. :( | 15:47 |
| *** david-lyle has quit IRC | 15:47 | |
| thiagop | dtantsur: Hi! Saw your comments on that patch to filter the version of oneviewclient. I'd like your (experienced) input about that | 15:47 |
| *** kozhukalov_ has joined #openstack-ironic | 15:47 | |
| dtantsur | TheJulia, how do I do it in an engine-independent way? Oo | 15:47 |
| *** mgoddard has quit IRC | 15:48 | |
| thiagop | dtantsur: we plan to release the 2.1.X version only when OneView comes to the 2.1 version. But also, we are constantly trying to improve the oneviewclient to be the central point of communication between the driver (and other tools) and OneView | 15:48 |
| dtantsur | ok, why don't you want mitaka to use 2.1.x? | 15:48 |
| thiagop | I don't think they have an expectation to change version soon, then no. | 15:49 |
| TheJulia | dtantsur: good question, I have a suspicion with my config, going to toss it up to the gate and see if it works since I'm struggling to reproduce locally :( | 15:50 |
| *** meghal has joined #openstack-ironic | 15:51 | |
| openstackgerrit | Julia Kreger proposed openstack/bifrost: Initial support for installation of ironic-inspector https://review.openstack.org/229930 | 15:51 |
| thiagop | dtantsur: I'm thinking about pinning the version, but with all this improvement work ongoing, we may release again soon. And I wouldn't like to piss the cores off every time a new release is tagged for python-oneviewclient... | 15:51 |
| dtantsur | thiagop, that's why you should NOT put an upper constraint, right? otherwise we'll need a new patch each time you release a minor version | 15:52 |
| thiagop | dtantsur: <2.1 is just a failsafe for a long term future, but if you think it's unappropriate, I can remove that, yes | 15:52 |
| *** ijw has joined #openstack-ironic | 15:53 | |
| thiagop | we'll keep releasing 2.0.X until HP comes with a new version of OneView, that's far from happening | 15:53 |
| dtantsur | thiagop, if you follow the semver practice, then 2.1 must not break users of 2.0, so "<2.1" looks suspicious to me | 15:53 |
| dtantsur | thiagop, if you want to put "<3" it would be understandable | 15:53 |
| *** david-ly_ is now known as david-lyle | 15:54 | |
| thiagop | dtantsur: I think I got it. I need to plan these releases a little better. | 15:55 |
| * thiagop is in a brave new world | 15:55 | |
| dtantsur | :) | 15:55 |
| dtantsur | thanks | 15:56 |
| meghal | jroll: I had a question on how frequently is the hash ring reset right now ? Is it done during the sync_local_state periodic task ? | 15:56 |
| jroll | rloo: devananda: dtantsur: whoever else cares... jfyi, in the interest of release management free time, we're going to release 4.2.0 as liberty final, and release 4.2.1 on tuesday with the japanese translations and bugfixes | 15:56 |
| jroll | meghal: iirc, it's in that periodic task and also when a conductor joins or leaves the cluster | 15:57 |
| dtantsur | ack | 15:57 |
| devananda | jroll: ++ | 15:58 |
| rloo | jroll: i'm not sure what 'release management free time' means. would prefer if 4.2.1 was liberty final but whatever | 15:58 |
| jroll | rloo: "we're a bit late with these patches and thierry/doug/the gate are really %^&*ing busy", it would be a heroic effort to get 4.2.1 in | 15:59 |
| jroll | rloo: the last patches won't land until very late tonight, and the release will be completing around the same time | 15:59 |
| *** mgoddard_ has quit IRC | 15:59 | |
| rloo | jroll: oh. yeah, thierry seems tired based on his emails. ok. sad that it takes 12 hours or more to land a patch. | 16:00 |
| jroll | rloo: yeah, release week always means slow gate | 16:00 |
| rloo | jroll: i thought thurs was deadline and i don't see why it can't be pushed to the following week but whatever. i don't really care, just wondering! | 16:01 |
| jroll | thursday 1000 utc or so appears to be the deadline | 16:01 |
| jroll | really 4.2.0 works great | 16:01 |
| jroll | unless your system is japanese | 16:02 |
| rloo | devananda should just go to bed and forget the release notes til later this week. | 16:02 |
| rloo | jroll: yup. | 16:02 |
| jroll | I think as long as we get 4.2.1 out before summit, I'm happy | 16:02 |
| jroll | I can handle release notes and such today. | 16:02 |
| *** jcoufal has joined #openstack-ironic | 16:02 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-python-agent: [WIP] Support Linux-IO in addition to tgtd https://review.openstack.org/234861 | 16:03 |
| dtantsur | jroll, if we have a "known issues" section, we can document problem with locales there | 16:04 |
| *** mgoddard has joined #openstack-ironic | 16:05 | |
| *** jistr has quit IRC | 16:06 | |
| dtantsur | ok, calling it a day. see you tomorrow | 16:06 |
| *** dtantsur is now known as dtantsur|afk | 16:06 | |
| openstackgerrit | Thiago Paiva Brito proposed openstack/ironic: Update python-oneviewclient version https://review.openstack.org/234428 | 16:07 |
| thiagop | good night dtantsur|afk | 16:07 |
| NobodyCam | night dtantsur|afk :) | 16:07 |
| *** thrash|biab is now known as thrash | 16:07 | |
| *** yog_ has quit IRC | 16:09 | |
| *** subscope has quit IRC | 16:09 | |
| *** baoli has joined #openstack-ironic | 16:10 | |
| *** cemason_ has quit IRC | 16:10 | |
| meghal | jroll: currently, I see reset being called over here only - https://github.com/openstack/ironic/blob/master/ironic/conductor/manager.py#L1408 is it called some place else as well ? reason I am asking is because the take_over method during sync_local_state calls deploy on the driver and build the pxe config again, which is something we do not want to do for active instances…because if someone reboots an active instance they | 16:11 |
| *** testing has joined #openstack-ironic | 16:11 | |
| meghal | shouldn't the logic in takeover be left to the deploy driver | 16:11 |
| meghal | and let it decide whether to call prepare | 16:11 |
| *** testing is now known as Guest81146 | 16:11 | |
| meghal | am talking about this logic here - https://github.com/openstack/ironic/blob/master/ironic/conductor/manager.py#L1365 | 16:11 |
| *** cemason has joined #openstack-ironic | 16:12 | |
| meghal | this could be something our deployment specific logic does, but we clean up the pxe configuration once an instance becomes active…so that next time node reboots, it boots from local disk which has the new image | 16:13 |
| jroll | meghal: your long message got cut off, but I don't see why the logic that exists today doesn't work... | 16:13 |
| jroll | oh | 16:13 |
| meghal | so we do not want to build the pxe config on take over | 16:13 |
| jroll | meghal: booting from local disk is an option, and shouldn't generate a pxe config for the instance image | 16:14 |
| meghal | jroll: yeah so different deploy drivers will have different requirements | 16:14 |
| meghal | so it should be upto takeover method of the deploy driver to control that logic | 16:14 |
| jroll | meghal: so, you have your own driver or are using one in tree | 16:14 |
| meghal | and conductor/manager should just call takeover | 16:14 |
| meghal | jroll: we are using pxe driver from juno branch | 16:15 |
| meghal | jroll: https://github.com/openstack/ironic/blob/stable/juno/ironic/drivers/modules/pxe.py#L416 | 16:15 |
| jroll | meghal: oh, I don't think juno supported booting from local disk | 16:15 |
| jroll | this is fully supported now (since kilo iirc) | 16:16 |
| jroll | and prepare/take_over doesn't break things booting from disk | 16:16 |
| *** penick has joined #openstack-ironic | 16:16 | |
| meghal | jroll: is that a config option ? let me take a look | 16:17 |
| jroll | meghal: it's a node thing | 16:17 |
| jroll | lemme dig it up | 16:17 |
| jroll | meghal: http://docs.openstack.org/developer/ironic/deploy/install-guide.html#local-boot-with-partition-images | 16:18 |
| *** romainh has left #openstack-ironic | 16:19 | |
| *** wshao has quit IRC | 16:19 | |
| *** lintan has quit IRC | 16:23 | |
| meghal | jroll: thanks for the link…so it is part of the node capabilities and the driver would read those…but idea is that on a first time deployment we need the pxe config but do not need it for subsequent take over or reboots (after node becomes active), looking over here - https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/pxe.py#L594 , it just cleans up pxe config, but then who will build it in first place | 16:23 |
| *** lintan has joined #openstack-ironic | 16:24 | |
| jroll | meghal: prepare_instance() will clean up pxe configs there so that the node will boot from disk, not pxe | 16:24 |
| jroll | meghal: this does the setup for the deploy ramdisk https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/pxe.py#L463 | 16:25 |
| *** kkoski has quit IRC | 16:26 | |
| *** kozhukalov_ has quit IRC | 16:26 | |
| meghal | jroll: ahh ok…and who calls that because I do not see prepare_ramdisk being called from conductor/maanger.py ? pardon my ignorance because I am still familiar with juno code | 16:26 |
| jroll | meghal: pxe_ipmitool now refers to https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/iscsi_deploy.py#L646 | 16:27 |
| jroll | meghal: that PXEBoot class you're looking at just handles boot things, ISCSIDeploy handles the actual deployment | 16:27 |
| jroll | we split the DeployInterface into Deploy and Boot this cycle | 16:27 |
| *** r-daneel has joined #openstack-ironic | 16:28 | |
| jroll | so now prepare() will do the right thing there | 16:28 |
| meghal | jroll: ahh ok…gives more clarity now | 16:28 |
| jroll | :) | 16:28 |
| meghal | jroll: right…so conductor/manager will call prepare during takeover but this guy here https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/iscsi_deploy.py#L715 will just call prepare instance which cleans up pxe config for local boot cases | 16:29 |
| meghal | jroll: cool…thanks | 16:29 |
| jroll | meghal: right :) | 16:29 |
| meghal | jroll: so for certain drivers take over could be a noop kind of a thing then sync local state would not help much (except for periodic hash ring reset) and I was wondering if there should be an option to disable the periodic task and free up those periodic workers…and hash ring reset could be done in periodic power sync or so | 16:34 |
| meghal | what do you think ? | 16:34 |
| jroll | meghal: something like https://github.com/openstack/ironic/blob/master/ironic/conductor/manager.py#L158-L159 ? | 16:36 |
| jroll | :D | 16:36 |
| *** Guest81146 has quit IRC | 16:37 | |
| *** testing_ has joined #openstack-ironic | 16:37 | |
| meghal | jroll: ahha…did not know about that…that is exactly what I was wondering | 16:37 |
| meghal | jroll: but will this affect the mapping calls in any way ? | 16:37 |
| meghal | because the hash ring might not be reset periodically and always return stale values | 16:37 |
| meghal | or that will not be the case ? | 16:38 |
| *** cemason has quit IRC | 16:38 | |
| meghal | talking about the logic here - https://github.com/openstack/ironic/blob/master/ironic/conductor/manager.py#L1443 | 16:38 |
| jroll | meghal: yeah, that's what I'm wondering now | 16:39 |
| *** jasondotstar_ has joined #openstack-ironic | 16:39 | |
| openstackgerrit | Merged openstack/bifrost: Add static DHCP for dynamic inventory https://review.openstack.org/217374 | 16:40 |
| *** bigjools_ has joined #openstack-ironic | 16:40 | |
| jroll | meghal: I think there are dragons there. hrm. | 16:40 |
| * jroll wishes the conductor just didn't have local state | 16:41 | |
| meghal | hmm | 16:42 |
| jroll | meghal: that is to say, I think the ring would get out of date | 16:43 |
| meghal | yeah…periodic reset of hash ring needs to be taken care of in the mapping call itself like calculating the cache age or something | 16:43 |
| meghal | and we can continue to have sync local state explicitly resetting it | 16:43 |
| jroll | obligatory https://github.com/openstack/ironic/blob/master/ironic/conductor/rpcapi.py#L99 | 16:43 |
| jroll | :) | 16:44 |
| meghal | jroll: hah :) | 16:44 |
| meghal | brb | 16:45 |
| * jroll also wants to benchmark the hash ring reset | 16:46 | |
| jroll | meghal: mind filing a bug for this? | 16:46 |
| *** garthb has quit IRC | 16:47 | |
| *** UForgotten_ has joined #openstack-ironic | 16:48 | |
| *** e0ne_ has joined #openstack-ironic | 16:49 | |
| *** e0ne has quit IRC | 16:50 | |
| *** jaypipes has quit IRC | 16:50 | |
| *** sbadia has quit IRC | 16:50 | |
| *** gabriel has quit IRC | 16:50 | |
| *** bigjools has quit IRC | 16:50 | |
| *** UForgotten has quit IRC | 16:50 | |
| *** jasondotstar has quit IRC | 16:50 | |
| *** pshige_ has quit IRC | 16:50 | |
| *** bapalm has quit IRC | 16:50 | |
| *** praneshp has quit IRC | 16:50 | |
| *** lifeless has quit IRC | 16:50 | |
| *** mtreinish has quit IRC | 16:50 | |
| *** degorenko is now known as _degorenko | 16:51 | |
| meghal | jroll: will do | 16:53 |
| jroll | thanks much | 16:53 |
| *** phschwartz is now known as phschwartz_aw | 16:54 | |
| *** baoli_ has joined #openstack-ironic | 16:54 | |
| *** e0ne_ is now known as e0ne | 16:55 | |
| *** cemason has joined #openstack-ironic | 16:56 | |
| *** bapalm has joined #openstack-ironic | 17:00 | |
| *** lifeless has joined #openstack-ironic | 17:00 | |
| *** mtreinish has joined #openstack-ironic | 17:00 | |
| *** kkoski has joined #openstack-ironic | 17:02 | |
| *** mrmoje has quit IRC | 17:03 | |
| *** mrmoje has joined #openstack-ironic | 17:04 | |
| *** Nisha_away has joined #openstack-ironic | 17:06 | |
| *** Nisha has quit IRC | 17:06 | |
| *** gabriel has joined #openstack-ironic | 17:06 | |
| *** bapalm has quit IRC | 17:07 | |
| *** lifeless has quit IRC | 17:07 | |
| *** mtreinish has quit IRC | 17:07 | |
| *** spandhe has joined #openstack-ironic | 17:07 | |
| *** jaypipes has joined #openstack-ironic | 17:07 | |
| *** kozhukalov_ has joined #openstack-ironic | 17:07 | |
| *** pshige_ has joined #openstack-ironic | 17:07 | |
| *** cemason has quit IRC | 17:08 | |
| *** derekh has quit IRC | 17:08 | |
| *** sbadia has joined #openstack-ironic | 17:10 | |
| *** bapalm has joined #openstack-ironic | 17:10 | |
| *** lifeless has joined #openstack-ironic | 17:10 | |
| *** mtreinish has joined #openstack-ironic | 17:10 | |
| *** ijw has quit IRC | 17:11 | |
| *** harshs has joined #openstack-ironic | 17:14 | |
| lucasagomes | folks I'm calling it a day | 17:14 |
| lucasagomes | have a great night everyone | 17:14 |
| *** lucasagomes is now known as lucas-dinner | 17:14 | |
| openstackgerrit | Sam Betts proposed openstack/ironic-python-agent: Add tinyipa to IPA imagebuild directory https://review.openstack.org/234902 | 17:14 |
| *** harshs has quit IRC | 17:15 | |
| jroll | night lucas-dinner \o | 17:15 |
| sambetts | o/ lucas-dinner | 17:15 |
| *** jcoufal has quit IRC | 17:15 | |
| *** baoli has quit IRC | 17:15 | |
| *** athomas has quit IRC | 17:16 | |
| *** achanda has joined #openstack-ironic | 17:17 | |
| *** garthb has joined #openstack-ironic | 17:18 | |
| *** dolphm has left #openstack-ironic | 17:18 | |
| *** cemason has joined #openstack-ironic | 17:23 | |
| *** phschwartz_aw is now known as phschwartz | 17:23 | |
| *** harshs has joined #openstack-ironic | 17:25 | |
| *** meghal has quit IRC | 17:27 | |
| *** baoli has joined #openstack-ironic | 17:28 | |
| jroll | devananda: rloo: do either of you have thoughts on which sessions should be fishbowls? | 17:29 |
| rloo | jroll: fishbowl == biggest rooms? No, but i can think about it now. | 17:30 |
| *** lazy_prince has quit IRC | 17:30 | |
| jroll | rloo: correct, I'm looking into it now as well | 17:30 |
| *** meghal has joined #openstack-ironic | 17:30 | |
| jroll | workrooms are ~30, fishbowls are ~100 iirc | 17:30 |
| jroll | nova is locked into a fishbowl, so we have three of those left | 17:30 |
| *** bmpbike has quit IRC | 17:31 | |
| jroll | thinking for fishbowl: third party CI, driver API, group management? | 17:31 |
| NobodyCam | night lucas-dinner | 17:31 |
| jroll | those seem like things operators / vendors will care most about | 17:31 |
| *** meghal has quit IRC | 17:31 | |
| rloo | jroll: yeah | 17:31 |
| jroll | if that makes sense to you, I'll just do it | 17:31 |
| NobodyCam | that seems good | 17:31 |
| NobodyCam | :) | 17:31 |
| jroll | and then it's just checking for conflicts with other teams or talks | 17:32 |
| rloo | jroll: driver composition may affect operators/vendors too | 17:32 |
| *** baoli has joined #openstack-ironic | 17:32 | |
| *** meghal has joined #openstack-ironic | 17:32 | |
| *** cemason has quit IRC | 17:32 | |
| jroll | rloo: oh, true | 17:32 |
| rloo | jroll: having said that, i need to finish reading/grokking that spec | 17:32 |
| *** trown is now known as trown|lunch | 17:32 | |
| testing_ | 1 | 17:34 |
| *** baoli_ has quit IRC | 17:35 | |
| rloo | jroll: we had discussed driver composition at the last summit. i don't recall a huge crowd then. i guess it depends what we do/how controversial/hard/whatever for operators to understand. | 17:35 |
| jroll | rloo: I can't remember if that was the focus of that session, though, or if it derailed to composition | 17:35 |
| *** meghal has quit IRC | 17:36 | |
| jroll | rloo: I do think vendor things will draw more people than ops things | 17:36 |
| rloo | jroll: even a 30-person room is large so i'm not too concerned. as long as we don't do a neutron thing. An army showed up then. | 17:37 |
| jroll | rloo: yeah, exactly | 17:37 |
| *** Marga_ has quit IRC | 17:38 | |
| *** Marga_ has joined #openstack-ironic | 17:38 | |
| rloo | jroll: the nova one is about the nova/ironic driver right? | 17:38 |
| rloo | jroll: #3? | 17:38 |
| jroll | rloo: yes, correct | 17:38 |
| jroll | nova folks will also be there for the most part | 17:38 |
| *** testing_ has quit IRC | 17:39 | |
| rloo | jroll: ok, annotated on the etherpad. | 17:39 |
| *** testing has joined #openstack-ironic | 17:39 | |
| *** jasondotstar_ is now known as jasondotstar | 17:39 | |
| *** testing is now known as Guest39846 | 17:39 | |
| *** meghal has joined #openstack-ironic | 17:39 | |
| jroll | thanks | 17:40 |
| *** Guest39846 has quit IRC | 17:40 | |
| TheJulia | dtantsur|afk: http://docs.sqlalchemy.org/en/rel_1_0/dialects/mysql.html?highlight=charset#create-table-arguments-including-storage-engines | 17:41 |
| jroll | rloo: thinking driver composition in a workroom, I'd like to mostly talk about the internals there | 17:42 |
| jroll | it will affect ops/vendors but we have to get through the technical bits first | 17:42 |
| jroll | so beyond that... I know naohirot has a talk, not sure who else does | 17:43 |
| jroll | and ops conflicts with our wednesday sessions | 17:43 |
| *** mkovacik has quit IRC | 17:43 | |
| rloo | jroll: that is fine to do driver composition in a workroom. | 17:43 |
| rloo | jroll: yeah, those ops conflicts mean we won't have many ops folks. | 17:44 |
| *** praneshp has joined #openstack-ironic | 17:44 | |
| *** Marga_ has quit IRC | 17:44 | |
| jroll | rloo: yeah, so I'm going to put the driver API thing on thursday, under the assumption that operators may have a tendency to write a downstream driver | 17:45 |
| jroll | as opposed to group management / third party CI, those will be vendor-heavy if anything | 17:45 |
| *** pelix has quit IRC | 17:45 | |
| jroll | ops may care about group management, but in the sense that they want to do it, not that they care about how it works under the hood | 17:45 |
| rloo | jroll: i would think vendors want to write downstream drivers, not operators. but i don't know. | 17:48 |
| rloo | jroll: well, vendors want upstream drivers too. | 17:48 |
| jroll | rloo: operators may have some to implement hacks :) | 17:49 |
| jroll | it's more that, operators will probably care about downstream drivers more than group management or third party CI | 17:49 |
| jroll | but maybe that's wrong | 17:49 |
| jroll | meghal: penick: ^ thoughts? | 17:49 |
| penick | *backscroll* | 17:50 |
| rloo | jroll: well, i am the wrong person to comment on that! | 17:50 |
| jroll | penick: summit sessions on "driver API for downstream drivers" vs "group management of nodes" vs "third party CI" - which would you rather be in? | 17:50 |
| jroll | context is conflicts with ops sessions | 17:50 |
| penick | One of the first two for sure. Probably the driver one. What does group management of nodes mean? | 17:51 |
| jroll | rloo: me too, given I don't care about group management at all with my ops hat on, that's what for loops are for | 17:51 |
| jroll | penick: ironic node-update --tag rack-f4-21 blablabla; or manage all things in a blade maybe. | 17:51 |
| *** Sukhdev has joined #openstack-ironic | 17:51 | |
| rloo | penick: here's the etherpad: https://etherpad.openstack.org/p/mitaka-ironic-design-summit-ideas | 17:52 |
| jroll | penick: the first is managing multiple nodes by tags, not adding tags | 17:52 |
| jroll | but I suspect that session is going to be about vendor-specific grouping things :( | 17:52 |
| *** Sukhdev has quit IRC | 17:52 | |
| penick | Well, admittedly the group thing does have some value to us. We’re trying to come up with a flow to handle hardware as it comes in to the DC, or as one of our tenants needs to manage a large number of hosts at once. The ability to say “let me reimage these (40-1500) hosts right now” is a use case we do hit. But if I have t pick between the driver API and the grouping thing, i’ll go to the driver thing | 17:54 |
| penick | I’ll send one of our other ops folks to the grouping thing | 17:55 |
| jroll | penick: yeah, the grouping thing is well-understood for "normal" hardware, I think. use tags. coming soon (TM) | 17:56 |
| jroll | maybe that isn't sufficient, idk | 17:57 |
| *** Sukhdev has joined #openstack-ironic | 17:57 | |
| jroll | alright, I'll do driver API on thursday | 17:58 |
| Madasi | I would use tags, makes it easier than having to pull an inventory list from somewhere to figure out which servers/node ids correspond to a given rack/upstream router/etc | 17:58 |
| jroll | now I just need to talk to wanyen and naohirot | 17:59 |
| jroll | Madasi: yeah, vendors want to use other things like larger chassis and CMDBs, I suspect | 17:59 |
| Madasi | yeah, pluggable source of truth makes sense | 17:59 |
| jroll | "larger chassis" meaning blades, or the groupings like our OCP stuff has | 18:00 |
| *** kkoski has quit IRC | 18:00 | |
| *** baoli has quit IRC | 18:00 | |
| *** Marga_ has joined #openstack-ironic | 18:00 | |
| *** sinval has joined #openstack-ironic | 18:00 | |
| Madasi | but they are wanting to do so through something other than freeform tags? | 18:00 |
| Madasi | like dedicated fields for that type of thing? | 18:01 |
| Madasi | that would get messy | 18:01 |
| openstackgerrit | Julia Kreger proposed openstack/bifrost: Initial support for installation of ironic-inspector https://review.openstack.org/229930 | 18:01 |
| jroll | Madasi: line 66 https://etherpad.openstack.org/p/mitaka-ironic-design-summit-ideas | 18:01 |
| jroll | maybe not special fields, dunno | 18:01 |
| jroll | rloo: so now, driver composition before or after driver API :) | 18:02 |
| jroll | re-worded, early morning or late afternoon | 18:03 |
| openstackgerrit | Julia Kreger proposed openstack/bifrost: Initial addition of an inspection module https://review.openstack.org/232745 | 18:03 |
| rloo | jroll: hmm. both bad times. | 18:03 |
| jroll | hehe | 18:03 |
| jroll | I vote 9:50 I think | 18:03 |
| openstackgerrit | Julia Kreger proposed openstack/bifrost: WIP: Additional inspection role and adding role to NV job script https://review.openstack.org/232746 | 18:04 |
| rloo | jroll: i think driver composition before. i don't think we'll be able to do detailed stuff wrt driver api so that can wait til the aft when i'm tired. | 18:04 |
| jroll | rloo: ok, that's what I did. my proposed schedule is on the pad if you want to take a look | 18:05 |
| rloo | i don't think we can have group management if it conflicts with wanyen's talk | 18:05 |
| rloo | jroll: i thought wanyen was going to propose a spec for group management | 18:06 |
| jroll | oh yeah, meant to switch those two | 18:06 |
| jroll | yep | 18:06 |
| jroll | just switched them | 18:06 |
| rloo | jroll: what is naohirot talking about? | 18:06 |
| jroll | rloo: making ironic truly open and reliable | 18:07 |
| jroll | https://openstacksummitoctober2015tokyo.sched.org/event/92e6d1c440398cbe3898e009023a416c | 18:07 |
| Madasi | that spec is so vague, and different from the talk in the etherpad | 18:07 |
| jroll | Madasi: the group management thing? yeah, that one is a *third* proposal on how to do it | 18:07 |
| jroll | Madasi: HP is going to put up a spec for hw-specific group management stuff | 18:08 |
| Madasi | the part about hardware support in place of tags makes some sense, as source of truth, but has some pitfalls as well if you plan on having more than one hardware type in your fleet | 18:08 |
| jroll | yeah, totally | 18:09 |
| Madasi | i'll wait until you get something more solid after the summit, before I start meddling :) | 18:10 |
| jroll | heh | 18:10 |
| Madasi | but I'd totally go to that session | 18:10 |
| jroll | I suspect we will do tags + vendor stuff | 18:10 |
| rloo | jroll: the schedule seems fine. i was wondering what naohirot might want to attend, but that's the only non-ops time for the driver API. | 18:12 |
| rloo | jroll: and anyway, naohirot's driver is upstream. | 18:12 |
| jroll | rloo: yeah, not much we can do there | 18:12 |
| *** ijw has joined #openstack-ironic | 18:14 | |
| *** e0ne has quit IRC | 18:15 | |
| *** ukalifon1 has quit IRC | 18:17 | |
| *** ijw_ has joined #openstack-ironic | 18:18 | |
| *** ifarkas has quit IRC | 18:18 | |
| *** ijw has quit IRC | 18:19 | |
| *** ukalifon has joined #openstack-ironic | 18:19 | |
| *** e0ne has joined #openstack-ironic | 18:20 | |
| *** achanda has quit IRC | 18:25 | |
| sambetts | Night all o/ | 18:27 |
| *** ukalifon has quit IRC | 18:27 | |
| NobodyCam | night sambetts | 18:28 |
| *** trown|lunch is now known as trown | 18:28 | |
| *** linggao has joined #openstack-ironic | 18:29 | |
| *** mehip has joined #openstack-ironic | 18:34 | |
| linggao | Hi ironic, what are the best docs to use to install vanilla openstack and ironic (not devstack)? Start with http://docs.openstack.org/kilo/install-guide/install/apt/content/ and then do http://docs.openstack.org/developer/ironic/deploy/install-guide.html ? | 18:35 |
| *** david-lyle has quit IRC | 18:35 | |
| jroll | linggao: I'm not 100% sure, but yes that sounds like a good start | 18:36 |
| *** mehip has quit IRC | 18:38 | |
| linggao | ok, it seems quite long :-). | 18:38 |
| jroll | it's a complex system :) | 18:38 |
| linggao | yes. true. | 18:39 |
| *** david-lyle has joined #openstack-ironic | 18:39 | |
| *** mehip has joined #openstack-ironic | 18:39 | |
| *** harshs has quit IRC | 18:40 | |
| *** mehip has quit IRC | 18:40 | |
| *** david-ly_ has joined #openstack-ironic | 18:40 | |
| openstackgerrit | Merged openstack/ironic: Fixes logging of failure in deletion of swift temporary object https://review.openstack.org/232440 | 18:41 |
| *** mehip has joined #openstack-ironic | 18:41 | |
| NobodyCam | linggao: would love to review a doc patch on whats required to get things set up (/shameless plug) | 18:41 |
| linggao | NobodyCam, really wish installation of openstack is easy. :-) | 18:43 |
| *** david-ly_ has quit IRC | 18:43 | |
| NobodyCam | :) | 18:43 |
| *** kkoski has joined #openstack-ironic | 18:44 | |
| *** david-ly_ has joined #openstack-ironic | 18:44 | |
| *** david-lyle has quit IRC | 18:44 | |
| *** lucas-dinner has quit IRC | 18:44 | |
| *** david-ly_ is now known as david-lyle | 18:45 | |
| *** david-lyle has quit IRC | 18:45 | |
| *** david-lyle has joined #openstack-ironic | 18:47 | |
| *** trown is now known as trown|mtg | 18:48 | |
| openstackgerrit | Allan Vidal proposed stackforge/pyghmi: Fix TS inventory parsing bugs (DIMM, PSU). https://review.openstack.org/234930 | 18:50 |
| openstackgerrit | Allan Vidal proposed stackforge/pyghmi: Add support for event ack timeout setting. https://review.openstack.org/234931 | 18:50 |
| praneshp | linggao: you can try https://github.com/stackforge/anvil if you are using rpm based installs, there is an irc room at #openstack-anvil | 18:54 |
| linggao | praneshp, I will take a look. thanks! | 18:56 |
| *** harshs has joined #openstack-ironic | 18:56 | |
| *** Nisha_away has quit IRC | 18:56 | |
| jroll | wheeeeeeeeeeeeeeee. https://etherpad.openstack.org/p/summit-mitaka-ironic | 18:57 |
| jroll | all things summit. | 18:57 |
| *** Nisha_away has joined #openstack-ironic | 18:57 | |
| jroll | sched.org also has descriptions and etherpads for everything | 18:57 |
| *** mehip has quit IRC | 18:57 | |
| *** penick has quit IRC | 18:58 | |
| *** david-lyle has quit IRC | 19:00 | |
| NobodyCam | jroll: awesome!!!!! | 19:03 |
| jroll | :) | 19:04 |
| * jroll afks a while | 19:06 | |
| *** david-lyle has joined #openstack-ironic | 19:08 | |
| *** achanda has joined #openstack-ironic | 19:09 | |
| *** hongbin has joined #openstack-ironic | 19:11 | |
| *** vilobhmm11 has joined #openstack-ironic | 19:12 | |
| *** penick has joined #openstack-ironic | 19:25 | |
| *** david-lyle has quit IRC | 19:25 | |
| *** david-ly_ has joined #openstack-ironic | 19:25 | |
| *** mehip has joined #openstack-ironic | 19:29 | |
| *** Marga_ has quit IRC | 19:33 | |
| *** david-ly_ has quit IRC | 19:38 | |
| *** david-lyle has joined #openstack-ironic | 19:38 | |
| *** david-lyle has quit IRC | 19:39 | |
| *** david-lyle has joined #openstack-ironic | 19:41 | |
| *** trown|mtg is now known as trown | 19:43 | |
| *** priteau has quit IRC | 19:43 | |
| *** e0ne has quit IRC | 19:49 | |
| *** dprince has quit IRC | 19:51 | |
| openstackgerrit | Merged openstack/ironic: Updated from global requirements https://review.openstack.org/234664 | 19:54 |
| openstackgerrit | Merged openstack/ironic: Fix an annoying detail in the developer quick-start https://review.openstack.org/234783 | 19:54 |
| openstackgerrit | Merged openstack/bifrost: Fix boot.ipxe config behavior to ensure proper use https://review.openstack.org/233039 | 19:54 |
| *** david-lyle has quit IRC | 19:56 | |
| openstackgerrit | Allan Vidal proposed stackforge/pyghmi: Fix TS inventory parsing bugs (DIMM, PSU) https://review.openstack.org/234953 | 19:56 |
| jlvillal | TheJulia: This made me think of Bifrost: https://openstacksummitoctober2015tokyo.sched.org/event/8cdc4e5a816f22bf2978e8869173de27 | 19:58 |
| *** david-lyle has joined #openstack-ironic | 19:58 | |
| *** achanda has quit IRC | 20:00 | |
| *** meghal has quit IRC | 20:00 | |
| *** david-ly_ has joined #openstack-ironic | 20:01 | |
| *** david-lyle has quit IRC | 20:03 | |
| *** david-ly_ is now known as david-lyle__ | 20:03 | |
| *** david-lyle__ is now known as david-lyle___ | 20:03 | |
| *** meghal has joined #openstack-ironic | 20:03 | |
| *** Marga_ has joined #openstack-ironic | 20:03 | |
| *** hongbin has left #openstack-ironic | 20:04 | |
| *** Nisha_away has quit IRC | 20:05 | |
| *** david-lyle has joined #openstack-ironic | 20:06 | |
| *** david-lyle___ has quit IRC | 20:06 | |
| *** Sukhdev has quit IRC | 20:07 | |
| *** Marga_ has quit IRC | 20:09 | |
| *** meghal has left #openstack-ironic | 20:10 | |
| openstackgerrit | Thiago Paiva Brito proposed openstack/ironic: [WIP] Power off the machine before set_boot_devic https://review.openstack.org/232186 | 20:13 |
| openstackgerrit | Allan Vidal proposed stackforge/pyghmi: Change acknowledgement_timeout to acknowledge_timeout https://review.openstack.org/234962 | 20:13 |
| mrda | Morning Ironic | 20:15 |
| thiagop | morning mrda | 20:16 |
| *** ionutbalutoiu has joined #openstack-ironic | 20:16 | |
| *** sambetts has quit IRC | 20:17 | |
| mrda | o/ | 20:17 |
| *** sambetts has joined #openstack-ironic | 20:18 | |
| NobodyCam | mornign mrda | 20:22 |
| thiagop | Guys, quick question: | 20:23 |
| thiagop | I'm validating a node that has "deploy_kernel" and "deploy_ramdisk" on driver_info | 20:24 |
| thiagop | the validation says that the node should have "kernel", "ramdisk" and "image_source" | 20:25 |
| thiagop | by the installation guide, it shouldn't be happening when I have deploy_kernel/deploy_ramdisk. Right? | 20:26 |
| mrda | thiagop: so you can get incorrect error messages if an exception gets thrown in validation | 20:29 |
| mrda | if you expect that deploy_kernel and deploy_ramdisk is all you need, verify that the paths for these are correct with right permissions | 20:30 |
| *** baoli has joined #openstack-ironic | 20:32 | |
| *** Sukhdev has joined #openstack-ironic | 20:33 | |
| thiagop | mrda: looking at pxe code now, I think it shouldn't be calling both _parse_driver_info and _parse_instance_info this way | 20:33 |
| *** baoli_ has joined #openstack-ironic | 20:33 | |
| thiagop | it'll never validate the node if it's not on a deploy process... | 20:34 |
| *** ijw_ has quit IRC | 20:36 | |
| *** baoli has quit IRC | 20:37 | |
| *** achanda has joined #openstack-ironic | 20:41 | |
| *** achanda has quit IRC | 20:41 | |
| *** achanda has joined #openstack-ironic | 20:44 | |
| *** afaranha has quit IRC | 20:45 | |
| openstackgerrit | Chris Krelle proposed openstack/bifrost: Follow up to 233039 https://review.openstack.org/234977 | 20:46 |
| spandhe | Hi folks, have a question about nova/virt/ironic/driver | 20:51 |
| *** Marga_ has joined #openstack-ironic | 20:51 | |
| spandhe | if deploy fails due to timeout, we call detroy: https://github.com/openstack/nova/blob/master/nova/virt/ironic/driver.py#L836 | 20:51 |
| spandhe | any reason why need to do that in the driver? nova/compute/manager can call _shutdown_instance on seeing RescheduleException, which can trigger destroy. | 20:52 |
| spandhe | https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L2186-L2190 | 20:53 |
| spandhe | rloo: ^ any idea about this? | 20:54 |
| *** Marga_ has quit IRC | 20:55 | |
| openstackgerrit | Monty Taylor proposed openstack/python-ironicclient: Remove lxml requirement https://review.openstack.org/234979 | 20:56 |
| *** achanda has quit IRC | 20:56 | |
| *** harshs has quit IRC | 20:57 | |
| *** mehip has quit IRC | 20:57 | |
| openstackgerrit | Sean Dague proposed openstack/ironic: remove lxml requirement https://review.openstack.org/234981 | 20:57 |
| rloo | spandhe: I don't know, although it seems to make sense to me that if the driver if checking for timeout, that it destroys. | 20:57 |
| rloo | spandhe: why do you think the nova/compute/manager should handle it? | 20:58 |
| *** kozhukalov_ has quit IRC | 20:58 | |
| spandhe | rloo: https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L2186-L2190 | 20:59 |
| *** mehip has joined #openstack-ironic | 20:59 | |
| rloo | spandhe: I don't know enough about nova and the interface between nova and the driver, to know which/where that should be done. do other drivers let nova/compute/manager handle it? | 21:01 |
| spandhe | rloo: checking that now. but basically destroy will be called multiple times due to save_and_reraise | 21:02 |
| rloo | spandhe: which could be the case of 'if you ain't sure it will be called up the stack, call it now, cuz at least it'll be called once instead of being missed'? | 21:02 |
| jroll | can I get a core to land this silly patch please? :) https://review.openstack.org/#/c/234979/ | 21:03 |
| spandhe | rloo: thats what I am trying to understand.. is there a case where nova compute wont call destroy? | 21:03 |
| rloo | spandhe: i dunno. maybe ask the nova folks? | 21:04 |
| jroll | spandhe: that's a good question, hm | 21:05 |
| *** mrmoje has quit IRC | 21:07 | |
| spandhe | rloo: jroll I will ask nova folks.. was just trying to see if anyone already knows why destroy was being called over there | 21:07 |
| *** harshs has joined #openstack-ironic | 21:08 | |
| jroll | spandhe: I know it's called, I'm not sure as to why :) | 21:08 |
| *** trown is now known as trown|outttypeww | 21:08 | |
| rloo | spandhe: my quick look seems to indicate that they are two different 'reasons'. i mean, it seems to make sense to destroy in both those cases. | 21:09 |
| *** praneshp_ has joined #openstack-ironic | 21:09 | |
| spandhe | rloo: nova compute is going to raise RescheduleException for all ironic driver exceptions/ | 21:10 |
| spandhe | rloo: and RescheduleException is always going to trigger _shutdown_instance | 21:10 |
| *** ionutbalutoiu has quit IRC | 21:10 | |
| rloo | spandhe: i go back to the 'agreement' or whatever between nova and the drivers. | 21:10 |
| *** praneshp has quit IRC | 21:10 | |
| jroll | spandhe: my best guess is that it's there to make darn sure we clean up if something goes wrong, as nova does a whole bunch of stuff in _shutdown_instance() before calling destroy(). but IMBW | 21:11 |
| openstackgerrit | Merged stackforge/pyghmi: Change acknowledgement_timeout to acknowledge_timeout https://review.openstack.org/234962 | 21:12 |
| *** praneshp has joined #openstack-ironic | 21:12 | |
| *** achanda has joined #openstack-ironic | 21:12 | |
| *** praneshp_ has quit IRC | 21:13 | |
| spandhe | jroll: ok.. let me dig a bit more into this. Ill get back. My problem is, I have patched shutdown_instance to deallocate network first and then call destroy. But when the driver triggers destroy on its own, my patch gets skipped and the network stays.. | 21:15 |
| jroll | spandhe: we re-raise the exception, so shutdown_instance should still get called, right? | 21:16 |
| jroll | oh, it isn't a rescheduledException | 21:16 |
| jroll | bla | 21:16 |
| spandhe | jroll: we reraise in virt driver. nova compute catches that and raises rescheduleException | 21:17 |
| jroll | spandhe: then it should still call shutdown_instance, right? | 21:17 |
| *** hblixt has joined #openstack-ironic | 21:18 | |
| spandhe | jroll: so nova compute does call shutdown_instance. but destroy has already happened | 21:18 |
| spandhe | once | 21:18 |
| spandhe | and it will happen again with shutdown_instance | 21:18 |
| jroll | yeah, something is wrong here | 21:19 |
| jroll | this does totally work | 21:19 |
| jroll | but now I don't know why :/ | 21:19 |
| *** linggao has quit IRC | 21:20 | |
| *** mrmoje has joined #openstack-ironic | 21:23 | |
| spandhe | jroll: ill find out :) | 21:24 |
| jroll | spandhe: please do file a bug though, this needs to be cleaned up. file it in nova with an ironic tag | 21:25 |
| spandhe | jroll: ok.. will do.. thanks! | 21:25 |
| jroll | thank you for pointing this out :) | 21:25 |
| jroll | rloo: one more :P https://review.openstack.org/#/c/234981/ | 21:26 |
| rloo | jroll: i was wondering when that one would appear :) | 21:27 |
| jroll | hah | 21:27 |
| jroll | I got distracted | 21:27 |
| *** penick has quit IRC | 21:31 | |
| *** penick has joined #openstack-ironic | 21:31 | |
| *** mehip has quit IRC | 21:31 | |
| thiagop | guys, I'm calling it a day | 21:31 |
| thiagop | see you tomorrow | 21:32 |
| *** penick has quit IRC | 21:32 | |
| *** thiagop has quit IRC | 21:32 | |
| *** penick has joined #openstack-ironic | 21:34 | |
| *** harshs has quit IRC | 21:34 | |
| *** Marga_ has joined #openstack-ironic | 21:36 | |
| mrda | yay for removing unneeded requirements! | 21:37 |
| NobodyCam | boo for gate failures | 21:41 |
| mrda | :( | 21:41 |
| *** praneshp has quit IRC | 21:41 | |
| * jroll rechecks | 21:43 | |
| *** harshs has joined #openstack-ironic | 21:43 | |
| * jroll needs to figure out what that intermittent bug is | 21:43 | |
| *** penick has quit IRC | 21:43 | |
| Madasi | spandhe: it looks like the place you indicated in the compute manager calls _shutdown_instance only in a very small window of build time, does anything call it in https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L2023-L2070 where other build exceptions are caught? | 21:44 |
| *** praneshp has joined #openstack-ironic | 21:45 | |
| openstackgerrit | Stephanie Miller proposed openstack/bifrost: Initial support for installation of ironic-inspector https://review.openstack.org/229930 | 21:46 |
| *** thrash is now known as thrash|g0ne | 21:48 | |
| *** achanda has quit IRC | 21:48 | |
| jroll | jlvillal: hey, is it cool if I mark you QA liasion here? https://wiki.openstack.org/wiki/CrossProjectLiaisons#QA | 21:49 |
| *** praneshp has quit IRC | 21:49 | |
| jlvillal | jroll: Sounds reasonable to me. Thanks. | 21:50 |
| jroll | thank YOU | 21:50 |
| *** vilobhmm11 has left #openstack-ironic | 21:51 | |
| *** bigjools_ is now known as bigjools | 21:52 | |
| *** bigjools has quit IRC | 21:52 | |
| *** bigjools has joined #openstack-ironic | 21:52 | |
| rloo | spandhe: is this useful: https://review.openstack.org/#/c/99519/ | 21:52 |
| spandhe | rloo: it does. Thanks! | 21:54 |
| rloo | spandhe: i sometimes wonder if we should add more comments in the code but anyway. | 21:54 |
| jroll | oh wow, I touched that | 21:55 |
| * jroll runs away | 21:55 | |
| jroll | (that bug anyway) | 21:56 |
| rloo | jroll: you can run, but you cannot hide... | 21:56 |
| jroll | yes I can | 21:57 |
| *** jroll has left #openstack-ironic | 21:57 | |
| *** jroll has joined #openstack-ironic | 21:57 | |
| jroll | see. | 21:57 |
| rloo | jroll: yup, i see you :) | 21:59 |
| *** baoli_ has quit IRC | 21:59 | |
| *** baoli has joined #openstack-ironic | 22:00 | |
| *** baoli has quit IRC | 22:00 | |
| *** baoli has joined #openstack-ironic | 22:01 | |
| *** [1]cdearborn has quit IRC | 22:02 | |
| openstackgerrit | Merged openstack/ironic: ucs_hostname changed to ucs_address https://review.openstack.org/232834 | 22:03 |
| openstackgerrit | Merged openstack/ironic: Remove functions: _cleanse_dict & format_message https://review.openstack.org/233718 | 22:03 |
| spandhe | Madasi: are any of those exception raised by ironic driver? | 22:03 |
| *** meghal1 has joined #openstack-ironic | 22:07 | |
| *** jamielennox|away is now known as jamielennox | 22:07 | |
| spandhe | rloo: jroll , in that review, I think shutdown_instance was not happening because we were not catching exception. maybe nova code was fixed after that to catch all exceptions, call _shutdown_instance and reraise | 22:09 |
| rloo | spandhe: would need to go through history/whatever to see what happened. | 22:10 |
| rloo | spandhe: oh, i have to take off now. later. | 22:10 |
| spandhe | rloo: ok.. talk to you tomorrow! | 22:10 |
| jroll | yeah, I have no clue myself :) | 22:10 |
| openstackgerrit | Merged openstack/python-ironicclient: Remove lxml requirement https://review.openstack.org/234979 | 22:10 |
| NobodyCam | have a good night rloo :) | 22:10 |
| spandhe | jroll: ill open a bug.. I see that till Icehouse, nova-compute was not calling shutdown_instance https://github.com/openstack/nova/blob/2014.1.5/nova/compute/manager.py#L2116-L2118 | 22:14 |
| spandhe | jroll: seems like this needs some discussion | 22:14 |
| jroll | spandhe: agree, thanks | 22:14 |
| *** achanda has joined #openstack-ironic | 22:15 | |
| * jlvillal wishes he could have days without meetings.... | 22:24 | |
| * jlvillal thinks five hours is too much for one day. | 22:25 | |
| Madasi | spandhe: a few. | 22:25 |
| Madasi | https://github.com/openstack/nova/blob/master/nova/virt/ironic/driver.py#L823-L827 calls _wait_for_active which can raise InstanceDeployFailure https://github.com/openstack/nova/blob/master/nova/virt/ironic/driver.py#L436 | 22:26 |
| spandhe | Madasi: ok.. ironic virt driver calls destroy beforehand only in one case - callback timeout. | 22:26 |
| Madasi | which would be caught and trigger driver's destroy, then reraised | 22:27 |
| spandhe | Madasi: right. | 22:27 |
| Madasi | hmm, that isn't the one I thought it was :( | 22:27 |
| Madasi | wait, this was it | 22:28 |
| spandhe | Madasi: yes :) The exception tha you pointed out, gets handled in nova compute: https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L2173-L2191 | 22:28 |
| jroll | spandhe: are you sure? it looks like ironic driver calls it on any exception | 22:28 |
| Madasi | _wait_for_Active can also throw InstanceNotFound - https://github.com/openstack/nova/blob/master/nova/virt/ironic/driver.py#L449-L454 | 22:28 |
| Madasi | that one is caught and just creates a create.end notification - https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L2023 | 22:29 |
| *** harshs has quit IRC | 22:29 | |
| spandhe | jroll: any exception that happens during deployment: https://github.com/openstack/nova/blob/master/nova/virt/ironic/driver.py#L826-L836 but thats because the deployment has already started by then. | 22:29 |
| *** Sukhdev has quit IRC | 22:30 | |
| Madasi | yes, that's the only time driver calls destroy | 22:30 |
| jroll | spandhe: any of these https://github.com/openstack/nova/blob/master/nova/virt/ironic/driver.py#L431-L462 | 22:30 |
| spandhe | Madasi: ok.. lets take the scenario youmentioned. | 22:30 |
| spandhe | InstanceNotFound | 22:30 |
| jroll | InstanceDeployFailure, InstanceNotFound, I assume connection failures may be raised, etc | 22:30 |
| Madasi | spandhe: it gets caught there, in the resources yield, but what about later like https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L2016 | 22:31 |
| spandhe | jroll: right.. all of these will be caught in yield | 22:32 |
| spandhe | https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L2173-L2191 | 22:32 |
| *** harshs has joined #openstack-ironic | 22:32 | |
| spandhe | Madasi: when it gets caught in resouces yield, shutdown_instance gets called | 22:32 |
| spandhe | Madasi: right? So shutdown happened. which means destroy also happened.. | 22:33 |
| Madasi | spandhe: don't think so | 22:33 |
| Madasi | that yield exception catching is part of _build_resources | 22:33 |
| spandhe | Madasi: jroll my point is, because of https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L2173 , all the exceptions now get caught and shutdown instance always gets called | 22:33 |
| jroll | ah, you may be right | 22:34 |
| jroll | ok I'd agree it is a bug | 22:34 |
| Madasi | https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L2001 is the call to spawn, and isn't using resources, only what was already retrieved from it | 22:34 |
| Madasi | and the wait_for_active is part of spawn | 22:34 |
| jroll | build_and_run_instance calls spawn() within the _build_resources context manager though | 22:35 |
| Madasi | hmm, true | 22:36 |
| jroll | so yeah, I agree this is a bug | 22:36 |
| * jroll got confused with all the indirection | 22:36 | |
| Madasi | so, why the long except block at https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L2023-L2070 | 22:36 |
| Madasi | seems weird then | 22:36 |
| spandhe | jroll: It is very confusing.. | 22:36 |
| jroll | :) | 22:37 |
| spandhe | Madasi: I am guessing for other drivers? | 22:37 |
| spandhe | Madasi: exceptions like InstanceNotFound will still be reraised, right? Even thought shutdown n all happened? | 22:38 |
| spandhe | And looks like thats the only places that closes the create loop with notification: end/error etc | 22:39 |
| Madasi | will it? | 22:41 |
| Madasi | i mean it will, but changed to a BuildAbortException, right? https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L2191-L2195 | 22:42 |
| spandhe | Madasi: I think so.. because yield resouces will reraise the exception after shutdown_instance.. right? | 22:42 |
| spandhe | Madasi: that is if shutdown_instance also fails | 22:42 |
| Madasi | wait, yeah | 22:42 |
| Madasi | so yes | 22:42 |
| spandhe | Madasi: thats another problem actually.. so if shutdown_instance also fails, original exception gets masked | 22:43 |
| Madasi | you are correctm b/c initial handling is with save_and_reraise | 22:43 |
| Madasi | spandhe: not just masked, also aborts build attempt | 22:43 |
| Madasi | hence the not reschedulign message | 22:43 |
| Madasi | which is why it is changing the exception type i'm sure | 22:44 |
| spandhe | Madasi: but the problem is, we dont find out why original deployment failed in the first place.. that can be annoying if you are booting hundreds of node concurrently | 22:45 |
| spandhe | when looking at logs is not an option | 22:45 |
| *** achanda has quit IRC | 22:45 | |
| * jroll wonders out loud to Madasi about what makes this work properly in our production (getting the actual exception and such) | 22:46 | |
| *** openstackgerrit has quit IRC | 22:46 | |
| *** openstackgerrit has joined #openstack-ironic | 22:47 | |
| *** achanda has joined #openstack-ironic | 22:47 | |
| *** achanda has quit IRC | 22:47 | |
| Madasi | jroll: it probably doesn't | 22:47 |
| jroll | Madasi: we do get the exception ironic through back | 22:47 |
| Madasi | that's probably the second error we see when someone deletes while building | 22:47 |
| jroll | threw* | 22:47 |
| jroll | omg. it all makes sense. | 22:47 |
| jroll | so yeah, the second delete would trigger "can't delete node while in cleaning" and such. | 22:48 |
| Madasi | this is why node cleanup failures are fatal to the build probably, and 99% of the ones we see are delete while building | 22:50 |
| jroll | yeah | 22:51 |
| Madasi | guess I know the next section of ironic/nova code I should try to make sense of | 22:52 |
| Madasi | spandhe: nice find | 22:54 |
| spandhe | Madasi: thanks :) Will create a bug for destroy issue.. Should I also create a nova bug for exception issue? | 22:56 |
| *** baoli has quit IRC | 22:57 | |
| jroll | create a bug for everything you think should be fixed :) | 22:58 |
| Madasi | yeah, I imagine the exception one might get some debate, since the exception changing is probably deliberate | 22:59 |
| Madasi | but doesn't mean it shouldn't be filed | 22:59 |
| *** jobewan has joined #openstack-ironic | 23:00 | |
| *** Sukhdev has joined #openstack-ironic | 23:01 | |
| spandhe | jroll: Madasi ok :) Created https://bugs.launchpad.net/nova-hyper/+bug/1506234 | 23:04 |
| openstack | Launchpad bug 1506234 in nova-hyper "Ironic virt driver in Nova calls destroy unnecessarily if spawn fails" [Undecided,New] | 23:04 |
| openstackgerrit | Stephanie Miller proposed openstack/bifrost: Initial support for installation of ironic-inspector https://review.openstack.org/229930 | 23:05 |
| *** achanda has joined #openstack-ironic | 23:09 | |
| jroll | spandhe: nova-hyper? | 23:12 |
| spandhe | jroll: changed that to Nova :) | 23:12 |
| jroll | heh, cool | 23:12 |
| jroll | thanks! | 23:12 |
| spandhe | jroll: does Ironic team look after the driver or nova team? Who will be reviewing the patchset for this? | 23:14 |
| jroll | spandhe: some of both; only nova team can approve it but we'll also review it | 23:14 |
| jroll | we have ironic folks that watch over nova bugs etc, it will get on our radar | 23:15 |
| spandhe | jroll: ok.. good to know.. thanks! | 23:15 |
| jroll | welcome :) | 23:15 |
| *** hblixt has quit IRC | 23:15 | |
| *** dims_ has joined #openstack-ironic | 23:19 | |
| *** dims has quit IRC | 23:22 | |
| *** [1]cdearborn has joined #openstack-ironic | 23:34 | |
| *** priteau has joined #openstack-ironic | 23:44 | |
| *** harshs has quit IRC | 23:47 | |
| *** harshs has joined #openstack-ironic | 23:48 | |
| *** priteau has quit IRC | 23:48 | |
| *** harshs has quit IRC | 23:48 | |
| *** penick has joined #openstack-ironic | 23:56 | |
| *** smoriya has joined #openstack-ironic | 23:57 | |
| *** mgoddard_ has joined #openstack-ironic | 23:57 | |
| *** mgoddard has quit IRC | 23:57 | |
| *** jobewan has quit IRC | 23:58 | |
| *** Haomeng|2 has joined #openstack-ironic | 23:58 | |
| *** praneshp has joined #openstack-ironic | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!