| *** david-lyle is now known as david-lyle_afk | 00:01 | |
| openstackgerrit | Alex Weeks proposed openstack/ironic-python-agent: Added extension to support rescue mode https://review.openstack.org/145385 | 00:02 |
|---|---|---|
| Shrews | NobodyCam: apparently i should get food off my brain since i just +A'd before the tests... *sigh* i guess gerrit will catch it properly should something break. | 00:04 |
| * Shrews goes for dinner | 00:04 | |
| NobodyCam | lol | 00:04 |
| *** mrda-away is now known as mrda | 00:04 | |
| NobodyCam | yea it still goes through all tests | 00:04 |
| mrda | . | 00:05 |
| NobodyCam | have a good dinner Shrews :) | 00:05 |
| NobodyCam | morning mrda | 00:05 |
| mrda | hey NobodyCam | 00:05 |
| JayF | mrda: I noticed today you hadn't re-reviewed https://review.openstack.org/#/c/141188/ yet, just wanted to make sure it was still on your radar | 00:05 |
| JayF | mrda: clif is at rookie-o so I'm shilling his patch :P | 00:06 |
| mrda | thanks JayF - but from holidays as of this minute :) | 00:06 |
| mrda | s/but/back/ | 00:06 |
| *** smoriya has joined #openstack-ironic | 00:06 | |
| JayF | I forget, you Aussies get what; 300 days of vacation a year :P | 00:07 |
| NobodyCam | lol | 00:07 |
| NobodyCam | no I think it only 298 | 00:07 |
| mrda | lol. I've only taken 5 days of something. I worked up to Christmas Eve | 00:07 |
| mrda | s/of/or | 00:07 |
| NobodyCam | uggh | 00:07 |
| NobodyCam | :) | 00:07 |
| mrda | but we are blessed with having a few holidays - of course we work hard to earn them :) | 00:08 |
| NobodyCam | yes!!!! :) hehhe | 00:09 |
| mrda | We also have some nasty bushfires here, which compensates for the holiadays (http://www.abc.net.au/news/2015-01-06/up-to-38-houses-destroyed-by-sa-bushfires/6002952) | 00:10 |
| rloo_afk | NobodyCam: wrt https://review.openstack.org/#/c/139217/, will you approve it? | 00:10 |
| devananda | mrda: are you doing LCA ? | 00:10 |
| mrda | devananda: of course :) I'll be in AKL from Sunday | 00:11 |
| mrda | but still working as I can | 00:11 |
| mrda | devananda: are you coming down? | 00:11 |
| rloo_afk | NobodyCam: wrt that last patch I added, it's fine it is isn't rebased now. I can do it later or not. It's just to show that the fsm code isn't quite correct yet. | 00:11 |
| devananda | mrda: not this time. too much travel booked alraedy | 00:11 |
| devananda | mrda: see you next year, though :) | 00:12 |
| *** ryanpetrello has quit IRC | 00:12 | |
| mrda | devananda: rats, would have enjoyed hanging out again! | 00:12 |
| devananda | mrda: as would I. but I'll also enjoy the time at home ;) | 00:12 |
| mrda | perhaps we could do an ironic midcycle the week before/after LCA in AU, so you all can see Australia and experience LCA :) | 00:12 |
| mrda | devananda: I totally understand. | 00:13 |
| NobodyCam | rloo_afk: with 3 +2's and 2 +1 I will approve 139217 (as soon a Mr J is done :-p) | 00:13 |
| *** ijw has joined #openstack-ironic | 00:13 | |
| rloo_afk | thx NobodyCam | 00:14 |
| NobodyCam | no no thank you :) | 00:14 |
| devananda | rloo_afk: for 142303, do you know offhand why it failed tempest? looks like REBUILD ? | 00:14 |
| rloo_afk | devananda: oh, I didn't even look at tempest. It was failing some unit tests, so I stopped there. | 00:15 |
| rloo_afk | devananda: I'll take a look tomorrow maybe. | 00:15 |
| devananda | rloo_afk: ok. well, you added / changed unit tests, so I'm not surprised that it fails | 00:16 |
| rloo_afk | devananda: probably a good time to rebase then. it is possible it was failing cuz a bunch of things may have failed at that time due to ... pick-your-poison... (ie don't remember) | 00:16 |
| devananda | rloo_afk: the question on my mind is -- are these failures problems in the state machine itself, or areas where the current (pre-fsm.py) code behaves in non-obvious ways? | 00:16 |
| rloo_afk | devananda: the state machine, or how we represent it with the target_state stuff. | 00:17 |
| * mrda is sweltering, it's 110F here today. | 00:17 | |
| NobodyCam | mrda: thats awesome... | 00:18 |
| mrda | NobodyCam, really? | 00:18 |
| devananda | rloo_afk: my goal was to represent what we *actually* do today with states, even where that is illogical | 00:18 |
| rloo_afk | devananda: wrt the fsm's _target_state and the nodes_target_provision_state; they aren't always in sync. | 00:19 |
| rloo_afk | devananda: yeah, so that's one of the problems. with the fsm stuff, the node's target-provision-state isn't what it was w/o the fsm stuff. | 00:19 |
| NobodyCam | mrda: not really .. It gets that hot down here.. but I am long gone before that time | 00:19 |
| devananda | rloo_afk: :( | 00:19 |
| mrda | :) | 00:20 |
| rloo_afk | devananda: I don't think it is a difficult problem. We just need to tweak it (our fsm) a bit or something. | 00:20 |
| NobodyCam | 139217 come on down your the next approved patch | 00:20 |
| devananda | rloo_afk: k. let me see what I can do in an hour | 00:20 |
| * NobodyCam tries to sound likethe "proce is right announser" | 00:21 | |
| NobodyCam | ok with that entire line type-oh-ed /me steps away from the keyboard | 00:22 |
| rloo_afk | devananda: ok. good luck ;) I'm going to take off soon, will see how far you get tomorrow. | 00:22 |
| openstackgerrit | Josh Gachnang proposed openstack/ironic-python-agent: Add command metrics to IPA API https://review.openstack.org/119981 | 00:22 |
| *** ryanpetrello has joined #openstack-ironic | 00:22 | |
| *** andreykurilin has quit IRC | 00:25 | |
| openstackgerrit | Ruby Loo proposed openstack/ironic: Issue with FSM's _target_state https://review.openstack.org/142303 | 00:26 |
| mrda | devananda: does it look like there will be a meetup in SFO the week after midcycle in Grenoble? | 00:27 |
| devananda | mrda: it's looking that way. there's a lot of interest ... and jroll / JayF have offered to host | 00:27 |
| mrda | and that would be a valid excuse to go visit these guys \o/ | 00:28 |
| *** rloo_afk has quit IRC | 00:29 | |
| jroll | could be fun, we're in a good spot :) | 00:34 |
| *** Masahiro has joined #openstack-ironic | 00:35 | |
| *** ryanpetrello has quit IRC | 00:37 | |
| naohirot | good morning to all | 00:39 |
| *** ijw has quit IRC | 00:39 | |
| mrda | hi naohirot | 00:39 |
| naohirot | mrda: hi | 00:39 |
| *** Masahiro has quit IRC | 00:39 | |
| jroll | devananda: mrda: so basically, if there isn't an official midcycle in SF area, we're going to host an unofficial one. you should come either way. | 00:43 |
| mrda | jroll: that's probably enough for me to start talking to the powers that be | 00:44 |
| * naohirot mrda: I just watched Yahoo News about Wild Fire | 00:44 | |
| jroll | mrda: cool, I'd hold off on hotel until devananda confirms SF and not, say, sunnyvale, but other than that do it up | 00:48 |
| mrda | naohirot: The fire is only 10 mile away from me | 00:49 |
| devananda | jroll: by official vs. unofficial, is there really anything more than me saying something on the mailing list? | 00:49 |
| devananda | jroll: also, if you want to recommend a hotel and sturff, we should, you know, say that, too | 00:49 |
| naohirot | mrda: really? | 00:49 |
| jroll | devananda: heh, idk. | 00:49 |
| mrda | devananda: jroll: so we're looking the week of Feb 9 - Feb 13 ish | 00:50 |
| JayF | devananda: let me see if we have a standard list of places or smth | 00:50 |
| jroll | russell_h: you have hotel recs around here? | 00:50 |
| devananda | if you guys want it to be official and such, it'll get more attention | 00:50 |
| mrda | naohirot: yep, it's pretty close. I can smell and see smoke from my front door | 00:50 |
| devananda | and I will need to know max headcount, things like that | 00:50 |
| jroll | right, I guess what I'm saying is if you say no meetup here, we're going to have our own :P | 00:50 |
| JayF | jroll: I was going to ask $recruiter_people since they likely do reccomendations for incoming folks? | 00:50 |
| jroll | JayF: that might work | 00:51 |
| jroll | or ask clif_h where he stayed | 00:51 |
| devananda | jroll: JayF: since it's the bay area, I would generally expect a lot of drop-ins as well | 00:51 |
| jroll | devananda: yep. lemme grab info | 00:51 |
| devananda | do ya'll *want* 20 or 30 people, if that happens? | 00:51 |
| naohirot | mrda: Is it happening in east coast? | 00:51 |
| jroll | russell_h: ^ you should talk to deva | 00:51 |
| jroll | devananda: I'm thinking probably 30 max? | 00:52 |
| russell_h | jroll: we usually use the marriott courtyard across the street, its not amazing but it works | 00:52 |
| * naohirot picture I watched is Cudlee Creek in South Australia | 00:52 | |
| devananda | russell_h: please send me all the infos and/or edit https://wiki.openstack.org/wiki/Sprints/IronicKiloSprint to add it directly | 00:52 |
| russell_h | ok, so confirmed we're doing it here then? | 00:53 |
| jroll | I guess so | 00:53 |
| devananda | russell_h: I believe jroll just told me that you will do it whether I say yes or no | 00:53 |
| russell_h | or, doing a thing here anyway, not "it" | 00:53 |
| devananda | russell_h: so the question becomes: do you want me to announce it or not? | 00:53 |
| devananda | also, ++ on JFDI | 00:53 |
| russell_h | heh | 00:53 |
| JayF | From a certain perspective; we have mid-cycle meetups daily ;) | 00:53 |
| jroll | devananda: what I meant was is, if nothing happens in the bay area, we're going to have one. if you'd rather host it at HP or something, we're ok with that too | 00:54 |
| jroll | but we have space | 00:54 |
| jroll | let's jfdi | 00:54 |
| devananda | jroll: I love it when other people host. I'm waaay better as a guest at a party than as the host. | 00:54 |
| jroll | heh | 00:54 |
| * naohirot mrda: near melbourne | 00:55 | |
| russell_h | devananda: ok, let me confirm I've got the space, add the info to that wiki, then you announce | 00:55 |
| devananda | russell_h: sounds good | 00:56 |
| *** Masahiro has joined #openstack-ironic | 00:58 | |
| mrda | naohirot: Cudlee Creek is the other side of the bushfire to where I am. I'm on the Golden Grove side of the fire. The fire is about 200 km in diameter | 00:58 |
| mrda | naohirot: there's another fire in Victoria near Melbourne. | 00:59 |
| mrda | naohirot: but I am in Adelaide | 00:59 |
| naohirot | mrda: I see, I haven't been in Australia. so I'm looking at Google map right now | 01:00 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Refactor async helper methods in conductor/manager.py https://review.openstack.org/139217 | 01:07 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Convert drivers to use process_event() https://review.openstack.org/140883 | 01:07 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Enable async callbacks from task.process_event() https://review.openstack.org/140868 | 01:07 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Convert check_deploy_timeout to use process_event https://review.openstack.org/140869 | 01:07 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: WIP: fix some things in the state machine https://review.openstack.org/145389 | 01:07 |
| devananda | gah! | 01:07 |
| devananda | didn' tmean to rebase all that | 01:07 |
| devananda | anyhow, gotta run now | 01:07 |
| devananda | that last one seems to fix some issues that rloo found (but not all) | 01:08 |
| devananda | see ya'll tmw | 01:08 |
| * naohirot mrda: now I understood where the fires are happening :) | 01:08 | |
| JayF | well 137217 was landing | 01:10 |
| JayF | I thought | 01:10 |
| JayF | wonder if it's not now | 01:10 |
| JayF | yeah someone needs to reapprove 137217 | 01:10 |
| * JayF would do it but he only has the big hammer for specs and ipa | 01:11 | |
| * naohirot mrda Adelaide is a beautiful city, isn't it. I'm walking a street in Google Map :) | 01:12 | |
| mrda | naohirot: that's a kind thing to say. I choose to live here :) | 01:13 |
| naohirot | mrda: Yeah, it's very nice! | 01:13 |
| naohirot | mrda: I'd like to visit there some day, is there a plan to have open stack summit in Adelaide? :) | 01:16 |
| *** ijw has joined #openstack-ironic | 01:17 | |
| mrda | naohirot: that would be great! But I think a sprint is more likely. We held linux.conf.au here 10 years ago :) | 01:17 |
| naohirot | mrda: I see. Yeah I think one of fun is attending International conference is site seeing near around. I believe Spring is the best :) | 01:20 |
| openstackgerrit | Alex Weeks proposed openstack/ironic-python-agent: Added extension to support rescue mode https://review.openstack.org/145385 | 01:20 |
| naohirot | mrda: s/site/sight/ | 01:20 |
| mrda | :) | 01:25 |
| *** nosnos has joined #openstack-ironic | 01:27 | |
| naohirot | mrda: (^_^) popular smiley here in Japan | 01:28 |
| *** ijw has quit IRC | 01:39 | |
| *** ryanpetrello has joined #openstack-ironic | 01:48 | |
| *** ijw has joined #openstack-ironic | 01:53 | |
| *** ijw has quit IRC | 01:55 | |
| *** ryanpetrello has quit IRC | 01:58 | |
| *** lucasagomes has quit IRC | 02:24 | |
| openstackgerrit | Merged openstack/ironic: Hide oslo.messaging DEBUG logs by default https://review.openstack.org/145361 | 02:27 |
| *** jerryz has joined #openstack-ironic | 02:29 | |
| openstackgerrit | Naohiro Tamura proposed openstack/ironic-specs: iRMC Virtual Media Deploy Driver for Ironic https://review.openstack.org/134865 | 02:32 |
| *** ramineni has joined #openstack-ironic | 02:40 | |
| openstackgerrit | Naohiro Tamura proposed openstack/ironic-specs: iRMC Management Driver for Ironic https://review.openstack.org/136020 | 02:42 |
| *** dlaube has quit IRC | 02:55 | |
| *** nosnos has quit IRC | 03:20 | |
| *** rlr0329 has quit IRC | 03:23 | |
| *** naohirot has quit IRC | 03:29 | |
| *** chlong has quit IRC | 03:41 | |
| *** chlong has joined #openstack-ironic | 03:42 | |
| *** chlong has quit IRC | 03:42 | |
| *** Nisha has joined #openstack-ironic | 03:44 | |
| *** chlong has joined #openstack-ironic | 03:44 | |
| *** pcrews has quit IRC | 03:44 | |
| *** chlong has quit IRC | 03:48 | |
| *** chlong has joined #openstack-ironic | 03:48 | |
| *** naohirot has joined #openstack-ironic | 04:04 | |
| *** teju has joined #openstack-ironic | 04:13 | |
| *** nosnos has joined #openstack-ironic | 04:15 | |
| *** Marga_ has quit IRC | 04:15 | |
| *** Marga_ has joined #openstack-ironic | 04:16 | |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic-specs: Discover node properties for iLO drivers https://review.openstack.org/103007 | 04:17 |
| openstackgerrit | Anusha Ramineni proposed openstack/ironic-specs: Implement Cleaning Operations for iLO drivers https://review.openstack.org/142401 | 04:20 |
| *** Marga_ has quit IRC | 04:20 | |
| *** rameshg87 has joined #openstack-ironic | 04:23 | |
| rameshg87 | good morning ironic | 04:24 |
| openstackgerrit | Merged openstack/ironic: Refactor async helper methods in conductor/manager.py https://review.openstack.org/139217 | 04:32 |
| rameshg87 | naohirot, hi | 04:41 |
| naohirot | rameshg87: hi | 04:41 |
| rameshg87 | naohirot, regarding https://review.openstack.org/#/c/144901/4 | 04:43 |
| naohirot | rameshg87: thanks for reviewing it | 04:44 |
| rameshg87 | naohirot, have you thought about moving scci.py to a separate module ? | 04:44 |
| rameshg87 | naohirot, current we have all the ilo-related functionalities in a separate module called proliantutils | 04:44 |
| naohirot | rameshg87: such as proliant library? | 04:45 |
| rameshg87 | naohirot, yeah | 04:45 |
| rameshg87 | naohirot, it allows someone else to reuse your library outside ironic as well | 04:45 |
| naohirot | rameshg87: Yes, you are right | 04:46 |
| rameshg87 | naohirot, and it's clean if those irmc-hardware-specific things are outside ironic and ironic just uses an api provided by your client | 04:46 |
| naohirot | rameshg87: in case of Fujitsu, we are not matured to contribute to open source. | 04:47 |
| naohirot | rameshg87: If I try to do so, I need to do a lot of chore. | 04:47 |
| rameshg87 | naohirot, yeah, but it's the same way as contributing to ironic | 04:48 |
| rameshg87 | naohirot, anyone can create a new repository in github and then publish a new module in PyPI | 04:48 |
| rameshg87 | naohirot, that's all required | 04:48 |
| rameshg87 | naohirot, this is another hardware vendor: https://github.com/seamicro/python-seamicroclient | 04:48 |
| rameshg87 | naohirot, seamicro has done the same too | 04:48 |
| naohirot | rameshg87: let me clarify one thing. | 04:49 |
| naohirot | rameshg87: In case of seamicro, did they create the python-seamicroclient after or at the same time developing the ironic code? | 04:50 |
| rameshg87 | naohirot, at the time actually | 04:50 |
| rameshg87 | naohirot, we too did it at the time of developing ironic code :) | 04:50 |
| naohirot | rameshg87: I see. How about proliantutils? | 04:51 |
| rameshg87 | naohirot, the same .. | 04:51 |
| rameshg87 | naohirot, a quick look | 04:51 |
| naohirot | rameshg87: I see. I thought that proliantutils had existed before Ironic. | 04:51 |
| rameshg87 | naohirot, if we refactor irmc-specific-things outside | 04:51 |
| rameshg87 | naohirot, scci.py, test_scci.py is moved outside which 1000-lines less in ironic :) | 04:52 |
| rameshg87 | naohirot, and you could maintain it outside ironic | 04:52 |
| naohirot | rameshg87: I understood. | 04:53 |
| rameshg87 | naohirot, proliantutils was developer in parallel while developing ilo driver in ironic | 04:53 |
| rameshg87 | s/developer/developed/ | 04:53 |
| naohirot | rameshg87: Regarding the line of code, is that the matter of Ironic? | 04:54 |
| rameshg87 | naohirot, idk actually .. | 04:54 |
| rameshg87 | naohirot, but i would also say it's a lot cleaner if it's outside ironic since it is irmc-hardware-specific | 04:55 |
| naohirot | rameshg87: I agree it's cleaner. | 04:56 |
| naohirot | rameshg87: For future activity, I'd like to know the way of thinking regarding the line of code in Ironic or OpenStack. | 04:57 |
| rameshg87 | naohirot, i didn't get | 04:57 |
| naohirot | rameshg87: You said that scci.py and test_sccy.py total number of lines is 1000. | 04:58 |
| rameshg87 | naohirot, ah that was just a number | 04:58 |
| naohirot | rameshg87: Increasing 1000 lines of code is too much? | 04:59 |
| rameshg87 | naohirot, i just wanted to convey that 1/4th of the code (1000 out of 4000) was specific to irmc-hardware and could be moved out :) | 04:59 |
| naohirot | rameshg87: I see. I understood. | 05:00 |
| naohirot | rameshg87: How about test data xml. It's total 2500 lines of data. | 05:00 |
| rameshg87 | naohirot, yeah but it was just test data for your test_scci.py right ? | 05:01 |
| rameshg87 | naohirot, so that too could be moved out, right ? | 05:01 |
| naohirot | rameshg87: At this moment, yes. | 05:01 |
| rameshg87 | naohirot, yeah | 05:02 |
| naohirot | rameshg87: But that data is used in Management Interface | 05:02 |
| rameshg87 | naohirot, yeah but in reality it could be moved out of management interface, right ? | 05:03 |
| naohirot | rameshg87: So I needs to check in XML data when I check in iRMC Management Driver. | 05:03 |
| rameshg87 | naohirot, all management interface needs to know if how to set the boot device and your python-irmcclient could provide an api to do that | 05:03 |
| rameshg87 | naohirot, how python-irmcclient does that is it's job | 05:04 |
| naohirot | rameshg87: I'm talking about get_sensor_data(). | 05:04 |
| naohirot | rameshg87: XML data has sensor data inside. | 05:04 |
| rameshg87 | naohirot, yeah, but it can be ideally part of python-irmcclient | 05:05 |
| rameshg87 | naohirot, and irmcclient could just provide an api for irmc ironic driver to do the job | 05:05 |
| rameshg87 | naohirot, irmc driver in ironic needn't worry about xmls if irmcclient can actually give the data in form of python dictionaries | 05:07 |
| *** Marga_ has joined #openstack-ironic | 05:07 | |
| naohirot | rameshg87: Okay, I'll try to do test in irmcclient package as much as possible | 05:07 |
| *** Marga_ has quit IRC | 05:08 | |
| *** Marga_ has joined #openstack-ironic | 05:08 | |
| rameshg87 | naohirot, yeah, what i meant was to move as much as possible irmc-hardware-specific into irmcclient | 05:08 |
| rameshg87 | naohirot, in my opinion irmc driver in ironic just needs a few APIs from irmcclient to do the job | 05:09 |
| naohirot | rameshg87: Yes, I understood. | 05:09 |
| rameshg87 | naohirot, APIs like client.set_power_state(), client.get_power_state(), client.get_boot_device(), etc | 05:09 |
| rameshg87 | naohirot, yeah :) | 05:09 |
| rameshg87 | naohirot, and one more thing | 05:09 |
| rameshg87 | naohirot, i see https://review.openstack.org/#/c/134865/ is still not rebased as a dependent of management interface | 05:10 |
| rameshg87 | naohirot, i guess that's what JayF meant | 05:11 |
| naohirot | rameshg87: really? | 05:11 |
| naohirot | rameshg87: I did rebase this morning. How did you know rebase is not enough? | 05:11 |
| rameshg87 | naohirot, it should come as a dependency | 05:12 |
| * rameshg87 gets a example | 05:12 | |
| rameshg87 | naohirot, have a look at https://review.openstack.org/#/c/142401/ | 05:12 |
| rameshg87 | naohirot, in the gerrit review you will see this review depends on "Implement Cleaning States" which is https://review.openstack.org/#/c/102685/22 | 05:13 |
| *** krtaylor has quit IRC | 05:13 | |
| rameshg87 | naohirot, and this is how we do it : https://wiki.openstack.org/wiki/Gerrit_Workflow#Add_dependency | 05:13 |
| rameshg87 | naohirot, if you add A as gerrit dependency of B, gerrit will make sure that B doesn't get merged before A | 05:14 |
| naohirot | rameshg87: Aha, I got it. As I just started to code iRMC Management driver, I created the dependency between iRMC Power driver code. | 05:15 |
| rameshg87 | naohirot, okay :) | 05:16 |
| naohirot | rameshg87: I need to do the same in case of spec. | 05:16 |
| rameshg87 | naohirot, yes | 05:16 |
| openstackgerrit | Anusha Ramineni proposed openstack/ironic-specs: Add Zapping support to iLO drivers https://review.openstack.org/145404 | 05:17 |
| naohirot | rameshg87: How should I create dependency? Deploy depends on Management, Management depends on Power, is it Okay? | 05:17 |
| *** pradipta_away is now known as pradipta | 05:20 | |
| naohirot | rameshg87: master <- Power <- Management <- Deploy, this is what I'm thinking. | 05:20 |
| rameshg87 | naohirot, but power is already merged right | 05:20 |
| rameshg87 | naohirot, and your management in independent | 05:21 |
| rameshg87 | naohirot, so it's just deploy depends on management | 05:21 |
| naohirot | rameshg87: Oh, Ok. that's simpler. | 05:21 |
| naohirot | rameshg87: I'll fix that. | 05:21 |
| rameshg87 | naohirot, okay | 05:22 |
| *** Haomeng|2 has joined #openstack-ironic | 05:23 | |
| *** Haomeng has quit IRC | 05:23 | |
| naohirot | rameshg87: I'll take a look how to creat PyPI package. Maybe I'll ask you if there is question how to do. | 05:23 |
| rameshg87 | naohirot, sure | 05:23 |
| naohirot | rameshg87: thanks! | 05:24 |
| rameshg87 | naohirot, wc | 05:24 |
| *** krtaylor has joined #openstack-ironic | 05:25 | |
| teju | rameshg87 : hi. can u please help me with https://ask.openstack.org/en/question/56956/failed-to-deploy-error-cannot-call-save-on-orphaned-node-object/? | 05:31 |
| *** pensu has joined #openstack-ironic | 05:33 | |
| rameshg87 | teju, first question is why are you trying to set node.provision_state = states.NOSTATE at the end of deploy | 05:34 |
| rameshg87 | teju, it should be actually node.provision_state = states.ACTIVE, right ? | 05:34 |
| teju | rameshg87 : if node.provision_state = states.ACTIVE, then we cannot provision the same node again.right? | 05:35 |
| rameshg87 | teju, you will need to tear_down the node before provisioning it again | 05:36 |
| rameshg87 | teju, by doing ironic node-set-provision-state <node-uuid> deleted | 05:36 |
| teju | rameshg87 : hmmm...ok | 05:37 |
| teju | rameshg87 : i dont see the comamnd "node-set-provision-state" ...is it added in juno? | 05:39 |
| rameshg87 | teju, ironic node-set-provision-state | 05:39 |
| rameshg87 | teju, yeah may be it was added in juno | 05:39 |
| teju | rameshg87 : ok. | 05:42 |
| teju | rameshg87 : in the pxe.py in https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/pxe.py , i see node.provision_state = states.ACTIVE node.target_provision_state = states.NOSTATE. i thought I can use the same idea in my driver code | 05:45 |
| rameshg87 | teju, yeah | 05:45 |
| rameshg87 | teju, that's the same thing i meant | 05:45 |
| rameshg87 | teju, node.provision_state = states.ACTIVE at the end of deploy if deploy was successful | 05:45 |
| naohirot | rameshg87: can I ask about https://github.com/hpproliant repository? | 05:50 |
| rameshg87 | naohirot, sure | 05:50 |
| naohirot | rameshg87: Why doesn't "hpproliant" have hyper-like? Is it HP's account? | 05:51 |
| Nisha | /join #openstack-nova | 05:52 |
| naohirot | rameshg87: I know that ramineni checked in the code into https://github.com/hpproliant/proliantutils | 05:52 |
| naohirot | rameshg87: Sorry, each account's homepage itself doesn't have hyper-link uder the left-upper corner icon. | 05:54 |
| naohirot | rameshg87: s/uder/under/ | 05:54 |
| rameshg87 | naohirot, there are multiple people working on it | 05:55 |
| rameshg87 | naohirot, so we created one account for all of us | 05:55 |
| naohirot | rameshg87: Anyway I believe "hpproliant" is one of HP's account. | 05:55 |
| *** eghobo has joined #openstack-ironic | 05:56 | |
| naohirot | rameshg87: Okay, I have to start to searching my company's id :) | 05:57 |
| rameshg87 | naohirot, it's not hp's account. rather our group uses that account :) | 05:58 |
| naohirot | rameshg87: I see. I'm doubting that my company has such that flexibility :-) | 05:59 |
| rameshg87 | naohirot, :) | 06:00 |
| *** Marga_ has quit IRC | 06:31 | |
| *** Marga_ has joined #openstack-ironic | 06:31 | |
| *** eghobo has quit IRC | 06:33 | |
| *** lintan has quit IRC | 06:43 | |
| openstackgerrit | Naohiro Tamura proposed openstack/ironic-specs: iRMC Virtual Media Deploy Driver for Ironic https://review.openstack.org/134865 | 06:57 |
| *** Masahiro has quit IRC | 07:03 | |
| *** Masahiro has joined #openstack-ironic | 07:03 | |
| *** dlpartain has joined #openstack-ironic | 07:16 | |
| *** Masahiro has quit IRC | 07:24 | |
| *** Masahiro has joined #openstack-ironic | 07:30 | |
| *** Masahiro has quit IRC | 07:33 | |
| *** dlpartain has left #openstack-ironic | 07:45 | |
| *** ifarkas has joined #openstack-ironic | 07:45 | |
| *** subscope has joined #openstack-ironic | 07:57 | |
| *** pcaruana|afk| is now known as pcaruana | 08:04 | |
| *** Nisha has quit IRC | 08:05 | |
| *** chlong has quit IRC | 08:07 | |
| *** yuanying has joined #openstack-ironic | 08:16 | |
| *** yuanying_ has quit IRC | 08:17 | |
| *** Nisha has joined #openstack-ironic | 08:29 | |
| *** kongfy has joined #openstack-ironic | 08:33 | |
| *** dtantsur|afk is now known as dtantsur | 08:42 | |
| dtantsur | Morning Ironic | 08:43 |
| *** Masahiro_ has joined #openstack-ironic | 08:57 | |
| *** pcaruana has quit IRC | 08:59 | |
| *** nosnos has quit IRC | 09:00 | |
| *** erwan_taf has joined #openstack-ironic | 09:00 | |
| *** nosnos has joined #openstack-ironic | 09:01 | |
| *** jcoufal has joined #openstack-ironic | 09:07 | |
| Nisha | dtantsur, morning | 09:07 |
| dtantsur | Nisha, o/ | 09:07 |
| *** nosnos has quit IRC | 09:08 | |
| Nisha | dtantsur, https://review.openstack.org/#/c/103007/27 ready for review :) | 09:08 |
| dtantsur | yeah on my radar | 09:09 |
| Nisha | :) | 09:09 |
| *** nosnos has joined #openstack-ironic | 09:10 | |
| *** pcaruana has joined #openstack-ironic | 09:13 | |
| *** romcheg has joined #openstack-ironic | 09:13 | |
| *** jistr has joined #openstack-ironic | 09:14 | |
| *** smoriya has quit IRC | 09:16 | |
| *** athomas has joined #openstack-ironic | 09:17 | |
| *** jiangfei has quit IRC | 09:23 | |
| *** GheRivero has quit IRC | 09:23 | |
| *** GheRivero has joined #openstack-ironic | 09:23 | |
| *** jiangfei has joined #openstack-ironic | 09:23 | |
| rameshg87 | erwan_taf, hi | 09:45 |
| *** kongfy has quit IRC | 09:50 | |
| erwan_taf | hey rameshg87 | 09:53 |
| erwan_taf | rameshg87: just commented the use case ;) | 09:53 |
| rameshg87 | erwan_taf, got it :) | 09:54 |
| erwan_taf | hope it makes sense for you | 09:54 |
| rameshg87 | erwan_taf, yeah, i am open to adding it. | 09:54 |
| erwan_taf | you didn't commented about adding the model name | 09:55 |
| rameshg87 | erwan_taf, doesn't vendor serve that use-case ? | 09:55 |
| rameshg87 | erwan_taf, or you would like to have precise model name as well ? | 09:56 |
| rameshg87 | erwan_taf, but i am afraid the more we add more vendor-specific things - the more we are limiting the applicability of machine-independent configuration | 09:56 |
| rameshg87 | erwan_taf, like firmware version and model name are specific to specific-hardware | 09:57 |
| rameshg87 | erwan_taf, like i will start needing to provide separate raid information for hp machines, separate one for dell machines | 09:57 |
| rameshg87 | erwan_taf, that worries me :( | 09:58 |
| rameshg87 | dtantsur, ifarkas, have some thoughts on the above ^^^ | 09:58 |
| erwan_taf | I do agree that collecting those information will vary from vendors but if you look at the use case, that's generic | 09:59 |
| rameshg87 | dtantsur, ifarkas, adding model name and firmware version also to the selection criteria | 09:59 |
| erwan_taf | I mean, an ATA-7 device is featuring thoses information | 09:59 |
| *** MattMan has joined #openstack-ironic | 10:00 | |
| ifarkas | rameshg87, I see the value of adding firmware and model name. If we want Ironic to make a smart decision, it should be as broad as it can be | 10:03 |
| ifarkas | rameshg87, maybe we should make this matching more generic to support arbitrary properties... | 10:03 |
| dtantsur | I'm not that familiar with RAID, sorry :) | 10:04 |
| *** derekh has joined #openstack-ironic | 10:05 | |
| rameshg87 | ifarkas, but then matching of arbitarary properties would be driver dependent | 10:07 |
| rameshg87 | ifarkas, erwan_taf, for now what ever we have added every raid controller (hopefully) should be able to do | 10:07 |
| rameshg87 | ifarkas, erwan_taf, i agree firmware and model name should be findable by every raid controller | 10:08 |
| rameshg87 | ifarkas, erwan_taf, but if we support arbitarary things - it might work for one driver and might not work for another driver | 10:08 |
| rameshg87 | ifarkas, erwan_taf, and that might defeat our purpose, right ? | 10:09 |
| *** andreykurilin has joined #openstack-ironic | 10:09 | |
| ifarkas | rameshg87, yeah, that's a good point | 10:13 |
| rameshg87 | ifarkas, erwan_taf, may be if the raid picks up, let's add support for more in increments | 10:14 |
| *** bigjools has quit IRC | 10:14 | |
| rameshg87 | ifarkas, erwan_taf, we could start versioning input raid info - like whatever we support till now (say interface_type, disk_type, vendor, etc) is raid info version 1 | 10:15 |
| *** jiangfei has quit IRC | 10:15 | |
| rameshg87 | ifarkas, erwan_taf, when there is a popular demand for a couple of more features from raid_info, we could add more and increment raid info version | 10:16 |
| *** jiangfei has joined #openstack-ironic | 10:16 | |
| ifarkas | rameshg87, works for me | 10:17 |
| *** naohirot has quit IRC | 10:17 | |
| rameshg87 | erwan_taf, what do you say ? | 10:18 |
| erwan_taf | I'm not really minded the same way and I know it is not very popular here. I mean, if you miss that information at the time you really needs it ... it will requires long time to get it onboard... | 10:20 |
| erwan_taf | while it doesn't really cost adding it right now | 10:20 |
| *** bigjools has joined #openstack-ironic | 10:26 | |
| *** Nisha_away has joined #openstack-ironic | 10:32 | |
| *** Nisha has quit IRC | 10:36 | |
| *** Nisha_away has quit IRC | 10:36 | |
| *** chlong has joined #openstack-ironic | 10:38 | |
| *** Masahiro_ has quit IRC | 10:41 | |
| *** Masahiro has joined #openstack-ironic | 10:43 | |
| openstackgerrit | Imre Farkas proposed openstack/ironic: DRAC specific RAID methods https://review.openstack.org/145464 | 10:43 |
| *** andreykurilin has quit IRC | 10:52 | |
| *** andreykurilin has joined #openstack-ironic | 10:52 | |
| openstackgerrit | Imre Farkas proposed openstack/ironic: DRAC specific RAID methods https://review.openstack.org/145464 | 10:53 |
| *** ramineni has quit IRC | 10:57 | |
| *** Masahiro has quit IRC | 11:00 | |
| *** Masahiro has joined #openstack-ironic | 11:00 | |
| *** lucasagomes has joined #openstack-ironic | 11:05 | |
| erwan_taf | rameshg87: did you understood the share_physical_disks use case ? I'm totally lost with it | 11:06 |
| *** Masahiro has quit IRC | 11:06 | |
| rameshg87 | erwan_taf, o dodm | 11:06 |
| rameshg87 | erwan_taf, i didn't see you had replied | 11:07 |
| rameshg87 | erwan_taf, because it didn't alert me | 11:07 |
| rameshg87 | erwan_taf, it's almost time for me to leave for home | 11:08 |
| rameshg87 | erwan_taf, i will be connecting back later | 11:08 |
| erwan_taf | sure | 11:08 |
| erwan_taf | see ya | 11:08 |
| rameshg87 | erwan_taf, bye | 11:08 |
| *** chlong has quit IRC | 11:08 | |
| *** rakesh_hs2 has joined #openstack-ironic | 11:08 | |
| *** rameshg87 has quit IRC | 11:09 | |
| *** pelix has joined #openstack-ironic | 11:09 | |
| *** andreykurilin has quit IRC | 11:10 | |
| *** GheRivero has quit IRC | 11:15 | |
| *** GheRivero has joined #openstack-ironic | 11:21 | |
| openstackgerrit | Merged stackforge/ironic-discoverd: Make database a required configuration option https://review.openstack.org/145231 | 11:26 |
| *** GheRivero has quit IRC | 11:33 | |
| *** GheRivero has joined #openstack-ironic | 11:34 | |
| *** pradipta is now known as pradipta_away | 11:41 | |
| *** GheRivero has quit IRC | 11:44 | |
| *** jcoufal_ has joined #openstack-ironic | 11:45 | |
| *** Masahiro has joined #openstack-ironic | 11:47 | |
| *** Masahiro has quit IRC | 11:47 | |
| *** Masahiro has joined #openstack-ironic | 11:48 | |
| *** jcoufal has quit IRC | 11:49 | |
| *** GheRivero has joined #openstack-ironic | 11:50 | |
| *** romcheg has quit IRC | 11:54 | |
| *** alexpilotti has joined #openstack-ironic | 11:57 | |
| *** david-lyle_afk is now known as david-lyle | 11:58 | |
| *** rakesh_hs2 has quit IRC | 12:17 | |
| *** romcheg has joined #openstack-ironic | 12:19 | |
| *** teju has quit IRC | 12:25 | |
| *** Masahiro has quit IRC | 12:26 | |
| *** pensu has quit IRC | 12:30 | |
| *** dprince has joined #openstack-ironic | 12:33 | |
| *** athomas has quit IRC | 12:35 | |
| *** athomas has joined #openstack-ironic | 12:38 | |
| *** lazy_prince is now known as killer_prince | 12:38 | |
| *** athomas has quit IRC | 12:38 | |
| *** athomas has joined #openstack-ironic | 12:40 | |
| *** nosnos has quit IRC | 12:43 | |
| *** ryanpetrello has joined #openstack-ironic | 12:44 | |
| *** MattMan is now known as MattMan-Away | 12:48 | |
| *** GheRivero has quit IRC | 12:53 | |
| *** GheRivero has joined #openstack-ironic | 12:53 | |
| *** dlpartain has joined #openstack-ironic | 13:00 | |
| *** david-lyle is now known as david-lyle_afk | 13:08 | |
| *** rameshg87 has joined #openstack-ironic | 13:11 | |
| *** naohirot has joined #openstack-ironic | 13:12 | |
| rameshg87 | erwan_taf, hi | 13:14 |
| erwan_taf | hey | 13:16 |
| erwan_taf | did you understood the "share_physical_disks" case ? | 13:16 |
| erwan_taf | I wasn't able to figure what it really means | 13:16 |
| rameshg87 | erwan_taf, yeah as far as i understand | 13:16 |
| rameshg87 | erwan_taf, we could create a single array with say 5 disks | 13:17 |
| rameshg87 | erwan_taf, and then create multiple logical disks of different raid levels in the same array until the array runs out of space | 13:17 |
| rameshg87 | erwan_taf, the logical disks part of the same raid array will share same physical disks | 13:17 |
| openstackgerrit | Matthew Gilliard proposed openstack/ironic: Check glance for new kernel and ramdisk on prepare https://review.openstack.org/142782 | 13:19 |
| *** GheRivero has quit IRC | 13:20 | |
| erwan_taf | rameshg87: I didn't understood why we need an option for that | 13:21 |
| erwan_taf | rameshg87: I mean, is it a property required at creation time ? | 13:21 |
| rameshg87 | erwan_taf, it will be a hint to the driver which selects the physical disks for the logical disk | 13:22 |
| rameshg87 | erwan_taf, for example, say if we have 4x600 GB disks P1, P2, P3, P4 | 13:23 |
| rameshg87 | erwan_taf, and we need 2 logical disks of raid 1 L1 and L2 | 13:23 |
| rameshg87 | erwan_taf, say logical disks of 200 GB | 13:23 |
| rameshg87 | erwan_taf, if share_physical_disks = true, then we can create a raid array A with P1 and P2, and create both logical disks L1 and L2 on raid array A | 13:24 |
| openstackgerrit | Tan Lin proposed openstack/ironic: Regard 'method' as a mandatory argument for VendorInterface.validate() https://review.openstack.org/144265 | 13:25 |
| rameshg87 | erwan_taf, if share_physical_disks = false, then we have to create two raid arrays - raid array A with P1 and P2; raid array B with P3 and P4; then create L1 on raid array A and L2 on raid array B | 13:25 |
| rameshg87 | erwan_taf, share_physical_disks = false should perform better because simultaneous writes to 2 logical disks will be done on different disks | 13:26 |
| rameshg87 | erwan_taf, it is just a hint to the driver on how to choose the physical disk | 13:26 |
| *** GheRivero has joined #openstack-ironic | 13:26 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-specs: In-band hardware properties inspection via ironic-discoverd https://review.openstack.org/135605 | 13:27 |
| *** Masahiro has joined #openstack-ironic | 13:27 | |
| *** andreykurilin has joined #openstack-ironic | 13:27 | |
| rameshg87 | lucasagomes, hi | 13:28 |
| lucasagomes | rameshg87, hi there | 13:30 |
| rameshg87 | lucasagomes, regarding https://review.openstack.org/#/c/130812/4/specs/kilo/new-ipxe-driver.rst | 13:30 |
| rameshg87 | lucasagomes, was just wondering what happens to it when we have a non-glance image | 13:30 |
| *** jcoufal_ has quit IRC | 13:31 | |
| *** jcoufal has joined #openstack-ironic | 13:31 | |
| lucasagomes | rameshg87, the ramdisks and kernels you mean? | 13:31 |
| rameshg87 | lucasagomes, yeah | 13:31 |
| *** Masahiro has quit IRC | 13:32 | |
| rameshg87 | lucasagomes, https://review.openstack.org/#/c/136741/ | 13:32 |
| rameshg87 | lucasagomes, in the above we are adding support for providing kernel/ramdisk from both http:// and file:// | 13:33 |
| lucasagomes | rameshg87, yeah probably won't be supported, since we need a HTTP server | 13:33 |
| rameshg87 | lucasagomes, http:// is fine, ipxe will gladly accept it :) | 13:33 |
| rameshg87 | lucasagomes, but file:// will fail | 13:33 |
| lucasagomes | when I wrote that spec the non-glance thing wasn't merged yet | 13:33 |
| lucasagomes | yeah | 13:33 |
| lucasagomes | unless we make those images available using the Ironic API | 13:33 |
| lucasagomes | but I don't think we want to do that | 13:33 |
| lucasagomes | maybe it's better to not support the file:// for the ramdisks | 13:34 |
| rameshg87 | lucasagomes, yeah, but the support is being added in a generic way | 13:34 |
| rameshg87 | lucasagomes, if file:// works for os-image, then file:// works for ramdisks too :) | 13:35 |
| rameshg87 | lucasagomes, may be we can just document that it doesn't work with file:// and leave it, right ? | 13:35 |
| lucasagomes | I believe so, we can fail validation if images in that driver uses file:// | 13:35 |
| rameshg87 | lucasagomes, yeah | 13:36 |
| lucasagomes | since the idea is be stateless we can't have local iamges | 13:36 |
| lucasagomes | images* | 13:36 |
| rameshg87 | lucasagomes, :) | 13:36 |
| * lucasagomes wants to have more time to work on that driver :( | 13:36 | |
| lucasagomes | rameshg87, but yeah, I have to update that spec | 13:37 |
| lucasagomes | and include that observation, at the time I wrote it non-glance images wasn't a thing yet | 13:37 |
| rameshg87 | lucasagomes, yeah | 13:37 |
| rameshg87 | lucasagomes, but it's a great change. you don't need to cache kernel/ramdisk anymore on the conductor after that :) | 13:38 |
| lucasagomes | yup, not pxe config files | 13:38 |
| lucasagomes | anything | 13:38 |
| rameshg87 | lucasagomes, no more failures if conductor goes down :) | 13:38 |
| lucasagomes | yeah, the take over will be straight forward | 13:38 |
| rameshg87 | lucasagomes, unless it goes down during deploy | 13:38 |
| lucasagomes | yeah but then there's the retry filter in nova | 13:39 |
| lucasagomes | that will resend the deploy request and this time (in case conductor is now down) will go to another conductor | 13:39 |
| rameshg87 | lucasagomes, oh okay | 13:41 |
| erwan_taf | rameshg87: doesn't sounds really determinastic at my taste... | 13:44 |
| rameshg87 | erwan_taf, what do you mean by that ? | 13:45 |
| *** dlpartain has left #openstack-ironic | 13:45 | |
| *** lintan_ has joined #openstack-ironic | 13:49 | |
| erwan_taf | rameshg87: oh that means that you can have several logical_disks pointed the same physical_disk if share_physical_disks is true ? | 13:50 |
| *** kbyrne has quit IRC | 13:50 | |
| rameshg87 | erwan_taf, yeah | 13:51 |
| erwan_taf | rameshg87: can we have some sample config for Backing physical disk hints ? | 13:53 |
| erwan_taf | I don't really get how you'll use all those information to create configuration | 13:53 |
| rameshg87 | erwan_taf, i guess it's there in spec | 13:53 |
| rameshg87 | erwan_taf, http://docs-draft.openstack.org/99/135899/15/check/gate-ironic-specs-docs/3701904//doc/build/html/specs/kilo/ironic-generic-raid-interface.html | 13:54 |
| rameshg87 | erwan_taf, look for "Some examples" | 13:54 |
| *** jjohnson2 has joined #openstack-ironic | 13:55 | |
| erwan_taf | rameshg87: I saw them but didn't understood as I should | 13:57 |
| erwan_taf | rameshg87: I don't get why we cannot merge both approach | 13:57 |
| erwan_taf | rameshg87: in fact we are defining another "physical_disks" structure withtout saying so | 13:57 |
| rameshg87 | erwan_taf, in example 1, we use machine-independent configuration. if i had 10 machines, i could apply same on all 10 machines without inspecting them | 13:58 |
| erwan_taf | number_of_physical_disks is in fact the size of a possible "physical_disks" struct | 13:58 |
| rameshg87 | erwan_taf, in example 2, operator will need to inspect each of 10 machine to create a configuration | 13:58 |
| erwan_taf | and 'interface_type, vendor etc... is just another way of selecting the required physical disk | 13:58 |
| erwan_taf | in one case we use the location as input, in the other case it's based on vendor/size/model information | 13:59 |
| rameshg87 | erwan_taf, yes | 13:59 |
| rameshg87 | erwan_taf, we could do it without vendor/size/model too if we want | 13:59 |
| erwan_taf | I mean, we could have a single way of writting this stuff instead of 2 | 14:00 |
| rameshg87 | erwan_taf, it will still do with just 2 inputs - disk_size_gb and raid_level | 14:00 |
| erwan_taf | this is just another way of describing the criteria of physical_disks | 14:00 |
| rameshg87 | erwan_taf, i just wanted to give 2 different examples of doing it | 14:01 |
| rameshg87 | erwan_taf, may be i still didn't get your question | 14:04 |
| erwan_taf | rameshg87: I mean, we can input almost the same information in different format | 14:08 |
| rameshg87 | erwan_taf, can you give me an example of yours ? | 14:08 |
| erwan_taf | 'physical_disks': [ ...] | 14:09 |
| erwan_taf | we do that, the amount of disks is implicit | 14:09 |
| erwan_taf | while when we want to give "hints" we have to provide The 'number_of_physical_disks' | 14:09 |
| rameshg87 | erwan_taf, yes | 14:09 |
| *** killer_prince is now known as lazy_prince | 14:10 | |
| *** MattMan-Away is now known as MattMan | 14:10 | |
| erwan_taf | If I use the "Backing physical disks" method, I cannot mix different vendors/model/fw | 14:10 |
| erwan_taf | +hints | 14:10 |
| rameshg87 | erwan_taf, i guess you don't need to | 14:11 |
| erwan_taf | why not ? | 14:11 |
| rameshg87 | erwan_taf, because you have already specified which disks you want, you no longer need to provide hints | 14:11 |
| erwan_taf | does physical_disks allow me to select a disk by vendor/size/model ? | 14:11 |
| erwan_taf | I only saw by location | 14:12 |
| rameshg87 | erwan_taf, ah it's not location actually | 14:12 |
| erwan_taf | that was my point | 14:12 |
| rameshg87 | erwan_taf, it's supposed to be a identifier for the driver to identify the exact disk | 14:12 |
| erwan_taf | so we could need differents hints to select different disks of the same array | 14:13 |
| rameshg87 | erwan_taf, in most cases it happens to be the location | 14:13 |
| erwan_taf | this is why we need an array of information | 14:13 |
| erwan_taf | the location doesn't guarantee anything... | 14:13 |
| rameshg87 | erwan_taf, the location identifies an exact disk, right ? | 14:13 |
| erwan_taf | just a location, not properties | 14:14 |
| rameshg87 | erwan_taf, for example in a hp box | 14:14 |
| rameshg87 | erwan_taf, if i mean controller="Smart Array P822 in Slot 2" and physical disk = "5I:1:1" in it | 14:14 |
| rameshg87 | erwan_taf, there is only one disk | 14:14 |
| rameshg87 | erwan_taf, and you want to use that disk on the machine | 14:15 |
| rameshg87 | erwan_taf, so you don't need to provide any more hints because you are asking the driver to use the specific disks | 14:15 |
| erwan_taf | In my mind, the location is the result of a detection of some disks properties | 14:15 |
| erwan_taf | "I want to create a raid L5 of X disks which have Y size and Z vendor and W model" | 14:16 |
| rameshg87 | http://paste.openstack.org/show/155849/ | 14:19 |
| rameshg87 | erwan_taf, you can have something like this as input for the above ^^^ | 14:19 |
| rameshg87 | erwan_taf, right ? | 14:19 |
| *** lucasagomes is now known as lucas-hungry | 14:20 | |
| erwan_taf | rameshg87: sorry, I meant, we want to create raid based on properties. The location isn't really important at my taste | 14:21 |
| rameshg87 | erwan_taf, yeah but there are people who want location too | 14:21 |
| erwan_taf | now, if I want to create a array mixing various vendors, without knowning the location, in the current state of the propoal, I can't | 14:21 |
| rameshg87 | erwan_taf, that's where the location-based input can help you | 14:22 |
| *** rloo has joined #openstack-ironic | 14:22 | |
| rameshg87 | erwan_taf, you could write your own script (say you want mix various vendors) to find physical disks of a machine and create a raid configuration | 14:22 |
| erwan_taf | so I meant that in fact we do want an array like physical_disks[] which provides the flexibility but where every disk can be defined by various methods (location, vendor, name, etc..) | 14:22 |
| rameshg87 | erwan_taf, we can have a large number of ways to find disks and each one just opens up use-cases for different user. i agree. | 14:23 |
| rameshg87 | erwan_taf, but for a starting we wanted to limit the number of ways to find the disk and make it simple for operator. | 14:24 |
| rameshg87 | erwan_taf, and do things on a incremental basis | 14:24 |
| rameshg87 | erwan_taf, rather than doing everything in the first shot | 14:24 |
| rameshg87 | erwan_taf, if we want to have more properties OR we want it to be generic, we can always do it in a second attempt | 14:24 |
| rameshg87 | erwan_taf, or a third attempt | 14:24 |
| rameshg87 | erwan_taf, but for a start this was my proposal | 14:25 |
| rameshg87 | erwan_taf, does that make sense ? | 14:25 |
| rameshg87 | erwan_taf, i need to go now . would be back in a hour or so after dinner. we can continue to discuss if we want | 14:28 |
| erwan_taf | I do understand the goal. My fear is getting something that only works for easy cases which will require to rework it later :/ | 14:28 |
| rameshg87 | erwan_taf, yeah but we could do it incrementally .. | 14:28 |
| rameshg87 | erwan_taf, start with an easy one, and get bigger step-by-step | 14:29 |
| *** jerryz has quit IRC | 14:30 | |
| openstackgerrit | Naohiro Tamura proposed openstack/ironic: Add iRMC Driver and its iRMC Power module https://review.openstack.org/144901 | 14:35 |
| *** rameshg87 is now known as rameshg87-dinner | 14:36 | |
| *** naohirot has quit IRC | 14:38 | |
| *** jerryz has joined #openstack-ironic | 14:42 | |
| *** rameshg87-dinner has quit IRC | 14:51 | |
| *** kbyrne has joined #openstack-ironic | 14:52 | |
| *** jjohnson2 has quit IRC | 14:55 | |
| *** jjohnson2_ has joined #openstack-ironic | 14:55 | |
| *** rlr0329 has joined #openstack-ironic | 14:56 | |
| *** jistr has quit IRC | 14:57 | |
| *** lucas-hungry is now known as lucasagomes | 14:57 | |
| *** jistr has joined #openstack-ironic | 14:59 | |
| NobodyCam | good morning | 15:06 |
| NobodyCam | :) | 15:06 |
| *** Marga_ has quit IRC | 15:10 | |
| Shrews | morning NobodyCam | 15:10 |
| NobodyCam | morning Shrews :) | 15:10 |
| dtantsur | NobodyCam, Shrews, morning | 15:10 |
| NobodyCam | morning dtantsur :) | 15:11 |
| Shrews | hey dtantsur | 15:12 |
| *** lazy_prince is now known as killer_prince | 15:14 | |
| *** Masahiro has joined #openstack-ironic | 15:16 | |
| openstackgerrit | Merged openstack/ironic: Add a fsm state -> dot diagram generator https://review.openstack.org/142887 | 15:16 |
| *** ryanpetrello_ has joined #openstack-ironic | 15:17 | |
| *** ryanpetrello has quit IRC | 15:19 | |
| *** ryanpetrello_ is now known as ryanpetrello | 15:19 | |
| *** Masahiro has quit IRC | 15:21 | |
| NobodyCam | rloo: are you about? | 15:23 |
| rloo | hi NobodyCam. Morning. | 15:24 |
| NobodyCam | morning :) I have a quick question: | 15:24 |
| openstackgerrit | Matthew Gilliard proposed openstack/ironic: Check glance for new kernel and ramdisk on prepare https://review.openstack.org/142782 | 15:25 |
| NobodyCam | re: the comment here: https://review.openstack.org/#/c/140868/6/ironic/conductor/task_manager.py | 15:25 |
| NobodyCam | you looking for wording changes in the comemnts? | 15:25 |
| rloo | yeah, wording change. | 15:25 |
| rloo | the 'Note that this does not alter the Node in any way' -- refers to if InvalidState is raised, right? | 15:26 |
| NobodyCam | :) ack thats my understanding | 15:26 |
| rloo | NobodyCam: I'm not even sure if that note is necessary | 15:27 |
| rloo | NobodyCam: I suppose it could say 'If InvalidState is raised, the Node won't be altered.'? | 15:28 |
| NobodyCam | thats what I was thinking too | 15:28 |
| NobodyCam | :) | 15:28 |
| dtantsur | rloo, morning | 15:29 |
| rloo | afternoon dtantsur | 15:29 |
| NobodyCam | brb | 15:33 |
| *** dkingshott has joined #openstack-ironic | 15:35 | |
| *** zz_jgrimm is now known as jgrimm | 15:36 | |
| * NobodyCam needs to run to starbuck as he is out of coffee.... bbiafm | 15:37 | |
| rloo | hey dtantsur, so DiscoverdInspect will start the ball rolling wrt invoking ironic-discoverd, and the node will be in 'inspect' state. during that time, someone could issue an ironic command to update the properties, as well as ironic-discoverd issuing an ironic command to update the properties? | 15:41 |
| dtantsur | rloo, yep | 15:42 |
| dtantsur | rloo, 3rdparty service is not different from a user | 15:42 |
| rloo | dtantsur: so ironic-discoverd's ironic commands might not get through if the node is locked. | 15:42 |
| dtantsur | rloo, exactly. that's why we need to release lock before calling to discoverd. | 15:43 |
| rloo | dtantsur: well, I mean, if someone issues some ironic command that locks the node. | 15:43 |
| dtantsur | rloo, well yeah. new state machine should prevent too much crazyness, for parallel update retries inside discoverd should be enough | 15:44 |
| *** subscope has quit IRC | 15:44 | |
| dtantsur | by crazyness I mean someone scheduling deploy :) | 15:44 |
| rloo | dtantsur: interesting. A deploy can't be scheduled cuz it is in inspect state, but we'll need to be careful about what is allowed in that state. | 15:45 |
| dtantsur | rloo, I guess it should be only update and power operations | 15:45 |
| rloo | dtantsur: I guess you'll find out :-) | 15:45 |
| dtantsur | :) | 15:45 |
| *** Marga_ has joined #openstack-ironic | 15:49 | |
| openstackgerrit | Steven Dake proposed openstack/ironic-specs: Override boot options via glance property https://review.openstack.org/144235 | 15:56 |
| lucasagomes | NobodyCam, rloo g'morning :D | 15:58 |
| rloo | hi lucasagomes | 15:58 |
| *** bradjones has quit IRC | 16:01 | |
| *** bradjones has joined #openstack-ironic | 16:01 | |
| NobodyCam | morning lucasagomes | 16:02 |
| *** andreykurilin has quit IRC | 16:02 | |
| rloo | NobodyCam: were you going to update 140868? | 16:08 |
| *** pcrews has joined #openstack-ironic | 16:09 | |
| openstackgerrit | Steven Dake proposed openstack/ironic: Add documentation to create in RegionOne https://review.openstack.org/139842 | 16:10 |
| NobodyCam | rloo: doing it now | 16:11 |
| NobodyCam | had to run to get coffee | 16:11 |
| rloo | NobodyCam: I had some comments in a test file too | 16:12 |
| NobodyCam | rloo: yep taken care of them now | 16:13 |
| rloo | thx NobodyCam | 16:13 |
| NobodyCam | :) | 16:14 |
| rlr0329 | I am trying to configure the agent_ipmitool driver. Some questions. | 16:15 |
| rlr0329 | 1. What node driver_info properties does the agent driver require? pxe_deploy_ramdisk and pxe_deploy_kernel? Are these images created with the diskimage-builder deploy-ironic element? | 16:16 |
| *** jistr has quit IRC | 16:18 | |
| *** jrist has quit IRC | 16:18 | |
| NobodyCam | rlr0329: there is a ironic-agent element too | 16:21 |
| NobodyCam | rlr0329: Bare Metal Provisioning | Status: http://bit.ly/ironic | 16:22 |
| rlr0329 | NobodyCam: I missed that :-( | 16:22 |
| NobodyCam | gah | 16:22 |
| NobodyCam | https://github.com/openstack/diskimage-builder/tree/master/elements/ironic-agent | 16:22 |
| *** jrist has joined #openstack-ironic | 16:24 | |
| NobodyCam | I would refer you to yuriyz who wrote that element | 16:25 |
| *** jgrimm is now known as zz_jgrimm | 16:25 | |
| rlr0329 | NobodyCam: oh yeah, I have that. Not much in the README. I have that ramdisk configured and can boot a server with it. Now I just need to integrated it with ironic and swift | 16:25 |
| NobodyCam | :) | 16:26 |
| *** jrist has quit IRC | 16:27 | |
| *** sdake has joined #openstack-ironic | 16:27 | |
| *** jrist has joined #openstack-ironic | 16:27 | |
| sdake | does anyone happen to know, if you boot an ironic node and the image has cloud init installed, can cloud init actually get at the metadata service? | 16:27 |
| rlr0329 | NobodyCam: I had downloaded the ironic-python-agent project but I had missed the diskimage-builder element. Looking at that | 16:27 |
| sdake | typically the metadata is stored in a config drive, but not sure about bare metal booting with ironic | 16:28 |
| mjturek | hey devanada, I'm looking at a bug you marked as confirmed https://bugs.launchpad.net/ironic/+bug/1310131 and had a question. What is the expected behavior? I mean, there is an error in the logs already though it is a bit vague. Let me know if you have any insight! | 16:28 |
| NobodyCam | sdake: yes cloud init can work, if meta data service is setup | 16:28 |
| openstackgerrit | Matthew Gilliard proposed openstack/ironic: Check glance for new kernel and ramdisk on prepare https://review.openstack.org/142782 | 16:29 |
| NobodyCam | j* are there any good first timer docs for creating the agent deploy ramdisk? | 16:29 |
| *** jerryz has quit IRC | 16:30 | |
| rlr0329 | NobodyCam: Looks like the ironic-agent element is for building your own ramdisk with Fedora. I can use the prebuilt CoreOS ramdisk for now | 16:31 |
| NobodyCam | rlr0329: yep :) | 16:31 |
| lucasagomes | sdake, for configdrive we have patches up for that too | 16:33 |
| lucasagomes | sdake, nova: https://review.openstack.org/#/c/145235/ (last patch in the series); Ironic: https://review.openstack.org/#/c/145235/ (for PXE drivers, agent already supports it) | 16:34 |
| *** jistr has joined #openstack-ironic | 16:34 | |
| *** Nisha has joined #openstack-ironic | 16:36 | |
| openstackgerrit | Dmitry Tantsur proposed stackforge/ironic-discoverd: Add 'default' argument to conf.get* functions https://review.openstack.org/145536 | 16:36 |
| *** krtaylor has quit IRC | 16:37 | |
| rlr0329 | I understand that the agent driver requires swift. I have swift running. Anything specific needs to be configured for IPA driver to use swift, i.e. ironic.conf? | 16:39 |
| sdake | lucasgomes lets say I want to use Juno, does something need to be configured for the metadata service since config drive sin't available? | 16:40 |
| *** dtantsur is now known as dtantsur|afk | 16:41 | |
| *** ChuckC has quit IRC | 16:42 | |
| lucasagomes | sdake, I don't recall any specific configuration for the metadata service | 16:43 |
| * lucasagomes looks a bit more | 16:44 | |
| lucasagomes | https://bugs.launchpad.net/tripleo/+bug/1239481 | 16:45 |
| sdake | lucasgomes thanks :) | 16:46 |
| NobodyCam | rlr0329: you have temp url support enabled in swift? | 16:46 |
| rlr0329 | NobodyCam: I think so | 16:47 |
| sdake | lucasgomes bummer, so its busted | 16:48 |
| sdake | seems like a major flaw :( | 16:48 |
| lucasagomes | indeed :/ | 16:48 |
| *** krtaylor has joined #openstack-ironic | 16:50 | |
| *** coolsvap is now known as coolsvap|afk | 16:51 | |
| *** coolsvap|afk is now known as coolsvap | 16:51 | |
| *** coolsvap is now known as coolsvap|afk | 16:52 | |
| sdake | interesting, need metadata server running on the host | 16:52 |
| sdake | rahter nuetorn-metadata service | 16:52 |
| *** Marga_ has quit IRC | 16:59 | |
| *** Marga_ has joined #openstack-ironic | 17:00 | |
| *** jcoufal has quit IRC | 17:05 | |
| *** Masahiro has joined #openstack-ironic | 17:05 | |
| openstackgerrit | Ruby Loo proposed openstack/ironic: Issue with FSM's _target_state https://review.openstack.org/142303 | 17:07 |
| rlr0329 | I'm still confused as to which ramdisk and kernel the agent driver uses. | 17:09 |
| rlr0329 | The pxe driver uses an image built with the deploy-ironic element to create the ramdisk that will expose the machine primary disk over iSCSI and reboot once Ironic signals it is finished. Is the agent ramdisk used instead of that deploy-ironic image? | 17:09 |
| JayF | dtantsur|afk: thanks for spelling out the DHCP thing to me in the spec, I feel a little thickheaded for not understanding that right away :) | 17:09 |
| JayF | rlr0329: no | 17:09 |
| JayF | https://github.com/openstack/ironic-python-agent is the agent/ramdisk for the IPA driver | 17:09 |
| JayF | https://github.com/openstack/ironic-python-agent/tree/master/imagebuild/coreos is where the ramdisk builder lives, and there are links to prebuilt agents in that README | 17:09 |
| *** Masahiro has quit IRC | 17:10 | |
| rlr0329 | JayF: I have downloaded the prebuilt CoreOS agent ramdisk and I can boot that | 17:11 |
| *** sbadia has quit IRC | 17:12 | |
| JayF | rlr0329: that's the proper ramdisk for the agent driver :) | 17:12 |
| JayF | rlr0329: you can also build your own (really simple and documented in that earlier link) to embed utilities, custom hardware managers (to do cooler things with cooler hardware), and other stuff | 17:12 |
| *** penick has joined #openstack-ironic | 17:14 | |
| rlr0329 | JayF: Yup. I have looked around that ramdisk. I am just confused about the next step and how that links to the actual image I want to load on the hardware, i.e. what image is used for the node properties so I can use the nova boot? | 17:14 |
| sdake | rlr0329 I explain how the agent deploys a node in my blog post here: http://sdake.io/2014/12/09/isnt-it-atomic-on-openstack-ironic-dont-you-think/ | 17:15 |
| sdake | rather how ironic deploys a node | 17:15 |
| sdake | there is a nice 12 step process :) | 17:15 |
| sdake | it probably needs to be actually documented, I think it was suggested last year I should do that, maybe I will :) | 17:15 |
| rlr0329 | sdake: Cool! Taking a look... | 17:15 |
| sdake | its actually not particularly obvious unless you sepnd alot of time working with ironic, like atleast 30 hours | 17:16 |
| JayF | That describes the flow with the pxe_* drivers | 17:17 |
| JayF | now with the agent_* drivers | 17:17 |
| *** Marga_ has quit IRC | 17:17 | |
| *** sbadia has joined #openstack-ironic | 17:17 | |
| *** Marga_ has joined #openstack-ironic | 17:17 | |
| sdake | true | 17:18 |
| JayF | rlr0329: the agent ramdisk is a transport to getting your own image on the machine. Just like with virt cloud, you have to supply an image to write to the disk to boot. In devstack, this is always cirros (which is a very minimal, only really useful for testing OS). Any full-disk image should work for the agent_* drivers, and partial disk images are what pxe_* uses today | 17:18 |
| sdake | I assume they are similar jayf | 17:18 |
| JayF | sdake: that's not a good assumption :) | 17:18 |
| sdake | :) | 17:18 |
| *** rloo has quit IRC | 17:18 | |
| JayF | sdake: in the agent, Ironic passes an http URL to IPA (Ironic python agent) which is downloaded *on the node* and written to the disk | 17:18 |
| JayF | sdake: this means you don't have a single bottleneck of network IO to a conductor ... but on the flip side, it also makes caching images more difficult | 17:19 |
| sdake | wow no iscsi | 17:19 |
| sdake | that rocks | 17:19 |
| JayF | sdake: (right now the "http url" sent to IPA is generally a Swift temporary URL) | 17:19 |
| JayF | sdake: exactly. | 17:19 |
| JayF | sdake: agent_* ironic driver + IPA is what we're using at Rackspace today for OnMetal | 17:19 |
| sdake | you have to install an agent in the host vm? | 17:20 |
| JayF | No; the agent runs in a ramdisk | 17:20 |
| sdake | nice! | 17:20 |
| JayF | basically with pxe_*, you have a "dumb" ramdisk which does nothing but expose iscsi targets | 17:20 |
| sdake | I'll have to try that out | 17:20 |
| JayF | with agent_*, we move more of the logic/work into the ramdisk (which includes a full python agent) in order to distribute the work | 17:20 |
| sdake | what builds the agent ramdisk, dib? | 17:21 |
| JayF | There's a set of DIB elements for IPA in the works; but right now it requires 3GB of ram to even boot :( | 17:21 |
| *** Marga_ has quit IRC | 17:21 | |
| JayF | The preferred builder for that (and the one we gate against) is here -> https://github.com/openstack/ironic-python-agent/tree/master/imagebuild/coreos | 17:21 |
| sdake | lot of ram | 17:21 |
| rlr0329 | JayF: What info , i.e. properties are used to create that URL so IPA knows which image to use? | 17:22 |
| *** Marga_ has joined #openstack-ironic | 17:22 | |
| JayF | rlr0329: it's sent in from nova, on the nova boot request | 17:22 |
| JayF | sdake: the builder I linked is based on CoreOS, and that one needs more like 1G of ram to work :) | 17:22 |
| *** krtaylor has quit IRC | 17:23 | |
| JayF | sdake: plus has SSH, and some other useful stuff baked in if you care to be able to observe the agent at work (of course, you have to get an SSH key into the image if you do that, which means either adding a cloud-config-url to your pxe boot or building a custom image) | 17:23 |
| *** Marga_ has quit IRC | 17:23 | |
| *** Marga_ has joined #openstack-ironic | 17:24 | |
| rlr0329 | JayF: IPA just uses the --image parameter to the nova boot command, i.e. no properties need to set on the node definitiion | 17:24 |
| rlr0329 | JayF: Is that correct? | 17:24 |
| *** Marga_ has quit IRC | 17:24 | |
| JayF | rlr0329: yes, but that image still has to be in glance backed by swift to work with the agent (so you can make a swift temp url for it) and there are some config settings around what users and such to talk to swift with to create the temp url | 17:24 |
| JayF | jroll: WTB docs :( | 17:25 |
| *** Marga_ has joined #openstack-ironic | 17:25 | |
| jroll | JayF: patches welcome :P | 17:25 |
| jroll | WTB time | 17:25 |
| JayF | time welcome :P | 17:26 |
| *** jistr has quit IRC | 17:26 | |
| *** penick has quit IRC | 17:26 | |
| rlr0329 | JayF: That makes sense. I have the images in glance backed by swift so sounds like I'm almost there.... Thanks for the clarification! | 17:26 |
| jroll | rlr0329: there are some config options to set in ironic wrt swift, as well as you need to set node.driver to agent_ipmitool or whatever power driver you're using | 17:27 |
| rlr0329 | JayF: I have the node.driver set. Do you know what the ironic config options for swift are that need to be set? | 17:28 |
| JayF | they should be in the sample config under the [agent] section | 17:29 |
| JayF | I only had to install Ironic once so I don't remember off the top of my head ;) | 17:29 |
| jroll | not [agent] section afaik | 17:29 |
| jroll | rlr0329: sec, I'll grab them for you | 17:29 |
| *** penick has joined #openstack-ironic | 17:30 | |
| lucasagomes | sdake, just reviewed 139842, it looks good but there's a rst syntax problem there | 17:30 |
| sdake | ok i'll fix | 17:30 |
| lucasagomes | if u want I can put a new patch set fixing it | 17:30 |
| lucasagomes | it's very very minimal | 17:30 |
| lucasagomes | ok | 17:30 |
| lucasagomes | thanks | 17:30 |
| rlr0329 | jroll: thanks! | 17:30 |
| sdake | lucasgomes interestingly enough it built with tox :) | 17:31 |
| sdake | I thought I tried it before review before | 17:31 |
| NobodyCam | woo hoo all hands call... :-p | 17:31 |
| JayF | A good trick is to do what lucasagomes did in the review, and click on the docs build when jenkins runs it on your change to go read the built doc and make sure it rendered right | 17:31 |
| JayF | IDK how to do that locally though :) | 17:31 |
| sdake | cool i'll try that | 17:32 |
| lucasagomes | JayF, you can do locally by opening... docs/build/html | 17:32 |
| sdake | wasn't quite sure what was wrong | 17:32 |
| lucasagomes | 1 sec | 17:32 |
| lucasagomes | sdake, JayF http://docs.openstack.org/developer/ironic/dev/dev-quickstart.html#building-developer-documentation | 17:32 |
| lucasagomes | firefox ironic/doc/build/html/index.html | 17:33 |
| jroll | rlr0329: you need to set up swift with a temp url key (swift post -m "Temp-URL-Key: $SWIFT_TEMPURL_KEY"), and then in glance section, set: swift_temp_url_key, swift_endpoint_url, swift_api_version, swift_account, swift_container, swift_temp_url_duration. then all that's left is agent section: heartbeat_timeout. also node.driver_info['deploy_kernel'] rather than 'pxe_deploy_kernel' | 17:33 |
| jroll | (same for ramdisk) | 17:33 |
| *** Marga_ has quit IRC | 17:35 | |
| *** chuckC_ has joined #openstack-ironic | 17:35 | |
| rlr0329 | jroll: I see those swift properties and that makes sense. So for the node.driver_info Those properties refer to the kernel and ramdisk of the full disk image? | 17:35 |
| *** Marga_ has joined #openstack-ironic | 17:35 | |
| jroll | rlr0329: no, of the agent ramdisk | 17:36 |
| jroll | a "full disk image" means the kernel/ramdisk are embedded in the image | 17:36 |
| openstackgerrit | Chris Krelle proposed openstack/ironic: Enable async callbacks from task.process_event() https://review.openstack.org/140868 | 17:36 |
| NobodyCam | rloo: around? | 17:37 |
| rlr0329 | jroll: so I need to load the agent ramdisk and kernel into glance/swift as well? | 17:37 |
| *** ifarkas has quit IRC | 17:37 | |
| jroll | rlr0329: yes, it's the same with the pxe driver | 17:37 |
| rlr0329 | jroll: ok. Thanks! | 17:38 |
| *** Marga_ has quit IRC | 17:39 | |
| *** Marga_ has joined #openstack-ironic | 17:40 | |
| openstackgerrit | Steven Dake proposed openstack/ironic: Add documentation to create in RegionOne https://review.openstack.org/139842 | 17:40 |
| *** Marga_ has quit IRC | 17:41 | |
| *** Marga_ has joined #openstack-ironic | 17:41 | |
| NobodyCam | anyone got a second to review my updates to https://review.openstack.org/140868 | 17:44 |
| *** pensu has joined #openstack-ironic | 17:44 | |
| *** EmilienM is now known as EmilienM|afk | 17:45 | |
| *** PaulCzar has quit IRC | 17:46 | |
| *** david-lyle_afk is now known as david-lyle | 17:47 | |
| *** krtaylor has joined #openstack-ironic | 17:47 | |
| devananda | morning, all | 17:53 |
| BadCub_ | morning Devananda! | 17:59 |
| NobodyCam | good morning devananda | 17:59 |
| *** athomas has quit IRC | 18:00 | |
| *** derekh has quit IRC | 18:02 | |
| devananda | looks like 140883 does indeed have some problems, some of which rloo pointed out in her follow on patch, 142303 | 18:10 |
| *** romcheg has quit IRC | 18:10 | |
| devananda | my rushed fixes last night (145389) passed tempest -- I'm going to sqash those back into 140838, which clearly needs them | 18:11 |
| NobodyCam | devananda: thank you... next on my list is 140869 | 18:11 |
| devananda | basically, rloo spotted the lack of this line: self._target_state = self._states[state]['target'] | 18:11 |
| devananda | adding that fixes the behavior, and then I had to fix a bunch of unit tests accordingly | 18:12 |
| devananda | also two of the state transitions are wrong | 18:12 |
| devananda | actually, what do you think of rebasing the whole stack on top of 145389 | 18:13 |
| *** bradjones has quit IRC | 18:13 | |
| devananda | since it's fixing states.py and fsm.py | 18:14 |
| devananda | (after I clean up the commit message) | 18:14 |
| NobodyCam | thats prob a good thing | 18:14 |
| devananda | oy .. in my haste last night, I munged the topic, too :( | 18:17 |
| devananda | I really sholuldn't be pushing code while running out of the office | 18:17 |
| *** pelix has quit IRC | 18:17 | |
| devananda | NobodyCam: any other patches you've updated in that series, aside from 140868? | 18:18 |
| *** rloo has joined #openstack-ironic | 18:18 | |
| *** bradjones has joined #openstack-ironic | 18:19 | |
| *** bradjones has quit IRC | 18:19 | |
| *** bradjones has joined #openstack-ironic | 18:19 | |
| NobodyCam | no I was just looking at 140869 while on the all hands call | 18:23 |
| devananda | ok - rebasing | 18:33 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Convert drivers to use process_event() https://review.openstack.org/140883 | 18:34 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Enable async callbacks from task.process_event() https://review.openstack.org/140868 | 18:34 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Convert check_deploy_timeout to use process_event https://review.openstack.org/140869 | 18:34 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Minor fixes to state model https://review.openstack.org/145389 | 18:34 |
| devananda | afk a while | 18:34 |
| NobodyCam | ack... thank you devananda :) | 18:34 |
| rloo | hi NobodyCam. I'm back. I guess there were some more changes to the fsm stuff? | 18:41 |
| NobodyCam | WB rloo :) | 18:41 |
| NobodyCam | yep | 18:42 |
| *** eghobo has joined #openstack-ironic | 18:42 | |
| rloo | NobodyCam: I'll try to get to them later today. | 18:42 |
| NobodyCam | devananda just rebased the set adding https://review.openstack.org/#/c/145389 | 18:42 |
| rloo | NobodyCam: thx | 18:43 |
| NobodyCam | :) | 18:44 |
| *** Nisha has quit IRC | 18:47 | |
| *** Marga_ has quit IRC | 18:49 | |
| *** Marga_ has joined #openstack-ironic | 18:50 | |
| *** Marga__ has joined #openstack-ironic | 18:53 | |
| *** Marga_ has quit IRC | 18:53 | |
| *** Masahiro has joined #openstack-ironic | 18:53 | |
| *** Masahiro has quit IRC | 18:58 | |
| *** subscope has joined #openstack-ironic | 19:05 | |
| *** ryanpetrello_ has joined #openstack-ironic | 19:16 | |
| NobodyCam | brb quick walkies | 19:17 |
| *** ryanpetrello has quit IRC | 19:18 | |
| *** ryanpetrello_ is now known as ryanpetrello | 19:18 | |
| *** GheRivero has quit IRC | 19:21 | |
| *** zz_jgrimm is now known as jgrimm | 19:22 | |
| russell_h | devananda: ok, space is confirmed and wiki is updated: https://wiki.openstack.org/wiki/Sprints/IronicKiloSprint | 19:25 |
| russell_h | devananda: we're going to add some info about accomodations this afternoon | 19:26 |
| devananda | russell_h: cheers | 19:29 |
| devananda | rloo: I started going through your patch, looking at what issues you found. got through 2 of them, which turned into https://review.openstack.org/145389 | 19:30 |
| devananda | rloo: and then I rebased the remaining patches on top of that fix | 19:30 |
| devananda | rloo: I haven't looked through all the new unit tests you added yet, though | 19:31 |
| rloo | devananda: ah, thx for explaining. I hope to get to it in an hour or so. (and I'll update/remove stuff from my patch after that...) | 19:31 |
| devananda | rloo: cool. I'll probably be in a meeting the rest of today. will try to stay on IRC, but we'll see | 19:33 |
| jroll | russell_h: Wednesday, February 13th - Friday, February 15th, 2015. (3 days) is actually friday-sunday | 19:37 |
| *** dank has joined #openstack-ironic | 19:37 | |
| NobodyCam | its 11th thu 13th no? | 19:38 |
| *** dkingshott has quit IRC | 19:41 | |
| Shrews | the state machine spec mentions an intermediate DEPLOYED state, but this isn't a valid state in the current state machine code. i wonder if the spec needs updated.... | 19:42 |
| Shrews | or else i'm missing something | 19:42 |
| *** GheRivero has joined #openstack-ironic | 19:43 | |
| jroll | NobodyCam: I would assume so, but the wiki says 13-15 | 19:44 |
| *** dlaube has joined #openstack-ironic | 19:44 | |
| openstackgerrit | Victor Howard proposed openstack/ironic: display error logging should be improved https://review.openstack.org/144771 | 19:45 |
| NobodyCam | Shrews: the spec is for NEW state machine deva's patches are switch the code to use a FSM and current states may (will) not match the new states | 19:45 |
| NobodyCam | Shrews: or am I missing something? | 19:46 |
| Shrews | NobodyCam: oh, no. you're right. i'm jumping ahead | 19:46 |
| NobodyCam | :) | 19:46 |
| Shrews | just finished battling a dead laptop and rebuilding an older one, so my mind was off of s/w and on h/w for too long, obviously | 19:47 |
| NobodyCam | :) | 19:47 |
| devananda | Shrews: fwiw, I made a similar mistake when implementing the /current/ state machine. 145389 corrects that | 19:47 |
| Shrews | :) | 19:48 |
| devananda | we used DELETED and DEPLOYDONE as a return value from certain functions, but these were never saved to node.provision_state | 19:48 |
| NobodyCam | oh devananda while your here | 19:49 |
| *** andreykurilin has joined #openstack-ironic | 19:50 | |
| NobodyCam | on 145389 states.py you just commented out (as opposed to deleting) the unused states (is this because we will be using them?) | 19:50 |
| *** GheRivero has quit IRC | 19:51 | |
| lucasagomes | I will call it a day, have a good night everyone! | 19:52 |
| NobodyCam | have a good night lucasagomes :) | 19:52 |
| *** lucasagomes is now known as lucas-dinner | 19:52 | |
| lucas-dinner | cheers :D | 19:52 |
| *** lucas-dinner has quit IRC | 19:52 | |
| Shrews | NobodyCam: i would assume so, b/c that's what sparked my question :) | 19:54 |
| devananda | NobodyCam: those match the spec, but not the current state ... so yea. they're there as reference. we could remove them, but I"ll be adding them back later | 19:54 |
| devananda | ah, so, this is incorrect (entirely my fault) | 19:59 |
| devananda | http://specs.openstack.org/openstack/ironic-specs/specs/kilo/new-ironic-state-machine.html#proposed-change | 19:59 |
| *** GheRivero has joined #openstack-ironic | 20:03 | |
| *** ryanpetrello_ has joined #openstack-ironic | 20:08 | |
| *** ryanpetrello has quit IRC | 20:08 | |
| *** ryanpetrello_ is now known as ryanpetrello | 20:08 | |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic-specs: Correct the diagram for the "current" state machine https://review.openstack.org/145569 | 20:16 |
| devananda | Shrews: that ^ should now match what I've implemented -- and more importantly, the actual states in the Juno release | 20:16 |
| devananda | it is clearly not following our guidelines for the new states ... | 20:17 |
| Shrews | devananda: cool | 20:18 |
| devananda | oh, there's a difference | 20:19 |
| devananda | arrgh | 20:20 |
| devananda | this isn't correct: machine.add_state(DEPLOYING, target=ACTIVE, | 20:22 |
| devananda | it works - but it changes API-visible behavior | 20:23 |
| *** EmilienM|afk is now known as EmilienM | 20:25 | |
| Shrews | devananda: should that stay the way it current is then? with target=DEPLOYDONE? | 20:26 |
| Shrews | currently | 20:27 |
| NobodyCam | brb | 20:27 |
| devananda | maybe, maybe not | 20:28 |
| devananda | preserving the current (illogical) behavior with the fsm.py module is going to require a couple more work arounds, I think | 20:28 |
| devananda | we could do that | 20:28 |
| devananda | but we want to change the state progression *anyway* | 20:28 |
| devananda | is there a need to add some hacks just to remove them? | 20:28 |
| *** pensu has quit IRC | 20:29 | |
| devananda | or jsut document: hey, this patch changes teh values of target_provision_state in this way | 20:29 |
| devananda | and then document the next API-visible change in the same way, and so on | 20:29 |
| devananda | since we're goign to be making several | 20:29 |
| devananda | jroll, JayF - I'd like your input on this too | 20:30 |
| jroll | devananda: on what, specifically? | 20:31 |
| Shrews | devananda: So, is the problem that the current fsm code doesn't really handle temporary transitory states (like DEPLOYDONE) well (thus the need for the hack)? So we'd either need to change it to automatically move from those to the true final state, or just model the real transition to the final state by jumping ahead a bit. | 20:31 |
| Shrews | just trying to make sure i understand the issue | 20:32 |
| devananda | Shrews: see the change to the spec I posted above | 20:32 |
| devananda | eg, right now, teh API shows "DEPLOYING/DEPLOYDONE" and then immediatly "ACTIVE/NONE" | 20:33 |
| Shrews | right | 20:33 |
| devananda | whereas fsm.py doesn't present a mechanism for skipping the target state | 20:33 |
| devananda | if current never == target, it's not "done" | 20:33 |
| devananda | but in the current code, current state is never "deploydone" | 20:33 |
| Shrews | right. it's a temporary state that is never assigned currently. we just go straight to active | 20:34 |
| Shrews | (i think) | 20:35 |
| devananda | right | 20:36 |
| Shrews | ok. i understood correctly, just bad with word things | 20:36 |
| devananda | fsm.py clears the target_state only when current_state == target_state | 20:37 |
| devananda | so if the target_state is DEPLOYDONE, changing teh current_state from DEPLOYING to ACTIVE will not clear it | 20:37 |
| NobodyCam | yep: https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/pxe.py#L481 | 20:38 |
| devananda | i can work around that -- i'm just asking whether we should work around it and then change the states, or whether we just change the states now. | 20:38 |
| *** GheRivero has quit IRC | 20:38 | |
| Shrews | I'm all for changing the states for that transition now, rather than put in a workaround to be removed later. But others may disagree. | 20:39 |
| NobodyCam | devananda: are we 100% confident that the new sates will be fully in place by end of cycke | 20:39 |
| NobodyCam | *cycle | 20:39 |
| devananda | NobodyCam: of course not. code's not written for that yet | 20:39 |
| devananda | but we srsly need it to be | 20:39 |
| devananda | Shrews: cool | 20:40 |
| NobodyCam | then I then we may need to do the workaround to ensure current behaviour continues as is | 20:40 |
| NobodyCam | then I then = then I think | 20:40 |
| devananda | NobodyCam: to note, this doesn't affect Nova, which only keys off of current state | 20:40 |
| devananda | it only changes the representation of target state | 20:41 |
| devananda | but I think rackspace *is* using that, which is why I pinged jroll | 20:41 |
| *** subscope has quit IRC | 20:41 | |
| jroll | devananda: ok, so that specifically. I've fixed that bug in our internal tooling :) | 20:41 |
| jroll | "bug" | 20:41 |
| *** Masahiro has joined #openstack-ironic | 20:42 | |
| devananda | jroll: so, changing what target_state the API shows won't affect ya'll? | 20:42 |
| jroll | devananda: correct | 20:42 |
| devananda | neat | 20:43 |
| devananda | afaik, no one else is keying off that either | 20:43 |
| NobodyCam | :) | 20:43 |
| jroll | devananda: I'd still post a warning to the list, but other than that I say jfdi | 20:43 |
| Shrews | Just Feed Dogs Icecream? :) | 20:44 |
| NobodyCam | lol | 20:44 |
| rloo | will the changed target state be changed again with the new fsm? | 20:45 |
| jroll | :P | 20:45 |
| jroll | rloo: I believe so, yes | 20:46 |
| jroll | well, certainly, there's new states | 20:46 |
| rloo | if we're talking about changing deploy->deploydone to deploy->active now, the new fsm has deploy->active. so wow, we'll be ahead of the game. | 20:46 |
| jroll | ah, yeah | 20:46 |
| devananda | :) | 20:47 |
| *** Masahiro has quit IRC | 20:47 | |
| Shrews | or will the new one go from deploy -> deployed, then deployed -> active? | 20:48 |
| devananda | Shrews: target state will be active (not deploydone) | 20:49 |
| rloo | Shrews: in the new world, I think it is deploying -> deployed -> active. But target is active according to the diagram. | 20:49 |
| Shrews | i guess i'm misreading the new state machine | 20:50 |
| Shrews | "If the steps taken during the active (-ING) state succeed, Ironic will automatically transition to the momentary (-ED) state and then to the next indicated state on the graph." | 20:50 |
| rloo | but I have wondered how the code will know to go from deployed->active. | 20:50 |
| devananda | rloo: when current == target | 20:51 |
| devananda | it will trigger the advancement to the next state | 20:51 |
| devananda | and clear the target | 20:51 |
| rloo | devananda: so start: current=deploying, target = active. | 20:52 |
| rloo | next, deploying is done, current = deployed, target = active ?? | 20:52 |
| rloo | then i think I need to review the code... | 20:52 |
| devananda | rloo: the Juno code does not ever expose DEPLOYDONE as a current state | 20:56 |
| *** GheRivero has joined #openstack-ironic | 20:56 | |
| devananda | the new state machine will expose current=DEPLOYED | 20:56 |
| rloo | devananda: yeah, I'm fine with that. Just wondering with the new fsm, how we get from *ED -> passive/target state. but maybe i need to review the fsm code first. | 20:56 |
| devananda | rloo: this may help -- https://review.openstack.org/#/c/145569/ | 20:57 |
| devananda | going through this caused me to realize the spec had a mistake in it | 20:57 |
| rloo | you just want me to review that patch now ;) | 20:58 |
| openstackgerrit | Victor Howard proposed openstack/ironic: display error logging should be improved https://review.openstack.org/144771 | 20:58 |
| NobodyCam | brb ... /me needs food | 20:58 |
| * devananda whistles innocently | 20:58 | |
| *** GheRivero has quit IRC | 21:00 | |
| *** GheRivero has joined #openstack-ironic | 21:00 | |
| *** openstackgerrit has quit IRC | 21:05 | |
| *** GheRivero has quit IRC | 21:05 | |
| *** openstackgerrit has joined #openstack-ironic | 21:06 | |
| NobodyCam | humm I keep wanting to change the title line of the FSM spec. basicly remove "A proposal for the " | 21:13 |
| mrda | Morning Ironic | 21:15 |
| rloo | NobodyCam: +1 | 21:15 |
| rloo | morning mrda | 21:15 |
| NobodyCam | morning mrda | 21:15 |
| NobodyCam | :) | 21:15 |
| mrda | o/ everyone | 21:21 |
| *** Hefeweizen has quit IRC | 21:21 | |
| *** GheRivero has joined #openstack-ironic | 21:22 | |
| *** GheRivero has quit IRC | 21:33 | |
| *** GheRivero has joined #openstack-ironic | 21:34 | |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic-specs: Correct the diagram for the "current" state machine https://review.openstack.org/145569 | 21:37 |
| *** bradjones has quit IRC | 21:40 | |
| *** bradjones has joined #openstack-ironic | 21:43 | |
| *** bradjones has quit IRC | 21:43 | |
| *** bradjones has joined #openstack-ironic | 21:43 | |
| *** dprince has quit IRC | 21:47 | |
| JayF | rloo: thanks for the review :D I'm working on your comments now | 21:48 |
| rloo | JayF: thx | 21:48 |
| NobodyCam | omg a new style of RV... I want one.... https://www.facebook.com/video.php?v=757792197632896 | 21:51 |
| jroll | wow | 21:52 |
| NobodyCam | :) | 21:52 |
| *** GheRivero has quit IRC | 21:56 | |
| *** chlong has joined #openstack-ironic | 21:56 | |
| *** ijw has joined #openstack-ironic | 21:58 | |
| mrda | wow | 21:59 |
| mrda | do you need a second truck to carry the furniture to fill one of these? | 22:00 |
| NobodyCam | i'm not sure | 22:00 |
| NobodyCam | I would hope it would fit inthe same truck. | 22:00 |
| NobodyCam | lol just thought it was kinda kewl | 22:01 |
| *** krtaylor has quit IRC | 22:01 | |
| mrda | I guess it's for more than a single night in a location :) | 22:01 |
| *** GheRivero has joined #openstack-ironic | 22:02 | |
| *** chlong has quit IRC | 22:03 | |
| *** jjohnson2_ has quit IRC | 22:04 | |
| *** ChuckC has joined #openstack-ironic | 22:06 | |
| NobodyCam | ya.. | 22:11 |
| *** GheRivero has quit IRC | 22:11 | |
| *** ijw has quit IRC | 22:12 | |
| *** ijw has joined #openstack-ironic | 22:12 | |
| *** GheRivero has joined #openstack-ironic | 22:17 | |
| NobodyCam | brb | 22:18 |
| jroll | devananda: going to review state machine series, I should skip the first one yes? | 22:22 |
| devananda | jroll: don't skip. I need to fix the commit message | 22:25 |
| devananda | butthe change is important | 22:25 |
| jroll | devananda: oh, it's marked wip | 22:25 |
| jroll | https://review.openstack.org/#/c/145389/ | 22:25 |
| jroll | to be clear | 22:25 |
| devananda | yah. I'm distracted right now | 22:25 |
| jroll | ok | 22:25 |
| jroll | I guess this is the one that chages target state behavior? | 22:25 |
| devananda | right | 22:26 |
| devananda | and teh commit message doesn't say why | 22:26 |
| devananda | or rather, says it poorly | 22:26 |
| jroll | k | 22:26 |
| devananda | feel free to update the mssage if you want to | 22:26 |
| * devananda returns to distraction | 22:26 | |
| *** Masahiro has joined #openstack-ironic | 22:31 | |
| *** GheRivero has quit IRC | 22:32 | |
| russell_h | devananda: wow, thanks for catching that date thing | 22:34 |
| russell_h | might have had people showing up some weird days | 22:34 |
| *** Masahiro has quit IRC | 22:36 | |
| *** GheRivero has joined #openstack-ironic | 22:38 | |
| *** ijw has quit IRC | 22:42 | |
| *** chuckC_ has quit IRC | 22:48 | |
| openstackgerrit | Jay Faulkner proposed openstack/ironic-python-agent: Allow use of multiple simultaneous HW managers https://review.openstack.org/143193 | 22:51 |
| JayF | rloo: jroll ^ that addresses your comments | 22:52 |
| JayF | thanks for the snippet to make that much cleaner jroll | 22:52 |
| jroll | np | 22:52 |
| *** GheRivero has quit IRC | 22:57 | |
| JayF | RFR -> https://review.openstack.org/#/c/145621/ PR to remove Python 2.6 from IPA jobs. | 23:01 |
| JayF | We have no ramdisks that use python 2.6, and there's no need in keeping those tests running. Once that merges, I'll PR removing that environment from tox.ini. | 23:02 |
| JoshNang | JayF: yess | 23:02 |
| JoshNang | JayF: we should try and add py33 again sometime | 23:02 |
| JayF | All the oslo modules we depend on don't work with py3 | 23:02 |
| JayF | so we can't work with py3 | 23:02 |
| JoshNang | :( | 23:02 |
| JayF | IIRC LoopingCall used eventlet in a way that broke our py3 tests | 23:03 |
| JayF | VERY early on, so we had to pull that, and pypy for a similar reason | 23:03 |
| jroll | eventlet may be fixed by now | 23:03 |
| jroll | s/may be/is likely | 23:04 |
| jroll | it would be worth running just to see :) | 23:04 |
| JayF | rloo: for errors.py #271, self.message is used by RESTError to set errors in the response | 23:04 |
| openstackgerrit | Alex Weeks proposed openstack/ironic-python-agent: Added extension to support rescue mode https://review.openstack.org/145385 | 23:05 |
| JayF | rloo: I was trying to keep the short-error (message) from having an erroneous {0} in it | 23:05 |
| JayF | because those are supposed to be static | 23:05 |
| JayF | whereas details is allowed to change based on the circumstances of the error | 23:05 |
| JayF | iirc | 23:05 |
| rloo | JayF: ugh. ok. | 23:05 |
| JayF | +1 on the ugh | 23:05 |
| rloo | JayF: you sure? I'll trust you on that. you sure? :-) | 23:05 |
| JayF | I don't love it either, but it seemed silly not to include the method that failed in the log back to Ironic | 23:05 |
| JayF | rloo: please don't trust me ;) https://github.com/openstack/ironic-python-agent/blob/master/ironic_python_agent/errors.py#L18 | 23:06 |
| *** ijw has joined #openstack-ironic | 23:06 | |
| NobodyCam | JayF: on https://review.openstack.org/#/c/143193/16/ironic_python_agent/errors.py class IncompatibleHardwareMethodError is why not provide the method and node info too? | 23:08 |
| JayF | NobodyCam: I don't think we have any node info to send across in that case; we don't pass a big node object around in IPA like Ironic does | 23:09 |
| jroll | NobodyCam: in IPA, there is only one node :P | 23:09 |
| jroll | ironic logs the node uuid when IPA returns an error | 23:09 |
| JayF | NobodyCam: in the method, "message" aiui should be static, for all instances of that exception, but details can change | 23:09 |
| JayF | that's why I put the method in details, and not in error | 23:09 |
| NobodyCam | ahh | 23:10 |
| NobodyCam | TY | 23:10 |
| rloo | JayF: gotta run. Will remove the -1 later | 23:10 |
| openstackgerrit | Jay Faulkner proposed openstack/ironic-python-agent: Allow use of multiple simultaneous HW managers https://review.openstack.org/143193 | 23:11 |
| *** rloo is now known as rloo_afk | 23:11 | |
| JayF | rloo: ^ I just did :) | 23:11 |
| JayF | lol | 23:11 |
| NobodyCam | hehehe | 23:11 |
| *** krtaylor has joined #openstack-ironic | 23:12 | |
| openstackgerrit | Josh Gachnang proposed openstack/ironic: Improve IPA client library https://review.openstack.org/111118 | 23:13 |
| *** Marga__ has quit IRC | 23:16 | |
| *** Marga_ has joined #openstack-ironic | 23:16 | |
| NobodyCam | lol love the var: homeless_agent | 23:22 |
| *** chlong has joined #openstack-ironic | 23:22 | |
| *** Hefeweizen has joined #openstack-ironic | 23:23 | |
| JayF | That was from ellenh iirc, when she implemented that super helpful ip autodetect features | 23:23 |
| JayF | NobodyCam: revisiting the comments about the exceptions / RESTError in IPA, it's really important to remember with IPA that most folks using it will never see IPA logs directly, so whatever results are important to send up have to be in the details of the RESTError | 23:26 |
| *** jgrimm is now known as zz_jgrimm | 23:26 | |
| openstackgerrit | Jay Faulkner proposed openstack/ironic-python-agent: Remove python 2.6 from tox.ini https://review.openstack.org/145631 | 23:30 |
| *** ChuckC has quit IRC | 23:37 | |
| NobodyCam | JayF: in tests/multi_hardware.py .. def test_both_succeed i'm missing something it looks like the test is that the first is called and the second is not. | 23:37 |
| JayF | NobodyCam: that's exactly how it's supposed to work | 23:38 |
| *** ChuckC has joined #openstack-ironic | 23:38 | |
| NobodyCam | thats ensuring both succceed? | 23:38 |
| JayF | NobodyCam: the specific hardware manager had the method, and it worked, so the generic was never called | 23:38 |
| JayF | NobodyCam: it's showing a case where both *could* succeed | 23:38 |
| JayF | NobodyCam: in fact, that's specifically what I'm asserting in the test: that MainlineHardareManager.both_succeed got one call, and Generic got none | 23:40 |
| *** ChuckC has quit IRC | 23:40 | |
| *** naohirot has joined #openstack-ironic | 23:41 | |
| naohirot | good morning ironic | 23:41 |
| *** eghobo has quit IRC | 23:42 | |
| NobodyCam | morning naohirot | 23:42 |
| JayF | morning | 23:42 |
| naohirot | NobodyCam: JayF: good evening | 23:42 |
| NobodyCam | JayF: ok... i think its just the name that caused my brain to loop on it | 23:42 |
| JayF | NobodyCam: yeah, naming those test methods was mildly difficult. I think that's the third name that one has had | 23:43 |
| JayF | lol | 23:43 |
| NobodyCam | lol | 23:43 |
| NobodyCam | test_mainline_succeed_generic_not_called | 23:44 |
| NobodyCam | lol | 23:44 |
| NobodyCam | ya don't listen to me. | 23:44 |
| *** Marga_ has quit IRC | 23:44 | |
| NobodyCam | thou I'd prob ask for a small comment explaining that | 23:44 |
| NobodyCam | lol | 23:45 |
| *** Marga_ has joined #openstack-ironic | 23:45 | |
| *** Marga_ has quit IRC | 23:46 | |
| *** Marga_ has joined #openstack-ironic | 23:47 | |
| *** alex_xu has quit IRC | 23:48 | |
| *** Marga_ has quit IRC | 23:48 | |
| *** Marga_ has joined #openstack-ironic | 23:48 | |
| *** Marga_ has quit IRC | 23:49 | |
| naohirot | rloo_afk: If you back, I'd like to hear your opinion regarding python-scciclient. | 23:49 |
| *** Marga_ has joined #openstack-ironic | 23:49 | |
| *** Marga_ has quit IRC | 23:49 | |
| *** Marga_ has joined #openstack-ironic | 23:50 | |
| *** andreykurilin has quit IRC | 23:50 | |
| *** alex_xu has joined #openstack-ironic | 23:50 | |
| naohirot | NobodyCam: can I hear your opinion too regarding python-scciclient, gathering all opinion earlier is better to proceed. | 23:52 |
| naohirot | NobodyCam: please look at https://review.openstack.org/#/c/134865/ and https://review.openstack.org/#/c/144901/ | 23:53 |
| naohirot | NobodyCam: Yesterday I had conversation with rameshg87 regarding moving scci module into external package. | 23:55 |
| NobodyCam | naohirot: will take a look | 23:55 |
| NobodyCam | JayF: reviewed | 23:56 |
| naohirot | NobodyCam: thanks | 23:56 |
| *** penick has quit IRC | 23:56 | |
| naohirot | JayF: I' sorry if I interrupt you | 23:57 |
| mrda | hi naohirot | 23:57 |
| naohirot | mrda: hi good morning | 23:57 |
| mrda | naohirot: Just FYI, we've had 2cm of rain overnight, fires under control \o/ | 23:59 |
| naohirot | mrda: that's really good | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!