| *** ijw has quit IRC | 00:00 | |
| devananda | NobodyCam: was it paris? o.0 | 00:01 |
|---|---|---|
| NobodyCam | I think so :-p | 00:01 |
| *** ijw has joined #openstack-ironic | 00:01 | |
| NobodyCam | shutting down... time to make the drive back to the RV :) | 00:02 |
| NobodyCam | see everyone tomorrow | 00:02 |
| *** Sukhdev has quit IRC | 00:03 | |
| JayF | o/ | 00:03 |
| JayF | enjoy gilroy | 00:04 |
| JayF | get some garlic fries | 00:04 |
| *** garthb has quit IRC | 00:10 | |
| *** anteaya has quit IRC | 00:10 | |
| *** rama_y has quit IRC | 00:13 | |
| *** milan has joined #openstack-ironic | 00:19 | |
| *** yuanying has quit IRC | 00:30 | |
| *** david-lyle has quit IRC | 00:31 | |
| *** yibo has joined #openstack-ironic | 00:41 | |
| *** jcoufal has joined #openstack-ironic | 00:42 | |
| *** hoangcx has joined #openstack-ironic | 00:45 | |
| *** ijw has quit IRC | 00:51 | |
| *** ijw has joined #openstack-ironic | 00:53 | |
| *** tiendc has joined #openstack-ironic | 00:53 | |
| tiendc | good morning, ironic | 00:55 |
| *** joprovost has joined #openstack-ironic | 00:59 | |
| rloo | harlowja: thx! :) | 01:01 |
| harlowja | xnp | 01:01 |
| harlowja | *np | 01:01 |
| *** anteaya has joined #openstack-ironic | 01:02 | |
| *** phuongnh has joined #openstack-ironic | 01:06 | |
| openstackgerrit | Dao Cong Tien proposed openstack/ironic: Nova-compatible serial console: socat console_utils https://review.openstack.org/328168 | 01:09 |
| *** pece has quit IRC | 01:09 | |
| *** piet has quit IRC | 01:11 | |
| openstackgerrit | Dao Cong Tien proposed openstack/ironic: IPMITool: add IPMISocatConsole class https://review.openstack.org/293873 | 01:11 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Follow-up fixes to 206244 https://review.openstack.org/341892 | 01:13 |
| devananda | rloo: incorporated most of your comments there ^ (all the ones I agreed with :) ) | 01:14 |
| rloo | devananda: thx! (I think ;)) | 01:14 |
| *** piet has joined #openstack-ironic | 01:16 | |
| *** ijw has quit IRC | 01:20 | |
| *** ijw has joined #openstack-ironic | 01:21 | |
| *** piet has quit IRC | 01:28 | |
| *** caoshufeng has joined #openstack-ironic | 01:35 | |
| *** yuanying has joined #openstack-ironic | 01:36 | |
| *** aNupoisc has quit IRC | 01:46 | |
| tiendc | rloo: vdrok: lucas|afk: hi, my patch just got +2 from Jim, need another +2 and a wf +1 https://review.openstack.org/#/c/328168/ | 01:50 |
| tiendc | rloo: vdrok: lucas|afk: could you help to review it? thanks | 01:51 |
| jroll | tiendc: why didn't you fix the _LE() thing when you rebased? :( | 01:51 |
| tiendc | I answered to your comment that I think _() is more reasonable. | 01:52 |
| tiendc | jroll: please see my comment inline | 01:52 |
| lazy_prince | is there a way to configure iPXE enabled per node..? if not, will such thing be encouraged by community.. | 01:52 |
| jroll | tiendc: oops, you're right. okay. I missed the raise there | 01:54 |
| jroll | lazy_prince: there is not, what's the use case? | 01:54 |
| *** ChrisAusten has joined #openstack-ironic | 01:54 | |
| *** watanabe_isao has quit IRC | 01:54 | |
| *** watanabe_isao has joined #openstack-ironic | 01:55 | |
| lazy_prince | lets say I have ironic with ipxe enabled. but for some reason a handful of servers do not work with iPXE. so I would like those node to use non-ipxe boot.. | 01:55 |
| tiendc | jroll: thanks for giving it +2 | 01:55 |
| jroll | tiendc: no problem | 01:55 |
| *** rloo has quit IRC | 01:56 | |
| lazy_prince | jroll: and since iPXE right now is a global config option, this is not possible unless, we make it as per node config | 01:57 |
| jroll | lazy_prince: right, I understand that, why doesn't ipxe work? | 01:57 |
| lazy_prince | jroll: I am not sure there.. may be buggy fw or buggy ipxe.. | 01:58 |
| jroll | lazy_prince: I'm sure the community isn't opposed to it, we love per-node configs :/ | 01:59 |
| lazy_prince | to be frank, ipxe with uefi seems not to work with HP ProLiant G9 Blades.. | 02:00 |
| jroll | that's really nufortunate, I'm curious what the problem is | 02:01 |
| lazy_prince | jroll: do we need a spec for this or just patch should do.. | 02:01 |
| jroll | lazy_prince: at least an RFE, maybe a small spec | 02:01 |
| lazy_prince | okay.. I will push them in a day or two... | 02:02 |
| jroll | great, thank you | 02:02 |
| * jroll signs off for the night | 02:02 | |
| *** yuanying has quit IRC | 02:03 | |
| *** yuanying has joined #openstack-ironic | 02:03 | |
| *** yuanying has quit IRC | 02:04 | |
| *** ijw has quit IRC | 02:05 | |
| *** yuanying has joined #openstack-ironic | 02:08 | |
| *** yuanying has quit IRC | 02:12 | |
| *** amotoki has joined #openstack-ironic | 02:13 | |
| *** yuanying has joined #openstack-ironic | 02:14 | |
| *** yuanying has quit IRC | 02:26 | |
| *** bvandewa_ has joined #openstack-ironic | 02:34 | |
| *** bvandewa has quit IRC | 02:37 | |
| *** bvandewa_ has quit IRC | 02:38 | |
| *** hoangcx has quit IRC | 02:39 | |
| *** adu has joined #openstack-ironic | 02:48 | |
| openstackgerrit | Yuiko Takada Mori proposed openstack/ironic: IPMITool: add IPMISocatConsole class https://review.openstack.org/293873 | 02:50 |
| *** sdake has joined #openstack-ironic | 02:59 | |
| *** joprovost1 has joined #openstack-ironic | 03:02 | |
| *** joprovost has quit IRC | 03:02 | |
| *** joprovost1 is now known as joprovost | 03:02 | |
| *** yuanying has joined #openstack-ironic | 03:13 | |
| *** mordred has quit IRC | 03:14 | |
| *** rajinir has quit IRC | 03:15 | |
| *** yuanying has quit IRC | 03:18 | |
| *** yuanying has joined #openstack-ironic | 03:18 | |
| *** yuanying has quit IRC | 03:23 | |
| *** hoangcx has joined #openstack-ironic | 03:28 | |
| *** jaybeale has joined #openstack-ironic | 03:34 | |
| *** jaybeale has quit IRC | 03:36 | |
| *** ijw has joined #openstack-ironic | 03:41 | |
| *** joprovost1 has joined #openstack-ironic | 03:43 | |
| *** joprovost has quit IRC | 03:44 | |
| *** joprovost1 is now known as joprovost | 03:44 | |
| *** joprovost has quit IRC | 03:46 | |
| *** joprovost has joined #openstack-ironic | 03:46 | |
| *** watanabe_isao has quit IRC | 03:54 | |
| *** joprovost has quit IRC | 03:54 | |
| openstackgerrit | Merged openstack/ironic: Follow-up fixes to 206244 https://review.openstack.org/341892 | 03:58 |
| *** yuanying has joined #openstack-ironic | 03:59 | |
| *** phuongnh has quit IRC | 04:01 | |
| *** phuongnh has joined #openstack-ironic | 04:01 | |
| *** ijw has quit IRC | 04:03 | |
| *** adu has quit IRC | 04:05 | |
| *** sacharya has joined #openstack-ironic | 04:08 | |
| *** sacharya has quit IRC | 04:08 | |
| *** harlowja has quit IRC | 04:13 | |
| *** vmud213 has joined #openstack-ironic | 04:19 | |
| *** ChrisAusten has quit IRC | 04:24 | |
| *** links has joined #openstack-ironic | 04:27 | |
| *** jcoufal has quit IRC | 04:37 | |
| *** bharathk has joined #openstack-ironic | 04:37 | |
| *** PollyZ has joined #openstack-ironic | 04:50 | |
| *** yuanying has quit IRC | 04:58 | |
| *** adu has joined #openstack-ironic | 05:08 | |
| *** yuanying has joined #openstack-ironic | 05:08 | |
| *** HenryG has quit IRC | 05:12 | |
| *** moshele has joined #openstack-ironic | 05:12 | |
| *** piet has joined #openstack-ironic | 05:14 | |
| *** HenryG has joined #openstack-ironic | 05:15 | |
| *** PollyZ_ has joined #openstack-ironic | 05:16 | |
| *** PollyZ has quit IRC | 05:18 | |
| *** garthb has joined #openstack-ironic | 05:20 | |
| *** jcoufal has joined #openstack-ironic | 05:23 | |
| *** sdake has quit IRC | 05:26 | |
| *** krtaylor has quit IRC | 05:33 | |
| *** adu has quit IRC | 05:33 | |
| *** moshele has quit IRC | 05:36 | |
| *** edand has joined #openstack-ironic | 05:37 | |
| *** ijw has joined #openstack-ironic | 05:38 | |
| *** tiendc has quit IRC | 05:38 | |
| *** bvandewa has joined #openstack-ironic | 05:41 | |
| *** ccamacho has quit IRC | 05:46 | |
| *** Goneri has joined #openstack-ironic | 05:54 | |
| *** yuanying has quit IRC | 05:56 | |
| *** yuanying has joined #openstack-ironic | 05:58 | |
| *** ijw has quit IRC | 05:59 | |
| *** rcernin has joined #openstack-ironic | 06:01 | |
| *** ijw has joined #openstack-ironic | 06:01 | |
| *** garthb has quit IRC | 06:06 | |
| *** ijw has quit IRC | 06:08 | |
| *** fragatina has joined #openstack-ironic | 06:12 | |
| *** fragatina has quit IRC | 06:12 | |
| *** ijw has joined #openstack-ironic | 06:12 | |
| *** piet has quit IRC | 06:13 | |
| *** fragatina has joined #openstack-ironic | 06:13 | |
| *** fragatina has quit IRC | 06:18 | |
| *** ijw has quit IRC | 06:19 | |
| *** mjura has joined #openstack-ironic | 06:22 | |
| openstackgerrit | Ramamani Yeleswarapu proposed openstack/ironic: Centralize config options - [neutron] https://review.openstack.org/304838 | 06:24 |
| *** jcoufal has quit IRC | 06:26 | |
| *** moshele has joined #openstack-ironic | 06:27 | |
| *** Goneri has quit IRC | 06:28 | |
| *** mjura has quit IRC | 06:29 | |
| *** mjura has joined #openstack-ironic | 06:29 | |
| *** fragatina has joined #openstack-ironic | 06:33 | |
| *** fragatina has quit IRC | 06:34 | |
| *** fragatina has joined #openstack-ironic | 06:35 | |
| openstackgerrit | Tuan proposed openstack/ironic: Add Python 3.5 classifier and venv https://review.openstack.org/341990 | 06:36 |
| *** PollyZ_ has quit IRC | 06:37 | |
| *** Goneri has joined #openstack-ironic | 06:38 | |
| *** pcaruana has joined #openstack-ironic | 06:40 | |
| *** yonglihe has joined #openstack-ironic | 06:47 | |
| *** hoangcx has quit IRC | 06:50 | |
| *** hoangcx_ has joined #openstack-ironic | 06:50 | |
| *** rbartal has joined #openstack-ironic | 06:51 | |
| *** tshefi has joined #openstack-ironic | 06:54 | |
| *** trown|outtypewww has quit IRC | 06:57 | |
| *** e0ne has joined #openstack-ironic | 06:58 | |
| *** trown has joined #openstack-ironic | 06:59 | |
| *** rcernin has quit IRC | 07:05 | |
| *** athomas has joined #openstack-ironic | 07:06 | |
| *** loki_ has joined #openstack-ironic | 07:09 | |
| *** tesseract- has joined #openstack-ironic | 07:10 | |
| *** ccamacho has joined #openstack-ironic | 07:11 | |
| lazy_prince | does anyone have the nova patch handy needed for VLAN in Ironic..? | 07:14 |
| openstackgerrit | Tuan proposed openstack/ironic: Update min tox version to 2.0 https://review.openstack.org/342008 | 07:19 |
| *** rcernin has joined #openstack-ironic | 07:20 | |
| *** e0ne has quit IRC | 07:21 | |
| openstackgerrit | Ramamani Yeleswarapu proposed openstack/ironic: Centralize config options - [DEFAULT] https://review.openstack.org/309070 | 07:21 |
| openstackgerrit | Anton Arefiev proposed openstack/python-ironic-inspector-client: Fix unit tests https://review.openstack.org/342012 | 07:26 |
| *** ohamada has joined #openstack-ironic | 07:28 | |
| *** hoangcx_ has quit IRC | 07:28 | |
| *** lazy_prince has quit IRC | 07:30 | |
| *** ishant|away has quit IRC | 07:30 | |
| *** kirankv has quit IRC | 07:31 | |
| *** lazy_prince has joined #openstack-ironic | 07:32 | |
| *** ishant|away has joined #openstack-ironic | 07:33 | |
| *** kirankv has joined #openstack-ironic | 07:33 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/python-ironic-inspector-client: Add a test dependency on requests-mock https://review.openstack.org/342017 | 07:34 |
| lazy_prince | nvm.. got the link https://review.openstack.org/#/c/297895/3 | 07:34 |
| *** fragatina has quit IRC | 07:36 | |
| *** fragatina has joined #openstack-ironic | 07:36 | |
| openstackgerrit | Swapnil Kulkarni (coolsvap) proposed openstack/ironic: [WIP] Testing latest u-c https://review.openstack.org/318440 | 07:39 |
| *** yuanying has quit IRC | 07:42 | |
| *** yuanying has joined #openstack-ironic | 07:43 | |
| *** fragatina has quit IRC | 07:52 | |
| *** bvandewa has quit IRC | 07:54 | |
| milan | #pixiesay Morning, Ironic! -m flexing | 07:56 |
| PixieBoots | ᕙʕ⇀ᴥ⇀ʔᕗ: Morning, Ironic! | 07:56 |
| *** d0ugal has joined #openstack-ironic | 07:59 | |
| *** jrist has joined #openstack-ironic | 07:59 | |
| *** zzzeek has quit IRC | 08:00 | |
| *** zzzeek has joined #openstack-ironic | 08:00 | |
| *** rcernin has quit IRC | 08:00 | |
| *** bvandewa has joined #openstack-ironic | 08:01 | |
| *** Nisha_away has joined #openstack-ironic | 08:02 | |
| *** jpich has joined #openstack-ironic | 08:02 | |
| *** bvandewa has quit IRC | 08:03 | |
| *** bvandewa has joined #openstack-ironic | 08:10 | |
| *** hamzy has quit IRC | 08:11 | |
| *** jistr is now known as jistr|off | 08:11 | |
| *** derekh has joined #openstack-ironic | 08:12 | |
| *** rcernin has joined #openstack-ironic | 08:13 | |
| aarefiev | morning all! | 08:13 |
| *** athomas has quit IRC | 08:16 | |
| aarefiev | morning milan, welcome back, how was your PTO | 08:17 |
| milan | morning aarefiev! thanks, it was very nice indeed :) | 08:17 |
| *** hoangcx has joined #openstack-ironic | 08:17 | |
| milan | aarefiev, how about your PTO plans? | 08:18 |
| *** athomas has joined #openstack-ironic | 08:19 | |
| *** ifarkas has joined #openstack-ironic | 08:19 | |
| *** MattMan has quit IRC | 08:19 | |
| *** MattMan has joined #openstack-ironic | 08:19 | |
| aarefiev | milan: nice :), yeah I have plans in a week, it's pretty hot here, so I go to the sea | 08:20 |
| *** amotoki has quit IRC | 08:20 | |
| milan | aarefiev, good :) I was in Croatia, water a bit colder but still nice | 08:21 |
| openstackgerrit | zhengchuanhu proposed openstack/python-ironicclient: Add Python 3.5 classifier and venv https://review.openstack.org/342037 | 08:21 |
| aarefiev | milan: cool, I've never been there, I thought about Bulgaria :) | 08:23 |
| milan | aarefiev, that's where my sis is right now :) | 08:24 |
| aarefiev | :) | 08:25 |
| *** Romanenko_K has joined #openstack-ironic | 08:28 | |
| *** lucas|afk is now known as lucasagomes | 08:29 | |
| lucasagomes | morning all | 08:29 |
| *** hamzy has joined #openstack-ironic | 08:30 | |
| aarefiev | morning lucasagomes | 08:30 |
| *** yuanying has quit IRC | 08:31 | |
| *** amotoki has joined #openstack-ironic | 08:34 | |
| openstackgerrit | Merged openstack/python-ironic-inspector-client: Add a test dependency on requests-mock https://review.openstack.org/342017 | 08:36 |
| *** yuanying has joined #openstack-ironic | 08:37 | |
| milan | morning lucasagomes :) | 08:37 |
| *** amotoki has quit IRC | 08:42 | |
| *** Linkmark has joined #openstack-ironic | 08:48 | |
| *** e0ne has joined #openstack-ironic | 08:51 | |
| *** yuanying has quit IRC | 08:53 | |
| openstackgerrit | Kyrylo Romanenko proposed openstack/python-ironicclient: Tests for testing chassis-create command https://review.openstack.org/293634 | 08:56 |
| *** milan has quit IRC | 08:58 | |
| vsaienk0 | Morning aarefiev, milan, amatoki, lucasagomes, dtantsur and all | 08:58 |
| *** e0ne has quit IRC | 08:59 | |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic: Collect deployment logs from IPA https://review.openstack.org/336102 | 09:00 |
| *** jtomasek has joined #openstack-ironic | 09:00 | |
| *** amotoki has joined #openstack-ironic | 09:01 | |
| *** e0ne has joined #openstack-ironic | 09:01 | |
| *** bvandewa has quit IRC | 09:04 | |
| *** vmud213 has quit IRC | 09:08 | |
| *** vmud213 has joined #openstack-ironic | 09:09 | |
| *** sambetts|afk is now known as sambetts | 09:13 | |
| sambetts | Morning all | 09:13 |
| *** mgould|afk is now known as mgould | 09:14 | |
| *** fragatina has joined #openstack-ironic | 09:15 | |
| mgould | morning lazy_prince milan aarefiev lucasagomes vsaienk0 dtantsur|afk sambetts and all other Ironicers! | 09:15 |
| sambetts | \o/ Multitenancy patches merged! | 09:15 |
| vsaienk0 | morning mgould, sambetts | 09:15 |
| vsaienk0 | sambetts: only 1 left | 09:15 |
| lucasagomes | sambetts, w00t | 09:16 |
| sambetts | then we just need to convince the nova guys to let us go ahead with merging the patch on that side | 09:16 |
| *** hoangcx has quit IRC | 09:17 | |
| openstackgerrit | Sergii Turivnyi proposed openstack/python-ironicclient: Negative tests for testing actions with Node. https://review.openstack.org/342060 | 09:19 |
| openstackgerrit | Sergii Turivnyi proposed openstack/python-ironicclient: Negative tests for testing actions with Node. https://review.openstack.org/342060 | 09:20 |
| *** clenimar_ has quit IRC | 09:30 | |
| *** amotoki has quit IRC | 09:33 | |
| *** sambetts has quit IRC | 09:36 | |
| *** mbound has joined #openstack-ironic | 09:38 | |
| mgould | sambetts: yay! | 09:38 |
| mgould | 10:38 - that's too early for champagne, right? | 09:39 |
| *** sambetts has joined #openstack-ironic | 09:39 | |
| sambetts | Well I just had a power cut... | 09:39 |
| *** milan has joined #openstack-ironic | 09:40 | |
| *** appu103_ has joined #openstack-ironic | 09:43 | |
| *** mbound has quit IRC | 09:43 | |
| *** appu103_ is now known as aparnav | 09:43 | |
| mgould | definitely champagne time, then | 09:44 |
| * sambetts really needs a UPS ... | 09:45 | |
| *** amotoki has joined #openstack-ironic | 09:46 | |
| sambetts | I really never know what to buy with a UPS, there are so many numbers to concider... | 09:49 |
| sambetts | I wish they just said will power a desktop PC for X minutes :-P | 09:49 |
| * mgould remembers the time at Edinburgh Uni when the power went out | 09:51 | |
| mgould | my building had some epic UPS that kept all the desktop PCs on | 09:51 |
| mgould | unfortunately, our home directories were on an AFS share hosted in another building, which had a much less epic UPS | 09:51 |
| sambetts | might have had its own backup generators with battery, I know our Uni had a setup like that | 09:52 |
| mgould | yeah, I think that's what it was | 09:52 |
| mgould | but we couldn't log in or do any work because of the AFS thing, so it was kinda pointless :-( | 09:52 |
| sambetts | One of Ciscos data centers has a really cool UPS system that uses a giant fly wheel | 09:52 |
| mgould | nice | 09:53 |
| sambetts | works better than batterys switching over while you wait for the backup power to kick in apprently | 09:54 |
| mgould | interesting | 09:56 |
| *** mbound has joined #openstack-ironic | 09:57 | |
| sambetts | fly wheels are cheaper in terms of materials and energy to make over batterys too | 09:58 |
| mgould | huh, TIL | 10:00 |
| openstackgerrit | Sivaramakrishna Garimella proposed openstack/ironic: Expose node's network_interface field in API https://review.openstack.org/317392 | 10:09 |
| *** dtantsur|afk is now known as dtantsur | 10:12 | |
| openstackgerrit | Sivaramakrishna Garimella proposed openstack/ironic: Update Ironic VM network connection https://review.openstack.org/256366 | 10:12 |
| dtantsur | Morning Ironic | 10:12 |
| openstackgerrit | Sivaramakrishna Garimella proposed openstack/ironic: Add configure_provision_network function https://review.openstack.org/256367 | 10:13 |
| *** sbog has quit IRC | 10:15 | |
| *** sbog has joined #openstack-ironic | 10:16 | |
| loki_ | dtantsur review please https://review.openstack.org/#/c/272658/ | 10:16 |
| dtantsur | I'll get to that eventually, but I can't promise how soon | 10:17 |
| loki_ | ok | 10:17 |
| aparnav | hey dtantsur, Can you take a look at this RFE https://bugs.launchpad.net/ironic/+bug/1593194 ? | 10:20 |
| openstack | Launchpad bug 1593194 in Ironic "[RFE]Enable hardware disk erase in ironic" [Undecided,Incomplete] | 10:20 |
| *** yibo has quit IRC | 10:20 | |
| dtantsur | aparnav, you need to limit your RFE only to things that are NOT implemented right now | 10:20 |
| dtantsur | I can't see anything new in it | 10:20 |
| openstackgerrit | Sivaramakrishna Garimella proposed openstack/ironic: Update Ironic VM network connection https://review.openstack.org/256366 | 10:21 |
| dtantsur | please describe what we have and what you want to have instead | 10:21 |
| openstackgerrit | Sivaramakrishna Garimella proposed openstack/ironic: Add configure_provision_network function https://review.openstack.org/256367 | 10:22 |
| openstackgerrit | Yuiko Takada Mori proposed openstack/ironic: IPMITool: add IPMISocatConsole class https://review.openstack.org/293873 | 10:26 |
| aparnav | dtantsur, ok. will update the description | 10:26 |
| openstackgerrit | Sivaramakrishna Garimella proposed openstack/ironic: Add configure_provision_network function https://review.openstack.org/256367 | 10:27 |
| openstackgerrit | Sivaramakrishna Garimella proposed openstack/ironic: Add configure_provision_network function https://review.openstack.org/256367 | 10:34 |
| openstackgerrit | Sivaramakrishna Garimella proposed openstack/python-ironicclient: Updates supporting ironic-neutron integration https://review.openstack.org/206144 | 10:40 |
| openstackgerrit | Sivaramakrishna Garimella proposed openstack/python-ironicclient: Add possibility to work with portgroups https://review.openstack.org/335964 | 10:41 |
| *** clenimar_ has joined #openstack-ironic | 10:41 | |
| *** caoshufeng has quit IRC | 10:44 | |
| *** phuongnh has quit IRC | 10:47 | |
| *** tshefi has quit IRC | 10:49 | |
| * sambetts wrote many words just now XD | 10:53 | |
| sambetts | having a english melt down | 10:53 |
| *** Nisha_away has quit IRC | 10:55 | |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Add test to verify ironic multitenancy https://review.openstack.org/269157 | 11:01 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Add configure_provision_network function https://review.openstack.org/256367 | 11:01 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Update Ironic VM network connection https://review.openstack.org/256366 | 11:01 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: DO NOT REVIEW https://review.openstack.org/296432 | 11:01 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Add Ironic/Neutron integration documentation https://review.openstack.org/258596 | 11:01 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Expose node's network_interface field in API https://review.openstack.org/317392 | 11:01 |
| *** hoangcx has joined #openstack-ironic | 11:03 | |
| vsaienk0 | dtantusr, sambetts, lucasagomes could you please check comments from devananda https://review.openstack.org/#/c/317392/19/ironic/api/controllers/v1/node.py | 11:04 |
| *** yuriyz has joined #openstack-ironic | 11:05 | |
| sambetts | vsaienk0: is that trace from when a node is in enroll or availible | 11:08 |
| sambetts | ? | 11:08 |
| vsaienk0 | sambetts: the trace is when node was enrolled with invalid driver, and then I update/delete it | 11:09 |
| sambetts | doesn't the API have access to the config file and can read enabled_network_interfaces from that? | 11:09 |
| vsaienk0 | sambetts: so the node is in enroll state | 11:09 |
| vsaienk0 | config file may be different on API and Conductor, we can't rely on that value | 11:10 |
| sambetts | how do we handle it for drivers now? | 11:10 |
| vsaienk0 | when conductor starts it register drivers, and they put to DB. API reads data from db when enrolling node | 11:11 |
| sambetts | vsaienk0: if thats the case that is a configuration error on the deployers part | 11:11 |
| sambetts | I think :/ | 11:11 |
| vsaienk0 | so what is your suggestion? get list of enabled_network_interfaces on API from config file, and compare during node enrollment? | 11:12 |
| milan | morning mgould, vsaienk0, sambetts :) | 11:13 |
| vsaienk0 | o/ milan | 11:13 |
| milan | sambetts, where is the |cat gone? ;) | 11:13 |
| sambetts | milan: I'm fully here today, not in a quantum state ;) | 11:13 |
| milan | :D | 11:13 |
| milan | lol | 11:13 |
| dtantsur | lol | 11:14 |
| milan | [sambetts|cat] to be more obvious ;) | 11:14 |
| milan | dtantsur, the patch please? ;) | 11:15 |
| sambetts | vsaienk0: tbh I'm inclined to allow nodes in enroll to have interfaces set that are invalid, and then use validation to check it when you try to move the node to managable, because thats what enroll is for | 11:15 |
| dtantsur | milan, yeah, sorry, it's https://review.openstack.org/336532 | 11:15 |
| milan | dtantsur, thx | 11:16 |
| dtantsur | my IPA API patches are in conflict.. again... >_< | 11:16 |
| sambetts | vsaienk0: although disabled interfaces I guess is probably something we want to catch early | 11:16 |
| vsaienk0 | sambetts: so when we add node with invalid driver, we will not able to update network driver to correct value for the node | 11:16 |
| sambetts | yeah, I'm just thinking it through | 11:17 |
| sambetts | vsaienk0: maybe we need to do DB cache part of driver comp work for network_interfaces ealier so we can check them without talking to the conductors? | 11:18 |
| * sambetts is trying to find where we handle it for drviers | 11:18 | |
| dtantsur | sambetts, all deps for your LLDP patch have merged \o/ | 11:21 |
| sambetts | \o/ | 11:21 |
| * sambetts wants to try it all out but his test bed is still in use by a different team | 11:21 | |
| dtantsur | we have downstream people dreaming of LLDP discovery for unrelated reasons as well | 11:21 |
| dtantsur | I think they want TripleO UI to draw network topology to a user :) | 11:21 |
| dtantsur | sambetts, does LLDP make any sense in virtual environment? I wonder if we need a gate with it enabled | 11:22 |
| * dtantsur will enable it in tripleo CI btw | 11:22 | |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic: Deprecate [ilo]/clean_priority_erase_devices config https://review.openstack.org/342114 | 11:22 |
| sambetts | ooo nice, hehe thats actually the sort of thing that I orignially wrote curvature for (the new wibbly wobbly horizon network view) | 11:22 |
| lucasagomes | vsaienk0, will take a look in a min, will get some lunch first | 11:22 |
| *** lucasagomes is now known as lucas-hungry | 11:23 | |
| dtantsur | interesting, I haven't used horizon at all :D | 11:23 |
| vsaienk0 | lucasagomes, thank you, and have a nice lunch! | 11:23 |
| * dtantsur is a CLI guy | 11:23 | |
| sambetts | :-P I've not used it since I stopped working on it | 11:23 |
| sambetts | dtantsur: Not sure if it works in a virutal env, I don't even know if things like OVS can do it | 11:24 |
| dtantsur | hmm, I see | 11:24 |
| sambetts | dtantsur: do you know where in the Ironic code we check if the driver you are trying to create a node with is valid or not? | 11:26 |
| sambetts | dtantsur: I can't see it in the nodes controller, but maybe I'm being blide | 11:26 |
| sambetts | blind* ] | 11:26 |
| sambetts | milan: more like sambetts|inabox | 11:26 |
| dtantsur | sambetts, when finding a matching controller maybe? not sure | 11:27 |
| sambetts | maybe :/ I thought there would be an if driver in enabled_drivers in the post function for the node or something | 11:28 |
| milan | sambetts|shroeding | 11:28 |
| sambetts | milan: yeah I wanted it but it doesn't fit :( | 11:28 |
| milan | aaah yeah | 11:28 |
| sambetts | dtantsur: I think I might have found it | 11:29 |
| sambetts | dtantsur: https://github.com/openstack/ironic/blob/master/ironic/conductor/rpcapi.py#L115 | 11:30 |
| vsaienk0 | sambetts: https://github.com/openstack/ironic/blob/master/ironic/api/controllers/v1/node.py#L1237 https://github.com/openstack/ironic/blob/master/ironic/api/controllers/v1/node.py#L1237 | 11:30 |
| sambetts | vsaienk0: yup | 11:31 |
| dtantsur | correct | 11:32 |
| sambetts | vsaienk0, dtantsur: I think either the Ironic API needs to contain the list of all network_interfaces enabled, or we probably should be taking network_interfaces into account in the hash ring because we currently don't enforce that different conductors to have the same network_interfaces loaded :/ | 11:32 |
| *** amotoki has quit IRC | 11:33 | |
| dtantsur | I don't think we can realistically switch to taking interfaces in account... | 11:34 |
| dtantsur | that's gonna have interesting upgrade consequences | 11:34 |
| sambetts | I think we need to enforce all conductors have the same loaded interfaces then, otherwise we're going to end up with a lot of stuck nodes | 11:35 |
| * sambetts bangs head on desk | 11:36 | |
| * milan cannot stop thinking about shroedinger's cat abbreviations such as |shroe~ |shro~ |shr~ |shroeow |catinabox |[meow] |[cat] | 11:37 | |
| sambetts | haha | 11:37 |
| *** jcoufal has joined #openstack-ironic | 11:44 | |
| milan | |[O.o] | 11:46 |
| sambetts | [???] | 11:47 |
| milan | i like mine more :P | 11:47 |
| sambetts | or [~~] | 11:48 |
| milan | heh that's a nice one :) | 11:48 |
| *** amotoki has joined #openstack-ironic | 11:49 | |
| openstackgerrit | Merged openstack/python-ironicclient: Add test for ironic node-list command https://review.openstack.org/329056 | 11:50 |
| dtantsur | sambetts, we can't enforce it.. | 11:52 |
| sambetts | no because rolling upgrades | 11:52 |
| dtantsur | yep | 11:52 |
| TheJulia | good morning | 11:53 |
| dtantsur | morning TheJulia | 11:53 |
| milan | morning TheJulia | 11:53 |
| sambetts | Morning morning TheJulia | 11:53 |
| sambetts | dtantsur: problem is with the current implemention in the network patches, depending on which conductor you end up on and which network interfaces are enabled on that conductor the node create will fail or pass | 11:55 |
| dtantsur | yeah.. | 11:55 |
| dtantsur | also wonderful: I've got a merge conflict with some patch, and git shows like half of file | 11:55 |
| sambetts | so you could make the same call twice and once it pass and once it fail | 11:55 |
| TheJulia | that should be changing, slightly so it shifts to validate | 11:56 |
| dtantsur | so I'll probably remove someone's work >_< | 11:56 |
| *** vmud213 has quit IRC | 11:56 | |
| TheJulia | but I've not looked at it yet | 11:56 |
| dtantsur | lets hope the person did have sane unit tests | 11:56 |
| TheJulia | this morning that is | 11:56 |
| sambetts | dtantsur: I had a merge like that the other day | 11:56 |
| *** vmud213 has joined #openstack-ironic | 11:56 | |
| sambetts | basically had to just nuke the work and then go through my git diff and revert anything that I didn't remember changing | 11:56 |
| dtantsur | well, I've moved several functions | 11:57 |
| dtantsur | and someone seemingly ended up modifying them | 11:57 |
| sambetts | dtantsur: yup thats the problem I had with this https://review.openstack.org/#/c/341086/ | 11:57 |
| sambetts | https://review.openstack.org/#/c/340567/ | 11:58 |
| sambetts | this ^ | 11:58 |
| sambetts | not that other one :-P | 11:58 |
| sambetts | TheJulia: oh? I was looking at this patch with vsaienk0 https://review.openstack.org/#/c/317392/19/ironic/api/controllers/v1/node.py and it seemed to not be in validate still | 11:58 |
| dtantsur | I think it's https://review.openstack.org/#/c/213262/126/ironic/drivers/modules/agent_base_vendor.py | 11:58 |
| sambetts | ah | 11:59 |
| sambetts | of course the network patches add stuff there | 11:59 |
| TheJulia | ugh, looks like it is basically still there | 12:00 |
| TheJulia | sambetts: I'm going to go get some coffee and shower real quick since I am in no possible way... awake | 12:01 |
| *** amotoki has quit IRC | 12:01 | |
| sambetts | :-p sure | 12:01 |
| *** amotoki has joined #openstack-ironic | 12:04 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic: Implement new heartbeat for AgentDeploy https://review.openstack.org/333964 | 12:06 |
| dtantsur | folks plz review ^^^ | 12:06 |
| dtantsur | I'm a bit tired of rebasing this chain.. | 12:06 |
| *** amotoki has quit IRC | 12:17 | |
| TheJulia | sambetts: I remember the discussion of the different enabled network drivers came up and there was an agreement that the configuration essentially had to be the same across all the conductors in term of drivers. I think we need the ability to pin hosts to a group of conductors, but I think heads will explode then | 12:23 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic: Promote agent vendor passthru to core API https://review.openstack.org/330018 | 12:23 |
| openstackgerrit | zhengchuanhu proposed openstack/python-ironicclient: Add Python 3.5 classifier and venv https://review.openstack.org/342037 | 12:24 |
| *** lucas-hungry is now known as lucasagomes | 12:24 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-python-agent: Use new agent API if available https://review.openstack.org/334523 | 12:24 |
| * TheJulia reads vasyl's reply on rev19 | 12:24 | |
| sambetts | TheJulia: right, its a crazy problem I think we need to solve, I can't remeber how driver reform does it, because I wonder if deployers might want to have some conductors that work with a hardware_type on one set of interfaces, and another conductor that works with the same hardware_type but on a different set of interfaces | 12:25 |
| sambetts | e.g. same hardware but one pool is configured to work with just IPMI and another pool is configured to work with CustomOOBThing, and should the nodes for them be able to land on different conductors | 12:27 |
| TheJulia | I think consistency across all conductors has been the the consensus for a while, but yeah...if I have 3 conductors over here, and 3 over there, and they can't see each other's pools of hardware.... (where I'm essentially using the API as a single pane of glass), things just don't work that way right now :( | 12:27 |
| sambetts | TheJulia: yeah thats the part thats missing | 12:27 |
| *** clenimar_ has quit IRC | 12:28 | |
| TheJulia | so for the immediate problem, not any of the long term problems... for validate, can't we check to see if the network driver is loaded/exists then before we try to call .validate() on the driver? | 12:28 |
| sambetts | if all conductors have to have the same configured interfaces the devas problem goes away because you just configure the API with the same list of interfaces | 12:28 |
| *** dprince has joined #openstack-ironic | 12:28 | |
| sambetts | or we write the enabled interfaces into the DB and read it from there in the API | 12:29 |
| *** baoli has joined #openstack-ironic | 12:32 | |
| sambetts | if ensure only interfaces that exist on all conductors into the DB then we can get a consensus, and it doesn't block conductors from starting with more/less interfaces than the whole pool, so rolling changes will still work | 12:32 |
| sambetts | e.g. conductor 1 has neutron, condutor 2 has neutron and flat, only neutron is written into the DB so only nodes with neutron network can be created | 12:33 |
| *** baoli_ has joined #openstack-ironic | 12:33 | |
| sambetts | then if conductor 1 adds flat, then both get written into the DB, so you can create nodes with both neutron and flat | 12:33 |
| sambetts | but if any 1 conductor removes an interface, then we have to remove it from the DB | 12:34 |
| TheJulia | hmm...hmmmm | 12:35 |
| openstackgerrit | zhengchuanhu proposed openstack/python-ironicclient: Add Python 3.5 classifier and venv https://review.openstack.org/342037 | 12:35 |
| TheJulia | sambetts: I worry that might be too much, and not a before friday thing | 12:36 |
| *** baoli has quit IRC | 12:36 | |
| TheJulia | but more thought required there at the moment | 12:37 |
| *** athomas has quit IRC | 12:37 | |
| jroll | we could get crazy and make the hash ring depend on enabled interfaces as well, right? | 12:37 |
| jroll | and then someone can do whatever | 12:37 |
| sambetts | jroll: yeah and I think that we should idealy | 12:38 |
| *** daemontool_ has quit IRC | 12:38 | |
| TheJulia | agreed | 12:38 |
| sambetts | #brainache | 12:38 |
| * jroll wonders how that'd work | 12:38 | |
| * TheJulia suspects jroll will now reach #brainache | 12:38 | |
| * jroll takes two steps backwards | 12:39 | |
| xavierr | good morning Ironicland :) | 12:39 |
| TheJulia | so why can't we add code to update_node and validate to handle the driver submitted being bad? | 12:39 |
| *** hoangcx has quit IRC | 12:40 | |
| TheJulia | because looking at the example of letting driver test1 through on network_interface, it bombs on trying to raise the exception, not the check to see if it is in the list of enabled drivers | 12:40 |
| *** loki_ has quit IRC | 12:40 | |
| *** links has quit IRC | 12:40 | |
| jroll | TheJulia: with the driver composition changes, we can, because enabled interfaces will be in the db | 12:41 |
| jroll | but, they aren't yet | 12:41 |
| TheJulia | yes, but driver composition is not here right now, this is to get is through until it exists | 12:41 |
| vsaienk0 | morning TheJulia, jroll | 12:42 |
| TheJulia | with big #TODO: Remove me once driver composition labels | 12:42 |
| TheJulia | good morning vsaienk0 | 12:42 |
| *** athomas has joined #openstack-ironic | 12:42 | |
| sambetts | Tbh I'm more concerned with the if I fire off 2 differnt nodecreates one might fail and one might not | 12:44 |
| sambetts | depending on the conductors | 12:44 |
| TheJulia | In a sense, that is how it is today | 12:45 |
| TheJulia | we cannot prevent bad configuration | 12:45 |
| jroll | well, deva was concerned with node create performance given the rpc call | 12:45 |
| sambetts | if I fire off 2 nodecreates with the same driver it isn't random to if it'll pass or fail today as far as I know | 12:45 |
| jroll | I am too | 12:45 |
| jroll | you should be able to spam node creates super fast | 12:45 |
| jroll | so I'd rather just handle it on validate for now | 12:46 |
| TheJulia | validate and update_node will both need to check and handle appropriately | 12:46 |
| jroll | yeah, agree | 12:46 |
| *** skramaja has quit IRC | 12:46 | |
| jroll | and we should document people should have that config the same on all conductors :) | 12:46 |
| TheJulia | and we just make sure our documentation explicitly states network config should be same on oll conductors | 12:47 |
| jroll | "anything else is unsupported" | 12:47 |
| TheJulia | heh | 12:47 |
| jroll | ^5 | 12:47 |
| TheJulia | yes yes | 12:47 |
| sambetts | well if we enforce the config to be the same on all conductors we don't need to call to the conductors at all and could just put it in the API config file right?? | 12:47 |
| TheJulia | so update_node and validate_driver_interfaces in manager.py | 12:47 |
| TheJulia | sambetts: in theory... yes | 12:48 |
| TheJulia | the downside is some people will not like that | 12:48 |
| TheJulia | that is actually a way to reduce the techdebt here | 12:48 |
| TheJulia | simplest path is to just force that config to be present for the API as well | 12:48 |
| TheJulia | it doesn't line up with driver composition exactly right now, and we remove it later once we've identified the proper way for handling that deep inside the conductor | 12:49 |
| jroll | yeah, I'm not super against that either, but it's an inconsistency and not very intuitive | 12:49 |
| jroll | we'll have a few support requests as to why this isn't working at all | 12:49 |
| jroll | (because folks won't realize it should be in api config) | 12:50 |
| dtantsur | a hint: time not spent on rebase the agent API thing will be spent on coding the driver composition | 12:51 |
| dtantsur | ;) | 12:51 |
| sambetts | hahaha | 12:51 |
| jroll | lolz | 12:51 |
| TheJulia | so... in the conductor, make something called _network_inteface_check that validates if the config is legit, then continue, if not raise an exception | 12:51 |
| *** bharathk has quit IRC | 12:52 | |
| TheJulia | add that to the beginning of validate_driver_interfaces and update_node.... and I think that will work *think* | 12:52 |
| TheJulia | heh | 12:52 |
| TheJulia | I must be thinking too much | 12:52 |
| jroll | yeah, I think that's reasonable | 12:52 |
| sambetts | so when a node moves from enroll to managable it'll fail or not | 12:53 |
| sambetts | if you've derped on node-create | 12:53 |
| TheJulia | it should I believe | 12:53 |
| jroll | right | 12:53 |
| TheJulia | fail that is | 12:53 |
| TheJulia | hmmmm | 12:53 |
| TheJulia | oh state machine, how do you force validation | 12:53 |
| *** d0ugal has quit IRC | 12:53 | |
| sambetts | does that mean you can set anything as the network_interface enabled/disable/non-existant while in enroll ? | 12:54 |
| TheJulia | so _do_node_verify as well to catch at the manageable state transition | 12:54 |
| TheJulia | in theory, yes | 12:55 |
| TheJulia | well, if it is non-existant, it defaults to flat, if it is defined to "meow", it bombs in logical steps forward | 12:55 |
| sambetts | by non-existant I mean the thing you set it too doesn't actualy exist | 12:55 |
| TheJulia | The meow network driver, where a cat comes to sit in your lap | 12:55 |
| jroll | and setting meow on node-update will fail during the api call | 12:55 |
| jroll | TheJulia: I think you mean where the cat goes and plays with the cabling | 12:56 |
| TheJulia | jroll: well, nibbling, but that too | 12:56 |
| TheJulia | its a feature... not a bug! | 12:56 |
| sambetts | ;) | 12:56 |
| jroll | hehe | 12:56 |
| jroll | but yeah, on enroll->manage or node-update, it will be validated | 12:56 |
| jroll | I think this is what deva suggested yesterday as well, I'm +1 on that | 12:57 |
| TheJulia | basically yeah | 12:57 |
| sambetts | so if I set network-interface to "thing-that-doesn't-exist" and then try to do a node-update it won't block me from updating the node? | 12:57 |
| TheJulia | sambetts: it would only block you if your not supplying a valid network driver | 12:58 |
| jroll | good question, I don't think it would | 12:58 |
| sambetts | thats the senario that happens right now, is that if you set a network_interface to a disabled or non-existant one it prevents all node-updates going forward | 12:59 |
| *** cdearborn has joined #openstack-ironic | 12:59 | |
| sambetts | even ones to correct the problem | 12:59 |
| sambetts | if you remove the check on node-create | 12:59 |
| jroll | ooo, interesting | 12:59 |
| TheJulia | yeah, we need to make sure it is a real driver in node-update as well | 12:59 |
| jroll | does it give a decent error? | 12:59 |
| TheJulia | http://paste.openstack.org/show/532762/ | 13:00 |
| sambetts | yes that ^ | 13:00 |
| sambetts | :-P | 13:00 |
| jroll | I mean via API, though | 13:00 |
| jroll | but yeah, that seems problematic :( | 13:00 |
| TheJulia | afaik it wouldn't get a reply right now | 13:00 |
| jroll | it'd be nice to s/driver/network interface/ in that error | 13:01 |
| sambetts | totally, I think thats a pretty simple change in the driver facotry base class | 13:01 |
| jroll | I feel like it comes from stevedore, but we can probably catch and reraise | 13:01 |
| jroll | oh no it's ours, cool | 13:02 |
| TheJulia | yeah, looks like it is getting raised when we try to raise an exception | 13:03 |
| * jroll does it | 13:03 | |
| lucasagomes | jroll, TheJulia sambetts if we have to enforce same config for all conductors we don't need a hashring | 13:04 |
| lucasagomes | sambetts, the update behavior sounds like a bug | 13:04 |
| jroll | lucasagomes: agree | 13:04 |
| lucasagomes | same for delete, based on vsaienk0 comment you can't update nor delete a node if the network_interface is disabled | 13:04 |
| lucasagomes | that's a bug for me | 13:04 |
| lucasagomes | per vsaienk0 comment i mean https://review.openstack.org/#/c/317392/19/ironic/api/controllers/v1/node.py | 13:05 |
| jroll | that traceback is odd :/ | 13:07 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Show what type of driver wasn't found in DriverNotFound https://review.openstack.org/342167 | 13:08 |
| jroll | TheJulia: sambetts: it'd be something like that, though I don't see where DriverNotFound could be raised there (as opposed to DriverNotFoundInEntrypoint) | 13:08 |
| *** mgoddard_ has joined #openstack-ironic | 13:09 | |
| *** rbudden has joined #openstack-ironic | 13:10 | |
| * jroll steps away for a bit | 13:11 | |
| *** mgoddard has quit IRC | 13:13 | |
| *** sdake__ has joined #openstack-ironic | 13:15 | |
| *** aparnav has quit IRC | 13:19 | |
| TheJulia | jroll: yeah, I'm kind of looking at it again going "wait a second, I just pulled down the latest rev and I'm trying to reproduce locally | 13:25 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic: Promote agent vendor passthru to core API https://review.openstack.org/330018 | 13:26 |
| *** amotoki has joined #openstack-ironic | 13:26 | |
| *** bharath has joined #openstack-ironic | 13:27 | |
| *** bharathk has joined #openstack-ironic | 13:31 | |
| *** bharath has quit IRC | 13:31 | |
| mgould | dtantsur: you can tell git to use the patience diff algorithm, that sometimes produces better diffs | 13:33 |
| *** ametts has joined #openstack-ironic | 13:33 | |
| mgould | git rebase -s recursive -Xpatience, I think | 13:34 |
| lucasagomes | random question: Is tempest branchless!? | 13:34 |
| dtantsur | mgould, interesting, never tried it | 13:35 |
| *** thiagop has joined #openstack-ironic | 13:37 | |
| thiagop | good morning! | 13:37 |
| mgould | thiagop: morning! | 13:38 |
| mgould | dtantsur: I think patience diff is slower but less likely to get confused by similar sections of code | 13:38 |
| mgould | http://stackoverflow.com/questions/4045017/what-is-git-diff-patience-for | 13:39 |
| TheJulia | http://paste.openstack.org/show/mkwGSdjRUgImHtzL9yRt/ at least it is kind of pretty.... | 13:39 |
| dtantsur | mgould, cool | 13:40 |
| TheJulia | so, all the periodic tasks start throwing stack traces if we don't validate the input early on | 13:41 |
| mgould | lucasagomes: looks like it! https://github.com/openstack/tempest/branches | 13:41 |
| TheJulia | jroll: I truly think we just need to base on the config file for now :( | 13:41 |
| lucasagomes | mgould, yeah oh dear | 13:42 |
| *** rbartal has quit IRC | 13:43 | |
| *** mbound has quit IRC | 13:43 | |
| TheJulia | lucasagomes: afaik tempest and devstack-gate are branchless | 13:44 |
| mgould | hand-hacked tests for version ranges woo yeah | 13:44 |
| lucasagomes | TheJulia, yeah, I just saw this: https://bugzilla.redhat.com/show_bug.cgi?id=1356092 | 13:45 |
| openstack | bugzilla.redhat.com bug 1356092 in openstack-ironic "Tempest test failing for cinder, sahara and neutron in RHOS-9" [High,New] - Assigned to lmartins | 13:45 |
| lucasagomes | (ignore the title) | 13:45 |
| lucasagomes | apparently due the branchless nature we have to make it backward compatible within the component plugin | 13:46 |
| TheJulia | yeouch | 13:46 |
| dtantsur | lucasagomes, the big was fixed upstream IIRC | 13:47 |
| dtantsur | including mitaka and liberty | 13:47 |
| *** bharathk has quit IRC | 13:47 | |
| dtantsur | lucasagomes, https://github.com/openstack/ironic/commit/3ff49f1d10f7acba33e00927e6c5180d3cb9a3be | 13:48 |
| lucasagomes | dtantsur, right, it was this: https://github.com/openstack/ironic/commit/b8f227e62e7aac87266cae53093ccb3e68bc6d32 | 13:48 |
| dtantsur | oh, or is it the opposite? | 13:48 |
| lucasagomes | dtantsur, the redhat bug is the other way around | 13:48 |
| dtantsur | right...... | 13:48 |
| lucasagomes | since tempest haven't been rebased it's failing | 13:48 |
| lucasagomes | we need to handle that AttributeError | 13:48 |
| lucasagomes | and call the old function | 13:48 |
| lucasagomes | *sigh* | 13:48 |
| *** daemontool has joined #openstack-ironic | 13:48 | |
| dtantsur | lucasagomes, isn't it easier to rebase tempest? | 13:48 |
| *** ElCoyote_ has joined #openstack-ironic | 13:48 | |
| lucasagomes | dtantsur, apparently that will break the other components not rebased yet | 13:49 |
| lucasagomes | I was talking to dmellado about it | 13:49 |
| lucasagomes | dtantsur, it's a ball of wax! | 13:49 |
| dtantsur | lucasagomes, then we need to revert the ironic change downstream | 13:49 |
| lucasagomes | dtantsur, yeah I was more inclined to make it backward compat within the ironic tempest plugin | 13:49 |
| lucasagomes | master and mitaka | 13:50 |
| dtantsur | lucasagomes, it does not constitute an appropriate backport to me | 13:50 |
| sambetts | less code in devstack-gate the better IMO | 13:50 |
| dtantsur | lucasagomes, it's our downstream problem | 13:50 |
| *** bharath has joined #openstack-ironic | 13:50 | |
| dtantsur | and downstream revert is MUCH faster FWIW | 13:50 |
| dtantsur | we can land it in 5 minutes | 13:51 |
| dtantsur | but dunno, maybe other distributions have similar problems | 13:51 |
| dtantsur | zigo, hi! do you guys ship tempest? how do you update it? | 13:51 |
| lucasagomes | dtantsur, yes I understand, I'm trying to figure whether it would be an upstream problem as well... the fact that the project is branchless it raises questions to me | 13:51 |
| lucasagomes | they should just remove something, at least make an alias to the new function | 13:52 |
| dtantsur | I agree | 13:53 |
| dtantsur | maybe worth chatting on #openstack-qa? | 13:53 |
| *** ametts has quit IRC | 13:54 | |
| *** ohamada has quit IRC | 13:55 | |
| *** mgoddard_ has quit IRC | 13:56 | |
| *** mgoddard has joined #openstack-ironic | 13:57 | |
| *** vmud213 has quit IRC | 13:57 | |
| *** bharath has quit IRC | 13:57 | |
| *** PollyZ has joined #openstack-ironic | 13:58 | |
| *** vishwanathj has joined #openstack-ironic | 13:58 | |
| lucasagomes | dtantsur, pinged it... for now I will submit a patch downstream to unblock things there | 14:01 |
| * lucasagomes is sad | 14:01 | |
| *** links has joined #openstack-ironic | 14:01 | |
| *** adu has joined #openstack-ironic | 14:03 | |
| *** amotoki has quit IRC | 14:04 | |
| *** ametts has joined #openstack-ironic | 14:07 | |
| *** amotoki has joined #openstack-ironic | 14:07 | |
| dtantsur | didn't we explicitly remove https://review.openstack.org/#/c/341253/2/nova/virt/ironic/driver.py ? | 14:13 |
| *** bharath has joined #openstack-ironic | 14:15 | |
| openstackgerrit | Sergii Turivnyi proposed openstack/python-ironicclient: Negative tests for testing actions with Node https://review.openstack.org/342060 | 14:15 |
| *** catintheroof has joined #openstack-ironic | 14:16 | |
| jroll | lucasagomes: dtantsur: fwiw, tempest does do releases at branching time, so you could use that downstream | 14:18 |
| jroll | dtantsur: we did remove that | 14:18 |
| jroll | TheJulia: ouchhhhh | 14:19 |
| *** joprovost has joined #openstack-ironic | 14:20 | |
| sambetts | jroll, TheJulia: :( | 14:21 |
| *** amotoki has quit IRC | 14:21 | |
| vsaienk0 | jroll, sambetts, lucasagomes, I'm sorry I was away. And didn't have opportunity to discuss https://review.openstack.org/#/c/317392/21/ironic/api/controllers/v1/node.py@1266 | 14:22 |
| *** ohamada has joined #openstack-ironic | 14:22 | |
| vsaienk0 | so how we are going to solve the problem, add network_interface to hasring? or guarantee that all conductors has the same configuration? | 14:23 |
| *** greghaynes has quit IRC | 14:23 | |
| dtantsur | jroll, yeah, but our gate seems to use tempest master, hence the downstream breakage | 14:23 |
| openstackgerrit | Merged openstack/ironic-python-agent: Updated from global requirements https://review.openstack.org/341860 | 14:24 |
| jroll | dtantsur: well, if you're using ironic master downstream, wouldn't you use tempest master as well? | 14:24 |
| dtantsur | jroll, we | 14:24 |
| dtantsur | we're not using ironic master | 14:24 |
| dtantsur | or we do, but not here | 14:25 |
| *** greghaynes has joined #openstack-ironic | 14:25 | |
| dtantsur | we've made this change on stable/mitaka and liberty because they had to work with tempest master | 14:25 |
| dtantsur | however tempest downstream was not updated | 14:25 |
| jroll | ohhhhh, I see | 14:25 |
| jroll | ouch | 14:25 |
| vsaienk0 | Thejulia: if we set network_interface to incorrect value, it will be impossible to update it to correct value also | 14:29 |
| jroll | so we have two options, right? | 14:30 |
| TheJulia | vsaienk0: not convinced that it is imposible, but I'm really thinking we need to read from the config file for now. | 14:30 |
| jroll | 1) eat the performance hit, only when network interface is passed in. it shouldn't be terrible but isn't ideal | 14:30 |
| *** ijw has joined #openstack-ironic | 14:30 | |
| jroll | (and remove that later) | 14:30 |
| sambetts | and document all conductors have the same interfaces loaded | 14:31 |
| jroll | 2) make people put enabled_net_int config in the api config file | 14:31 |
| jroll | (2) seems less annoying but more manual things for operators to do | 14:31 |
| jroll | (1) seems like it would be rare to happen but really annoying when it does | 14:31 |
| TheJulia | 2 seems more like a headache for configuration management systems | 14:31 |
| sambetts | and it still doesn't sovle the conductors with mixed interfaces problem | 14:32 |
| jroll | sambetts: either way we need to document that they need to be the same | 14:32 |
| TheJulia | and that it has been the case for a while now | 14:32 |
| jroll | TheJulia: well, people may need to change their thing, but I don't think it's too big of a deal, right? | 14:33 |
| TheJulia | shouldn't be as long as they understand the repercussion of doing so | 14:33 |
| jroll | repercussion of making the change? | 14:33 |
| TheJulia | that if they have nodes that are only supported on one conductor, and that conductor fails, etc | 14:34 |
| jroll | oh that, yeah, different interfaces between conductors is completely unsupported imo | 14:34 |
| TheJulia | Thinking about it... we need a sad pixieboots | 14:34 |
| jroll | different drivers/hardware types is ok | 14:34 |
| TheJulia | yeah | 14:34 |
| sambetts | we should and a spec for adding interfaces to the hashring? because in the end I think thats what we should support | 14:36 |
| lucasagomes | the thing of having the same config seems to kinda go against the current design :-/ I don't like it much | 14:37 |
| sambetts | yeah :/ | 14:38 |
| *** pcaruana has quit IRC | 14:39 | |
| *** vishwanathj has quit IRC | 14:40 | |
| thiagop | QQ about^: do we have a test on any pipeline that tests the Ironic hashring? | 14:40 |
| *** vishwanathj has joined #openstack-ironic | 14:40 | |
| vsaienk0 | sambetts: I don't know how hashring will be changed with driver-composition-reform but according to https://review.openstack.org/#/c/188370/20/specs/approved/driver-composition-reform.rst@223 it should be taken into account. | 14:40 |
| thiagop | or all our tests are considering single conductor? | 14:40 |
| sambetts | thiagop: no but we talked about using the multinode jobs to test it | 14:40 |
| jroll | I wish I had time to talk about this :( | 14:41 |
| vsaienk0 | In case of network_interface we should choose conductor with simultaneous match of driver and network_interface | 14:41 |
| thiagop | That may be a good course of action for QA in a near future | 14:41 |
| vsaienk0 | who know when driver-composition-reform will be implemented? | 14:41 |
| jroll | vsaienk0: hopefully this cycle | 14:42 |
| thiagop | Maybe we can rally people to do that in Barcelona as we did for the grenade in Austin :) | 14:42 |
| vsaienk0 | :) | 14:42 |
| *** mbound has joined #openstack-ironic | 14:44 | |
| clenimar | dtantsur: hey there | 14:45 |
| dtantsur | vsaienk0, the first step would be to land https://review.openstack.org/#/q/topic:bug/1570841 | 14:45 |
| dtantsur | clenimar, hi | 14:45 |
| clenimar | dtantsur: https://review.openstack.org/#/c/300154/ seems to be fine from identity's perspective. can you take a look at it? :) | 14:45 |
| dtantsur | looking | 14:45 |
| clenimar | dtantsur: thank you, sir | 14:46 |
| *** d0ugal has joined #openstack-ironic | 14:46 | |
| *** ijw has quit IRC | 14:46 | |
| vgadiraj | morning all | 14:47 |
| thiagop | morning vgadiraj | 14:47 |
| vsaienk0 | it seems that option 2 for now wins (have the same enabled_network_interfaces on conductors and API an compare network_interface with config value on API side) | 14:47 |
| *** amotoki has joined #openstack-ironic | 14:48 | |
| *** mbound has quit IRC | 14:48 | |
| vsaienk0 | I want make sure that we are ok to follow this way before uploading new patchset :) | 14:49 |
| jroll | I'm okay with that until driver comp stuff is done | 14:49 |
| vsaienk0 | jroll: I will mention it in relase note, should I add a note to config option also? | 14:50 |
| jroll | vsaienk0: yeah, that sounds good, probably docs as well | 14:51 |
| openstackgerrit | Kyrylo Romanenko proposed openstack/python-ironicclient: Add short driver argument for OSC node create command https://review.openstack.org/342215 | 14:52 |
| *** ijw has joined #openstack-ironic | 14:55 | |
| *** milan has quit IRC | 14:55 | |
| TheJulia | vsaienk0: thank you | 14:55 |
| sambetts | \o/ | 14:55 |
| *** edand has quit IRC | 14:56 | |
| *** mjura has quit IRC | 14:57 | |
| *** mgoddard_ has joined #openstack-ironic | 15:02 | |
| dtantsur | jroll, have you seen sambetts' mail on driver composition? | 15:04 |
| jroll | dtantsur: I've seen that it exists, I have not yet read it | 15:05 |
| dtantsur | aha | 15:05 |
| sambetts | jroll: its more or less just a cleaned up version of what I had on the etherpad yseterday | 15:05 |
| dtantsur | I just want to know if folks like it, cause I'm going to hack on driver composition | 15:05 |
| jroll | dtantsur: the implicit-ness bugs me a bit, otherwise it seems fine | 15:06 |
| jroll | but as defaults are available via API, not a huge deal | 15:06 |
| *** mgoddard has quit IRC | 15:06 | |
| dtantsur | sambetts, do you feel like proposing an amendment to the spec? | 15:06 |
| dtantsur | (I know you don't, but I guess you have to ;) | 15:06 |
| sambetts | ;) haha sure I' | 15:07 |
| sambetts | I | 15:07 |
| sambetts | -,- | 15:07 |
| sambetts | I'll add another post-it note to my monitor ;) | 15:07 |
| mgould | sambetts: can you still see the screen? :-) | 15:07 |
| dtantsur | hehe | 15:07 |
| dtantsur | ping me for review when you do | 15:07 |
| thiagop | lol | 15:07 |
| dtantsur | right now I'm a bit uncertain which version to hack on | 15:08 |
| sambetts | with all my TODOs its starting to get harder heh | 15:08 |
| sambetts | dtantsur: will do | 15:08 |
| JayF | I think we really underestimate how complex operators will find the "calculated defaults" pattern | 15:09 |
| TheJulia | they will find it horribly complex :( | 15:11 |
| dtantsur | maybe.. I personally find it the most logical thing we can do, but I can see people being confused by it.. | 15:12 |
| JayF | I'd rather an expected, fast fail with a good error message than "magic" trying to make something work, and then it works differently in small unexpected ways | 15:12 |
| jroll | I do still assert that 99% of operators do enrollment with a script and don't change drivers | 15:13 |
| *** liliars has joined #openstack-ironic | 15:13 | |
| * lucasagomes == JayF | 15:13 | |
| JayF | Which is exactly why we should ensure a minor config change or upgrade won't change the behavior of those scripts. | 15:13 |
| sambetts | then they have to explictly define all the interfaces intheir node crate | 15:14 |
| JayF | Or they are populated by inspector | 15:14 |
| JayF | both of those are reasonable things to expect | 15:14 |
| dtantsur | wait, inspector requires a node to be present and working :) | 15:14 |
| sambetts | because the currently defined defaults idea doesn't prevent vendors changeing their mind on which they want to be the default for thier hardware-tpye | 15:14 |
| dtantsur | or are you talking about discovery? | 15:14 |
| JayF | dtantsur: oooh. | 15:14 |
| JayF | dtantsur: that is tricky | 15:14 |
| dtantsur | discovery has its downsides | 15:14 |
| JayF | I thought inspector could choose a more capable driver for a node if it was available? | 15:15 |
| dtantsur | also discovery won't pick a network_interface for you | 15:15 |
| dtantsur | JayF, this is an idea for future improvements :) but still, we can't detect network_interface | 15:15 |
| dtantsur | and without vendor-provided defaults, we will have to store these defaults ourselves | 15:15 |
| JayF | Yeah; I guess. I would just suggest err'ing on the side of explictness rather than making a collection of implicit defaults. | 15:15 |
| dtantsur | not having defaults is a clear no-go.. so we have to figure our which defaults | 15:16 |
| *** d0ugal has quit IRC | 15:16 | |
| sambetts | if you want explict then we have to nuke all defaults in hardware_type | 15:16 |
| dtantsur | -2 | 15:16 |
| sambetts | exactly | 15:16 |
| dtantsur | :) | 15:16 |
| *** links has quit IRC | 15:17 | |
| *** moshele has left #openstack-ironic | 15:17 | |
| *** d0ugal has joined #openstack-ironic | 15:17 | |
| JayF | I'm just saying, we will give unexpected behavior to operators if we head down this path, and it's going to cause them pain | 15:17 |
| sambetts | if a user wants a specifc interface then they can just explictly ask for it in the API, and if its not enabled it will fail early for them | 15:18 |
| JayF | Is it less pain than the alternative? Maybe, I don't really know, but I know when I was a pure "sysadmin" this kind of default-setting-magic was one of my least favorite attributes in software I used. | 15:18 |
| sambetts | if they don't specify it then we pick one that will work for them | 15:18 |
| sambetts | I think doing that is better than endeding up with a node that isn't going to work | 15:19 |
| sambetts | and then having to know to override it anyway | 15:19 |
| JayF | I think ending up with a node that doesn't work is bad behavior; but is there no way to fail the node-create in those cases? | 15:19 |
| jroll | fwiw I fully expect that changing a default should include a deprecation cycle and release notes | 15:20 |
| lucasagomes | sambetts, it feels like we (devs) are saying that we are doing this to help operators but, we have an operator saying that it does not help :-) | 15:20 |
| JayF | lucasagomes: and I was a little worried I was alone until TheJulia agreed, lol | 15:20 |
| * lucasagomes likes predictability as well | 15:20 | |
| lucasagomes | JayF, right | 15:21 |
| *** mtanino has joined #openstack-ironic | 15:22 | |
| sambetts | JayF: wouldn't you prefer that what you specify in enabled_FOO_interfaces obeys what you ask for instead of implictly interfaces being enabled up to of it ? | 15:22 |
| *** ijw has quit IRC | 15:22 | |
| JayF | sambetts: I'm not sure I grok your questino/ | 15:23 |
| sambetts | right now hardware_type defaults are implictly enabled even if they might not work in your environment | 15:23 |
| JayF | so the implicitness I dislike is already present? | 15:23 |
| JayF | Well, not really, right? | 15:24 |
| JayF | Because when I enable a node with a given hardware type, I'm opting-in to that interface, right? | 15:24 |
| sambetts | so I can't deploy Cisco hardware if I only want to support ipmi power because its default power interface isn't ipmi? | 15:24 |
| jroll | the hardware types will have documented defaults that are also visible via the API, and deprecation periods to change those | 15:25 |
| *** fragatina has quit IRC | 15:26 | |
| jroll | JayF: an example of the case sambetts is concerned about: generic ipmi hardware, ipmi isn't allowed in the deployment, snmp is, because ipmitool is the default for the ipmi hardware type it will always be enabled and someone could forget to specify snmp power and wind up with an unusable node | 15:26 |
| JayF | jroll: my opinion would be that in that case, we should deny the node-create | 15:26 |
| JayF | jroll: fail it in the API at create time | 15:26 |
| jroll | JayF: well, because ipmitool is the default power interface, it's implicitly enabled | 15:26 |
| sambetts | even if it doesn't work in the deployment | 15:27 |
| JayF | As an operator, I want to be able to list interfaces I want to support in the config file | 15:27 |
| JayF | and *never ever ever* have a node-create succeed that requires an interface outside of that list | 15:27 |
| jroll | so here's a thought: | 15:28 |
| sambetts | but if you don't ask for a particular interface what does the node-create do? | 15:28 |
| JayF | It would use the default. If the default isn't enabled in my environment; fail the create | 15:28 |
| sambetts | fail if default is not in eabled_interfaces | 15:28 |
| jroll | if hardware type foo has a default power interface bar, and bar is not in enabled_power_interfaces, and power-interface isn't included in the node-create, fail the request | 15:28 |
| * jroll said the same thing as y'all | 15:28 | |
| JayF | ++ | 15:28 |
| JayF | TheJulia: ^ wdyt | 15:29 |
| * TheJulia context switches | 15:29 | |
| sambetts | which means I then have to overide that interface for every node I create with that hardware_type which doesn't align with the convience factor we discussed ysterday | 15:29 |
| dtantsur | jroll, it's good, except for: how will driver vendor passthru work in this case? | 15:30 |
| sambetts | so if I have all generic_ipmi_hardware but only want to support snmp then I always have to override | 15:30 |
| dtantsur | we always use the default vendor interface there | 15:30 |
| JayF | sambetts: As an operator, I don't find that convienient | 15:30 |
| jroll | dtantsur: I guess it wouldn't, we did want to fix that later, right? | 15:31 |
| dtantsur | jroll, yeah, maybe it's not a huge concern... | 15:31 |
| TheJulia | JayF: I guess I could go with that, although the concept is different w/r/t networking since we've always had a default, we just called it something else... kind of | 15:33 |
| *** clenimar_ has joined #openstack-ironic | 15:33 | |
| JayF | With networking, I'd imagine it'd be handled similarly, right? | 15:33 |
| JayF | One systemwide default which has to be explicitly overridden if someone wants a node that's different | 15:33 |
| TheJulia | that is essentially exactly what we have moving forward, although we have an underlying default if it is not defined | 15:35 |
| JayF | I thought what's being proposed is a list of defaults | 15:35 |
| sambetts | right, but we want all hardware_type interfaces to be consistnt, and having a config default for that breaks that idea | 15:35 |
| JayF | and we would default to a list of defaults | 15:35 |
| TheJulia | JayF: list of enabled drivers, single default that can be overriden | 15:36 |
| *** garthb has joined #openstack-ironic | 15:36 | |
| JayF | TheJulia: so if I enable neutron and flat, I'm guaranteed to never get a node on "flat" unless I explicitly configure it that way? | 15:36 |
| TheJulia | JayF: other way around, since flat was the pre-existing default | 15:37 |
| TheJulia | but you can override that | 15:37 |
| JayF | I'm not talking about a default config, I'm talking about a jay config | 15:37 |
| *** bharath has quit IRC | 15:37 | |
| TheJulia | yeah, I know :) | 15:37 |
| JayF | I just think of a world of "oh shit, I just deployed $valuable_data_server" onto a node that implicitly got swapped into flat instead of neutron networking | 15:38 |
| JayF | and I know we're trying to wrap some of that stuff in hardware_type, but that's why I disliked conflating environmental things into hardware_types from the start; it'd be easy for an operator to assume changing the "hardware" setting in a node would not impact how networking happened | 15:39 |
| sambetts | this is why I don't want defaults in the hardware_types | 15:39 |
| sambetts | I only want hardware_tpyes to specify supported implmentatiosn | 15:39 |
| sambetts | for that hardware | 15:39 |
| TheJulia | JayF: fyi, if you've not seen, https://review.openstack.org/#/c/285852/64/ironic/common/driver_factory.py@98 | 15:41 |
| *** krtaylor has joined #openstack-ironic | 15:43 | |
| *** vmud213 has joined #openstack-ironic | 15:43 | |
| *** rloo has joined #openstack-ironic | 15:46 | |
| lucasagomes | mariojv, around? | 15:48 |
| lucasagomes | mariojv, btw, replied to https://review.openstack.org/#/c/338264/5/ironic/tests/unit/drivers/modules/test_deploy_utils.py lemme know if that's fine | 15:49 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Expose node's network_interface field in API https://review.openstack.org/317392 | 15:49 |
| mariojv | lucasagomes: yeah, i just got back | 15:53 |
| mariojv | lucasagomes: +1'ed | 15:54 |
| lucasagomes | cool, thanks | 15:54 |
| *** mbound has joined #openstack-ironic | 15:57 | |
| alaski | I should know this but am not sure: is there anything about a Nova flavor that is specific to a baremetal instance? in other words is the flavor necessarily different from a flavor for a VM? | 15:58 |
| *** fragatina has joined #openstack-ironic | 15:59 | |
| thiagop | alaski: the cpu_arch parameter, basically | 16:00 |
| *** mbound has quit IRC | 16:02 | |
| alaski | thiagop: okay. and what does it look like for baremetal flavors? | 16:02 |
| vsaienk0 | alaski | 16:02 |
| thiagop | alaski: x86_64 | 16:03 |
| thiagop | VM's doesn't have this parameter, normally | 16:03 |
| thiagop | this is a parameter that we use not to trigger a build on any conductor, just the ones running the ironic virt driver | 16:04 |
| vsaienk0 | alaski: cpu_arch is an optional parameter. The flavor looks like nova flavor-create bm_flavor auto 3072 150 1 | 16:04 |
| *** mgoddard_ has quit IRC | 16:04 | |
| *** mgoddard_ has joined #openstack-ironic | 16:05 | |
| *** Linkmark has quit IRC | 16:05 | |
| alaski | so a VM flavor could have cpu_arch and typically doesn't, but a baremetal flavor must have it | 16:07 |
| vsaienk0 | alaski: no cpu_arch is an optional parameter for baremetal flavor | 16:08 |
| lucasagomes | +1 https://github.com/openstack/nova/blob/master/nova/virt/ironic/driver.py#L220-L221 | 16:08 |
| alaski | vsaienk0: that's helpful, thanks | 16:08 |
| alaski | thiagop: thanks for the info | 16:08 |
| *** Romanenko_K has quit IRC | 16:09 | |
| alaski | what I'm trying to determine is if there's a way for Nova to know whether a flavor is intended for Ironic or a another virt driver, and it doesn't seem like there is atm | 16:09 |
| vsaienk0 | alaski: in multi hypervisor environment nova availability zones may be used to define where to build an instance, or image property hypervisor_type=baremetal - for ironic case | 16:10 |
| thiagop | lucasagomes: L220-221 isn't for the cpu_arch on the node object? | 16:10 |
| rloo | wrt nova flavor, these are the 'instructions' we have: http://docs.openstack.org/developer/ironic/deploy/install-guide.html#flavor-creation | 16:11 |
| rloo | so cpu_arch in extra_specs is the only thing 'extra'. | 16:11 |
| alaski | vsaienk0: is the image property required, or just intended for use with an aggregate? | 16:11 |
| lucasagomes | thiagop, yes from ironic | 16:12 |
| thiagop | I was talking about flavors | 16:13 |
| alaski | rloo: that doc is useful, thanks | 16:13 |
| *** e0ne has quit IRC | 16:13 | |
| vsaienk0 | alaski, no it is not required parameter but in multi-typeyhypervisor setup it may be used during scheduling. | 16:13 |
| lucasagomes | thiagop, pointing out that having a cpu_arch is not required in Ironic, so shouldn't be in the flavor either | 16:13 |
| lucasagomes | it's optional | 16:13 |
| alaski | I may have artificially constrained myself to thinking about flavor. More generally is it possible within Nova to know whether a build is intended for a baremetal node, before it gets to n-cpu? | 16:14 |
| alaski | it seems like not, which on the one hand is good since VM boot and baremetal boots look generally the same \o/ | 16:15 |
| thiagop | lucasagomes: strange, last time I tried to boot an instance without this parameter, nova threw it to a VM | 16:15 |
| thiagop | maybe something changed in the virt driver since then | 16:15 |
| alaski | on the other hand having that distinction would help out with something I'm looking at | 16:15 |
| *** Nisha_away has joined #openstack-ironic | 16:16 | |
| jroll | alaski: I don't believe it is - the only way we do it at rackspace is by flavor classes | 16:16 |
| rloo | lucasagomes: just read what you said. We should change the wording in the install guide then. | 16:16 |
| *** milan has joined #openstack-ironic | 16:16 | |
| *** vmud213 has quit IRC | 16:16 | |
| lucasagomes | rloo, I quickly skimmed the code, maybe we should test it | 16:17 |
| lucasagomes | but I doesn't seem to be mandatory | 16:17 |
| alaski | jroll: gotcha | 16:17 |
| rloo | lucasagomes: yeah, if you have time to test. i'm swamped today (want to get as many networking patches reviewed before I go on vacation) | 16:18 |
| jroll | alaski: curious what you're hacking on :) | 16:18 |
| alaski | jaypipes had an idea for resource providers and ironic that some of us have been talking through, but it's predicated on being able to make this distinction | 16:18 |
| alaski | I think there are other approaches though | 16:19 |
| alaski | jroll: are you going to be at the nova midcycle/ | 16:19 |
| alaski | ? | 16:19 |
| jroll | alaski: yeah | 16:20 |
| *** sdake__ is now known as sdake | 16:20 | |
| JayF | I think jroll , me, devananda are going to be there repping ironic | 16:20 |
| rloo | there isn't anything to say that we cannot do something to flavors to make it unique to ironic, but i'd rather not if we don't see a good reason for it | 16:20 |
| lucasagomes | rloo, added a todo here | 16:20 |
| alaski | JayF: excellent | 16:20 |
| rloo | thx lucasagomes | 16:20 |
| alaski | rloo: I agree | 16:21 |
| rloo | alaski: we made sure to send our heavyweights to the nova midcycle :D | 16:21 |
| *** rcernin has quit IRC | 16:21 | |
| alaski | heh | 16:21 |
| jroll | hey, who you callin heavy | 16:22 |
| lucasagomes | rloo, | 9e7cf47c-7c52-49ea-bf7c-d37317dcddf9 | node-1 | 25a89246-57b2-448a-b77a-566392ba7775 | power off | deploying | False | | 16:22 |
| * rloo doesn't hear jroll's question | 16:23 | |
| lucasagomes | rloo, created a flavor without cpu_arch, removed the cpu_arch from one node and set the other in maintenance | 16:23 |
| lucasagomes | scheduler pick it and is deploying onto it, I will let you know if that works | 16:23 |
| lucasagomes | (note: that env is not that new, maybe ~2week old) | 16:23 |
| alaski | thanks for the info everyone. this helps me knock out one potential approach before we spend too much time on it | 16:24 |
| *** trown is now known as trown|lunch | 16:24 | |
| lucasagomes | rloo, | 9e7cf47c-7c52-49ea-bf7c-d37317dcddf9 | node-1 | 25a89246-57b2-448a-b77a-566392ba7775 | power on | active | False | | 16:25 |
| lucasagomes | thiagop, ^ works without cpu_arch (no cpu_arch in the flavor nor in the node) | 16:25 |
| thiagop | lucasagomes: cool, good to know | 16:25 |
| jroll | devananda: hey, just realized I'll be on the plane to portland during our meeting monday. will you be around to cover? | 16:25 |
| rloo | lucasagomes: thx for checking | 16:25 |
| thiagop | lucasagomes: but now I'm really confused. If I do not set an availability zone, users may end up deploying an Ironic instance instead of a VM without knowing? | 16:27 |
| *** piet has joined #openstack-ironic | 16:27 | |
| sambetts | thiagop: an AZ or host agregate | 16:28 |
| lucasagomes | thiagop, if you have a hybrid environment (vm and baremetals) you need to work around such things, cells, AZ, host agregate | 16:28 |
| lucasagomes | on the bright side, the nova Ironic driver is doing what it's suppose to do: making baremetal look&feel exactly like VMs for the end-user | 16:29 |
| sambetts | \o/ | 16:29 |
| sambetts | the only problem is the potential for scheduling a smaller flavor on to a big ironic node, but thats just down to schedular filters | 16:29 |
| *** ijw has joined #openstack-ironic | 16:30 | |
| thiagop | sambetts: Yeah, the virt driver uses that exact filters, right? | 16:30 |
| *** daemontool has quit IRC | 16:30 | |
| lucasagomes | sambetts, yeah, that's still a problem... user may ask for X and get Y depending on the fielter | 16:31 |
| sambetts | thiagop: with Ironic we suggest you use the exact filters to prevent that yup :) but if you use host aggregates to their full potential you can avoid it with those instead of the exact filders | 16:31 |
| lucasagomes | since with baremetal you always get 100% | 16:31 |
| thiagop | But I keep finding it strange anyway... | 16:31 |
| * sambetts is looking forward to multi nova compute + resource pooling for this sort of thing | 16:32 | |
| dtantsur | see you tomorrow | 16:32 |
| *** dtantsur is now known as dtantsur|afk | 16:32 | |
| sambetts | o/ dtantsur|afk | 16:32 |
| thiagop | night dtantsur|afk | 16:32 |
| *** ijw has quit IRC | 16:32 | |
| lucasagomes | dtantsur|afk, see ya | 16:33 |
| * lucasagomes is heading soon too | 16:33 | |
| sambetts | same | 16:33 |
| devananda | morning, all | 16:34 |
| thiagop | night lucasagomes sambetts , morning devananda | 16:34 |
| lucasagomes | devananda, morning | 16:35 |
| lucasagomes | calling it a day folks | 16:35 |
| *** ijw has joined #openstack-ironic | 16:35 | |
| lucasagomes | have a great evening all :-) | 16:35 |
| lucasagomes | talk soon | 16:35 |
| *** lucasagomes is now known as lucas|afk | 16:35 | |
| NobodyCam | Good morning Ironicers | 16:36 |
| thiagop | am I crazy or the devstack is logging less output these days? | 16:36 |
| thiagop | Morning NobodyCam | 16:36 |
| NobodyCam | morning thiagop | 16:37 |
| openstackgerrit | Merged openstack/ironic: Update Ironic VM network connection https://review.openstack.org/256366 | 16:39 |
| mgould | morning NobodyCam | 16:41 |
| *** vmud213 has joined #openstack-ironic | 16:41 | |
| NobodyCam | morning mogul :) | 16:41 |
| NobodyCam | gah | 16:41 |
| rloo | morning devananda. Would be worth getting your ack on the direction of https://review.openstack.org/#/c/317392 | 16:41 |
| NobodyCam | mgould: | 16:41 |
| jroll | devananda: hey, just realized I'll be on the plane to portland during our meeting monday. will you be around to cover? | 16:41 |
| mgould | morning devananda , night lucas|afk | 16:41 |
| NobodyCam | morning jroll rloo devananda TheJulia :) | 16:42 |
| jroll | mornign NobodyCam :) | 16:42 |
| rloo | morning NobodyCam | 16:42 |
| NobodyCam | :) | 16:43 |
| devananda | jroll: should be fine | 16:43 |
| jroll | devananda: awesome, thanks | 16:43 |
| JayF | I'll be around for the monday meeting too, I'm just driving down to portland tuesday morning | 16:43 |
| *** baoli_ has quit IRC | 16:48 | |
| *** d0ugal has quit IRC | 16:49 | |
| devananda | JayF: I'm driving down as well | 16:49 |
| JayF | devananda: any advice? I'm thinking since I'm going south from Tacoma it shouldn't take much longer to get there than google maps says | 16:49 |
| jroll | I'd assume there will be traffic at 8 or 9 am or whatever that might not show on google maps when you're leaving | 16:51 |
| *** derekh has quit IRC | 16:51 | |
| devananda | once your out of tacoma, the roads should be pretty clear until portland | 16:51 |
| JayF | Yeah; that's the impression I got from how rural it is on the way up | 16:52 |
| JayF | I'm kinda in the southwest part of tacoma (near university place), so I'm even clear of most of the tacoma traffic | 16:52 |
| *** baoli has joined #openstack-ironic | 16:57 | |
| *** jpich has quit IRC | 16:57 | |
| milan | sambetts, ifarkas, aarefiev, dtantsur|afk and inspector folks, our experimental grenade job doesn't execute smoke tests after upgrade; see https://review.openstack.org/#/c/342275/ please, review | 16:57 |
| *** ifarkas has quit IRC | 16:57 | |
| rloo | vsaienk0: apart from the merge conflict, is there more work to be done on https://review.openstack.org/#/c/332177/? wondering if i should review or wait | 16:58 |
| sambetts | milan: please review this if you have a chance, I'd like to get that in, and any changes to the Ironic related jobs in devstack-gate cause me painful merges :-P | 16:59 |
| *** sdake has quit IRC | 17:00 | |
| rloo | jroll, devananda, JayF: not sure what your 'agenda' if any, might be for nova midcycle next week. are there any patches that you think ought to be reviewed before then? | 17:00 |
| JayF | I was hoping the rescue stuff could've been ready by then, but the spec isn't even merged so I think that's off :( | 17:01 |
| milan | sambetts, which pull request you mean? | 17:01 |
| jroll | rloo: if we get networking or serial console stuff (including client changes) I'd love to try to bribe someone into working around the FF | 17:01 |
| JayF | The only other bit is compute interactions, which I'm getting somewhat up to speed on | 17:01 |
| rloo | JayF: sorry, i have been focusing on the network stuff. | 17:01 |
| sambetts | milan: https://review.openstack.org/#/c/340567/ | 17:01 |
| jroll | rloo: my primary goals are around multiple-compute stuff | 17:01 |
| sambetts | milan: sorry it didn't paste.. | 17:01 |
| rloo | jroll: OH. i can look at the console stuff. i +A'd one this AM. | 17:01 |
| openstackgerrit | Chris Krelle proposed openstack/ironic: Update devstack section of quickstart to use agent_ipmitool https://review.openstack.org/341801 | 17:01 |
| milan | sambetts, OK will do | 17:02 |
| jroll | rloo: awesome \o/ | 17:02 |
| *** sdake has joined #openstack-ironic | 17:02 | |
| NobodyCam | ok pep8 checks pass locally | 17:02 |
| milan | sambetts, tomorrow :P | 17:02 |
| sambetts | milan: hehe ;) | 17:03 |
| *** rama_y has joined #openstack-ironic | 17:03 | |
| *** jtomasek has quit IRC | 17:04 | |
| *** jrist has quit IRC | 17:04 | |
| jlvillal | sambetts: Your interface names remind me of a Lady Gaga song. | 17:05 |
| sambetts | jlvillal: haha ;) | 17:05 |
| jlvillal | sambetts: BAR, RAR, RAR, RAR.. | 17:05 |
| jlvillal | sambetts: http://www.azlyrics.com/lyrics/ladygaga/badromance.html :) | 17:06 |
| jlvillal | sambetts: And thanks for the email. Makes sense at first read to me. | 17:07 |
| jlvillal | sambetts: Though what happens if one says: [FOO, SPAM] and the other list is [SPAM, FOO] ? How is the order decided? | 17:08 |
| sambetts | jlvillal: the order only comes from the hardware_type list not the config file list | 17:08 |
| jlvillal | sambetts: Okay. Thanks. | 17:08 |
| jlvillal | milan: I feel bad for not paying enough attention to devstack-gate-test. I'll try to look at your pull requests while I am on vacation! | 17:10 |
| *** athomas has quit IRC | 17:14 | |
| *** vishwanathj has quit IRC | 17:15 | |
| *** PollyZ has quit IRC | 17:15 | |
| *** mgoddard_ has quit IRC | 17:16 | |
| *** Nisha_away has quit IRC | 17:17 | |
| *** PollyZ has joined #openstack-ironic | 17:17 | |
| thiagop | Found the problem with devstack | 17:18 |
| *** e0ne has joined #openstack-ironic | 17:19 | |
| *** ohamada has quit IRC | 17:19 | |
| thiagop | jlvillal: append the last phrase with "drinking piña colada by the beach" :) | 17:19 |
| jlvillal | thiagop: Uh. Not too many beaches in Moscow ;) | 17:20 |
| jlvillal | Maybe "river beaches" :) | 17:20 |
| thiagop | jlvillal: joke /fail | 17:20 |
| thiagop | lol | 17:20 |
| jlvillal | heh | 17:20 |
| * sambetts -> home | 17:22 | |
| sambetts | night everyone | 17:22 |
| sambetts | o/ | 17:22 |
| *** sambetts is now known as sambetts|afk | 17:22 | |
| *** PollyZ has quit IRC | 17:25 | |
| *** harlowja has joined #openstack-ironic | 17:28 | |
| *** e0ne has quit IRC | 17:29 | |
| *** PollyZ has joined #openstack-ironic | 17:30 | |
| NobodyCam | night sambetts|afk | 17:31 |
| *** rcernin has joined #openstack-ironic | 17:32 | |
| *** PollyZ has quit IRC | 17:35 | |
| *** sabeen has joined #openstack-ironic | 17:41 | |
| *** e0ne has joined #openstack-ironic | 17:41 | |
| *** sdake_ has joined #openstack-ironic | 17:42 | |
| *** sabeen1 has joined #openstack-ironic | 17:43 | |
| *** tesseract- has quit IRC | 17:44 | |
| *** PollyZ has joined #openstack-ironic | 17:44 | |
| *** sdake has quit IRC | 17:44 | |
| *** sabeen has quit IRC | 17:45 | |
| NobodyCam | jroll: looks like only the first ones needed to be escaped ... | 17:45 |
| *** trown|lunch is now known as trown | 17:47 | |
| *** sabeen1 has quit IRC | 17:47 | |
| *** PollyZ has quit IRC | 17:49 | |
| *** mbound has joined #openstack-ironic | 17:49 | |
| jroll | NobodyCam: weird, I'm wondering if the pxe_ part is going to try to turn into a link | 17:54 |
| NobodyCam | I looked at the rendered version and it looked ok | 17:55 |
| jroll | rather generate a warning | 17:55 |
| NobodyCam | I put the link in my reply | 17:55 |
| * jroll looks | 17:55 | |
| jroll | oh we don't error on warning in ironic, yay | 17:55 |
| jroll | (there's currently a bug where that setting doesn't work, is why I'm wondering) | 17:56 |
| jroll | +2 | 17:56 |
| NobodyCam | :) | 17:56 |
| *** PollyZ has joined #openstack-ironic | 17:57 | |
| *** catintheroof has quit IRC | 17:58 | |
| rloo | jroll: I took a look at the console patch, needs some work (and if you read my comment you can see I am a bit annoyed): https://review.openstack.org/#/c/293873 | 18:00 |
| rloo | jroll: nothing that someone else cannot revise so it could be ready by next Tuesday, but I won't be here to monitor/baby it along. | 18:01 |
| *** sdake has joined #openstack-ironic | 18:02 | |
| rloo | jroll: the nova side of the console stuff i am ok with: https://review.openstack.org/#/c/328157/ | 18:02 |
| jroll | rloo: cool, I will look out for it tomorrow | 18:03 |
| jroll | thank you :) | 18:03 |
| *** sdake_ has quit IRC | 18:03 | |
| jroll | sorry for the annoyance, that follow-on bothered me as well | 18:03 |
| *** rajinir has joined #openstack-ironic | 18:10 | |
| rloo | jroll: i saw at least 2 people asking about combining the follow-on, so i see no reason why they didn't do it | 18:10 |
| *** vmud213 has quit IRC | 18:11 | |
| jroll | rloo: looks like they did some but not all | 18:14 |
| jroll | well, maybe nothing but the reno | 18:14 |
| jroll | :( | 18:14 |
| rloo | jroll: that is one of my points, that *we* should not be spending our time figuring out what *they* did or did not do | 18:15 |
| rloo | jroll: I was tempted to -2 and tell them not to ping me until they were absolutely sure it was ok | 18:15 |
| jroll | rloo: I don't think you should be spending time making sure other reviewer's comments were addressed - that's for the other reviewer to do | 18:15 |
| rloo | jroll: i know but sometimes i like to see what others have asked and i like to see a reply | 18:16 |
| jroll | rloo: though, it's important and we're generally bad at following up on our own -1s, so I see why you're doing that :/ | 18:16 |
| jroll | yeah | 18:16 |
| *** sdake_ has joined #openstack-ironic | 18:16 | |
| rloo | jroll: and anyway, i was annoyed. so... | 18:16 |
| *** PollyZ_ has joined #openstack-ironic | 18:16 | |
| jroll | yeah | 18:17 |
| *** sdake has quit IRC | 18:18 | |
| rajinir | thiago, sambetts, updated https://wiki.openstack.org/wiki/Ironic/Testing | 18:19 |
| *** PollyZ has quit IRC | 18:19 | |
| * thiagop looks | 18:20 | |
| rajinir | thiago, sambetts, watanabe_isao please review | 18:20 |
| openstackgerrit | Rajini Ram proposed openstack/ironic: Thirdparty CI link https://review.openstack.org/342336 | 18:22 |
| thiagop | rajinir: I think the example is inverted | 18:22 |
| thiagop | rajinir: should be "vendor-recheck" | 18:22 |
| rajinir | thiago: fixed. good catch | 18:23 |
| rajinir | thiago: Also review https://review.openstack.org/342336 | 18:23 |
| thiagop | rajinir: I think there is something broken with the wiki syntax too. I'm seeing a lot of broken code areas after "delete the line" | 18:24 |
| *** mgould is now known as mgould|afk | 18:24 | |
| *** jaybeale has joined #openstack-ironic | 18:24 | |
| rajinir | thiagop: let me fix those syntax , there is only spaces. wierd | 18:24 |
| thiagop | maybe a terminator that wasn't added... | 18:25 |
| *** ijw has quit IRC | 18:26 | |
| *** PollyZ_ has quit IRC | 18:26 | |
| rajinir | thiagop: now | 18:28 |
| *** ijw has joined #openstack-ironic | 18:29 | |
| *** ijw has quit IRC | 18:29 | |
| thiagop | rajinir: gone. Thanks | 18:30 |
| *** ijw has joined #openstack-ironic | 18:30 | |
| rloo | jroll: wrt https://review.openstack.org/#/c/339197/, were you going to update it? or would you like me to? | 18:31 |
| jroll | rloo: oops, I forgot about it, I can | 18:31 |
| jroll | or rather, I thought I updated it | 18:31 |
| rloo | jroll: pepp8 failure. | 18:32 |
| thiagop | rajinir: that hack for require-approval is awesome, but zuul will wait for a verified to start running or it'll run only if the patchset was approved before the patchset arrives? | 18:32 |
| jroll | rloo: yeah, I see | 18:32 |
| rloo | jroll: hardware.py, line 363 too long | 18:32 |
| rajinir | thiago: Have to give credit to watanabe_isao for that tip | 18:32 |
| rajinir | thiago: Not sure what the zuul behavior will be, I would imagine it will get notified only after Jenkins +1s | 18:34 |
| thiagop | rajinir: but the comment message will not be the same, don't know... | 18:34 |
| thiagop | rajinir: I think it worth adding an example of how the regex for comments should be. See http://paste.openstack.org/show/532905/ | 18:35 |
| *** wajdi has joined #openstack-ironic | 18:35 | |
| *** bvandewa has joined #openstack-ironic | 18:36 | |
| thiagop | rajinir: and in the last line, some can use devstack-gate to build it too | 18:36 |
| thiagop | rajinir: good job, I wouldn't write all that good stuff from my head:) | 18:36 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic-python-agent: Fix doc warnings https://review.openstack.org/339197 | 18:37 |
| jroll | rloo: done | 18:37 |
| rloo | jroll: looking... | 18:37 |
| *** bvandewa has quit IRC | 18:37 | |
| rajinir | thiagop: updated your regex comments | 18:38 |
| wajdi | good afternoon! Been a while since I've been here. Hope you're all well! :) | 18:38 |
| thiagop | Hello wajdi | 18:39 |
| rajinir | thiagop: where does the use devstack-gate build go? Can you explain | 18:39 |
| NobodyCam | hey hey wajdi WB | 18:40 |
| wajdi | thanks! :) | 18:40 |
| *** Sukhdev has joined #openstack-ironic | 18:41 | |
| *** mbound_ has joined #openstack-ironic | 18:42 | |
| *** mbound has quit IRC | 18:43 | |
| *** Sukhdev has quit IRC | 18:44 | |
| *** Sukhdev has joined #openstack-ironic | 18:45 | |
| thiagop | rajinir: where you suggest using devstack plugin, in the very last line | 18:49 |
| thiagop | one can use devstack-gate too | 18:49 |
| rloo | jroll, TheJulia, NobodyCam, does anyone remember. If we deprecate a config option, do we log a msg if that config is being used during deprecation period? | 18:50 |
| *** mbound_ has quit IRC | 18:50 | |
| NobodyCam | I think that would be the 'nice' thing to do.. | 18:51 |
| rajinir | thiago: got it, updated. I didn't touch that faq before | 18:51 |
| jroll | rloo: oslo.config will automatically log it if it is set, I believe | 18:51 |
| jroll | rloo: s/log it/warn/ | 18:52 |
| rloo | jroll: ok | 18:52 |
| *** PollyZ has joined #openstack-ironic | 18:53 | |
| *** sambetts|afk has quit IRC | 18:54 | |
| *** bvandewa has joined #openstack-ironic | 18:54 | |
| *** dprince has quit IRC | 18:54 | |
| *** sambetts has joined #openstack-ironic | 18:56 | |
| *** e0ne has quit IRC | 18:57 | |
| *** jaybeale has quit IRC | 18:57 | |
| *** adu has quit IRC | 18:58 | |
| *** e0ne has joined #openstack-ironic | 18:58 | |
| *** sukhdev_ has joined #openstack-ironic | 19:03 | |
| mariojv | question about rloo's comment on https://review.openstack.org/#/c/339852/ | 19:07 |
| mariojv | is recheck different from reverify? i've seen both | 19:07 |
| rama_y | jroll, rloo: there are 3 toctree warnings that are not addressed in this patch: https://review.openstack.org/#/c/339852/ | 19:08 |
| jroll | mariojv: nope, it used to be though | 19:08 |
| mariojv | nice | 19:09 |
| rama_y | jroll, rloo: doc/source/api/autoindex.rst, doc/source/deploy/radosgw.rst, doc/source/releasenotes/index.rst | 19:09 |
| jroll | rama_y: should I block it from landing? :) | 19:09 |
| jroll | rama_y: you can also block it, by voting W-1 | 19:09 |
| JayF | I mean, why not let it land and followup those 3 fixes? | 19:10 |
| jroll | that's fine too | 19:10 |
| rama_y | jroll: this patch can land; since I wasn't clear on where to index these three .rst files, I didn't include them in this patch. | 19:10 |
| jroll | rama_y: ah, I see, that's a good question (e.g. releasenotes is published elsewhere) | 19:11 |
| *** rloo has quit IRC | 19:13 | |
| *** rloo has joined #openstack-ironic | 19:14 | |
| *** catintheroof has joined #openstack-ironic | 19:18 | |
| *** sdake__ has joined #openstack-ironic | 19:19 | |
| *** sdake_ has quit IRC | 19:21 | |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Expose node's network_interface field in API https://review.openstack.org/317392 | 19:21 |
| JayF | https://review.openstack.org/#/c/171878/ (rescue spec) has 1x+2 and 3x+1 if someone could have a look and help me land it I'd be very grateful. | 19:21 |
| *** mbound has joined #openstack-ironic | 19:22 | |
| *** e0ne has quit IRC | 19:24 | |
| *** sdake_ has joined #openstack-ironic | 19:25 | |
| *** mbound has quit IRC | 19:26 | |
| *** PollyZ has quit IRC | 19:26 | |
| rloo | NobodyCam: if you are around, let me know about my question on https://review.openstack.org/#/c/341801/7/doc/source/dev/dev-quickstart.rst | 19:27 |
| NobodyCam | my me take a look | 19:28 |
| *** sdake__ has quit IRC | 19:28 | |
| NobodyCam | rloo: not really sure. but when I ran pep checks I got: doc/source/dev/dev-quickstart.rst:379: D000 Unknown target name: "agent". | 19:30 |
| *** PollyZ has joined #openstack-ironic | 19:31 | |
| rajinir | thiago: https://review.openstack.org/#/c/342281/ | 19:31 |
| rajinir | thiagop:https://review.openstack.org/#/c/342281/ your comment on this review, can you elaborate. We encountered this and that's why we submitted the patch | 19:32 |
| *** fragatina has quit IRC | 19:32 | |
| rloo | NobodyCam: I'm not an .rst expert. seems odd though. | 19:32 |
| rajinir | thiagop: Its in the plugin , but how does it get used in the manager.py | 19:33 |
| NobodyCam | I'm happy to change, but this appears to render correctly | 19:34 |
| *** _vsaienko has joined #openstack-ironic | 19:34 | |
| *** _vs_ has joined #openstack-ironic | 19:34 | |
| rloo | NobodyCam: i don't know whether to approve or not, and i don't have time to look into it so will leave for someone else to approve or not. i mean, it is only doc, but i don't like putting in something if we don't know why. | 19:35 |
| NobodyCam | I'm about to run and grab lunch. I will try and look into when I'm back | 19:36 |
| *** _vs_ has left #openstack-ironic | 19:37 | |
| * devananda runs errands and lunches late | 19:37 | |
| _vsaienko | morning NobodyCam, rloo, devananda | 19:38 |
| rloo | hi _vsaienko | 19:38 |
| _vsaienko | rloo I have resolved you comments https://review.openstack.org/#/c/317392/, once you have a time please have look at the new version | 19:38 |
| rloo | _vsaienko: thx. will try to look soon. | 19:39 |
| _vsaienko | rloo thank you! | 19:39 |
| *** e0ne has joined #openstack-ironic | 19:39 | |
| rloo | _vsaienko: do you know the status of the following patch https://review.openstack.org/#/c/332177/ | 19:39 |
| *** e0ne has quit IRC | 19:39 | |
| *** e0ne has joined #openstack-ironic | 19:40 | |
| *** fragatina has joined #openstack-ironic | 19:40 | |
| openstackgerrit | Rajini Ram proposed openstack/ironic: Thirdparty CI link https://review.openstack.org/342336 | 19:40 |
| _vsaienko | rloo it needs rebasing, and testing | 19:42 |
| rloo | _vsaienko: so the testing part is why it is WIP? | 19:43 |
| *** Goneri has quit IRC | 19:43 | |
| _vsaienko | rloo I don't know, vdrok split portgroup part to single patch, probably it is WIP due to lack of testing from his side | 19:44 |
| rloo | _vsaienko: ok. i probably don't have time to review that today anyway. we're getting close! | 19:45 |
| _vsaienko | rloo, I will try to look on it tomorrow | 19:45 |
| rloo | _vsaienko: that would be good, so as not to prevent people from reviewing (if they want to review). I am away after today, back next Thurs. | 19:46 |
| *** adu has joined #openstack-ironic | 19:49 | |
| *** adu has quit IRC | 19:53 | |
| *** _vsaienko has quit IRC | 19:53 | |
| *** _vsaienko has joined #openstack-ironic | 19:53 | |
| * TheJulia re-appears from being outside in the crazy North Carolina heat | 19:58 | |
| *** ametts has quit IRC | 20:00 | |
| *** PollyZ has quit IRC | 20:03 | |
| *** fragatina has quit IRC | 20:04 | |
| *** fragatina has joined #openstack-ironic | 20:06 | |
| *** ijw has quit IRC | 20:09 | |
| *** adu has joined #openstack-ironic | 20:09 | |
| *** ametts has joined #openstack-ironic | 20:15 | |
| *** _vsaienko has quit IRC | 20:15 | |
| *** _vsaienko has joined #openstack-ironic | 20:16 | |
| *** _vsaienko has quit IRC | 20:25 | |
| *** _vsaienko has joined #openstack-ironic | 20:26 | |
| *** ijw has joined #openstack-ironic | 20:26 | |
| openstackgerrit | Akilan Pughazhendi proposed openstack/ironic: Grammar fix in code contribution guide https://review.openstack.org/342393 | 20:27 |
| *** _vsaienko has quit IRC | 20:29 | |
| openstackgerrit | Merged openstack/ironic: Deprecate [ilo]/clean_priority_erase_devices config https://review.openstack.org/342114 | 20:33 |
| thiagop | rajinir: Hi. Answered on the review | 20:38 |
| thiagop | rajinir: The tempest plugin mimics all the tempest module/class structure | 20:39 |
| rajinir | thiagop: Is it enabled using enable "enable_plugin ironic git://git.openstack.org/openstack/ironic" | 20:40 |
| *** e0ne has quit IRC | 20:40 | |
| rajinir | thiagop: I see, if you don't enable plugin, it is running the tests form tempest code right | 20:41 |
| thiagop | rajinir: I think tempest somehow discovers the plugins in all projects that are installed and superseed over its own when available | 20:42 |
| rajinir | thiagop: I don't see that happening in our builds. So we are missing something | 20:43 |
| jroll | rajinir: that line enables the devstack plugin | 20:43 |
| jroll | rajinir: tempest plugins are enabled by the package (in this case, ironic) being installed alongside tempest | 20:43 |
| jroll | in the normal devstack case, tempest runs in a venv with nothing else installed, so you need ironic installed in that venv | 20:44 |
| thiagop | the "enable_plugin" enables the plugin of ironic for devstack. Without this line, you keep running the devstack/lib/ironic script to build the environment | 20:44 |
| *** ijw has quit IRC | 20:45 | |
| jroll | rajinir: you need something like this https://github.com/openstack-infra/project-config/blob/master/jenkins/jobs/devstack-gate.yaml#L723-L730 | 20:45 |
| rajinir | jroll: I have it enabled, but still the devstack-gate was running the code under tempest | 20:45 |
| thiagop | the takeaway is that either devstack/lib/ironic and the baremetal tests of tempest will not be maintained, right jroll ? | 20:45 |
| jroll | thiagop: devstack/lib/ironic is already gone from devstack, we're working on removing our tests from tempest | 20:45 |
| thiagop | jroll: ack | 20:46 |
| thiagop | rajinir: that's strange... | 20:46 |
| rajinir | jroll,thiagop if it is not maintained, it might be a good idea to remove | 20:47 |
| thiagop | rajinir: I think it's there so long as we keep running tests for the past supported versions of openstack | 20:47 |
| jroll | rajinir: devstack/lib/ironic is already gone from devstack, we're working on removing our tests from tempest | 20:47 |
| rajinir | jroll, thiagop: ok, thanks | 20:48 |
| thiagop | not at all | 20:48 |
| jroll | no problem | 20:48 |
| * thiagop thought that the plugin discovery was something more magic... | 20:50 | |
| jroll | nope, it just loads everything under a given entrypoint | 20:52 |
| jroll | thiagop: https://github.com/openstack/tempest/blob/master/tempest/test_discover/plugins.py#L74 | 20:52 |
| thiagop | rajinir: just another hint: there is no way to pass the deploywait_timeout through devstack-gate, as we do with active_timeout (that is implied from DEVSTACK_GATE_TEMPEST_BAREMETAL_BUILD_TIMEOUT) | 20:56 |
| thiagop | rajinir: you have to put a post-extra directive to set it directly on tempest.conf on your DEVSTACK_LOCAL_CONFIG | 20:57 |
| rajinir | thiagop: ok, thanks. That's what I have odne | 20:57 |
| thiagop | post-extra because post-config runs before tempest is installed | 20:57 |
| thiagop | jroll: besides the fact that we may need the tests for the stable branches of Ironic, what do we need to wipe them from tempest? | 20:59 |
| jroll | thiagop: we don't need them for stable anymore, we use the master branch tempest plugin | 21:00 |
| jroll | but there's two other pieces | 21:00 |
| jroll | 1) dropping the tests themselves, easy | 21:00 |
| jroll | 2) dropping a bunch of hacks we have in there, not easy | 21:00 |
| jroll | like, hacks in normal nova tests | 21:00 |
| jroll | we need to convert them to feature flags and such instead of "if ironic:" | 21:00 |
| thiagop | uhmmm | 21:01 |
| *** caiobo-afk has quit IRC | 21:02 | |
| jroll | ? | 21:02 |
| jroll | thiagop: does that make sense? | 21:03 |
| thiagop | looking... | 21:03 |
| jroll | ah | 21:03 |
| thiagop | I think I need to understand better how these tests depends from each other to put a hand on this | 21:06 |
| *** piet has quit IRC | 21:09 | |
| *** trown is now known as trown|outtypewww | 21:11 | |
| apughazh | Hey guys, just submitted my first patch! | 21:14 |
| apughazh | https://review.openstack.org/#/c/342393/ | 21:14 |
| *** moshele has joined #openstack-ironic | 21:14 | |
| jroll | apughazh: \o/ | 21:16 |
| jroll | +A | 21:16 |
| *** adu has quit IRC | 21:16 | |
| NobodyCam | apughazh: awesome welcome to the community | 21:17 |
| apughazh | Thanks guys! | 21:17 |
| jroll | and gals :) | 21:18 |
| thiagop | sambetts: for when you come back: are tempest-dsvm-ironic-pxe_iscsi_cimc and tempest-dsvm-ironic-pxe_ucs running for doc patches on purpose? | 21:19 |
| thiagop | sambetts: saw here: https://review.openstack.org/#/c/342336/ | 21:19 |
| *** moshele has quit IRC | 21:22 | |
| *** ccamacho is now known as ccamacho|out | 21:22 | |
| *** rcernin has quit IRC | 21:24 | |
| *** bsilverman has joined #openstack-ironic | 21:24 | |
| *** bsilverman has left #openstack-ironic | 21:25 | |
| *** [1]cdearborn has joined #openstack-ironic | 21:28 | |
| *** hparekh has quit IRC | 21:30 | |
| *** cdearborn has quit IRC | 21:31 | |
| *** adu has joined #openstack-ironic | 21:33 | |
| *** ijw has joined #openstack-ironic | 21:35 | |
| *** ijw has quit IRC | 21:36 | |
| *** ijw has joined #openstack-ironic | 21:37 | |
| *** catintheroof has quit IRC | 21:40 | |
| *** wajdi_ has joined #openstack-ironic | 21:43 | |
| *** hparekh has joined #openstack-ironic | 21:43 | |
| *** wajdi has quit IRC | 21:46 | |
| *** wajdi_ has quit IRC | 21:48 | |
| *** wajdi has joined #openstack-ironic | 21:49 | |
| *** amotoki has quit IRC | 21:53 | |
| openstackgerrit | Merged openstack/ironic: Fix docstring warnings https://review.openstack.org/339852 | 21:56 |
| openstackgerrit | Merged openstack/ironic: Grammar fix in code contribution guide https://review.openstack.org/342393 | 21:56 |
| *** wajdi_ has joined #openstack-ironic | 21:58 | |
| *** ametts has quit IRC | 22:01 | |
| *** wajdi has quit IRC | 22:01 | |
| *** wajdi has joined #openstack-ironic | 22:01 | |
| *** openstackgerrit has quit IRC | 22:03 | |
| *** wajdi__ has joined #openstack-ironic | 22:03 | |
| *** openstackgerrit has joined #openstack-ironic | 22:03 | |
| *** wajdi_ has quit IRC | 22:05 | |
| *** wajdi has quit IRC | 22:07 | |
| *** wajdi__ has quit IRC | 22:07 | |
| *** piet has joined #openstack-ironic | 22:08 | |
| *** adu has quit IRC | 22:11 | |
| *** devananda has quit IRC | 22:20 | |
| *** devananda has joined #openstack-ironic | 22:21 | |
| *** ijw has quit IRC | 22:22 | |
| *** sukhdev_ has quit IRC | 22:23 | |
| *** rbudden has quit IRC | 22:45 | |
| *** [1]cdearborn has quit IRC | 22:47 | |
| *** adu has joined #openstack-ironic | 22:50 | |
| *** jcoufal has quit IRC | 23:06 | |
| *** jcoufal has joined #openstack-ironic | 23:06 | |
| *** fragatina has quit IRC | 23:07 | |
| *** jerrygb has joined #openstack-ironic | 23:08 | |
| *** jcoufal has quit IRC | 23:08 | |
| *** thiagop has quit IRC | 23:13 | |
| *** Nisha_away has joined #openstack-ironic | 23:13 | |
| *** Nisha_away has quit IRC | 23:18 | |
| *** praneshp has joined #openstack-ironic | 23:19 | |
| *** PollyZ has joined #openstack-ironic | 23:22 | |
| *** jerrygb has quit IRC | 23:24 | |
| *** PollyZ has quit IRC | 23:25 | |
| *** rook- has quit IRC | 23:28 | |
| *** ijw has joined #openstack-ironic | 23:29 | |
| *** ijw has quit IRC | 23:34 | |
| *** praneshp has quit IRC | 23:39 | |
| *** _Fdaisuke_ has quit IRC | 23:47 | |
| *** piet has quit IRC | 23:48 | |
| *** mtanino has quit IRC | 23:55 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!