| *** krtaylor has joined #openstack-ironic | 00:08 | |
| *** yuanying has joined #openstack-ironic | 00:12 | |
| *** datajerk has joined #openstack-ironic | 01:08 | |
| *** Haomeng|2 has joined #openstack-ironic | 01:18 | |
| *** Haomeng has quit IRC | 01:20 | |
| *** yonglihe_ has quit IRC | 01:21 | |
| *** datajerk has quit IRC | 01:24 | |
| *** penick has joined #openstack-ironic | 01:26 | |
| *** chenglch has joined #openstack-ironic | 01:28 | |
| *** penick has quit IRC | 02:25 | |
| *** chenglch has quit IRC | 02:30 | |
| *** chenglch_ has joined #openstack-ironic | 02:30 | |
| *** ramineni has joined #openstack-ironic | 02:50 | |
| *** lazy_prince is now known as killer_prince | 03:21 | |
| *** shakamunyi has joined #openstack-ironic | 03:25 | |
| *** pcrews has joined #openstack-ironic | 03:26 | |
| *** pcrews has quit IRC | 03:42 | |
| *** nikunj2512 has joined #openstack-ironic | 03:50 | |
| *** killer_prince is now known as lazy_prince | 03:58 | |
| *** shakamunyi has quit IRC | 04:00 | |
| *** pensu has joined #openstack-ironic | 04:01 | |
| *** Poornima has joined #openstack-ironic | 04:08 | |
| *** vinbs has joined #openstack-ironic | 04:13 | |
| *** lazy_prince is now known as killer_prince | 04:17 | |
| *** nikunj2512 has quit IRC | 04:30 | |
| *** nikunj2512 has joined #openstack-ironic | 04:30 | |
| *** rushiagr_away is now known as rushiagr | 04:40 | |
| *** chenglch_ has quit IRC | 05:24 | |
| *** chenglch has joined #openstack-ironic | 05:26 | |
| vinbs | Morning Ironic! | 05:49 |
|---|---|---|
| *** Nisha has joined #openstack-ironic | 05:51 | |
| *** rameshg87 has joined #openstack-ironic | 05:58 | |
| *** k4n0 has joined #openstack-ironic | 06:20 | |
| *** killer_prince is now known as lazy_prince | 06:37 | |
| *** ifarkas has joined #openstack-ironic | 06:52 | |
| *** andreykurilin_ has joined #openstack-ironic | 06:59 | |
| *** dtantsur|afk is now known as dtantsur | 07:05 | |
| dtantsur | Good monday morning, Ironic | 07:05 |
| * dtantsur tries to wake up | 07:06 | |
| *** jcoufal has joined #openstack-ironic | 07:10 | |
| *** Chao_Yan has joined #openstack-ironic | 07:13 | |
| Haomeng|2 | vinbs: morning | 07:18 |
| Haomeng|2 | dtantsur: morning | 07:18 |
| Haomeng|2 | vinbs: see your offline words, any question? | 07:19 |
| *** lazy_prince has quit IRC | 07:22 | |
| *** bluex-pl has joined #openstack-ironic | 07:33 | |
| *** bluex-pl has quit IRC | 07:33 | |
| *** bluex-pl has joined #openstack-ironic | 07:34 | |
| *** killer_prince has joined #openstack-ironic | 07:36 | |
| *** killer_prince is now known as lazy_prince | 07:36 | |
| *** andreykurilin_ has quit IRC | 07:44 | |
| *** jistr has joined #openstack-ironic | 07:53 | |
| Nisha | dtantsur: good morning | 07:56 |
| dtantsur | Nisha, Haomeng|2, morning | 07:56 |
| Haomeng|2 | dtantsur: :) | 07:56 |
| *** lazy_prince has quit IRC | 07:57 | |
| *** lazy_prince has joined #openstack-ironic | 07:58 | |
| *** bluex-pl has quit IRC | 08:01 | |
| *** bluex-pl has joined #openstack-ironic | 08:03 | |
| *** derekh has joined #openstack-ironic | 08:04 | |
| *** ndipanov has joined #openstack-ironic | 08:07 | |
| *** athomas has joined #openstack-ironic | 08:10 | |
| *** yuanying has quit IRC | 08:12 | |
| *** yuanying has joined #openstack-ironic | 08:13 | |
| vinbs | Haomeng, yes | 08:14 |
| vinbs | Haomeng|2, yes | 08:14 |
| Haomeng|2 | vinbs: :) | 08:14 |
| vinbs | Haomeng|2, I have an issue with git | 08:14 |
| vinbs | Haomeng|2, https://review.openstack.org/#/c/118614/ | 08:14 |
| Haomeng|2 | vinbs: you mena you encountered "Merge Failed." case ,right? | 08:15 |
| vinbs | Haomeng|2, yes | 08:15 |
| Haomeng|2 | vinbs: I understand for some case, we have to merge code by manually | 08:15 |
| Haomeng|2 | vinbs: because git dont know how to merge code if same line changed | 08:16 |
| Haomeng|2 | vinbs: so go ahead to manually rebase it in your git working dir and commit code again | 08:16 |
| vinbs | Haomeng|2, So I manually add or delete lines to match it with the latest code right? | 08:17 |
| *** yuanying has quit IRC | 08:17 | |
| vinbs | Haomeng|2, and then add in my changes | 08:17 |
| vinbs | Haomeng|2, and commit | 08:17 |
| vinbs | Haomeng|2, is that right? | 08:17 |
| Haomeng|2 | vinbs: yes, after the code rebased, if there is some confit to be fixed by manually, we can find some ">>>" in the code, and git prompt that we have to merge by manually and run "git add" command and commit the review again | 08:19 |
| vinbs | Haomeng|2, okay let me do that.. thanks! :) | 08:20 |
| Haomeng|2 | vinbs: yw | 08:20 |
| *** lazy_prince has quit IRC | 08:23 | |
| *** MattMan has joined #openstack-ironic | 08:30 | |
| *** k4n0 has quit IRC | 08:33 | |
| *** lucasagomes has joined #openstack-ironic | 08:41 | |
| *** Poornima has quit IRC | 08:41 | |
| *** yuanying has joined #openstack-ironic | 08:43 | |
| *** rushiagr is now known as rushiagr_away | 08:44 | |
| *** rushiagr_away is now known as rushiagr | 08:45 | |
| *** k4n0 has joined #openstack-ironic | 08:51 | |
| *** rushiagr is now known as rushiagr_away | 08:56 | |
| rameshg87 | dtantsur, hi | 09:06 |
| dtantsur | rameshg87, hi | 09:06 |
| rameshg87 | dtantsur, just one thing on this comment: https://review.openstack.org/#/c/123047/1/ironic/drivers/modules/agent_config.template | 09:07 |
| dtantsur | right | 09:08 |
| rameshg87 | dtantsur, i am not sure why ipa guys did the same for 'ipa-api-url' argument | 09:08 |
| rameshg87 | dtantsur, i agree that there is no need of 'if' there | 09:08 |
| dtantsur | rameshg87, let's sync with jroll on it? I don't see point in both | 09:08 |
| rameshg87 | dtantsur, even i don't :) | 09:08 |
| *** datajerk has joined #openstack-ironic | 09:09 | |
| rameshg87 | dtantsur, anyways i will remove off the 'if' from both and raise the review and get comments from jroll on it | 09:09 |
| dtantsur | ack | 09:10 |
| *** bluex-pl has quit IRC | 09:11 | |
| vinbs | dtantsur, Haomeng|2, I see a new reviewer for my patch - IBM xCAT CI | 09:11 |
| dtantsur | it's a 3rd party CI system, not a reviewer | 09:11 |
| vinbs | dtantsur, I see | 09:12 |
| vinbs | dtantsur, incase I need to commit a new driver which is specific to a vendor's hardware | 09:14 |
| vinbs | dtantsur, I would need to setup that 3rd party CI system similar to IBM xCAT CI? | 09:15 |
| dtantsur | vinbs, probably not. At least we don't do it right now. Of course it would be _much_ better if you do. | 09:16 |
| *** yuanying has quit IRC | 09:18 | |
| *** bluex-pl has joined #openstack-ironic | 09:18 | |
| vinbs | dtantsur, Is there any guide on setting up 3rd party CI System? | 09:20 |
| dtantsur | vinbs, I'm not aware of any, sorry. You'd better ask on #openstack-infra IMO | 09:21 |
| vinbs | dtantsur, cool.. thanks | 09:22 |
| *** rushiagr_away is now known as rushiagr | 09:23 | |
| *** pensu has quit IRC | 09:36 | |
| *** k4n0 has quit IRC | 09:44 | |
| *** pensu has joined #openstack-ironic | 09:52 | |
| *** romcheg has joined #openstack-ironic | 10:05 | |
| *** athomas has quit IRC | 10:10 | |
| *** yuanying has joined #openstack-ironic | 10:13 | |
| *** Chao_Yan has quit IRC | 10:16 | |
| *** athomas has joined #openstack-ironic | 10:19 | |
| Haomeng|2 | vinbs: yes, IBMxCAT CI will not vote our patch | 10:19 |
| *** mitz_ has joined #openstack-ironic | 10:22 | |
| *** vinbs has quit IRC | 10:23 | |
| *** vinbs has joined #openstack-ironic | 10:23 | |
| *** Nisha has quit IRC | 10:23 | |
| vinbs | Haomeng|2, got it! :) | 10:29 |
| *** k4n0 has joined #openstack-ironic | 10:30 | |
| *** bluex-pl has quit IRC | 10:38 | |
| *** bluex-pl has joined #openstack-ironic | 10:39 | |
| *** bluex-pl has quit IRC | 10:40 | |
| *** bluex-pl has joined #openstack-ironic | 10:44 | |
| *** yuanying has quit IRC | 10:47 | |
| *** Poornima has joined #openstack-ironic | 10:47 | |
| Haomeng|2 | vinbs: :) | 10:48 |
| *** zigo has quit IRC | 10:50 | |
| *** zigo has joined #openstack-ironic | 10:52 | |
| *** vdrok_ has joined #openstack-ironic | 11:03 | |
| *** ramineni has quit IRC | 11:05 | |
| *** rameshg87 has quit IRC | 11:06 | |
| *** bluex-pl has quit IRC | 11:06 | |
| *** bluex-pl has joined #openstack-ironic | 11:07 | |
| *** bluex-pl has quit IRC | 11:07 | |
| *** bluex-pl has joined #openstack-ironic | 11:08 | |
| *** dtantsur is now known as dtantsur|brb | 11:08 | |
| *** shakamunyi has joined #openstack-ironic | 11:18 | |
| *** shakamunyi has quit IRC | 11:20 | |
| *** yuanying has joined #openstack-ironic | 11:25 | |
| *** lucasagomes is now known as lucas-hungry | 11:27 | |
| *** nikunj2512 has quit IRC | 11:31 | |
| *** Haomeng has joined #openstack-ironic | 11:35 | |
| *** Haomeng|2 has quit IRC | 11:36 | |
| *** yuanying has quit IRC | 11:59 | |
| *** marzif__ has quit IRC | 12:02 | |
| *** bluex-pl has quit IRC | 12:07 | |
| *** bluex-pl has joined #openstack-ironic | 12:08 | |
| *** openstackgerrit has joined #openstack-ironic | 12:15 | |
| openstackgerrit | Dmitry Tantsur proposed a change to openstack/ironic: Neutron DHCP implementation to raise exception if no ports have VIF https://review.openstack.org/123108 | 12:27 |
| dtantsur|brb | https://review.openstack.org/#/c/122728/ has 2x +2, could someone approve? | 12:28 |
| *** dtantsur|brb is now known as dtantsur | 12:28 | |
| *** lucas-hungry is now known as lucasagomes | 12:34 | |
| *** rushiagr is now known as rushiagr_away | 12:35 | |
| openstackgerrit | Vladyslav Drok proposed a change to openstack/ironic-python-agent: Cleanup coreos-oem-inject.py https://review.openstack.org/122693 | 12:36 |
| *** Haomeng|2 has joined #openstack-ironic | 12:38 | |
| *** Haomeng has quit IRC | 12:39 | |
| Shrews | dtantsur: lol... assertCalleOnceWith()... how did that ever get approved? | 12:53 |
| dtantsur | Shrews, no idea :) I guess we need to pay more attention to reviewing tests... | 12:54 |
| *** yuanying has joined #openstack-ironic | 12:55 | |
| Shrews | dtantsur: lgtm | 12:57 |
| Shrews | +A'd | 12:57 |
| *** todd_dsm has joined #openstack-ironic | 12:58 | |
| *** dprince has joined #openstack-ironic | 12:58 | |
| dtantsur | thanks! | 12:59 |
| *** todd_dsm has quit IRC | 13:00 | |
| *** todd_dsm has joined #openstack-ironic | 13:01 | |
| *** killer_prince has joined #openstack-ironic | 13:11 | |
| *** killer_prince is now known as lazy_prince | 13:11 | |
| *** faizan has joined #openstack-ironic | 13:14 | |
| faizan | lucasagomes, dtantsur, Hi | 13:17 |
| faizan | lucasagomes, dtantsur, Thanks for your review and approval for add/rm_node_capability bug | 13:17 |
| lucasagomes | faizan, np :) | 13:17 |
| dtantsur | you're welcome :) | 13:17 |
| Shrews | lucasagomes: dtantsur: for https://review.openstack.org/122802, i'm not so sure a detailed explanation of ironic-dbsync in the quickstart guide is the right place. A guide is not "quick" if you have to explain every detail. I think we need a more thorough documenting of that command separately (which does not yet exist). WDYT? | 13:18 |
| Shrews | i'm happy to add that as a separate doc somewhere | 13:18 |
| dtantsur | Shrews, I think user should be at least aware of difference between 2 dbsync variants (one works for sqlite, the other works when database needs to be upgraded) | 13:19 |
| faizan | lucasagomes, dtantsur, I kindly request you to look at this one also - https://review.openstack.org/#/c/122819/3 | 13:19 |
| dtantsur | I don't think we need to go into details other than that | 13:19 |
| Shrews | dtantsur: but an upgrade is not even considered in the quickstart guide OR install guide. a fresh setup is assumed (based on all of the other procedures). so i find that confusing | 13:21 |
| lucasagomes | Shrews, hmm I see the point... I dunno where that should be documented tho :/ | 13:21 |
| * lucasagomes thinking | 13:21 | |
| Shrews | lucasagomes: separate docs for our utilities? | 13:21 |
| Shrews | similar to a man-page | 13:22 |
| *** linggao has joined #openstack-ironic | 13:22 | |
| lucasagomes | hmm yeah that would be useful/the right place indeed | 13:23 |
| Shrews | lucasagomes: dtantsur: i'll throw something together, then we can decide | 13:24 |
| lucasagomes | Shrews, I will re-review that patch | 13:24 |
| dtantsur | ok | 13:24 |
| lucasagomes | let's wait for viktors openion on that as well | 13:24 |
| *** yuanying has quit IRC | 13:29 | |
| lucasagomes | Shrews, it also makes me wonder whether we should verify if the db exist and use upgrade by default or if it doesn't just use create_schema by default | 13:29 |
| Shrews | lucasagomes: hmm, not sure i like changing a utility's default behavior based on an external condition. /me thinks about it | 13:32 |
| lucasagomes | Shrews, well it it's like "touch" if file is not there create it | 13:33 |
| lucasagomes | if not just touch/update it | 13:33 |
| *** bluex-pl has quit IRC | 13:33 | |
| lucasagomes | but I see that it may hide some problems too, since there's no default behavior per si | 13:33 |
| *** bluex-pl has joined #openstack-ironic | 13:33 | |
| lucasagomes | it depends on some conditions | 13:34 |
| *** faizan has quit IRC | 13:34 | |
| Shrews | "upgrade" will actually create a db if it doesn't exist, too, but it runs through the migration rather than just create based on the latest model. that's just an fyi, not sure if someone would actually prefer to do a create vs. upgrade in that situation (should be the same result) | 13:37 |
| Shrews | not for sqlite, of course.... anyway... more coffee | 13:37 |
| openstackgerrit | A change was merged to openstack/ironic: Do not cache auth token in Neutron DHCP provider https://review.openstack.org/122728 | 13:38 |
| *** vinbs has quit IRC | 13:41 | |
| viktors | lucasagomes: hi! Did I miss something? | 13:42 |
| * lucasagomes is in a call | 13:42 | |
| lucasagomes | viktors, Shrews 1 min will finish soon | 13:42 |
| viktors | ok | 13:42 |
| openstackgerrit | Vladyslav Drok proposed a change to openstack/ironic-python-agent: Cleanup coreos-oem-inject.py https://review.openstack.org/122693 | 13:50 |
| jroll | morning all :) | 13:57 |
| *** chenglch has quit IRC | 14:01 | |
| NobodyCam | Good Morning Ironic | 14:13 |
| dtantsur | morning, NobodyCam, jroll, viktors | 14:14 |
| *** ChuckC_ has quit IRC | 14:14 | |
| NobodyCam | morning dtantsur :) | 14:14 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Do not set the context twice when forming RPC objects https://review.openstack.org/122401 | 14:14 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Make context mandatory when instantiating a RPC object https://review.openstack.org/121923 | 14:14 |
| lucasagomes | NobodyCam, morning | 14:14 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Do not use the context parameter on refresh() https://review.openstack.org/122441 | 14:14 |
| lucasagomes | viktors, Shrews back | 14:14 |
| NobodyCam | morning lucasagomes | 14:14 |
| viktors | morning, NobodyCam | 14:15 |
| lucasagomes | viktors, we were just talking about that patch that Shrews added changing the documentation to explicity pass "create_schema" to the db-sync utility | 14:15 |
| NobodyCam | morning viktors and jroll | 14:15 |
| viktors | lucasagomes: yes, I left some comments to Shrews's patch | 14:15 |
| lucasagomes | and we were wondering if it would make sense to make the dbsync tool, ack differently by default when there's a database present we use 'update' by default, if not we do a 'create_schema' | 14:16 |
| lucasagomes | oh right I will take a look | 14:16 |
| jroll | heya dtantsur NobodyCam viktors Shrews lucasagomes | 14:17 |
| jroll | so many people :o | 14:17 |
| lucasagomes | jroll, morning | 14:17 |
| lucasagomes | heh indeed | 14:17 |
| viktors | lucasagomes: yes, I agree with it, but IMO it's would be nice also provide a detailed documentation, at least for user's compatibility | 14:17 |
| viktors | hi jroll! | 14:17 |
| * Shrews distracted by people installing a new door. not really present atm | 14:17 | |
| lucasagomes | viktors, oh definitely, the documentation is super needed | 14:17 |
| NobodyCam | oh morning Shrews :) | 14:19 |
| *** k4n0 has quit IRC | 14:19 | |
| *** todd_dsm has quit IRC | 14:20 | |
| *** todd_dsm has joined #openstack-ironic | 14:21 | |
| *** yuanying has joined #openstack-ironic | 14:25 | |
| jroll | vdrok: one more thing on https://review.openstack.org/#/c/122693/ :) | 14:27 |
| *** agordeev has quit IRC | 14:27 | |
| *** jistr has quit IRC | 14:30 | |
| openstackgerrit | Jim Rollenhagen proposed a change to openstack/ironic-python-agent: Add docstrings for api/ tree https://review.openstack.org/120661 | 14:31 |
| jroll | Shrews: fixed that thing, ty ^ | 14:31 |
| openstackgerrit | Jim Rollenhagen proposed a change to openstack/ironic-python-agent: Delete unused method get_agent_mac_address() https://review.openstack.org/121153 | 14:32 |
| *** jistr has joined #openstack-ironic | 14:33 | |
| *** agordeev has joined #openstack-ironic | 14:44 | |
| krtaylor | dtantsur, vinbs, you were asking for guidelines on setting up third-party CI testing - here is the link http://ci.openstack.org/ (see the HOWTOS box) | 14:44 |
| *** Poornima has quit IRC | 14:45 | |
| krtaylor | dtantsur, vinbs, also, we meet every Monday to discuss how to make CI better -> https://wiki.openstack.org/wiki/Meetings/ThirdParty | 14:45 |
| dtantsur | cool thanks | 14:45 |
| vdrok_ | jroll, morning :) will fix | 14:47 |
| vdrok_ | hi everyone! | 14:47 |
| jroll | good morning :) thanks! | 14:47 |
| openstackgerrit | A change was merged to openstack/ironic: Updated from global requirements https://review.openstack.org/122423 | 14:49 |
| NobodyCam | morning vdrok_ | 14:52 |
| vdrok_ | morning NobodyCam ! | 14:52 |
| *** Guest75828 has joined #openstack-ironic | 14:52 | |
| NobodyCam | :) | 14:54 |
| *** dlaube has joined #openstack-ironic | 14:55 | |
| *** yuanying has quit IRC | 14:59 | |
| openstackgerrit | Yuriy Zveryanskyy proposed a change to openstack/ironic: Do not return 'id' in REST API error messages https://review.openstack.org/123150 | 15:03 |
| *** bluex-pl has quit IRC | 15:07 | |
| *** foexle has joined #openstack-ironic | 15:12 | |
| *** jistr has quit IRC | 15:12 | |
| *** jistr has joined #openstack-ironic | 15:13 | |
| *** todd_dsm has quit IRC | 15:16 | |
| NobodyCam | so do we still need / want Outstanding Items for graduation on hte agenda? | 15:18 |
| dtantsur | I hope nobody is going to revoke our graduation :) | 15:19 |
| NobodyCam | lol yep, just wasn't sure if there was anything outsianding we wanted to talk about | 15:20 |
| NobodyCam | I removed that section | 15:21 |
| NobodyCam | :-p | 15:21 |
| *** pcrews has joined #openstack-ironic | 15:23 | |
| linggao | devananda, ping. | 15:24 |
| NobodyCam | and of course as always on monday's call for new agenda items. | 15:25 |
| NobodyCam | s/monday's/monday,/ | 15:25 |
| openstackgerrit | Vladyslav Drok proposed a change to openstack/ironic-python-agent: Cleanup coreos-oem-inject.py https://review.openstack.org/122693 | 15:28 |
| jroll | JayF: ^ you should look at that too | 15:31 |
| *** ChuckC_ has joined #openstack-ironic | 15:31 | |
| JayF | Yeah I saw the bug get picked up | 15:32 |
| zigo | Hi there! I got a few unit tests errors when building in Wheezy: https://juno.dev-debian.pkgs.enovance.com/job/ironic/3/console | 15:36 |
| zigo | Does anyone have an idea why? | 15:36 |
| *** Guest75828 is now known as annegentle | 15:38 | |
| lucasagomes | NobodyCam, oh btw, I won't be able to make to the meeting tonight :( | 15:38 |
| lucasagomes | NobodyCam, I've to solve some stuff after work so I may not arrive in time for it | 15:38 |
| jroll | zigo: what version of keystoneclient is that? | 15:38 |
| jroll | and what version of ironic | 15:39 |
| NobodyCam | lucasagomes: its ok ... is there anything you need me to chat about there? | 15:39 |
| jroll | and glanceclient :( | 15:39 |
| zigo | jroll: 0.10.1 for ksc, and Juno b3 for Ironic. 0.14.0 for glanceclient. | 15:40 |
| *** todd_dsm has joined #openstack-ironic | 15:40 | |
| jroll | zigo: these all remind me of things we fixed during this cycle where the clients broke ironic | 15:40 |
| zigo | jroll: The weird thing is that I don't have these in Sid, just in my Wheezy backport. | 15:40 |
| zigo | In Sid, I only have a single unit test error. | 15:40 |
| * jroll wonders what juno b3 is | 15:41 | |
| jroll | juno-3 I guess | 15:41 |
| Shrews | that "'Client' object has no attribute 'endpoint'" error was fixed long ago | 15:41 |
| zigo | Not even... | 15:41 |
| zigo | Nothing, all works ! | 15:41 |
| zigo | Oh, shit, I'm building the wrong version (eg: 2014.1, not Juno). | 15:41 |
| zigo | Sorry guys. | 15:41 |
| jroll | yeah, I don't claim to know anything about what packages are in which distro | 15:42 |
| zigo | Fixing my CI ! :) | 15:42 |
| jroll | lol | 15:42 |
| jroll | cool :) | 15:42 |
| lucasagomes | NobodyCam, hmm not off the top of my head | 15:42 |
| zigo | jroll: Well, for Debian (where I do all the packaging of OpenStack), I just follow the global-requirements.txt, and then, Ubuntu guys just take all of my work for non core packages (ie: all Python modules). | 15:44 |
| jroll | zigo: oh, cool, I didn't know that you did the packaging :) | 15:44 |
| zigo | They sometimes do small modifications, and I try to grab them when I have time, and when it's relevant... | 15:44 |
| zigo | jroll: http://qa.debian.org/developer.php?login=openstack-devel@lists.alioth.debian.org <--- 179 packages so far ... :) | 15:45 |
| jroll | wow, nice | 15:45 |
| *** rloo has joined #openstack-ironic | 15:46 | |
| jroll | JayF: also mind grabbing this? https://review.openstack.org/#/c/120661/ | 15:46 |
| jroll | morning rloo :) | 15:46 |
| rloo | morning jroll! | 15:46 |
| zigo | FYI, I'm planning to remove Ironic from Jessie, because it's going to be Icehouse, and there's a consensus that it's not useable enough there. I'll maintain it in subsequent releases, but it wont make it to the next Debian stable then... | 15:47 |
| *** ndipanov is now known as ndipanov_gone | 15:47 | |
| NobodyCam | lucasagomes: ack ... | 15:47 |
| zigo | How happy are you guys with Juno? | 15:47 |
| NobodyCam | morning rloo | 15:48 |
| JayF | jroll: should we talk to zigo about ipa packaging? Perhaps there'd be a perspective there worth getting | 15:48 |
| rloo | morning NobodyCam | 15:48 |
| zigo | JayF: What's ipa ? | 15:48 |
| NobodyCam | :) | 15:48 |
| JayF | zigo: ironic-python-agent ... it's a ramdisk agent used to provision machines in some Ironic deploy drivers | 15:49 |
| zigo | Ah ok. | 15:49 |
| JayF | zigo: we've been wondering how it's best to be packaged, because the average user will only use it in Ramdisk/kernel form (similar to ipxe or syslinux where the "image" is what is really wanted) | 15:49 |
| JayF | or will build their own | 15:49 |
| zigo | Well, if you need packages in Debian (or Ubuntu), I'm definitively the person to talk with! :) | 15:50 |
| JayF | so I wasn't sure the best way for us to 'release' IPA for packaging | 15:50 |
| JayF | I don't know :) | 15:50 |
| JayF | Was hoping you could help tell me that | 15:50 |
| *** r-daneel has joined #openstack-ironic | 15:50 | |
| JayF | jroll: docstring+unused methods both approved again | 15:50 |
| zigo | JayF: Is the plan to add some stuff to a "standard" debian initrd image? | 15:51 |
| *** Mike_D_wk has quit IRC | 15:51 | |
| jroll | thanks JayF | 15:51 |
| zigo | Because that's not hard to do ... | 15:51 |
| *** Mike_D_wk has joined #openstack-ironic | 15:51 | |
| zigo | See cloud-initramfstools for example. | 15:51 |
| JayF | zigo: We build the ramdisk, right now, by embedding it in a CoreOS pxe image | 15:51 |
| JayF | zigo: I don't think there's an intention to change the ramdisk's internals based on which distro it's being packaged for :) | 15:52 |
| zigo | JayF: Would that work to just package the tools to actually *build* the ramdisk, rather than trying to package the image itself? | 15:53 |
| JayF | zigo: the image build tools are here -> https://github.com/openstack/ironic-python-agent/tree/master/imagebuild | 15:53 |
| JayF | zigo: I'd suspect if someone was building it themselves (that's what Rackspace does for OnMetal), they'd pull the code in using git or pypi, not distro packages | 15:53 |
| zigo | JayF: I don't see why you wouldn't want to use packages if they are up-to-date and in good shape. | 15:54 |
| devananda | morning, all | 15:54 |
| *** yuanying has joined #openstack-ironic | 15:55 | |
| JayF | zigo: Because I can express a dependency on pypi or git using a requirements.txt | 15:56 |
| zigo | JayF: Well, I can express dependencies using debian/control ! :) | 15:56 |
| JayF | zigo: for instance, one of the cases for building your own agent is to add your own hardware manager (like, say, this one: https://github.com/rackerlabs/onmetal-ironic-hardware-manager) -- and in our case we build that by using requirements.txt to pull in our third-party code when building | 15:56 |
| dtantsur | devananda, morning | 15:56 |
| zigo | A large amount of my packaging time is spent translating requirements.txt into a valid debian/control file. | 15:56 |
| devananda | JayF: my concern with not having a release tagged (packaged or otherwise) is versioning. If someone, two months from now, wants to build an IPA image to match the Juno release of Ironic, can you guarantee that what they build from the git tree is going to work? | 15:57 |
| NobodyCam | good morning devananda | 15:57 |
| JayF | devananda: I am 100% in agreement at this point we should tag a release | 15:57 |
| devananda | JayF: glad to hear it :) | 15:57 |
| JayF | devananda: I'm mainly curious about if/how that translates to distro packaging... | 15:57 |
| devananda | JayF: and in that case, if there's a release, why not package it? | 15:57 |
| JayF | as a user in this case, I'd want the built image to be apt installable, and wouldn't care about the source (i.e. building my own; I'd use a source like git rather than a debian source image) | 15:58 |
| devananda | JayF: "awnt the built image to be apt instalable" -- I don't think I agree there | 15:58 |
| JayF | devananda: I know that if IPA is packaged in a hard to use way (like "here's some source and a Makefile, enjoy") that it won't be used very much | 15:58 |
| devananda | JayF: let's say I want to build an image using the Juno release of IPA, but I need add my-fancy-driver to it. | 15:59 |
| zigo | JayF: Well, let's just make it easy to use then! :) | 15:59 |
| devananda | JayF: also, by the way, this is the same set of problems which diskimage-builder has faced | 15:59 |
| devananda | which is one reason I've been suggesting ya'll just use that | 16:00 |
| JayF | zigo: devananda: how is the ironic pxe ramdisk packaged in distros? | 16:00 |
| devananda | JayF: diskimage-builder and tripleo-image-elements are released to PIP. zigo: are there packaged versions of those tools? | 16:01 |
| zigo | devananda: yes. | 16:02 |
| jroll | so the recommended way is to build the ramdisk | 16:02 |
| devananda | ah yep. just searched and found them in Jessie | 16:02 |
| zigo | I've updated diskimage-builder, though I probably also should update tripleo-image-elements. | 16:03 |
| jroll | so we should package a way to build IPA | 16:03 |
| jroll | but not IPA itself | 16:03 |
| JayF | How would a potential consumer of Ironic via distro packages discover how to do all the steps required, including building the pxe ramdisk? | 16:03 |
| zigo | JayF: By reading the official OpenStack install-manual? :) | 16:03 |
| zigo | s/install-manual/install-guide/ | 16:04 |
| jroll | docs, yo | 16:04 |
| jroll | we have docs for this, for the pxe driver | 16:04 |
| jroll | or they are in flight at least | 16:04 |
| devananda | yep | 16:04 |
| JayF | Looking for them now | 16:04 |
| JayF | not having much success finding anything but developer docs | 16:04 |
| devananda | well - we have developer docs. not being integrated in Juno has meant we don't have a presense in the official OpenStack Manual | 16:04 |
| devananda | also - we need a doc writer | 16:05 |
| zigo | I'd be very happy to do more collaborative work with you guys, and do more packaging work, so that Ironic is in a better shape in Debian. | 16:05 |
| zigo | I can package more stuff, and do corrections in existing ones. | 16:05 |
| zigo | Though what I can't really do is test it all... | 16:05 |
| zigo | I have not enough time for it. | 16:05 |
| jroll | JayF: at a minimum, https://review.openstack.org/#/c/118614/ | 16:06 |
| JayF | If this is the right place to be: http://docs.openstack.org/developer/ironic/ I don't see any docs at all that would help me discover how to install Ironic at all... I think I'm in the wrong place? :( | 16:06 |
| zigo | So your feedback is *very* important for me. | 16:06 |
| jroll | JayF: this is almost there: http://docs.openstack.org/developer/ironic/deploy/install-guide.html | 16:07 |
| JayF | zigo: I'm only curious from the perspective of wanting software I work on packaged well... Rackspace doesn't run our cloud out of debian packages :) | 16:07 |
| jroll | but nothing about building ramdisks etc | 16:07 |
| JayF | zigo: so I'm not exactly your customer | 16:07 |
| JayF | jroll: exactly what I was concerned about | 16:07 |
| jroll | so we need to improve our docs | 16:07 |
| jroll | even if we packaged a ramdisk image or whatever | 16:07 |
| JayF | I'm just going down the rabbithole because when we talk about how IPA should be packaged | 16:08 |
| jroll | there's still no docs on using that | 16:08 |
| JayF | really we're asking "how would someone use a distro package to setup ironic + ipa driver" | 16:08 |
| zigo | Got it. | 16:08 |
| JayF | and the answer appears to be we have *no* docs that cover everything needed to deploy ironic, with main deficiencies seeming to be around ramdisk building | 16:08 |
| devananda | jroll, JayF: right. and that hasn't been a priority of the development y'all have been doing, AIUI. | 16:08 |
| devananda | not that it shouldn't be :) | 16:08 |
| JayF | devananda: You're exactly right :) We've been prioritizing documenting how *we* install and use it over documenting how others can do it in a more standard way :) | 16:09 |
| devananda | zigo: you're on a bit of a timeline to get all these packaged, right? | 16:09 |
| *** datajerk has quit IRC | 16:09 | |
| zigo | devananda: I've finished packaging all dependencies for Juno early this week, and I'm finishing to upload Juno b3. I'd like to have everything ready for the 16 of October, on time for the release... | 16:10 |
| JayF | Given what I've heard; I'd say improve the docs in/around IPA itself on how to build a ramdisk image, then just package IPA in a way that it could be built | 16:10 |
| JayF | i.e. instead of putting it on the pythonpath, put the code and build infra in /usr/share/ somewhere so people can use the packaged software to build IPA | 16:10 |
| JayF | but aiui I don't think we'd be able to 'pip install ironic-python-agent' then use the result of that to build a ramdisk | 16:11 |
| devananda | zigo: gotcha, thanks | 16:11 |
| * zigo just uploaded Ironic Juno b3 to Debian experimental. | 16:11 | |
| devananda | JayF: I'm getting the impression you guys may need to change some of the structuring of IPA to accomodate packaging the build tools separately from the IPA code itself | 16:12 |
| zigo | FYI, everything for Juno, and probably Kilo, will be in Experimental durring the Jessie freeze. | 16:12 |
| devananda | JayF: since you don't actually *install* IPA itself | 16:12 |
| zigo | (that's the normal procedure during the freeze of Debian: you can't upload to Sid/Testing) | 16:12 |
| JayF | devananda: yeah; you had originally proposed that as something we talk about splitting out at the summit | 16:12 |
| devananda | JayF: yep | 16:12 |
| JayF | devananda: and I think we should split out all ironic ramdisk build utilities into their own repository--everything from DIB elements for pxe driver to all the various builders for IPA | 16:13 |
| JayF | but for J, what do we want to do? | 16:13 |
| devananda | JayF: is this something you want to try to do in the next week or two? or better to solve it for Kilo? | 16:13 |
| devananda | heh | 16:13 |
| * dtantsur steps afk until meeting time | 16:14 | |
| *** marcoemorais has joined #openstack-ironic | 16:15 | |
| openstackgerrit | Ramakrishnan G proposed a change to openstack/ironic: Pass ipa-driver-name to agent ramdisk https://review.openstack.org/123047 | 16:17 |
| devananda | zigo: looking at https://packages.debian.org/jessie/python-ironic I do not believe ironic depends on seamicroclient | 16:18 |
| devananda | or on python-pyghmi | 16:18 |
| zigo | devananda: That's the Icehouse release, I think it did ... | 16:18 |
| devananda | zigo: though i'm not sure we're communicating that properly -- these are optional dependencies for vendor drivers | 16:18 |
| devananda | zigo: ahh. it may have | 16:18 |
| devananda | zigo: ahh. it may have been in our requirements file back then | 16:18 |
| zigo | I just removed them for Juno indeed! :) | 16:19 |
| *** derekh has quit IRC | 16:19 | |
| zigo | devananda: Please try the package I just uploaded tomorrow, when it reaches the mirrors. | 16:19 |
| openstackgerrit | Vladyslav Drok proposed a change to openstack/ironic: Add parameter to override locale to utils.execute https://review.openstack.org/122743 | 16:19 |
| zigo | devananda: Do you think there's a chance that Ironic from Juno works together with the rest of Icehouse? | 16:20 |
| zigo | Or no way? | 16:20 |
| JayF | No way | 16:20 |
| JayF | because you wouldn't have our Nova driver | 16:20 |
| zigo | Ok. :) | 16:20 |
| JayF | We vendored it into Ironic for Icehouse, but in Juno it'll be in Nova | 16:21 |
| JayF | s/'ll be/is/ | 16:21 |
| devananda | zigo: yep. Ironic doesn't function within OpenStack unless you have the Nova "ironic" driver, and that depends on unstable/internal APIs in Nova | 16:21 |
| devananda | so it's unfortunately tightly coupled to the release of Nova for now | 16:22 |
| zigo | I was just trying to see if it was possible to still have Ironic in Jessie, never mind then. | 16:24 |
| rloo | Hi lucasagomes. qq on 121923. Do you know if the object versions need to be updated? eg line 29: https://review.openstack.org/#/c/121923/5/ironic/objects/chassis.py | 16:25 |
| lucasagomes | rloo, looking | 16:25 |
| lucasagomes | rloo, so, no because I haven't changed anything in the objects api | 16:26 |
| rloo | but context is mandatory now, no? | 16:26 |
| lucasagomes | to instantiate the object class | 16:27 |
| rloo | lucasagomes: right. | 16:27 |
| lucasagomes | it was already mandatory to pass the context for every classmethod | 16:27 |
| rloo | lucasagomes: is it mandatory if it is context=None? | 16:27 |
| lucasagomes | so there's no method signature being changed or anything | 16:27 |
| * zigo goes back to packaging the rest of Juno (eg: trove & designate still missing). | 16:27 | |
| zigo | Thanks for the chat guys! :) | 16:27 |
| zigo | And feel free to ping me again. | 16:27 |
| lucasagomes | rloo, that's for @remotable, the context=None is not used anymore | 16:28 |
| lucasagomes | and will be removed in the future, then I think we should update the version | 16:28 |
| lucasagomes | because we will change the method signature | 16:28 |
| lucasagomes | but for now, we just use the internal context, so the methods does change anything (same input same output) | 16:29 |
| *** yuanying has quit IRC | 16:29 | |
| *** r-daneel has quit IRC | 16:29 | |
| devananda | zigo: cheers, thanks for dropping in! | 16:30 |
| rloo | lucasagomes: ok, if you say so. thx. (I wasn't sure, will need to think about it.) | 16:30 |
| *** marcoemorais has quit IRC | 16:30 | |
| *** marcoemorais has joined #openstack-ironic | 16:31 | |
| lucasagomes | rloo, it's fine... because this api should be updated for @remotable[_classmethod]s | 16:31 |
| lucasagomes | because these are the methods that are executed over rpc | 16:31 |
| lucasagomes | the instantiation is local, so there's no need to update the version (afaiui) | 16:32 |
| *** r-daneel has joined #openstack-ironic | 16:33 | |
| JayF | zigo: ty | 16:33 |
| openstackgerrit | A change was merged to openstack/ironic-python-agent: Add docstrings for api/ tree https://review.openstack.org/120661 | 16:34 |
| *** yjiang5 has joined #openstack-ironic | 16:39 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Do not set the context twice when forming RPC objects https://review.openstack.org/122401 | 16:41 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Do not use the context parameter on refresh() https://review.openstack.org/122441 | 16:41 |
| openstackgerrit | A change was merged to openstack/ironic-python-agent: Delete unused method get_agent_mac_address() https://review.openstack.org/121153 | 16:41 |
| Shrews | jroll: got a sec? | 16:43 |
| lucasagomes | right folks I will call it a day a bit earlier today | 16:43 |
| *** jistr has quit IRC | 16:43 | |
| Shrews | night lucasagomes | 16:43 |
| jroll | Shrews: what's up? | 16:43 |
| lucasagomes | have a good night | 16:43 |
| Shrews | jroll: this... https://review.openstack.org/#/c/123108/1/ironic/tests/dhcp/test_neutron.py | 16:43 |
| NobodyCam | night lucasagomes | 16:43 |
| lucasagomes | see ye later! (I hope I can make it to the meeting) | 16:43 |
| Shrews | jroll: i'm failing to see how mock_updo is never called (either before or after the change). i mean, all update_dhcp() does is call that method | 16:44 |
| NobodyCam | bbt ... brb | 16:44 |
| jroll | Shrews: I had the same problem, can't remember right away what I missed :P | 16:44 |
| * jroll looks | 16:44 | |
| jroll | so mock_updo is update_port_dhcp_opts | 16:45 |
| Shrews | yup | 16:45 |
| jroll | which is called here: https://github.com/openstack/ironic/blob/master/ironic/dhcp/neutron.py#L156 | 16:45 |
| jroll | which is after the `if not vifs` check | 16:46 |
| Shrews | jroll: oh, wait | 16:46 |
| * Shrews just now notices "_port_" | 16:46 | |
| jroll | :) | 16:46 |
| jroll | yeah, I did the same thing | 16:46 |
| Shrews | ugh | 16:47 |
| Shrews | mondays | 16:47 |
| * jroll passes Shrews a coffee | 16:47 | |
| * Shrews finds it hard to concentrate with the sound of a sawsall going on 15ft away from him | 16:48 | |
| *** r-daneel_ has joined #openstack-ironic | 16:49 | |
| *** r-daneel has quit IRC | 16:49 | |
| * NobodyCam passes Shrews a pair of ear plugs | 16:49 | |
| *** rloo has quit IRC | 16:53 | |
| NobodyCam | didn't we land a fix for this: http://logs.openstack.org/51/120651/4/gate/gate-tempest-dsvm-ironic-pxe_ssh/08de6f4/console.html.gz#_2014-09-19_19_20_47_621 | 16:53 |
| *** rloo has joined #openstack-ironic | 16:53 | |
| Shrews | NobodyCam: yeah | 16:54 |
| Shrews | https://review.openstack.org/122832 | 16:54 |
| *** athomas has quit IRC | 16:54 | |
| NobodyCam | ahh yes TY Shrews | 16:54 |
| *** linggao has quit IRC | 16:59 | |
| *** ifarkas has quit IRC | 17:06 | |
| NobodyCam | quick walkies ...brb | 17:07 |
| *** eghobo has joined #openstack-ironic | 17:09 | |
| devananda | quick store run - bbiab | 17:10 |
| *** linggao has joined #openstack-ironic | 17:11 | |
| linggao | devananda, ping. | 17:15 |
| * NobodyCam is back | 17:17 | |
| *** todd_dsm has quit IRC | 17:20 | |
| *** harlowja_away is now known as harlowja | 17:21 | |
| openstackgerrit | A change was merged to openstack/ironic: Make context mandatory when instantiating a RPC object https://review.openstack.org/121923 | 17:23 |
| *** eghobo has quit IRC | 17:25 | |
| *** yuanying has joined #openstack-ironic | 17:25 | |
| NobodyCam | any one Vinay's irc handle? | 17:26 |
| NobodyCam | s/one/one know/ | 17:26 |
| dtantsur | NobodyCam, vinbs maybe? | 17:30 |
| NobodyCam | dtantsur: maybe | 17:30 |
| NobodyCam | :) | 17:30 |
| *** penick has joined #openstack-ironic | 17:30 | |
| NobodyCam | I'll just comment on the review | 17:30 |
| *** marcoemorais has quit IRC | 17:31 | |
| *** marcoemorais has joined #openstack-ironic | 17:32 | |
| *** krtaylor has quit IRC | 17:32 | |
| *** krtaylor has joined #openstack-ironic | 17:41 | |
| *** todd_dsm has joined #openstack-ironic | 17:50 | |
| * devananda is back | 17:53 | |
| devananda | linggao: pong | 17:53 |
| NobodyCam | WB devananda | 17:53 |
| linggao | Hi devananda, just a short question. I am looking into add petitboot support in Ironic for kilo release. But not sure if it is supported or not already. | 17:54 |
| JayF | whoa nice | 17:55 |
| JayF | jroll: ^^ look at petitboot | 17:55 |
| JayF | linggao: We have definate interest in looking at kexec reboots in the agent | 17:55 |
| JayF | linggao: lower provisioning times would be great :) and skipping a post helps that | 17:55 |
| devananda | linggao: No, it's not currently supported | 17:55 |
| NobodyCam | devananda: are there any post-graduation items that should be on the agenda for the meeting? i assume not but wanted to dbl check | 17:55 |
| devananda | the problem with kexec is hardware support | 17:56 |
| linggao | thanks, JayF and devananda. So is it like writing another Deployment driver? | 17:56 |
| devananda | we (Ironic) can't claim that kexec will work, because we (Ironic) don't control what hardware it's going to be run on, and a lot of hardware will fail in very bad ways | 17:56 |
| *** jcoufal has quit IRC | 17:56 | |
| linggao | like pxe | 17:57 |
| JayF | I don't know where something like this would integrate | 17:57 |
| devananda | but on specific hardware, where someone (perhaps a vendor, or perhaps the local operator) has tested it? Yep, it could be significantly faster for provisioning in some cases | 17:57 |
| devananda | so far, I've pushed back on adding kexec support to Ironic. Not because its out of scope, but because it's an optimization with more limited use cases | 17:59 |
| *** yuanying has quit IRC | 17:59 | |
| linggao | JayF and devananda, are you saying that Ironic has kexec, but not tested? | 17:59 |
| devananda | JayF: I suspect it would require changes inside the drivers, but probably not beyond that | 17:59 |
| JayF | devananda: I was thinking it'd be akin to the ipxe changes lucasagomes made? | 18:00 |
| JayF | for IPA it'd also need the agent to coorperate | 18:00 |
| linggao | devananda, JayF, IBM Power8 uses petitboot bookloader, that's why we'd like to add this support in Ironic. | 18:00 |
| JayF | linggao: ah, that makes sense. I'm interested from a perspective of kexec keeping us from needing an extra reboot | 18:01 |
| *** vdrok_ has quit IRC | 18:02 | |
| devananda | actually, it'll need a change to conductor/utils.py -- or to every driver which calls cnoductor.utils.node_power_action | 18:02 |
| devananda | linggao: Power8 is precisely one of those cases where a vendor can certify that kexec is safe on the hardware they are selling and packaging Ironic with | 18:02 |
| devananda | linggao: so I agree that it's a good fit there | 18:02 |
| devananda | linggao: but the implementation must be such that we are not recommending it as part of the default deployment process | 18:03 |
| linggao | devananda, thanks. I'll do some research and prototype. then write a design because I am not confident to write a design yet. | 18:04 |
| linggao | devananda, agree. it should not be a default. | 18:04 |
| linggao | devananda, maybe I can post a blueprint? | 18:05 |
| *** harlowja has quit IRC | 18:05 | |
| *** harlowja_ has joined #openstack-ironic | 18:05 | |
| devananda | linggao: a spec would be a good place for everyone to discuss the implementation :) | 18:06 |
| linggao | devananda, will do. thanks! | 18:06 |
| *** marcoemorais has quit IRC | 18:06 | |
| *** marcoemorais has joined #openstack-ironic | 18:06 | |
| NobodyCam | devananda: we are being careful with Translated string changes ATM how about a change like: https://review.openstack.org/#/c/123108/1/ironic/dhcp/neutron.py | 18:11 |
| *** penick has quit IRC | 18:13 | |
| *** yjiang5 has quit IRC | 18:14 | |
| NobodyCam | oh fresh coffee is ready.. /me refills cup | 18:14 |
| openstackgerrit | A change was merged to openstack/ironic: Imported Translations from Transifex https://review.openstack.org/120701 | 18:20 |
| *** linggao has quit IRC | 18:23 | |
| Shrews | NobodyCam: yeah, i thought of that too for that file. it would eventually get wrapped up in an existing message in _do_node_deploy() that goes to the log and to node.last_error. | 18:24 |
| Shrews | i'm unclear if that qualifies | 18:24 |
| Shrews | i don't *think* it does since it is neither an API error message or config option/desc change | 18:26 |
| rloo | Shrews, NobodyCam: it is the same string as before, and it was _LW() before, so it seems fine to me? It think it falls under bullet 4 of 'What changes are not affected..." | 18:26 |
| NobodyCam | Shrews: ya, and the actual text has not changed | 18:26 |
| rloo | https://wiki.openstack.org/wiki/StringFreeze | 18:26 |
| Shrews | rloo: yeah, it's the same, but where it ends up is different now (log only vs. log&node.last_error) | 18:27 |
| rloo | but the msg was marked for translation before. and it still is. so it shouldn't matter. | 18:28 |
| Shrews | yeah, guess that's true | 18:28 |
| rloo | Shrews: well, "shouldn't matter" to me, but then, I'm not quite sure what/how they are doing things ;) | 18:28 |
| Shrews | rloo: you know all and see all... | 18:29 |
| Shrews | :P | 18:29 |
| rloo | Shrews: ha ha. that would be very scary! | 18:29 |
| NobodyCam | lol I +2'd but did not +a | 18:30 |
| NobodyCam | :-p | 18:30 |
| Shrews | NobodyCam: i'd say go for it | 18:31 |
| *** linggao has joined #openstack-ironic | 18:32 | |
| NobodyCam | Shrews: you also have a +2 on that patch | 18:33 |
| Shrews | lol. lemme rephrase, "I'd say YOU go for it b/c I am chicken" | 18:34 |
| rloo | NobodyCam, Shrews: I would do it, but... are you two OK with the test? | 18:34 |
| devananda | afaik, it's not changing a string so it's fine | 18:35 |
| devananda | it is moving the string out of _LW though, which is new this cycle | 18:35 |
| devananda | and I'm not sure if/how that affects the i18n team | 18:35 |
| rloo | NobodyCam, Shrews, devananda: done. | 18:35 |
| devananda | :) | 18:35 |
| rloo | although I put a note in the bug itself. Not sure if this actually addresses it ;) | 18:36 |
| devananda | I'm troubled that, if I don't set neutron.url, it looks like it will query keystone every time | 18:40 |
| devananda | i'm also troubled that it looks like ironic's consumption of glance and swift don't ever call keystone.get_service_url and instead completely rely on teh config file (but I haven't dug in that much yet) | 18:41 |
| devananda | also, I'm only moderately troubled by this, but we should fix it at some point | 18:42 |
| * rloo is sad that devananda is troubled | 18:45 | |
| *** romcheg has quit IRC | 18:47 | |
| *** achanda has joined #openstack-ironic | 18:47 | |
| NobodyCam | thank you rloo :) | 18:47 |
| rloo | NobodyCam: yw ;) | 18:47 |
| NobodyCam | last call for agenda updates | 18:47 |
| *** openstack has joined #openstack-ironic | 18:56 | |
| -sendak.freenode.net- [freenode-info] if you're at a conference and other people are having trouble connecting, please mention it to staff: http://freenode.net/faq.shtml#gettinghelp | 18:56 | |
| * devananda makes a quick run for more coffee | 18:56 | |
| *** achanda has quit IRC | 19:05 | |
| *** rwsu has joined #openstack-ironic | 19:10 | |
| *** steveh1 has joined #openstack-ironic | 19:12 | |
| openstackgerrit | A change was merged to openstack/ironic: Do not set the context twice when forming RPC objects https://review.openstack.org/122401 | 19:13 |
| openstackgerrit | A change was merged to openstack/python-ironicclient: Updated from global requirements https://review.openstack.org/120651 | 19:13 |
| *** romcheg has joined #openstack-ironic | 19:14 | |
| *** penick has joined #openstack-ironic | 19:15 | |
| *** marcoemorais has quit IRC | 19:16 | |
| *** marcoemorais has joined #openstack-ironic | 19:16 | |
| devananda | greghaynes: hi! want to join the meeting? we're talking abotu the hash ring | 19:17 |
| *** achanda has joined #openstack-ironic | 19:19 | |
| *** rushiagr_away is now known as rushiagr | 19:28 | |
| openstackgerrit | A change was merged to openstack/ironic: Neutron DHCP implementation to raise exception if no ports have VIF https://review.openstack.org/123108 | 19:28 |
| *** yuanying has quit IRC | 19:29 | |
| *** achanda has quit IRC | 19:36 | |
| *** todd_dsm has quit IRC | 19:44 | |
| *** achanda has joined #openstack-ironic | 19:44 | |
| *** todd_dsm has joined #openstack-ironic | 19:49 | |
| *** yuanying has joined #openstack-ironic | 19:55 | |
| *** shakamunyi has joined #openstack-ironic | 19:57 | |
| NobodyCam | thank you all | 20:00 |
| devananda | jroll: NobodyCam: fwiw, we do have a problem right now with features between drivers | 20:00 |
| jroll | we do | 20:00 |
| devananda | specifically whole-disk-images && boot-from-local-disk | 20:00 |
| NobodyCam | :( | 20:00 |
| jroll | the top of my list is to document differences between pxe and ipa | 20:01 |
| devananda | those are not exposed in the API at all | 20:01 |
| jroll | and next is to clear out that list in kilo | 20:01 |
| devananda | and it's totally not obvious that if you use PXE you get XXX and if you use IPA you get YYY | 20:01 |
| jroll | right, which is why we need really good documentation, at a minimum | 20:01 |
| devananda | yup | 20:02 |
| devananda | and we'll STILL end up confusing users | 20:02 |
| *** dprince has quit IRC | 20:02 | |
| jroll | yep | 20:02 |
| NobodyCam | ya | 20:02 |
| jroll | :( | 20:02 |
| *** tatyana has joined #openstack-ironic | 20:02 | |
| jroll | although | 20:02 |
| JayF | I like what devananda noted about how those aren't exposed in the API | 20:02 |
| JayF | why not expose them in the API? | 20:02 |
| jroll | 1) boot from pxe/disk, I don't think is a problem, users don't need to know | 20:02 |
| linggao | devananda, the driver page https://wiki.openstack.org/wiki/Ironic/Drivers only lists the drivers that's in upstream, correct? | 20:02 |
| JayF | Obviously support for features is something that will evolve over time for the various drivers | 20:02 |
| devananda | linggao: correct. | 20:02 |
| JayF | why shouldn't it be documented dynamically? | 20:03 |
| linggao | devananda, ok. thanks. | 20:03 |
| dtantsur | g'night folks | 20:03 |
| devananda | dtantsur: g'night! | 20:03 |
| *** dtantsur is now known as dtantsur|afk | 20:03 | |
| jroll | 2) whole-disk versus split-disk images, I don't think is much of a problem unless someone is running BOTH drivers. otherwise it's only a problem for the poor guy making images | 20:03 |
| jroll | night dtantsur|afk | 20:03 |
| NobodyCam | night dtantsur|afk | 20:04 |
| jroll | and that's covered by good docs | 20:04 |
| jroll | maybe | 20:04 |
| devananda | JayF: if drivers behave differently, it's a failing of the abstraction layer that Ironic should provide | 20:04 |
| jroll | idk | 20:04 |
| devananda | jroll: it's a problem for users who don't understand the choice they are making when they select a driver | 20:04 |
| JayF | devananda: so you're saying we can't ever support a driver that, say, can't implement a Serial Console? | 20:04 |
| devananda | jroll: also Ironic explicitly should support >1 driver in a region. | 20:04 |
| devananda | JayF: nope. | 20:04 |
| devananda | JayF: I said "behave" -- not "expose features" | 20:04 |
| jroll | devananda: agree, agree | 20:05 |
| *** rushiagr is now known as rushiagr_away | 20:05 | |
| jroll | devananda: and this still supports >1 drivers, just need double the images to do so | 20:05 |
| devananda | JayF: where we expect drivers to differ in their features / capabilities, we need to consciously know that we want users to have to make a choice | 20:05 |
| devananda | JayF: and then expose that in the API in a reasonable and programatic way | 20:05 |
| JayF | That sounds like what I was trying to suggest :) | 20:06 |
| devananda | jroll: right. but then as a user (operator) I'm going to get different behavior on pxe-node vs ipa-node | 20:06 |
| jroll | devananda: what different behavior? | 20:06 |
| jroll | devananda: as a user, do I care how the node boots? | 20:06 |
| jroll | devananda: or are we talking deployer here (same question still applies, I think) | 20:07 |
| devananda | JayF: see the /v1/nodes/NNN/validate output | 20:07 |
| devananda | hm, that isn't well documented | 20:08 |
| devananda | it should include samples visible at http://docs.openstack.org/developer/ironic/webapi/v1.html#get--v1-nodes-validate | 20:08 |
| jroll | that's... a thing? | 20:08 |
| jroll | huh. | 20:08 |
| devananda | also, that should probably be mirrored in /drivers/ or something now | 20:08 |
| * jroll tries it out | 20:08 | |
| devananda | jroll: it tests each node.driver.XXX.validate() method | 20:08 |
| devananda | and then returns yes/no/NotImplemented | 20:09 |
| NobodyCam | brb | 20:09 |
| devananda | so that a user can discover which interfaces are[nt] supported | 20:09 |
| devananda | vs which aren't configured appropriately | 20:09 |
| jroll | devananda: hmmm | 20:09 |
| jroll | neat | 20:09 |
| devananda | yea, we need a driver-X command to do the same thing | 20:10 |
| devananda | we have driver-properties now, but no driver-capabilities | 20:10 |
| jroll | yep | 20:11 |
| devananda | heh. too late, really | 20:11 |
| devananda | we're commited to the current API | 20:11 |
| jroll | yeah, I'd want the same data as /node/validate, but for each method in each interface | 20:12 |
| jroll | {"management": {"set_boot_device": {"result": true}}} | 20:12 |
| jroll | etc | 20:12 |
| devananda | jroll: ah. so, i disagree on that quite strongly | 20:12 |
| devananda | jroll: if a driver implements an interface, it needs to implement ALL of it | 20:12 |
| devananda | a driver doesn't get to choose which parts of an interface it implements, or it's just broken the contract | 20:13 |
| jroll | ah, I see | 20:13 |
| jroll | yeah, maybe driver-capabilities is the right thing to do | 20:13 |
| * jroll needs to go afk for a bit | 20:13 | |
| devananda | this is where promoting methods onto the driver interfaces is fairly serious, and why I felt the discovery stuff needed to bake (a lot) more | 20:13 |
| jroll | yeah, that's fair | 20:14 |
| NobodyCam | back | 20:14 |
| devananda | 20:06:48 < jroll> devananda: as a user, do I care how the node boots? | 20:15 |
| devananda | to that question, yes | 20:15 |
| devananda | jroll: if using the PXE driver (net boot only) I can't update my own local kernel on the box | 20:15 |
| NobodyCam | esp you're loading your own image | 20:15 |
| jroll | ahh | 20:15 |
| jroll | shit | 20:15 |
| devananda | jroll: i would need to update the image in glance (which I may not be allowed to do) | 20:15 |
| jroll | yep, touché | 20:15 |
| JayF | If I had a technology time machine, I'd use it to go back in time and try to talk early-openstack into full-disk-images only :) | 20:16 |
| devananda | jroll: and if using the IPA driver, I can't do rebuild --preserve-ephemeral | 20:16 |
| devananda | jroll: which means it's not usable with tripleo | 20:16 |
| jroll | devananda: that's a different limitation, but yeah | 20:16 |
| devananda | jroll: right. that's my point. each of those drivers made different choices along two different dimensions, neither of which are directly visible or discoverable | 20:17 |
| devananda | but both impact users | 20:17 |
| devananda | our abstraction layer is broken along those two points right now | 20:17 |
| jroll | right, I don't disagree with that at all :) | 20:17 |
| devananda | and it's going to hurt users | 20:17 |
| devananda | :) | 20:18 |
| devananda | now, we aren't going to fix that in Juno, so doc'n will need to do. and I'll email the dev list a summary, which should at least help raise awareness | 20:18 |
| NobodyCam | we can put a positive spin on it, I'm sure | 20:19 |
| jroll | yeah | 20:20 |
| NobodyCam | jroll: https://review.openstack.org/#/c/97150/ :-p\ | 20:21 |
| jroll | NobodyCam: I know :P | 20:22 |
| * jroll really goes afk now | 20:23 | |
| * NobodyCam also steps away for a few minutes... | 20:24 | |
| *** yuanying has quit IRC | 20:29 | |
| *** lazy_prince is now known as killer_prince | 20:42 | |
| *** foexle has quit IRC | 20:42 | |
| *** linggao has left #openstack-ironic | 20:49 | |
| *** todd_dsm has quit IRC | 20:52 | |
| *** tatyana has quit IRC | 20:53 | |
| *** penick has quit IRC | 20:59 | |
| *** penick has joined #openstack-ironic | 21:03 | |
| *** todd_dsm has joined #openstack-ironic | 21:03 | |
| *** penick has quit IRC | 21:04 | |
| NobodyCam | ahhh food helps | 21:11 |
| *** andreykurilin has joined #openstack-ironic | 21:12 | |
| *** ChuckC_ has quit IRC | 21:20 | |
| *** penick has joined #openstack-ironic | 21:24 | |
| *** penick has quit IRC | 21:24 | |
| *** yuanying has joined #openstack-ironic | 21:26 | |
| *** penick has joined #openstack-ironic | 21:26 | |
| *** penick has quit IRC | 21:26 | |
| *** penick has joined #openstack-ironic | 21:36 | |
| *** pensu has quit IRC | 21:47 | |
| NobodyCam | gah I keep looking in ironic/common for the periodic_task.py | 21:49 |
| *** harlowja_ has quit IRC | 21:49 | |
| *** harlowja has joined #openstack-ironic | 21:49 | |
| NobodyCam | humm : https://github.com/openstack/ironic/blob/master/ironic/openstack/common/periodic_task.py#L28 | 21:52 |
| *** yuanying has quit IRC | 21:59 | |
| *** romcheg has quit IRC | 22:02 | |
| *** yjiang5 has joined #openstack-ironic | 22:06 | |
| *** zer0c00l has quit IRC | 22:10 | |
| *** ChuckC_ has joined #openstack-ironic | 22:13 | |
| *** andreykurilin has quit IRC | 22:17 | |
| *** todd_dsm has quit IRC | 22:45 | |
| openstackgerrit | Josh Gachnang proposed a change to openstack/ironic-python-agent: Adding support for decommissioning https://review.openstack.org/104379 | 22:50 |
| openstackgerrit | Josh Gachnang proposed a change to openstack/ironic-python-agent: Add versioning to Agent decommission https://review.openstack.org/106859 | 22:53 |
| openstackgerrit | Josh Gachnang proposed a change to openstack/ironic-python-agent: Adding support for decommissioning https://review.openstack.org/104379 | 22:53 |
| NobodyCam | brb walkies time | 22:55 |
| *** yuanying has joined #openstack-ironic | 22:55 | |
| *** penick has quit IRC | 23:02 | |
| yjiang5 | devananda: I'm trying to figure out how to use ironic to deploy my lab's physical machines. According to http://devstack.org/lib/baremetal.html and http://devstack.org/lib/ironic.html, seems now I can use baremetal/devstack to achieve it, but I can't use ironic/devstack yet, am I right? | 23:07 |
| jroll | yjiang5: why can't you use devstack? | 23:10 |
| jroll | err, ironic/devstack | 23:10 |
| NobodyCam | I've don'e it with TripleO | 23:11 |
| yjiang5 | jroll: thanks for reply. I just noticed that the betametal side will try to setup the dnsmasq for the pxe network, while ironic side seems missed that. (Currently I'm stuck on the dnsmasq) | 23:12 |
| jroll | ironic uses neutron for dhcp (dnsmasq) | 23:13 |
| jroll | yjiang5: ironic/devstack should work; if it doesn't, that's a bug | 23:13 |
| NobodyCam | there is also this: https://review.openstack.org/#/c/118614/ (though not quite ready to land) | 23:14 |
| jroll | ah yes!!! | 23:15 |
| jroll | yjiang5: ^ instructions for you | 23:15 |
| yjiang5 | jroll: thanks. I didn't find a clear doc talking about what shall I do for neutron side when I try to boot a ironic node. I try to "ps aux| grep dnsmaq", and it try to set the dnsmasq to a virtual device, not to the eth1 (which is for the PXE network). | 23:15 |
| yjiang5 | jroll: NobodyCam: thanks and I will have a look !! | 23:16 |
| NobodyCam | yjiang5: that should help just may be a little ruff around the edges | 23:16 |
| yjiang5 | NobodyCam: really thanks. I really should asking earlier in IRC other than struggling myself for so long time . Will have a careful reading :) | 23:18 |
| *** penick has joined #openstack-ironic | 23:18 | |
| NobodyCam | hummm | 23:23 |
| yjiang5 | NobodyCam: Hi, when I try to restart the neutron-plugin-openvsiwtch-agent, it failed because unregonized service. Do you have any hints on it? | 23:26 |
| NobodyCam | if I take a node and update it to have a Bad ssh user name it will get set to maintenance mode (this is expected) if I then set maintenance mode to false it never again gets set into maintenance, even though I have NOT corrected the bad username.... is this correct | 23:27 |
| NobodyCam | can someone check me on htat? | 23:27 |
| jroll | NobodyCam: that doesn't sound correct | 23:28 |
| *** yuanying has quit IRC | 23:29 | |
| NobodyCam | I just tested it and my node stayed not in moantenance mode for at least 15 minutes and I saw reservation get set to the conductors name several times | 23:29 |
| NobodyCam | *maintenance even | 23:29 |
| jroll | strange | 23:29 |
| *** Haomeng has joined #openstack-ironic | 23:30 | |
| *** lucasagomes has quit IRC | 23:31 | |
| *** Haomeng|2 has quit IRC | 23:31 | |
| yjiang5 | NobodyCam: I find the service in the end. SHould it be neutron-openvswitch-agent ? Or it's neutron-plugin-openvsiwtch-agen in the ubuntu distro? | 23:33 |
| NobodyCam | yjiang5: for the restart? (line #364) | 23:35 |
| yjiang5 | NobodyCam: yes | 23:35 |
| NobodyCam | yjiang5: neutron-plugin-openvswitch-agent is an ubuntu package | 23:37 |
| yjiang5 | NobodyCam: got it and thanks. | 23:37 |
| NobodyCam | :) | 23:38 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!