| wanyen_ | Deva: I want to continue the discussion of ilo driver completion status mechanism. With the mechanism to send the auth token in a floppy image, we need to upload the floppy umage to a web server for iLO to get access to it. | 00:05 |
|---|---|---|
| wanyen_ | Will it be okay from security perscpective? | 00:06 |
| wanyen_ | s/pserscpective/perspective | 00:06 |
| NobodyCam | grrr | 00:17 |
| NobodyCam | DuplicateOptError: duplicate option: rpc_backend | 00:18 |
| NobodyCam | seems celiometer had the same issue | 00:19 |
| devananda | wanyen_: what is this webserver? | 00:30 |
| *** matsuhashi has joined #openstack-ironic | 00:30 | |
| devananda | mrda: hi! are you still looking for things to do? :) | 00:30 |
| NobodyCam | devananda: got a second to give a quick look over a error? | 00:32 |
| devananda | NobodyCam: sure | 00:32 |
| NobodyCam | http://paste.openstack.org/show/EGd9ABNGPZNa9gzSddZG/ | 00:32 |
| devananda | mrda: if so, want to fix this patch up? https://review.openstack.org/#/c/76293/ | 00:32 |
| NobodyCam | top bit is the test code failing, I have 42 tests that fail all the same way | 00:33 |
| mrda | hey devananda, sure. I'm working on +bug/1271317 but I'll take more | 00:33 |
| mrda | devananda: I'll take a look, thanks | 00:33 |
| devananda | mrda: also, i don't understand this patch of yours: 76094 | 00:36 |
| devananda | mrda: seems to contradict several other patches we just landed explicitly allowing that field to be viewed and set | 00:36 |
| devananda | mrda: oooh. I see. | 00:38 |
| mrda | so this allows it to be set, but it is neither mandatory, nor has a default | 00:38 |
| mrda | I wasn't sure about Yuriy's comment though | 00:38 |
| NobodyCam | brb | 00:40 |
| mrda | devananda: do you have a concern with this? | 00:40 |
| devananda | mrda: no -- i can confirm lucas' bug is valid and your patch fixes it | 00:40 |
| devananda | mrda: also, if you're working on this, it'd be grand if you do two more things | 00:41 |
| mrda | sure | 00:41 |
| devananda | mrda: 1. remove 'maintenance' from the list of internal_attrs in v1/node.py so that we CAN change it | 00:41 |
| devananda | taht should be very simple | 00:42 |
| devananda | and then add a helper function in the API to check for the maintenance bit and reject certain [*] requests if it is set | 00:43 |
| devananda | [* exactly which ones isn't written down yet] | 00:43 |
| mrda | lol, you're not making this easy for me devanada :) | 00:44 |
| mrda | Did you want to provide some criteria for this? either as a review comment or in the bug itself? | 00:44 |
| wanyen_ | [Deva] iLO virtual media requires media location from http or https. Therefore, this is the web server for hosting virtual media images. | 00:45 |
| devananda | mrda: i think this is separate from the bug | 00:45 |
| devananda | mrda: bug: GET /v1/nodes/ returns maintenance field when it shouldn't. you've got a fix for that -- i'm about to +2 it | 00:46 |
| devananda | mrda: making 'maintenance' user-settable: small patch, folks will easily +2 that | 00:46 |
| mrda | ok, new patch, gotcha | 00:46 |
| devananda | then the work of "make maintenance DO something" can be started in earnest, and other -core folks are likely to dogpile on that too | 00:47 |
| mrda | lol, of course. Should be fun. | 00:47 |
| devananda | first thing for that would be "dont allow a new deploy to be started if a node is in maintenance" | 00:47 |
| * mrda wishes he was in CA next week to have that discussion f2f :( | 00:47 | |
| devananda | eg, NodeStatesController.provision should chec | 00:47 |
| devananda | check maintenance state and return error if it is set | 00:48 |
| devananda | and the requested state is not DELETED | 00:48 |
| devananda | mrda: oh! and i just found a bug in the API for this, which you'll find quickly too | 00:49 |
| devananda | mrda: as soon as I remove 'maintenance' from the internal_attrs list and try to change it, i get | 00:49 |
| devananda | Invalid input for field/attribute maintenance. Value: 'True'. Wrong type. Expected '<type 'bool'>', got '<type 'unicode'>' | 00:49 |
| devananda | ok - i need to run now :) | 00:50 |
| devananda | thanks! | 00:50 |
| mrda | thanks devananda! | 00:50 |
| devananda | wanyen_: what kind of web server? Ironic doesn't run a separate web server today, so are you introducing a new external dependency? is this part of ilo chassis manager? also - can you raise this on the mailing list and I promise to get back to you later tonight :) | 00:51 |
| devananda | wanyen_: also, could you upload it to glance, if it's just a disk image? | 00:51 |
| NobodyCam | :) we'll still be on irc | 00:55 |
| wanyen_ | [Deva] This web server can be a backend store for Glance. Users will upload the virtual media image to the web server and use glance image-create with <--location ulr> to register the image. iLO need access the image via http or https. It does not integrate with Glance. | 00:59 |
| wanyen_ | [Deva] I meant iLO Iornic driver uses glance to query image info but iLO RIBCL requires image location to be reference via http/https | 01:04 |
| *** killer_prince has quit IRC | 01:08 | |
| *** lazy_prince has joined #openstack-ironic | 01:14 | |
| * NobodyCam calls it a day... Good night All | 01:14 | |
| *** lazy_prince is now known as killer_prince | 01:14 | |
| openstackgerrit | Yongli He proposed a change to openstack/ironic: Remove extraneous vim configuration comments for ironic https://review.openstack.org/73160 | 01:22 |
| *** rloo has quit IRC | 01:25 | |
| *** lnxnut has quit IRC | 01:33 | |
| *** nosnos has joined #openstack-ironic | 01:41 | |
| *** rongze has joined #openstack-ironic | 02:00 | |
| openstackgerrit | Haomeng,Wang proposed a change to openstack/ironic: Re-enable lazy translation https://review.openstack.org/74626 | 02:12 |
| openstackgerrit | Jenkins proposed a change to openstack/ironic: Updated from global requirements https://review.openstack.org/75565 | 02:32 |
| Haomeng | devananda: ping | 02:35 |
| Haomeng | devananda: want to confirm with you about a bug - https://bugs.launchpad.net/ironic/+bug/1284781, I understand we should remove the mandatory fields definnation from API Object, because it is defineds in Obeject JsonPatchType class mandatory_attrs method already, right? | 02:38 |
| openstackgerrit | Jenkins proposed a change to openstack/python-ironicclient: Updated from global requirements https://review.openstack.org/75585 | 02:38 |
| Haomeng | devananda: I left comments in the bug already, check it if you have time, thanks:) | 02:46 |
| *** jcooley_ has quit IRC | 03:23 | |
| *** jcooley_ has joined #openstack-ironic | 03:27 | |
| *** matsuhashi has quit IRC | 03:37 | |
| *** harlowja is now known as harlowja_away | 03:57 | |
| *** jcooley_ has quit IRC | 04:27 | |
| *** KanagarajM_ has joined #openstack-ironic | 04:28 | |
| *** matsuhashi has joined #openstack-ironic | 04:32 | |
| *** ap has joined #openstack-ironic | 04:40 | |
| *** ap has quit IRC | 04:40 | |
| *** rongze has quit IRC | 04:44 | |
| *** rongze has joined #openstack-ironic | 04:44 | |
| *** rongze has quit IRC | 04:49 | |
| *** jcooley_ has joined #openstack-ironic | 04:55 | |
| *** jcooley_ has quit IRC | 05:00 | |
| *** jcooley_ has joined #openstack-ironic | 05:01 | |
| *** killer_prince has quit IRC | 05:09 | |
| *** coolsvap has joined #openstack-ironic | 05:11 | |
| *** jcooley_ has quit IRC | 05:12 | |
| *** lazy_prince has joined #openstack-ironic | 05:13 | |
| *** lazy_prince is now known as killer_prince | 05:13 | |
| *** rongze has joined #openstack-ironic | 05:15 | |
| *** rongze_ has joined #openstack-ironic | 05:16 | |
| *** rongze has quit IRC | 05:17 | |
| *** jcooley_ has joined #openstack-ironic | 05:18 | |
| *** killer_prince has quit IRC | 05:19 | |
| *** rongze_ has quit IRC | 05:21 | |
| *** jcooley_ has quit IRC | 05:21 | |
| *** lazy_prince has joined #openstack-ironic | 05:28 | |
| *** lazy_prince is now known as killer_prince | 05:28 | |
| *** killer_prince2 has joined #openstack-ironic | 05:37 | |
| *** killer_prince has quit IRC | 05:57 | |
| *** lazy_prince has joined #openstack-ironic | 05:58 | |
| *** lazy_prince is now known as killer_prince | 05:58 | |
| *** loki_ has joined #openstack-ironic | 06:00 | |
| openstackgerrit | Jenkins proposed a change to openstack/ironic: Imported Translations from Transifex https://review.openstack.org/71192 | 06:07 |
| *** vkozhukalov_ has joined #openstack-ironic | 06:10 | |
| *** rongze has joined #openstack-ironic | 06:15 | |
| *** vkozhukalov_ has quit IRC | 06:16 | |
| *** jcooley_ has joined #openstack-ironic | 06:19 | |
| *** rongze has quit IRC | 06:19 | |
| *** rongze has joined #openstack-ironic | 06:21 | |
| *** rwsu has quit IRC | 06:27 | |
| openstackgerrit | yangxurong proposed a change to openstack/ironic: Fix params order in assertEqual https://review.openstack.org/76077 | 06:31 |
| *** jcooley_ has quit IRC | 06:37 | |
| *** saju_m has joined #openstack-ironic | 06:50 | |
| *** jcooley_ has joined #openstack-ironic | 07:15 | |
| *** ifarkas has joined #openstack-ironic | 07:20 | |
| *** rameshg87 has joined #openstack-ironic | 07:21 | |
| *** jcooley_ has quit IRC | 07:21 | |
| *** rameshg871 has joined #openstack-ironic | 07:31 | |
| *** rameshg871 has left #openstack-ironic | 07:33 | |
| *** rameshg871 has joined #openstack-ironic | 07:33 | |
| *** rameshg87 has quit IRC | 07:34 | |
| rameshg871 | good morning Haomeng: | 07:34 |
| *** xyz__ has joined #openstack-ironic | 07:35 | |
| *** dshulyak has joined #openstack-ironic | 07:39 | |
| *** xyz__ has quit IRC | 07:39 | |
| *** vkozhukalov_ has joined #openstack-ironic | 07:48 | |
| Haomeng | rameshg871: morning:) | 07:52 |
| *** romcheg has joined #openstack-ironic | 08:03 | |
| *** KanagarajM_ has quit IRC | 08:27 | |
| *** Kanagaraj has joined #openstack-ironic | 08:27 | |
| *** rongze has quit IRC | 08:34 | |
| rameshg871 | Haomeng: i just wanted to talk regarding the review comments that you had for https://review.openstack.org/#/c/76144/ | 08:42 |
| Haomeng | rameshg871: sure | 08:43 |
| Haomeng | rameshg871: let me look into it first | 08:44 |
| rameshg871 | https://review.openstack.org/#/c/76144/1/ironic/drivers/modules/ilo_common.py | 08:44 |
| Haomeng | ok | 08:44 |
| rameshg871 | the error_msg variable doesn't have _() surrounding it, but it is passed to IloLicenseError exception which has i18n enabled | 08:45 |
| Haomeng | last two comments | 08:45 |
| rameshg871 | wouldn't that be fine ? | 08:45 |
| rameshg871 | yes, last two comments | 08:45 |
| Haomeng | ok | 08:45 |
| Haomeng | fine | 08:45 |
| Haomeng | I will reply you in the patch as inline comments:) | 08:45 |
| Haomeng | no worries:) | 08:45 |
| Haomeng | :) | 08:45 |
| rameshg871 | okay :-) | 08:46 |
| rameshg871 | thanks | 08:46 |
| Haomeng | rameshg871: welcome & thanks for your patch:) | 08:46 |
| rameshg871 | thanks :-) | 08:48 |
| *** jistr has joined #openstack-ironic | 08:49 | |
| Haomeng | rameshg871: one question | 08:50 |
| rameshg871 | yes Haomeng: | 08:50 |
| Haomeng | rameshg871: did you encounter such "tox -evenv -- echo 'done'" issue these days, looks it is not stable to setup the ut env to install test requirements, maybe it is our networking issue:) | 08:50 |
| *** killer_prince has quit IRC | 08:51 | |
| Haomeng | rameshg871: such as - Download error on https://pypi.python.org/simple/pbr/: [Errno 110] Connection timed out -- Some packages may not be found! | 08:51 |
| rameshg871 | Haomeng: yes, we encounter such issues some times .. | 08:51 |
| rameshg871 | not everytime .. | 08:51 |
| Haomeng | rameshg871: should be networking not stable I think | 08:52 |
| Haomeng | rameshg871: need luck:) | 08:52 |
| Haomeng | let me try again:) | 08:52 |
| rameshg871 | some of our machines are behind proxy server, and we were thinking that it was because of the proxy | 08:52 |
| Haomeng | yes | 08:52 |
| *** rsacharya has quit IRC | 08:52 | |
| *** rsacharya_ is now known as rsacharya | 08:52 | |
| *** ndipanov has joined #openstack-ironic | 08:53 | |
| rameshg871 | Haomeng: whenever we have such issue, we go into the virtual environment and make sure that we do "pip install <module>" for all the modules in requirements.txt and test-requirements.txt | 08:53 |
| rameshg871 | after activating the virtual environment of course :-) | 08:53 |
| rameshg871 | and then try tox again. it has worked | 08:54 |
| *** viktors has joined #openstack-ironic | 08:55 | |
| Haomeng | rameshg871: ok | 08:58 |
| Haomeng | great idea | 08:58 |
| Haomeng | install the missing libs by manually:) | 08:59 |
| rameshg871 | :-) | 08:59 |
| Haomeng | :) | 08:59 |
| *** aignatov_ is now known as aignatov | 09:00 | |
| *** lazy_prince has joined #openstack-ironic | 09:03 | |
| *** lazy_prince is now known as killer_pr | 09:03 | |
| *** killer_pr is now known as killer_prince | 09:03 | |
| *** jcooley_ has joined #openstack-ironic | 09:03 | |
| *** jcooley_ has quit IRC | 09:09 | |
| *** derekh has joined #openstack-ironic | 09:14 | |
| *** killer_prince has quit IRC | 09:18 | |
| *** killer_prince has joined #openstack-ironic | 09:19 | |
| *** jcooley_ has joined #openstack-ironic | 09:20 | |
| rameshg871 | Haomeng: what time you will be available on the irc usually ? | 09:21 |
| *** athomas has quit IRC | 09:22 | |
| *** pradipta has joined #openstack-ironic | 09:25 | |
| *** jcooley_ has quit IRC | 09:25 | |
| *** max_lobur has joined #openstack-ironic | 09:28 | |
| *** Kanagaraj has quit IRC | 09:36 | |
| *** athomas has joined #openstack-ironic | 09:37 | |
| *** martyntaylor has joined #openstack-ironic | 09:37 | |
| mdurnosvistov | Morning all! :) | 09:38 |
| *** lucasagomes has joined #openstack-ironic | 09:43 | |
| yuriyz | morning Ironic | 09:44 |
| *** mdurnosvistov_lt has joined #openstack-ironic | 09:44 | |
| openstackgerrit | lokesh s proposed a change to openstack/ironic: Ironic deployment docs https://review.openstack.org/74282 | 09:45 |
| Haomeng | rameshg871: my time is +8 | 09:48 |
| Haomeng | rameshg871: available from 8am - 10pm:) | 09:49 |
| Haomeng | rameshg871: what is your time zone | 09:49 |
| Haomeng | rameshg871: now, my local time is around 6pm:) | 09:50 |
| *** tatyana has joined #openstack-ironic | 09:54 | |
| *** mdurnosvistov_lt has quit IRC | 09:56 | |
| *** viktors has quit IRC | 09:57 | |
| *** jcooley_ has joined #openstack-ironic | 09:59 | |
| *** mdurnosvistov_lt has joined #openstack-ironic | 09:59 | |
| *** jcooley_ has quit IRC | 10:06 | |
| *** romcheg has quit IRC | 10:14 | |
| *** mdurnosvistov_lt has quit IRC | 10:16 | |
| *** mdurnosvistov_lt has joined #openstack-ironic | 10:18 | |
| openstackgerrit | Michael Davies proposed a change to openstack/ironic: Add option to sync node power state from DB https://review.openstack.org/76293 | 10:18 |
| *** mdurnosvistov_lt has quit IRC | 10:19 | |
| mrda | devananda: there's some unit tests added - let me know if you have any feedback :) | 10:20 |
| *** mrda is now known as mrda_away | 10:21 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Fix race condition when deleting a node https://review.openstack.org/76178 | 10:21 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Do not delete a Node which is not powered off https://review.openstack.org/76258 | 10:26 |
| *** aignatov is now known as aignatov_ | 10:47 | |
| *** mdurnosvistov_lt has joined #openstack-ironic | 10:53 | |
| *** mdurnosvistov_lt has quit IRC | 10:53 | |
| *** romcheg has joined #openstack-ironic | 10:54 | |
| rameshg871 | Haomeng: my time is +5.30, available from 10am-6pm. | 10:56 |
| Haomeng | rameshg871: :) | 10:57 |
| rameshg871 | good morning mdurnosvistov: yuriyz: | 10:57 |
| romcheg | Morningeveryone | 10:58 |
| lucasagomes | morning folks :) | 10:59 |
| rameshg871 | good morning romcheg: lucasagomes: | 10:59 |
| *** Xurong has quit IRC | 11:12 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Add provision_updated_at to node's resource https://review.openstack.org/76494 | 11:13 |
| *** pbrooko has joined #openstack-ironic | 11:17 | |
| *** aignatov_ is now known as aignatov | 11:24 | |
| openstackgerrit | Shuangtai Tian proposed a change to openstack/ironic: Change the import order in log_hangder https://review.openstack.org/76500 | 11:24 |
| *** killer_prince2 has quit IRC | 11:31 | |
| *** rameshg871 has left #openstack-ironic | 11:31 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: API: Expose a way to start/stop the console https://review.openstack.org/72998 | 11:33 |
| *** Alexei_987 has joined #openstack-ironic | 11:37 | |
| *** matsuhashi has quit IRC | 11:46 | |
| *** jcooley_ has joined #openstack-ironic | 11:48 | |
| openstackgerrit | Victor Sergeyev proposed a change to openstack/ironic: WIP: sync oslo.db code https://review.openstack.org/76204 | 11:53 |
| *** jcooley_ has quit IRC | 11:54 | |
| *** killer_prince has quit IRC | 11:57 | |
| *** coolsvap has quit IRC | 12:02 | |
| *** loki_ has quit IRC | 12:05 | |
| *** lucasagomes has quit IRC | 12:06 | |
| *** lucasagomes has joined #openstack-ironic | 12:07 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: API: Expose a way to start/stop the console https://review.openstack.org/72998 | 12:17 |
| *** lazy_prince has joined #openstack-ironic | 12:23 | |
| *** lazy_prince is now known as killer_prince | 12:23 | |
| *** lucasagomes is now known as lucas-hungry | 12:25 | |
| *** zul has joined #openstack-ironic | 12:32 | |
| openstackgerrit | Fengqian.gao proposed a change to openstack/ironic: Python 3: replace "im_self" by "__self__" https://review.openstack.org/76517 | 12:35 |
| *** jdob has joined #openstack-ironic | 12:44 | |
| *** viktors has joined #openstack-ironic | 12:45 | |
| viktors | lucas-hungry: hello | 12:46 |
| openstackgerrit | Victor Sergeyev proposed a change to openstack/ironic: Sync common db code from Oslo https://review.openstack.org/76204 | 12:49 |
| viktors | lucas-hungry: please see ^ | 12:50 |
| viktors | lucas-hungry: after lunch I guess :) | 12:51 |
| *** killer_prince is now known as lazy_prince | 13:05 | |
| *** killer_prince has joined #openstack-ironic | 13:09 | |
| *** lazy_prince has quit IRC | 13:09 | |
| openstackgerrit | Victor Sergeyev proposed a change to openstack/ironic: WIP: Use oslo.db module (DRAFT) https://review.openstack.org/42159 | 13:33 |
| openstackgerrit | Victor Sergeyev proposed a change to openstack/ironic: Sync common db code from Oslo https://review.openstack.org/76204 | 13:33 |
| *** jcooley_ has joined #openstack-ironic | 13:36 | |
| *** nosnos has quit IRC | 13:41 | |
| *** zul has quit IRC | 13:42 | |
| *** rloo has joined #openstack-ironic | 13:43 | |
| *** zul has joined #openstack-ironic | 13:44 | |
| *** jistr is now known as jistr|english | 13:44 | |
| *** lucas-hungry is now known as lucasagomes | 13:48 | |
| lucasagomes | viktors, ack :D | 13:49 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: API: Expose a way to start/stop the console https://review.openstack.org/72998 | 13:50 |
| viktors | lucasagomes: oh, you are not hungry now :) | 13:53 |
| viktors | lucasagomes: I've made a sync of oslo.db code to ironic. But this sync is not trivial, because we already removed global engine from oslo.db. Can you please take a look at this patch? | 13:53 |
| lucasagomes | viktors, sure will do | 13:53 |
| viktors | lucasagomes: thanks | 13:53 |
| *** jcooley_ has quit IRC | 14:16 | |
| *** matty_dubs|gone is now known as matty_dubs | 14:23 | |
| *** jbjohnso has joined #openstack-ironic | 14:35 | |
| *** mdurnosvistov_lt has joined #openstack-ironic | 14:36 | |
| *** mdurnosvistov_lt has quit IRC | 14:36 | |
| *** jistr|english is now known as jistr | 14:40 | |
| *** rameshg87 has joined #openstack-ironic | 14:46 | |
| *** rameshg87 has quit IRC | 14:59 | |
| *** saju_m has quit IRC | 15:02 | |
| *** linggao has joined #openstack-ironic | 15:02 | |
| NobodyCam | Good Morning Ironic | 15:08 |
| romcheg | Morning NobodyCam, lucasagomes and the others! | 15:09 |
| NobodyCam | morning romcheg :) | 15:09 |
| lucasagomes | morning NobodyCam romcheg :D | 15:10 |
| NobodyCam | morning lucasagomes :) | 15:12 |
| NobodyCam | lucasagomes: I broke the nova driver tests. but give me a second looks at scroll bac in OOO to see who pinged me lastnight | 15:13 |
| *** max_lobur has quit IRC | 15:13 | |
| lucasagomes | NobodyCam, hey oh I closed my irc client so I don't have the scrollback :( | 15:14 |
| lucasagomes | NobodyCam, no worries about the tests, we are going to change a good bit of the code | 15:15 |
| lucasagomes | after the changes in Ironic | 15:15 |
| NobodyCam | lol ... let me push up what I have | 15:15 |
| NobodyCam | i'm getting a reall odd error | 15:15 |
| NobodyCam | http://paste.openstack.org/show/EGd9ABNGPZNa9gzSddZG/ | 15:16 |
| *** jrist_ has quit IRC | 15:17 | |
| *** jrist_ has joined #openstack-ironic | 15:17 | |
| *** rwsu has joined #openstack-ironic | 15:18 | |
| *** max_lobur has joined #openstack-ironic | 15:19 | |
| *** rameshg87 has joined #openstack-ironic | 15:19 | |
| lucasagomes | hah lemme see | 15:20 |
| lucasagomes | NobodyCam, err what haha | 15:20 |
| NobodyCam | just pushed up what I had... I think... still need more coffee | 15:21 |
| lucasagomes | off the top of my head idk | 15:21 |
| lucasagomes | I gotta take a look to see why it's failing like that | 15:21 |
| NobodyCam | ya we dont touch rpc_backend | 15:21 |
| lucasagomes | yeah | 15:21 |
| lucasagomes | looks a bit unrelated | 15:21 |
| NobodyCam | I get 42 of theose errors | 15:21 |
| lucasagomes | btw pyghmi coming to fedora: http://pkgs.fedoraproject.org/cgit/python-pyghmi.git/ | 15:21 |
| lucasagomes | :) | 15:21 |
| NobodyCam | I rebsed | 15:21 |
| lucasagomes | and ironic is being worked as well :D | 15:21 |
| NobodyCam | nice!!!! | 15:21 |
| lucasagomes | 42!? | 15:22 |
| lucasagomes | jeez, yeah... it's not something in ur venv? | 15:22 |
| lucasagomes | did you rebuild it? | 15:22 |
| NobodyCam | yea | 15:24 |
| NobodyCam | watching zuul right now | 15:25 |
| lucasagomes | ack | 15:25 |
| NobodyCam | found this: (which scared me): http://lists.openstack.org/pipermail/openstack-dev/2013-February/005865.html | 15:25 |
| NobodyCam | see the fix attached to https://bugs.launchpad.net/ceilometer/+bug/1130952 | 15:26 |
| lucasagomes | hmm | 15:27 |
| NobodyCam | brb morning walies | 15:27 |
| romcheg | Have to go to my Spanish class | 15:30 |
| romcheg | Will be later | 15:30 |
| romcheg | davidlenwell: ^ | 15:30 |
| davidlenwell | k | 15:30 |
| davidlenwell | hasta la vista | 15:31 |
| romcheg | davidlenwell: (: Tengo que correr. Adios :) | 15:31 |
| davidlenwell | a qué hora va a regresar? | 15:32 |
| NobodyCam | oh it to early to do that to :-p | 15:32 |
| NobodyCam | morning davidlenwell :) | 15:32 |
| lucasagomes | :P I have a pretty rusty spanish | 15:33 |
| davidlenwell | morning NobodyCam | 15:33 |
| davidlenwell | I suck at it too .. but I've been helping my girlfriend study for her spanish class | 15:33 |
| romcheg | davidlenwell: voy a volver en dos horas | 15:33 |
| davidlenwell | bueno | 15:33 |
| lucasagomes | coming back in 2 hours | 15:33 |
| lucasagomes | :D | 15:33 |
| davidlenwell | si | 15:34 |
| NobodyCam | lucasagomes: lol seems I also have some pep9 issues in nova driver : ./nova/tests/virt/ironic/test_driver.py:232:68: E502 the backslash is redundant between brackets | 15:34 |
| NobodyCam | doh | 15:34 |
| romcheg | davidlenwell: I have the same story about my girlfriend and her piano classes :) | 15:34 |
| lucasagomes | heh my mother language is portuguese so I can understand a couple of things in spanish but I don't speak it at all | 15:34 |
| lucasagomes | NobodyCam, hah | 15:35 |
| lucasagomes | NobodyCam, that's an easy one | 15:35 |
| NobodyCam | :) | 15:35 |
| matty_dubs | lucasagomes: Portuguese always confuses me, because it _looks_ like Spanish, except everything is different | 15:35 |
| *** romcheg has quit IRC | 15:36 | |
| openstackgerrit | Imre Farkas proposed a change to openstack/ironic: Add support for custom libvirt uri https://review.openstack.org/74062 | 15:39 |
| *** aignatov is now known as aignatov_ | 15:41 | |
| lucasagomes | matty_dubs, lol | 15:42 |
| lucasagomes | matty_dubs, yeah, it's very similar | 15:42 |
| lucasagomes | but as u said, some words are the same but the meaning is complete different | 15:43 |
| *** lnxnut has joined #openstack-ironic | 15:44 | |
| *** romcheg has joined #openstack-ironic | 15:53 | |
| *** rloo_ has joined #openstack-ironic | 15:53 | |
| *** rloo has quit IRC | 15:53 | |
| *** rloo_ has quit IRC | 15:54 | |
| *** rloo has joined #openstack-ironic | 15:55 | |
| *** jrist_ is now known as jrist | 15:55 | |
| *** max_lobur has quit IRC | 15:56 | |
| davidlenwell | As soon as I finish installing ubuntu on my boot node.. I have a real 4 node cluster w/ipmi to work with! | 15:59 |
| davidlenwell | it was already done with 13.04 but I did a do-release-upgrade that hosed it somehow.. so im just starting it over from scratch | 16:00 |
| romcheg | Gah! Confused the time. | 16:01 |
| davidlenwell | so no class? | 16:02 |
| romcheg | davidlenwell: No, I'm awailable now | 16:02 |
| romcheg | It will be later, so we can have a chat, if you have time | 16:02 |
| davidlenwell | give me a minute to start this installation of ubuntu | 16:03 |
| romcheg | np | 16:05 |
| *** dtalton has joined #openstack-ironic | 16:06 | |
| *** dtalton has left #openstack-ironic | 16:07 | |
| davidlenwell | romcheg: Can you please link me to your review again too ? | 16:07 |
| openstackgerrit | Imre Farkas proposed a change to openstack/ironic: Add support for custom libvirt uri https://review.openstack.org/74062 | 16:07 |
| *** rameshg87 has quit IRC | 16:08 | |
| *** romcheg has quit IRC | 16:09 | |
| *** romcheg2 has joined #openstack-ironic | 16:11 | |
| romcheg2 | davidlenwell: https://review.openstack.org/#/c/75894/ | 16:11 |
| romcheg2 | davidlenwell: We can set up a g+/skype call if you want | 16:11 |
| davidlenwell | I've stopped doing skype since they're owned by ms | 16:12 |
| romcheg2 | g+? | 16:12 |
| davidlenwell | sure | 16:12 |
| davidlenwell | dlenwell@gmail | 16:12 |
| *** dshulyak has quit IRC | 16:13 | |
| romcheg2 | davidlenwell: I've sent you an invite | 16:22 |
| davidlenwell | really? | 16:23 |
| davidlenwell | to dlenwell@gmail.com? I have to recieved one | 16:23 |
| NobodyCam | lucasagomes: ok seems it mush have been something on my mac. tests seem to be passing for Mr J | 16:26 |
| *** jcooley_ has joined #openstack-ironic | 16:31 | |
| *** rloo has quit IRC | 16:32 | |
| *** ifarkas has quit IRC | 16:32 | |
| *** rloo has joined #openstack-ironic | 16:32 | |
| *** datajerk has quit IRC | 16:33 | |
| *** rsacharya has left #openstack-ironic | 16:37 | |
| *** aignatov_ is now known as aignatov | 16:38 | |
| *** aignatov is now known as aignatov_ | 16:41 | |
| shortstop | Warning: Review Nag :-p - Could I please get reviews on 73054? It'd be great to get some comments since I'd be able to incorporate any remaining changes sooner. | 16:42 |
| NobodyCam | :) | 16:43 |
| *** rloo has quit IRC | 16:49 | |
| *** rloo has joined #openstack-ironic | 16:49 | |
| *** coolsvap has joined #openstack-ironic | 16:56 | |
| *** rloo has quit IRC | 16:57 | |
| *** jistr has quit IRC | 16:57 | |
| *** yuriyz has quit IRC | 16:58 | |
| *** rloo has joined #openstack-ironic | 16:58 | |
| NobodyCam | shortstop: reviewed | 16:59 |
| shortstop | NobodyCam, awesome, thanks :) | 17:00 |
| *** killer_prince is now known as lazy_prince | 17:01 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: PXE clean_up() to remove the pxe_deploy_key parameter https://review.openstack.org/71879 | 17:05 |
| *** saju_m has joined #openstack-ironic | 17:06 | |
| *** tatyana has quit IRC | 17:08 | |
| *** lazy_prince is now known as killer_prince | 17:08 | |
| NobodyCam | bbt brb | 17:12 |
| *** coolsvap has quit IRC | 17:25 | |
| *** coolsvap has joined #openstack-ironic | 17:25 | |
| *** vkozhukalov_ has quit IRC | 17:32 | |
| *** coolsvap has quit IRC | 17:34 | |
| NobodyCam | lucasagomes: nova driver patches rebased and pushd up with some other work too | 17:35 |
| lucasagomes | NobodyCam, w00t | 17:35 |
| lucasagomes | I will take a look later | 17:35 |
| *** coolsvap has joined #openstack-ironic | 17:35 | |
| NobodyCam | walkies time bbiafm | 17:35 |
| NobodyCam | :) | 17:35 |
| *** Alexei_987 has quit IRC | 17:37 | |
| *** datajerk has joined #openstack-ironic | 17:39 | |
| NobodyCam | gah nova.conf.sample is not up to date. | 17:48 |
| NobodyCam | :-p | 17:48 |
| NobodyCam | fixing volume driver | 17:48 |
| *** matty_dubs is now known as matty_dubs|lunch | 17:49 | |
| *** jcooley_ has quit IRC | 17:52 | |
| *** jcooley_ has joined #openstack-ironic | 17:56 | |
| *** datajerk has quit IRC | 17:58 | |
| lucasagomes | NobodyCam, hah | 17:59 |
| lucasagomes | NobodyCam, btw, we should allow the power state operations in a node which is in maintenancement mode right? | 17:59 |
| NobodyCam | lucasagomes: I think as a operator I would want that | 18:00 |
| lucasagomes | NobodyCam, ack | 18:00 |
| *** jcooley_ has quit IRC | 18:01 | |
| devananda | good morning, all | 18:01 |
| NobodyCam | Good morning devananda :) | 18:02 |
| lucasagomes | morning devananda | 18:02 |
| * devananda reads scrollback | 18:03 | |
| * NobodyCam watches tox rebuild his test venv :-p | 18:03 | |
| devananda | lucasagomes: awesome to see pyghmi & ironic coming to fedora! | 18:04 |
| lucasagomes | devananda, yeah, ironic still under review | 18:05 |
| devananda | lucasagomes: review? | 18:05 |
| lucasagomes | I already pushed pyghmi to the testing repository | 18:05 |
| lucasagomes | devananda, yeah package review | 18:05 |
| lucasagomes | to check the spec file, see if it builds etc | 18:05 |
| devananda | ahh | 18:06 |
| *** killer_prince has quit IRC | 18:06 | |
| lucasagomes | devananda, e.g https://bugzilla.redhat.com/show_bug.cgi?id=1067445 | 18:06 |
| devananda | jbjohnso: ^^ | 18:08 |
| *** lazy_prince has joined #openstack-ironic | 18:11 | |
| *** lazy_prince is now known as killer_prince | 18:12 | |
| devananda | NobodyCam: you see https://etherpad.openstack.org/p/iLODriverIronicDevstack ? | 18:12 |
| * NobodyCam clicks | 18:12 | |
| NobodyCam | devananda: I need to wrap my head around the web server image stuff | 18:16 |
| devananda | NobodyCam: yea, that's a no-go IMO | 18:16 |
| devananda | they should be using glance | 18:16 |
| *** jcooley_ has joined #openstack-ironic | 18:20 | |
| *** derekh has quit IRC | 18:21 | |
| *** killer_prince has quit IRC | 18:27 | |
| *** killer_prince has joined #openstack-ironic | 18:28 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Allow client to set the maintenance mode https://review.openstack.org/76614 | 18:29 |
| lucasagomes | total wip ^ | 18:29 |
| lucasagomes | I'll have to run now | 18:29 |
| lucasagomes | have a good night devananda NobodyCam others :) | 18:29 |
| NobodyCam | devananda: do you know if iLo supports other then http(s) connections? | 18:29 |
| NobodyCam | night lucasagomes :) | 18:29 |
| *** rameshg87 has joined #openstack-ironic | 18:30 | |
| NobodyCam | ha /me click the ask It button on his corp windowws image :-p | 18:31 |
| *** harlowja_away is now known as harlowja | 18:34 | |
| *** saju_m has quit IRC | 18:35 | |
| *** rameshg87 has quit IRC | 18:35 | |
| *** matty_dubs|lunch is now known as matty_dubs | 18:38 | |
| *** ndipanov has quit IRC | 18:39 | |
| *** lucasagomes has quit IRC | 18:40 | |
| *** athomas has quit IRC | 18:40 | |
| wanyen_ | [nobodycam] ilo virtual meida only support http/https | 18:43 |
| NobodyCam | wanyen_: ahh | 18:43 |
| NobodyCam | it cann't pull tftp ? | 18:44 |
| devananda | wanyen_: hi! glance uses HTTP(S). Why do you need to put the ISO on a separate web server? | 18:44 |
| wanyen_ | [deva] this web server can be a backend store for glance but ilo needs to use http directly not glance api | 18:45 |
| *** ndipanov has joined #openstack-ironic | 18:45 | |
| *** aignatov_ is now known as aignatov | 18:46 | |
| devananda | wanyen_: so the normal work flow is that a user will select an image from glance and issue a request to nova, eg, "nova boot --image some-image-name --flavor some-baremetal-flavor my-instance-name" | 18:47 |
| wanyen_ | u[deva] users will post virtual media image to the web server and then register the image using glance image-create --location ,http url. to register the image | 18:47 |
| * NobodyCam wounders if it would be to hackish to add apahe like interface to vender_passthrough so iLo could request media thru it | 18:48 | |
| wanyen_ | the ilo driver will query glance for image info to get the http url and then pass that url to ilo RIBCL | 18:49 |
| devananda | wanyen_: http://docs.openstack.org/developer/glance/glanceapi.html#retrieve-raw-image-data | 18:49 |
| *** ndipanov has quit IRC | 18:49 | |
| devananda | wanyen_: I dont understand why there is a need to use a _separate_ web server for the ISO when glance already supports fetching an image directly via http(s) | 18:50 |
| wanyen_ | [deva] if user configure web server as a backend store for glance then they don't need to use a seperate web server. However, if user configure swift for backend store then they would need a a seperate web server | 18:51 |
| devananda | wanyen_: ilo does not need to implement all of the glance API, just needs to be able to authenticate with glance -- or the image needs to be public | 18:52 |
| wanyen_ | Proliant ilo does not understand/support glance authentication | 18:53 |
| *** rameshg87 has joined #openstack-ironic | 18:53 | |
| wanyen_ | ilo virtual media driver will use glance to retrieve image info based on the image id and get the http url from glance. So iLo virtualmedia driver is using glance but Proliant ilo needs to use http directly | 18:54 |
| wanyen_ | ,deva. The web server requiremnt is only for user using ilo virtual media driver. It is not required if users use ironic pxe driver. | 18:58 |
| * NobodyCam steps away from keyboard (prob ~30 minutes). | 19:03 | |
| devananda | wanyen_: does proliant ilo support http authentication with the webserver? | 19:05 |
| rameshg87 | devananda: yes, it supports basic authentication where it can provide a username and password to the webserver | 19:05 |
| wanyen_ | <deva> yes. I believe so. Ramesh can you fill in teh detail? | 19:05 |
| openstackgerrit | Victor Sergeyev proposed a change to openstack/ironic: Sync common db code from Oslo https://review.openstack.org/76204 | 19:06 |
| wanyen_ | deva, since user register their image to glance, the normal nova boot flow will work | 19:07 |
| rameshg87 | yes wanyen_: , proliant ilo supports basic authentication with the web server | 19:08 |
| devananda | wanyen_: the external dependency on a web server to provide images is what troubles me, since openstack already has a mechanism for this (glance) and you are duplicating that | 19:08 |
| wanyen_ | The only difference is that prolinst ilo does not download image from glance, it uses http/https to retrieve images | 19:08 |
| devananda | wanyen_: glance serves images over http/s already | 19:09 |
| wanyen_ | yes. if user configure web server as glance backend store then no addition web server is needed. | 19:10 |
| devananda | eg, for my local devstack, i can fetch an image like this | 19:10 |
| devananda | $ curl http://192.168.122.81:9292/v1/images/3def324c-8feb-4a83-b7b9-f4b2c26a3201 2>/dev/null | wc -c | 19:10 |
| devananda | 3714968 | 19:10 |
| devananda | there's no separate web server here -- glance-api streams the images over HTTP | 19:10 |
| rameshg87 | devananda: doesn't glance require authentication for retrieving the image ? | 19:11 |
| devananda | all i have done is disable keystone auth in glance as a quick proof of concept | 19:11 |
| devananda | rameshg87: yes, but that can be either disabled or modified | 19:11 |
| devananda | rameshg87: is it impossible for iLO to send X-Auth-Token header, if the token is supplied for you? | 19:12 |
| rameshg87 | devananda: yes, iLO can't send X-Auth-Token | 19:12 |
| devananda | ok | 19:12 |
| jroll | if you use the swift backend for glance, you can generate a temporary http url that does not require auth | 19:13 |
| wanyen_ | deva, ProLiant ilo does not understand glance and therefore cannot do the glance authentication. | 19:13 |
| devananda | jroll: yep, that was what I was about to test & suggest | 19:13 |
| rameshg87 | jroll: i read about this. is this by swift-proxy-server ? | 19:13 |
| jroll | :) | 19:13 |
| jroll | rameshg87: let me find a link | 19:13 |
| devananda | jroll: thanks :) | 19:14 |
| openstackgerrit | Victor Sergeyev proposed a change to openstack/ironic: Sync common db code from Oslo https://review.openstack.org/76204 | 19:14 |
| devananda | I think it's keystone v3 signed url stuff? | 19:14 |
| devananda | I was just talking about this with lifeless yesterday | 19:14 |
| jroll | rameshg87: https://swiftstack.com/docs/admin/middleware/tempurl.html | 19:14 |
| rameshg87 | jroll: thanks | 19:16 |
| jroll | rameshg87: http://torgomatic.us/blog/2013/05/08/an-introduction-to-tempurl-in-openstack-swift/ might also be helpful | 19:16 |
| jroll | no problem :) | 19:16 |
| lifeless | devananda: yes, heat depends on keystone V3 now, we had to upgrade our keystone on the cd-undercloud to keep it working, because it has this feature you need :) | 19:17 |
| rameshg87 | so devananda: , in such a setup the glance can use backing store for images as swift, right ? and ilodriver can get a temporary url for the object stored in swift, am i correct ? | 19:21 |
| devananda | rameshg87: I believe so, yes | 19:21 |
| jroll | rameshg87: that's correct | 19:21 |
| devananda | rameshg87: and since you are also now making a POST back to Ironic API to signal completion, instead of writing to a status file on the floppy | 19:21 |
| devananda | rameshg87: you could upload the floppy image to glance and use a tempurl for that as well (and of course, delete it when done just the same) | 19:22 |
| devananda | rameshg87: then there should be no dependency on a separate external web server | 19:22 |
| rameshg87 | devananda: yes, i think we can do so if the tempurl works... | 19:22 |
| devananda | great :) | 19:22 |
| wanyen_ | deva; ther are other glance backend store such as file, cinder, ...etc | 19:23 |
| wanyen_ | does tempurl also works for those backend stores? | 19:23 |
| devananda | jroll: ^ ? | 19:24 |
| jroll | I have no idea | 19:26 |
| jroll | I would guess no - I think it's a feature of swift | 19:26 |
| *** vkozhukalov_ has joined #openstack-ironic | 19:26 | |
| jroll | wanyen_: ^ | 19:27 |
| wanyen_ | so if users uses web server and swift as backend stores then the virtual meida image can be host there. However,if user uses file or cinder as backend store then virtual media images will still need to be hosted in a web server | 19:28 |
| jroll | right, but I think you can just document that as a dependency of the driver and move on :) | 19:29 |
| wanyen_ | sound fine. | 19:30 |
| devananda | ++ | 19:30 |
| rameshg87 | devananda: wanyen_: jroll: one question | 19:30 |
| rameshg87 | assume if we decide to tempurl mechanism, and use glance with swift for storing the images, the customers would still think okay to setup swift for using the features of the ilo driver, right ? | 19:32 |
| *** martyntaylor has left #openstack-ironic | 19:32 | |
| wanyen_ | [ramesh] I think we should allow users to use either swift or web servers as backend store for virtual media images | 19:33 |
| devananda | wanyen_: i would strongly prefer the iLO driver have a dependency on existing openstack components (eg, glance and swift) than to add a new external dependency which duplicates existing functionality. | 19:34 |
| wanyen_ | glance allows web servers as backend store so it's not duplicating or introducting new dependency | 19:35 |
| devananda | rameshg87, wanyen_ - if a customer is not concerned about security (eg, simple http auth is OK) because this is on a trusted private network, then they can use glance with auth disabled -- this is still better than adding a new external dependency | 19:35 |
| devananda | wanyen_: so there are three possible ways: glance by itself with no auth, glance with webserver with simple auth, glance with swift and signed urls. | 19:37 |
| devananda | this seems fine to me -- but saying that it only supports glance with a webserver is not OK, since taht is not how most deployments of glance are set up | 19:37 |
| wanyen_ | deva: agree. i would like ilo virtual media driver to work as many backend stores as possible, | 19:38 |
| *** coolsvap1 has joined #openstack-ironic | 19:38 | |
| wanyen_ | s/work/work with/ | 19:38 |
| rameshg87 | devananda: i had one more question. generally for ironic, are the baremetal node's management console's setup as part of the private provisioning network ? | 19:40 |
| rameshg87 | devananda: the reason is for ilo virtual media driver, the iLO and glance/swift need to be in the same network to retrieve the image. (as opposed to the pxe mechanism where the image is downloaded via the server's NIC) | 19:41 |
| wanyen_ | ramesh, do you want to discuss loading auth token in floppy image and upload to glance? | 19:41 |
| *** pradipta` has joined #openstack-ironic | 19:41 | |
| *** vkozhukalov_ has quit IRC | 19:41 | |
| rameshg87 | yes wanyen_:, i think we could discuss that as well. | 19:42 |
| devananda | rameshg87: generally speaking, data centers have a lot of incentive to physically isolate the BMC / management network from the data network | 19:42 |
| devananda | rameshg87: security is a very big reason for that separation. also, often the management network is slower (eg 1Gbit) rather than 10+ | 19:43 |
| devananda | rameshg87: pulling anything that contains secure data (eg, the keys contained on the floppy) should happen over the mgmt network | 19:43 |
| rameshg87 | devananda: but the virtual media solution is currently supported to pull the OS image also over the management network | 19:44 |
| devananda | rameshg87: ideally, the OS image would be pulled over the data network and then its signature verified (using the keys transferred over the mgmt network and/or keys embedded in the host) | 19:44 |
| rameshg87 | devananda: typo, i meant "currently supposed to pull" | 19:45 |
| devananda | rameshg87: yes, i understand that. just saying what I think would be best :) | 19:45 |
| *** dhellmann is now known as dhellmann_ | 19:45 | |
| devananda | rameshg87: I think it's OK, but let me enumerate my concerns | 19:45 |
| * NobodyCam is back | 19:45 | |
| jbjohnso | devananda, man I'm slow to look at my chat | 19:45 |
| rameshg87 | devananda: okay | 19:45 |
| devananda | * slower network, leads to more network congestion / longer deploys | 19:46 |
| jbjohnso | woo.. I can see things | 19:46 |
| jbjohnso | devananda, fyi, I made secure deployment payload with *everything* needed to establish trust for esxi and Linux, the footprint was about 800-ish kilobytes | 19:47 |
| *** coolsvap has quit IRC | 19:47 | |
| *** pradipta has quit IRC | 19:47 | |
| *** agordeev has quit IRC | 19:47 | |
| rameshg87 | devananda: okay. if at all we put the management and data network in the same subnet (the internal trusted n/w), do you see anything hindering from using ironic features (just asking because it was against the expectation) | 19:48 |
| jbjohnso | devananda, this included an iPXE build with https support (I patched it to have a entropy provider in EFI mode), a client certificate and the certificate chain for the deployment manager | 19:48 |
| devananda | rameshg87: you do not want the tenants to have access to the mgmt network after provisioning | 19:48 |
| devananda | rameshg87: well - i'm assuming you dont | 19:48 |
| *** vkozhukalov has joined #openstack-ironic | 19:48 | |
| rameshg87 | devananda: i meant ironic features like remote console, etc. i know it depends on implementation, but just a question | 19:49 |
| rameshg87 | devananda: okay | 19:49 |
| jbjohnso | devananda, well, there is another sticky problem. If you are a tenant getting baremetal after another tenant has had it, there is almost certainly the capacity for a rootkit to have been injected | 19:49 |
| devananda | rameshg87: so you need SDN-capable switch to prevent a tenant from accessing the mgmt network, interfering with other deploys, etc | 19:49 |
| devananda | jbjohnso: of course | 19:49 |
| rameshg87 | devananda: okay | 19:49 |
| devananda | jbjohnso: but that's a separate vector from "tenant has access to other tenant's BMC" | 19:50 |
| rameshg87 | devananda: won't keep you bugging for long :-); just one more thing that wanyen_: and myself wanted to discuss | 19:50 |
| jbjohnso | devananda, I didn't get to show you what I'm trying to open source, consider it a 'proxy' for consoles (text only, but who needs video) | 19:50 |
| devananda | rameshg87: console access via ipmitool / ipminative drivers works by SOL connection to the BMC. so it should be unaffected by whether or not other services (eg glance) are exposed on the control network | 19:51 |
| rameshg87 | devananda: okay | 19:51 |
| rameshg87 | devananda: so currently ilo driver, during deployment, takes the token, puts it into floppy image and (as per our new solution) uploads it to glance or swift for the deployment | 19:52 |
| rameshg87 | devananda: do you see any security concerns in doing so ? i mean putting the token temporarily on an external place | 19:53 |
| jbjohnso | question, is that all you are putting in via remot emedia? Some credential, or actually bootstrapping the whole shebang? | 19:53 |
| devananda | jbjohnso: see https://etherpad.openstack.org/p/iLODriverIronicDevstack | 19:53 |
| devananda | jbjohnso: your answers await there :) | 19:54 |
| jbjohnso | devananda, ok, fyi, my solution for that in our world was to use a floppy sized image uploaded to the service processor for he whole shebang | 19:54 |
| *** morgabra has quit IRC | 19:54 | |
| devananda | rameshg87: yea, that's less than ideal. but the same issue applied when you were uploading the floppy to a separate webserver | 19:55 |
| jbjohnso | devananda, I suspect remote on-thy-fly access was because the iso's being constructed were too big | 19:55 |
| devananda | rameshg87: the PXE driver is passing the token via TFTP right now, which is also less than ideal -- it's unencrypted and on the data network, not the control network | 19:55 |
| *** linggao_ has joined #openstack-ironic | 19:56 | |
| *** athomas has joined #openstack-ironic | 19:57 | |
| devananda | rameshg87: I think if you clearly document that the glance store should be secured because it will be accessible from the control plane AND hosting temporary secure tokens (both signed URL and keystone token), then it is OK. | 19:57 |
| *** morgabra has joined #openstack-ironic | 19:57 | |
| devananda | rameshg87: in early Juno, I would like Ironic to get signed / temp url support as well, so then there will be no need to pass a token | 19:57 |
| *** romcheg has joined #openstack-ironic | 19:58 | |
| *** agordeev has joined #openstack-ironic | 19:58 | |
| *** dkehn__ has joined #openstack-ironic | 19:58 | |
| *** romcheg has quit IRC | 19:59 | |
| *** agordeev has quit IRC | 19:59 | |
| *** aignatov2 has joined #openstack-ironic | 19:59 | |
| NobodyCam | humm somehow I broke the libvirt tests with our volume driver. | 19:59 |
| * NobodyCam looks to see why | 19:59 | |
| rameshg87 | devananda: okay | 19:59 |
| *** jbjohnso_ has joined #openstack-ironic | 20:01 | |
| *** aignatov has quit IRC | 20:01 | |
| *** romcheg2 has quit IRC | 20:01 | |
| *** linggao has quit IRC | 20:01 | |
| *** jbjohnso has quit IRC | 20:01 | |
| *** jdob has quit IRC | 20:01 | |
| *** jrist has quit IRC | 20:01 | |
| *** dkehn_ has quit IRC | 20:01 | |
| jbjohnso_ | boy, IRC server decided I should really shut up.. | 20:01 |
| *** aignatov2 is now known as aignatov | 20:01 | |
| *** jdob_ has joined #openstack-ironic | 20:01 | |
| NobodyCam | jbjohnso_: where you netsplit | 20:01 |
| JoshNang_ | devananda: I can help with the swift/glance temp url stuff when we get there. I just implemented it a couple weeks ago for something similar | 20:02 |
| *** jrist has joined #openstack-ironic | 20:02 | |
| jbjohnso_ | devananda, don't know how far my prattling got, but I'll shut up about it unless asked more about it.. .it's really in my mind a very tidy way of addressing this issue | 20:03 |
| *** martyntaylor1 has joined #openstack-ironic | 20:04 | |
| *** romcheg has joined #openstack-ironic | 20:05 | |
| *** agordeev has joined #openstack-ironic | 20:05 | |
| devananda | JoshNang_: awesome - it sounds like rameshg87 is looking to add support for swift tempurls to their iLO driver now | 20:05 |
| devananda | JoshNang_: and I would like ironic to be able to provide tempurls for deploy agents to POST data back to Ironic | 20:06 |
| devananda | rameshg87: if you're interested, i believe jbjohnso_ has implemented a similar system to what you are doing (passing secure tokens via the virtual media channel) and he probably has some good lessons-learned that he might share with you :) | 20:07 |
| jbjohnso_ | right, the biggest thing was getting to under 800 kilobytes for a firmware https capable downloader, client cert, and CA | 20:08 |
| rameshg87 | JoshNang_: great. thanks for the offer. it would be helpful to us | 20:08 |
| JoshNang_ | rameshg87: if you run into any issues or want a hand with temp urls, i'd love to help. I'm in a conference this week, but I can find some time to help. Otherwise, I'll be at the meetup next week as well. | 20:08 |
| NobodyCam | :) | 20:09 |
| *** dkehn__ is now known as dkehn_ | 20:09 | |
| rameshg87 | JoshNang_: thanks. did you mean you have already implemented and submitted some change which uses temp urls | 20:09 |
| rameshg87 | JoshNang_: if so, sharing the change you have made would be ideal to us. | 20:10 |
| JoshNang_ | devananda: what's the benefit of tempurls to post back to for deploy agents? i probably just missed the discussion on it. | 20:10 |
| JoshNang_ | rameshg87: it was for a different project unfortunately | 20:11 |
| devananda | JoshNang_: both PXE and iLO driver will signal to /v1/{NODE}/vendor_passthru/ when the deploy is complete | 20:11 |
| JoshNang_ | devananda: ooo gotcha. didn't realize that's what you meant. that definitely makes sense. | 20:12 |
| devananda | JoshNang_: this requires authentication, of course. it's best if we could only grant access to a specific URL by the deploy agent (and even better if we can revoke it when complete / timed out) | 20:12 |
| JoshNang_ | rameshg87: i'll gist the relevant code, one sec. | 20:13 |
| JoshNang_ | devananda: makes sense. are you thinking something like sending a token when the deploy agent starts and the deploy agent posting back with that key? | 20:15 |
| devananda | JoshNang_: that's what we do today | 20:15 |
| devananda | JoshNang_: keystone token is less than ideal, however, because it grants global access to what ever taht user can do (in this case, it must be an admin user) | 20:16 |
| wanyen_ | [deva] do we allow bug fixes after 03/06? | 20:16 |
| devananda | JoshNang_: AIUI, ironic needs keystone v3 support to be able to control per-resource access | 20:17 |
| devananda | wanyen_: yes. that time should be used for documentation, increasing test coverage, and fixing bugs | 20:17 |
| *** morgabra has quit IRC | 20:17 | |
| *** morgabra has joined #openstack-ironic | 20:18 | |
| * NobodyCam makes a bagel | 20:18 | |
| JoshNang_ | rameshg87: https://gist.github.com/pcsforeducation/9237670 so this is what we used. it worked in some limited testing | 20:21 |
| rameshg87 | JoshNang_: thanks. :-) | 20:22 |
| JoshNang_ | devananda: ok. will ironic not support keystone v2 and below then? or just use the less secure keystone token passing? | 20:23 |
| JoshNang_ | rameshg87: no problem! | 20:24 |
| rameshg87 | devananda: we were just putting together the alternatives. it just seems like one of the alternatives that we had discussed "glance with http backing store" wouldn't work because http backing store doesn't support adding images. it just supports retrieval of images. | 20:24 |
| rameshg87 | devananda: so we would be left with only 2 - glance w/o authentication, glance + swift temp urls | 20:25 |
| wanyen_ | [deva] ramesh just checked that glance web-server back-end store only supports image download but not image upload. We would like to allow users to use web server to host virtual media images as users of the virtual meida already have the web server set up to host the images for virtual media in their environment. | 20:25 |
| devananda | JoshNang_: is there a good reason to keep support of v2 after we add keystone v3? | 20:25 |
| JoshNang_ | devananda: i haven't looked into keystone enough to know | 20:26 |
| devananda | JoshNang_: nor have I yet. something to discuss at the summit :) | 20:27 |
| NobodyCam | devananda: I can see operators wanting to test with their current infrastructure which may not have v3? | 20:27 |
| JoshNang_ | devananda: haha at least I have something to read up on on the flight home | 20:27 |
| devananda | if we add keystone v3 support to improve security around deploys, and someone wants to test ironic without that, i think it's fair that they must do something explicit to do that | 20:29 |
| devananda | falling back to a less secure mode automatically doesn't seem appropriate | 20:30 |
| NobodyCam | devananda: like setting conf var to v2 Vs. v3? | 20:30 |
| NobodyCam | yes I agree no fallback | 20:31 |
| wanyen_ | [deva] the ilo virtual media driver will support swift with temp url but I believe it would be beneficial to users if we give them an option to upload iamge to a web server as they desire | 20:31 |
| *** morgabra has quit IRC | 20:32 | |
| wanyen_ | users will still register their image to glance after they upload it to webserver | 20:32 |
| jroll | devananda: +1 | 20:32 |
| *** morgabra has joined #openstack-ironic | 20:33 | |
| devananda | wanyen_: sure. so if a deployer chooses to deploy glance-backed-by-webserver, and also chooses to expose that webserver to users to allow them to upload images, I think that's their prerogative, though I also think it's unnecessary :) | 20:33 |
| devananda | wanyen_: but that can't be the only supported mechanism. I feel that the iLO driver must support working within the core components of openstack (eg, glance and swift/cinder/file) | 20:36 |
| devananda | wanyen_: integration with the core components should be the default and tested approach. deployer may vary from that, that is their choice, but it should not be the reference | 20:37 |
| *** martyntaylor1 has left #openstack-ironic | 20:38 | |
| wanyen_ | deva, yes i agree but I want to keep the webserver as backend store open. If user choose web server as backend store then ilo driver will need to upload floppy image to web server directly. For Swift backend, ilo driver willuse glance api to upload floppy image. | 20:38 |
| *** dhellmann_ is now known as dhellmann | 20:39 | |
| jroll | maybe create that as a separate driver on top of the ilo driver and ship it separately from ironic | 20:39 |
| jroll | just override only the "get image" bit | 20:40 |
| *** romcheg has left #openstack-ironic | 20:42 | |
| NobodyCam | jroll: that seems like a support nightmare | 20:42 |
| NobodyCam | :-p | 20:43 |
| jroll | NobodyCam: why? | 20:43 |
| NobodyCam | users would forget that step and call in why is this not working | 20:43 |
| jroll | maybe | 20:44 |
| wanyen_ | The whole reason to submit ilo driver upstream is to allow deployers to have access directly without the need to contact vendors. I really would like to have it all on Ironic | 20:44 |
| *** max_lobur has joined #openstack-ironic | 20:48 | |
| NobodyCam | grrr I am not getting the same failures | 20:49 |
| devananda | wanyen_: an alternative -- what if the OS image and floppy images were stored in separate places | 20:49 |
| devananda | wanyen_: eg, OS image is in glance (whether file, swift, or webserver backed) | 20:49 |
| wanyen_ | is this acceptable- If user chooses web server as backend store then ilo driver will upload floppy image to web server directly. For Swift backend, ilo driver willuse glance api to upload floppy image. | 20:50 |
| devananda | wanyen_: and the floppy image is served directly from the conductor | 20:50 |
| devananda | wanyen_: this also addresses rameshg87's concern about uploading the floppy (containing a keystone token) to glance -- which is a valid concern. | 20:51 |
| devananda | wanyen_: such security issues are significantly mitigated if the key transfer happens directly from teh conductor to the proliant iLO | 20:52 |
| wanyen_ | floppy iamge is built in ilo driver if that what you meant by served directly from conductor | 20:53 |
| devananda | wanyen_: right now, floppy image is build by the ilo driver on the iornic-conductor, but it is then uploaded somewhere else (web server, glance, etc) before being downloaded via HTTP by the proliant BMC | 20:53 |
| devananda | wanyen_: i'm suggesting that it doesn't need to be uploaded at all -- the ironic-conductor process already must have access to the control network, and could serve the floppy image directly to the BMC | 20:54 |
| devananda | or more specifically | 20:54 |
| devananda | the ilo driver could manage a local process which serves that image | 20:54 |
| devananda | the way that the PXE driver manages tftpd to serve kernel/ramdisk/token sets during a PXE deploy | 20:55 |
| wanyen_ | deva: the problem is that ilo can only access floppy image via http | 20:57 |
| rameshg87 | devananda: but then, since proliant BMC knows only virtual media over http, it should be a web-server on the ironic conductor node that serves the image. atleast a sort of minimal webserver that runs on the conductor node | 20:57 |
| devananda | rameshg87: exactly | 20:58 |
| wanyen_ | sounds good | 20:59 |
| rameshg87 | devananda: but even though the webserver is not external to the conductor node, it would still require the customer to setup and enable that http server on the conductor node; just like they would enable tftp on the conductor node for pxe i guess | 21:00 |
| rameshg87 | devananda: this could be used only for uploading the temporary floppy images | 21:01 |
| NobodyCam | devananda: rameshg87 would somehting like starting a wisgi process per deploy work? | 21:05 |
| NobodyCam | I really would like the users to not have to setup anyting external. | 21:05 |
| rameshg87 | NobodyCam: i guess that's a good idea. i think you meant something like this: https://code.google.com/p/wsgi-fileserver/ | 21:10 |
| rameshg87 | am i correct ? | 21:10 |
| NobodyCam | basicly... our api is running in a wisgi process (unless you've setup apache) | 21:11 |
| devananda | NobodyCam: this is regarding the conductor though, not the API | 21:11 |
| NobodyCam | it is basicly just a small foot print web server | 21:11 |
| devananda | NobodyCam: analogous to the tftpd process which deployers must start today on conductor nodes | 21:11 |
| *** aignatov is now known as aignatov_ | 21:12 | |
| NobodyCam | devananda: ya. but what if conductor spun up a wigi process only for the deploy then kills it after the deploy | 21:12 |
| devananda | so either the deployer would start a minimal webserver and set the file path as a CONF option, just like we do with the PXE driver | 21:13 |
| lifeless | wanyen_: rameshg87: So I'm curious, what advantage does iLO have over the PXE method then ? | 21:13 |
| lifeless | wanyen_: it seems like it will be slower (HTTP == userspace vs iSCSI== kernel or multicast == bandwidth efficient at massive scale) | 21:13 |
| lifeless | NobodyCam: PXE depends on tftpd; depending on httpd for an HTTP filesystem exposed area seems fine to me | 21:14 |
| devananda | lifeless: transfer keys over secure virtual media channel && single power cycle | 21:14 |
| lifeless | devananda: How do those features specifically work? | 21:15 |
| devananda | lifeless: PXE driver will always need to transfer keys over the data plane. iLO driver can do that over the control plane | 21:15 |
| devananda | lifeless: their notes are here - https://etherpad.openstack.org/p/iLODriverIronicDevstack | 21:15 |
| devananda | lifeless: tldr is | 21:15 |
| lifeless | Is the virtual media being referred to the floppy disk or the image content ? | 21:15 |
| devananda | both | 21:15 |
| wanyen_ | lifeless,: we have deployers who prefer virtual meida over pxe. Also virtual media cansupport live media for rescure and test drive purposes | 21:15 |
| devananda | lifeless: tldr is power off; mount ISO and boot it; mount floppy to get keys; copy image out of ISO onto local disk; POST completion notice to ironic; ironic does cleanup and powers on the machine | 21:16 |
| rameshg87 | lifeless: another interesting note i think is is virtual media installs are completely done through baremetal node. this would mean no conductor intervention for the deployment. it would need to transfer only the compressed image over the wire, and the image decompression and writing of image to disk will be done on the baremetal node itself | 21:16 |
| devananda | lifeless: at least that is my understanding from their etherpad and code review | 21:17 |
| lifeless | rameshg87: baremetalnode is the conductor. | 21:17 |
| *** dhellmann_ has joined #openstack-ironic | 21:17 | |
| lifeless | rameshg87: oh, you mean the target node being booted? | 21:17 |
| rameshg87 | lifeless: yes, the target node being booted | 21:17 |
| lifeless | rameshg87: but you're reading data off of an HTTP server. THats precisely as much I/O. | 21:18 |
| lifeless | rameshg87: on the conductor | 21:18 |
| devananda | lifeless: no. the ISO is pulled from glance, not from the conductor | 21:18 |
| lifeless | devananda: I got a different impression from the bit in this conversation where NobodyCam suggested a wsgi http app | 21:19 |
| devananda | lifeless: the only thing served from teh conductor would be the floppy (that's what i'm proposing -- right now, that is pulled from some other web server !glance) | 21:19 |
| lifeless | devananda: kk | 21:19 |
| rameshg87 | lifeless: yes, as devananda said, the ISO will be pulled directly from glance/swift backing store onto the target node | 21:19 |
| *** sbadia_ has joined #openstack-ironic | 21:20 | |
| rameshg87 | lifeless: the floppy image contains the information to be passed for the deployment. basically the proliant iLO supports two types of virtual media - virtual media cdrom and virtual media floppy | 21:20 |
| *** openstackgerrit has quit IRC | 21:20 | |
| *** dhellmann has quit IRC | 21:20 | |
| *** sbadia has quit IRC | 21:20 | |
| *** shortstop has quit IRC | 21:20 | |
| *** sirushti has joined #openstack-ironic | 21:20 | |
| *** sbadia_ is now known as sbadia | 21:20 | |
| rameshg87 | lifeless: the virtual media cdrom would server the compress OS image from the glance/swift backing store directly | 21:20 |
| *** dhellmann_ is now known as dhellmann | 21:21 | |
| lifeless | devananda: the etherpad shows two boots to userspace | 21:21 |
| rameshg87 | lifeless: the virtual media floppy would serve the configuration related stuffs for the deployment (like ironic api url, token for signalling completion of the installation) | 21:21 |
| *** jdob has joined #openstack-ironic | 21:21 | |
| *** jdob_ has quit IRC | 21:21 | |
| lifeless | devananda: one boot to deploy, but it boots a deploy kernel + ramdisk | 21:21 |
| *** openstackgerrit has joined #openstack-ironic | 21:21 | |
| devananda | lifeless: ooh. you're right | 21:21 |
| lifeless | devananda: so secure key path seems to be the only feature tied to iLO AFAICT | 21:22 |
| devananda | wanyen_: i thought you were implementing this with only one boot cycle? | 21:22 |
| lifeless | devananda: and thats cool - but we can do that and still use common code for everything else | 21:22 |
| lifeless | so HP machines will get the multicast stuff | 21:22 |
| lifeless | devananda: I am interested in this because I don't want to have to choose between increased security and speed :) | 21:22 |
| devananda | lifeless: ++ :) | 21:22 |
| *** coolsvap_ has joined #openstack-ironic | 21:23 | |
| lifeless | devananda: also | 21:24 |
| lifeless | devananda: as I read with fine toothed comb | 21:24 |
| lifeless | devananda: the floppy is transferred over HTTP | 21:24 |
| lifeless | devananda: not HTTPS | 21:24 |
| lifeless | devananda: so its no more secure, just more convoluted. | 21:24 |
| devananda | ... | 21:24 |
| lifeless | 5. Retrieve the OS image details from glance - http location using direct_url field | 21:25 |
| lifeless | sorry | 21:25 |
| lifeless | 3. Upload the floppy image to the http server using sftp | 21:25 |
| lifeless | 4. Attach the floppy image as floppy virtual media | 21:25 |
| lifeless | those two lines | 21:25 |
| *** mdurnosvistov_ has joined #openstack-ironic | 21:25 | |
| devananda | lifeless: oh. that's what we were discussing | 21:25 |
| devananda | lifeless: i proposed NOT uploading it to a separate http server | 21:26 |
| devananda | lifeless: but instead serving it from a locally-managed service, akin to the current dependency on tftpd | 21:26 |
| devananda | rameshg87: does proliant support mounting virtual floppy over HTTPS? | 21:26 |
| wanyen_ | yes. it does. | 21:27 |
| lifeless | ok, so if its HTTPS then and only then does it become better | 21:27 |
| *** mdurnosvistov has quit IRC | 21:28 | |
| wanyen_ | lifeless: it'sa bout deploy's choice as well. we do have a lot of requests to support non-pxe boot. | 21:29 |
| lifeless | wanyen_: sure, but do they mean 'local' or !pxe | 21:30 |
| lifeless | and if !pxe, what about pxe concerns them ? | 21:30 |
| rameshg87 | devananda: lifeless: yes https is supported | 21:30 |
| wanyen_ | As far as i know, some have security concerns and some said their environment is not set up for pxe | 21:31 |
| rameshg87 | https is supported for mounting virtual floppy | 21:31 |
| wanyen_ | reamesh: we should look into using https | 21:31 |
| lifeless | wanyen_: ok, so I'm not aganist virtual floppy | 21:31 |
| lifeless | wanyen_: it just seems like that is the only real difference | 21:31 |
| wanyen_ | lfeless: live cd can only used for rescure and test drive | 21:32 |
| wanyen_ | s/only/also | 21:32 |
| *** coolsvap has joined #openstack-ironic | 21:32 | |
| lifeless | wanyen_: sure, but we'll have ramdisk versions of those anyway, right? | 21:33 |
| wanyen_ | yes. that's in the plan for ironic right? | 21:34 |
| devananda | yep | 21:34 |
| *** dhellmann has quit IRC | 21:34 | |
| *** _sirushti has joined #openstack-ironic | 21:34 | |
| *** dhellmann has joined #openstack-ironic | 21:35 | |
| *** coolsvap1 has quit IRC | 21:35 | |
| *** jcooley_ has quit IRC | 21:35 | |
| *** sirushti has quit IRC | 21:35 | |
| *** _sirushti is now known as sirushti | 21:35 | |
| *** coolsvap_ has quit IRC | 21:35 | |
| *** jcooley_ has joined #openstack-ironic | 21:36 | |
| devananda | lifeless: one issue with PXE is requiring DHCP to be owned/managed. iLO driver, afaict, does not require that | 21:37 |
| *** derekh has joined #openstack-ironic | 21:38 | |
| *** pbrooko has quit IRC | 21:40 | |
| rameshg87 | devananda: lifeless: i just had a question on (HTTP == userspace vs iSCSI== kernel) | 21:40 |
| rameshg87 | is deployment using iSCSI happens completely in kernel space ? | 21:41 |
| lifeless | rameshg87: all the data mangling yeah - we use dd configured for non-cached writes | 21:42 |
| lifeless | rameshg87: AFAICT we saturated 10Gbps links doing the deploy | 21:42 |
| rameshg87 | lifeless: :-) | 21:43 |
| lifeless | rameshg87: and if the ISO with the image being copied from is on the BMC, that means we'll do data transfer about 10 times faster... | 21:43 |
| rameshg87 | lifeless: yeah i agree, i guess it all depends on how fast iLO can retrieve the data from the http server on the other side and provide it | 21:44 |
| devananda | most control plane networks that I've seen are 1Gb or slower. people tend to prefer to spend money making the data network fast :) | 21:46 |
| wanyen_ | pxe driver needs to download image from glance as well. iLO driver saves the dd part of data xfer. | 21:46 |
| devananda | congestion on the control plane seems like a significant limitation when trying to deploy a large quantity of nodes in parallel | 21:48 |
| devananda | *like it will be | 21:48 |
| wanyen_ | okay. To summarize ilo upload auth token to a web server on conductor node, This web server is only used for token passing. ilo driver should support glance/swift and glance/web server backend stores. right? | 21:51 |
| wanyen_ | Okay. Thanks! I am going off to lunch. | 21:53 |
| rameshg87 | devananda: lifeless: wanyen_: i will need to leave now. thanks for all your thoughts | 21:54 |
| *** rameshg87 has left #openstack-ironic | 21:54 | |
| NobodyCam | have a good night rameshg87 | 21:54 |
| *** jdob has quit IRC | 21:56 | |
| lifeless | wanyen_: pxe driver downloads on a network 10+ times faster than the mgmt network :) | 21:59 |
| lifeless | wanyen_: I'm not convinced that doing any bulk transfer on the iLO network makes sense | 22:00 |
| lifeless | wanyen_: I think it makes lots of sense to secure things | 22:04 |
| devananda | wanyen_: another aspect that I do not see covered -- how will even a rudimentary partition description (eg, size of root fs and swap space) be passed from nova down to the node? | 22:10 |
| *** mdurnosvistov_lt has joined #openstack-ironic | 22:16 | |
| lifeless | devananda: the way I'd do it is use the pxe data handover protocol, just embedded into the floppy boot sequence rather than pxelinux.cfg | 22:24 |
| devananda | lifeless: indeed, that looks quite appealing to me | 22:29 |
| *** matty_dubs is now known as matty_dubs|gone | 22:29 | |
| devananda | lifeless: though from a code perspective, i need to look closer to see what will need to be done | 22:29 |
| devananda | lifeless: the key passing is integral to the deploy driver at the moment. not mix-n-match | 22:30 |
| lifeless | devananda: helper function, done. :) | 22:30 |
| devananda | lifeless: something like taht :) | 22:31 |
| *** jbjohnso_ has quit IRC | 22:32 | |
| *** mrda_away is now known as mrda | 22:41 | |
| openstackgerrit | A change was merged to openstack/ironic: Fix race condition when deleting a node https://review.openstack.org/76178 | 22:49 |
| mrda | morning all | 22:51 |
| *** romcheg has joined #openstack-ironic | 22:55 | |
| *** lnxnut has quit IRC | 22:55 | |
| mrda | lucasagomes: I was going to write patch https://review.openstack.org/#/c/76614/1 today - you beat me to it overnight :) | 22:56 |
| NobodyCam | morning mrda | 22:58 |
| mrda | morning NobodyCam | 22:59 |
| devananda | mrda: g'morning! | 23:01 |
| devananda | mrda: I'm just fixing up 76293 right now | 23:02 |
| devananda | mrda: thanks for the unit tests :) | 23:02 |
| mrda | devananda: cool, hope that's ok | 23:02 |
| mrda | devananda: so you'll address Yuriy's comments in your new patch? | 23:03 |
| *** harlowja is now known as harlowja_away | 23:05 | |
| NobodyCam | quick walkies. brb | 23:08 |
| devananda | yep | 23:11 |
| devananda | just about finished | 23:11 |
| *** linggao_ has quit IRC | 23:12 | |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Add option to sync node power state from DB https://review.openstack.org/76293 | 23:14 |
| *** lazy_prince has joined #openstack-ironic | 23:20 | |
| *** killer_prince has quit IRC | 23:22 | |
| *** lazy_prince is now known as killer_prince | 23:22 | |
| openstackgerrit | Jenkins proposed a change to openstack/ironic: Updated from global requirements https://review.openstack.org/75565 | 23:27 |
| *** vkozhukalov has quit IRC | 23:28 | |
| openstackgerrit | Jenkins proposed a change to openstack/python-ironicclient: Updated from global requirements https://review.openstack.org/75585 | 23:34 |
| openstackgerrit | Michael Davies proposed a change to openstack/ironic: Prevent GET /v1/nodes returning maintenance field https://review.openstack.org/76094 | 23:39 |
| *** mdurnosvistov_lt has quit IRC | 23:39 | |
| mrda | devananda: any other things you'd like addressed during your night? :) | 23:40 |
| devananda | mrda: hmmm! lemme see :) | 23:40 |
| devananda | mrda: so, have you seen https://etherpad.openstack.org/p/IronicReviewDay ? | 23:41 |
| devananda | mrda: this is a running list that -core folks are using to keep track of the really-easy and the really-important patches | 23:41 |
| * devananda thinks | 23:43 | |
| devananda | mrda: sun jing's patch for serial console support is, IMO, really close to landing | 23:43 |
| mrda | devananda: I'll take a look | 23:44 |
| devananda | mrda: there's a little refactoring it needs. It looks like sun wasn't clear on how to do that, despite our comments | 23:44 |
| devananda | https://review.openstack.org/#/c/64100/ | 23:44 |
| NobodyCam | devananda: on 64100 I'd be ok with filing a bug for the http vs https thing | 23:46 |
| devananda | NobodyCam: ++ | 23:47 |
| devananda | NobodyCam: that's pretty much how I feel about everything in that patch aside from the unreleased-lock issue | 23:47 |
| devananda | not perfect but good enough to iterate on further | 23:47 |
| NobodyCam | even the mocking for the item being tested? | 23:48 |
| NobodyCam | devananda: ++ as long as the issues are tracked either via or etherpad | 23:48 |
| NobodyCam | *...via BUG or... | 23:49 |
| * NobodyCam knows he would forget at least ONE | 23:50 | |
| devananda | NobodyCam: via LP bugs. better visibility | 23:50 |
| openstackgerrit | A change was merged to openstack/python-ironicclient: Fix params order in assertEqual https://review.openstack.org/74745 | 23:50 |
| devananda | and at this point, yes. we need to merge serial console and ephemeral disk support THIS week | 23:51 |
| NobodyCam | so tomorrow really | 23:52 |
| devananda | yep | 23:56 |
| NobodyCam | devananda: line 323 for locking issue? #323 proc = subprocess.Popen(' '.join(x), shell=True) | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!