| *** baoli has joined #openstack-ironic | 00:05 | |
| *** Fdaisuke has quit IRC | 00:09 | |
| *** sdake has joined #openstack-ironic | 00:09 | |
| *** awiddersheim has quit IRC | 00:09 | |
| *** crushil has joined #openstack-ironic | 00:13 | |
| *** sdake has quit IRC | 00:14 | |
| *** crushil has quit IRC | 00:33 | |
| *** sdake has joined #openstack-ironic | 00:39 | |
| *** awiddersheim has joined #openstack-ironic | 01:10 | |
| *** amarao has quit IRC | 01:13 | |
| *** awiddersheim has quit IRC | 01:14 | |
| *** crushil has joined #openstack-ironic | 01:16 | |
| *** jkilpatr has quit IRC | 01:19 | |
| *** sdake has quit IRC | 01:28 | |
| *** foexle has quit IRC | 01:35 | |
| *** amotoki has joined #openstack-ironic | 01:53 | |
| *** crushil has quit IRC | 02:06 | |
| *** amotoki has quit IRC | 02:18 | |
| *** sdake has joined #openstack-ironic | 02:20 | |
| *** sdake has quit IRC | 02:20 | |
| *** amotoki has joined #openstack-ironic | 02:26 | |
| openstackgerrit | Naohiro Tamura proposed openstack/ironic: Update the existing APIs due to adding get_supported_power_states https://review.openstack.org/382194 | 02:27 |
|---|---|---|
| *** sdake has joined #openstack-ironic | 02:32 | |
| *** amotoki has quit IRC | 02:45 | |
| *** ElCoyote_ has quit IRC | 02:51 | |
| *** baoli has quit IRC | 02:58 | |
| *** Sukhdev has joined #openstack-ironic | 03:01 | |
| *** rcernin has joined #openstack-ironic | 03:11 | |
| *** awiddersheim has joined #openstack-ironic | 03:11 | |
| *** sdake has quit IRC | 03:12 | |
| *** sdake has joined #openstack-ironic | 03:13 | |
| *** rcernin has quit IRC | 03:13 | |
| *** rcernin has joined #openstack-ironic | 03:14 | |
| *** awiddersheim has quit IRC | 03:16 | |
| *** sdake has quit IRC | 03:22 | |
| *** rcernin has quit IRC | 03:25 | |
| *** amotoki has joined #openstack-ironic | 03:45 | |
| *** crushil has joined #openstack-ironic | 03:48 | |
| *** sdake has joined #openstack-ironic | 03:51 | |
| *** links has joined #openstack-ironic | 03:56 | |
| *** links has quit IRC | 03:57 | |
| *** baoli has joined #openstack-ironic | 03:57 | |
| *** Nisha_Agarwal has joined #openstack-ironic | 04:03 | |
| *** baoli has quit IRC | 04:04 | |
| *** crushil has quit IRC | 04:05 | |
| *** vikrant has joined #openstack-ironic | 04:10 | |
| *** baoli has joined #openstack-ironic | 04:10 | |
| *** baoli has quit IRC | 04:12 | |
| openstackgerrit | Nisha Agarwal proposed openstack/proliantutils: Discover iscsi_boot and iscsi_iqn https://review.openstack.org/341653 | 04:17 |
| *** sdake has quit IRC | 04:17 | |
| *** baoli has joined #openstack-ironic | 04:18 | |
| openstackgerrit | Naohiro Tamura proposed openstack/ironic: Follow-up iRMC power driver for soft reboot and soft power off https://review.openstack.org/425593 | 04:20 |
| *** sdake has joined #openstack-ironic | 04:25 | |
| *** nmathew has joined #openstack-ironic | 04:28 | |
| *** vinsh has joined #openstack-ironic | 04:31 | |
| *** baoli has quit IRC | 04:32 | |
| *** rbudden has quit IRC | 04:34 | |
| *** sdake_ has joined #openstack-ironic | 04:35 | |
| *** sdake has quit IRC | 04:38 | |
| *** yuanying has joined #openstack-ironic | 04:48 | |
| *** vinsh has quit IRC | 04:51 | |
| *** amotoki has quit IRC | 04:56 | |
| openstackgerrit | Naohiro Tamura proposed openstack/ironic: Add Inject NMI to api-ref https://review.openstack.org/423941 | 05:01 |
| openstackgerrit | Nisha Agarwal proposed openstack/proliantutils: Discover iscsi_boot and iscsi_iqn https://review.openstack.org/341653 | 05:12 |
| *** awiddersheim has joined #openstack-ironic | 05:13 | |
| *** sdake_ has quit IRC | 05:15 | |
| *** awiddersheim has quit IRC | 05:18 | |
| *** mtanino has quit IRC | 05:19 | |
| *** amotoki has joined #openstack-ironic | 05:25 | |
| openstackgerrit | Nisha Agarwal proposed openstack/proliantutils: Discover iscsi_boot and iscsi_iqn https://review.openstack.org/341653 | 05:31 |
| *** udesale has joined #openstack-ironic | 05:51 | |
| *** mtanino has joined #openstack-ironic | 06:13 | |
| *** awiddersheim has joined #openstack-ironic | 06:16 | |
| *** mtanino has quit IRC | 06:19 | |
| *** awiddersheim has quit IRC | 06:20 | |
| openstackgerrit | Naohiro Tamura proposed openstack/ironic: Add soft reboot, soft power off and power timeout to api-ref https://review.openstack.org/423986 | 06:24 |
| openstackgerrit | Madhuri Kumari proposed openstack/ironic: Update the outdated link in user-guide https://review.openstack.org/426647 | 06:24 |
| openstackgerrit | Aparna proposed openstack/proliantutils: Add HPSUM firmware update support https://review.openstack.org/422004 | 06:41 |
| *** Nisha_Agarwal has quit IRC | 06:56 | |
| *** rcernin has joined #openstack-ironic | 07:11 | |
| *** rcernin has quit IRC | 07:12 | |
| *** rcernin has joined #openstack-ironic | 07:12 | |
| vsaienk0 | Morning Ironic'ers! | 07:24 |
| *** karthiks has joined #openstack-ironic | 07:29 | |
| *** sayantani01 has quit IRC | 07:32 | |
| *** sayantani01 has joined #openstack-ironic | 07:33 | |
| karthiks | aarefiev, Can you please review the spec https://review.openstack.org/#/c/396147/ | 07:34 |
| patchbot | patch 396147 - ironic-inspector-specs - Spec for retrieving NUMA node information | 07:34 |
| *** vsaienko has joined #openstack-ironic | 07:35 | |
| *** lennyb has joined #openstack-ironic | 07:37 | |
| *** lennyb_ has joined #openstack-ironic | 07:37 | |
| *** lennyb_ has quit IRC | 07:37 | |
| *** sayantani01 has quit IRC | 07:41 | |
| *** vsaienko has quit IRC | 07:49 | |
| *** rbartal has joined #openstack-ironic | 07:49 | |
| *** tesseract has joined #openstack-ironic | 07:50 | |
| *** ChubYann has quit IRC | 08:00 | |
| *** Sukhdev has quit IRC | 08:02 | |
| *** vsaienko has joined #openstack-ironic | 08:03 | |
| *** Nisha_Agarwal has joined #openstack-ironic | 08:06 | |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Add Ironic standalone tests https://review.openstack.org/423556 | 08:07 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: [Devstack] Set DEFAULT_IMAGE_NAME variable https://review.openstack.org/425599 | 08:07 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: [Devstack] Download both disk and uec images https://review.openstack.org/420692 | 08:07 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: [Devstack] Add stack user to libvirt group https://review.openstack.org/426664 | 08:07 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: DNM/DNR: Temp patch to verify ironic standalone tests https://review.openstack.org/426665 | 08:07 |
| *** ccamacho has joined #openstack-ironic | 08:15 | |
| *** pcaruana has joined #openstack-ironic | 08:15 | |
| *** awiddersheim has joined #openstack-ironic | 08:18 | |
| *** jtomasek has joined #openstack-ironic | 08:20 | |
| *** awiddersheim has quit IRC | 08:22 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ironic-ui: Imported Translations from Zanata https://review.openstack.org/426667 | 08:26 |
| *** amarao has joined #openstack-ironic | 08:34 | |
| *** Nisha_Agarwal has quit IRC | 08:45 | |
| *** aparnav has joined #openstack-ironic | 08:46 | |
| *** mgoddard has joined #openstack-ironic | 08:48 | |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Add Ironic standalone tests https://review.openstack.org/423556 | 08:53 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: DNM/DNR: Temp patch to verify ironic standalone tests https://review.openstack.org/426665 | 08:53 |
| *** vsaienko has quit IRC | 08:57 | |
| *** amotoki has quit IRC | 08:59 | |
| *** ralonsoh has joined #openstack-ironic | 09:00 | |
| *** zzzeek has quit IRC | 09:00 | |
| *** zzzeek has joined #openstack-ironic | 09:00 | |
| *** amotoki has joined #openstack-ironic | 09:00 | |
| *** e0ne has joined #openstack-ironic | 09:01 | |
| aarefiev | karthiks: hey, sorry, I forgot to post comment, I'm still not sure about data format | 09:15 |
| *** sayantani01 has joined #openstack-ironic | 09:15 | |
| aarefiev | morning all! | 09:15 |
| *** sayantani01 has quit IRC | 09:20 | |
| *** Nisha_Agarwal has joined #openstack-ironic | 09:20 | |
| vdrok | morning ironic, vsaienk0 and aarefiev ! | 09:24 |
| karthiks | Thanks aarefiev :) | 09:26 |
| *** jschwarz is now known as jschwarz|lunch | 09:27 | |
| pas-ha | morning ironicers :) | 09:28 |
| *** lucas-afk is now known as lucasagomes | 09:28 | |
| lucasagomes | morning all, pas-ha aarefiev vdrok :D | 09:29 |
| vdrok | morning pas-ha and lucasagomes :) | 09:29 |
| aarefiev | morning vdrok,pas-ha, lucasagomes | 09:29 |
| *** rcernin has quit IRC | 09:31 | |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Add Ironic standalone tests https://review.openstack.org/423556 | 09:33 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: DNM/DNR: Temp patch to verify ironic standalone tests https://review.openstack.org/426665 | 09:33 |
| *** rcernin has joined #openstack-ironic | 09:33 | |
| *** derekh has joined #openstack-ironic | 09:40 | |
| *** mgoddard has quit IRC | 09:48 | |
| *** makowals has quit IRC | 09:48 | |
| *** makowals has joined #openstack-ironic | 09:49 | |
| *** jistr has quit IRC | 09:49 | |
| *** jistr has joined #openstack-ironic | 09:51 | |
| *** makowals has quit IRC | 09:54 | |
| *** makowals has joined #openstack-ironic | 09:55 | |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic: SNMP agent support for OOB inspection for iLO Drivers https://review.openstack.org/372428 | 09:57 |
| *** MattMan has quit IRC | 09:57 | |
| *** athomas has joined #openstack-ironic | 09:57 | |
| *** MattMan has joined #openstack-ironic | 09:57 | |
| vsaienk0 | morning vdrok, pas-ha, aarefiev, dtantsur, lucasagomes, milan, sambetts :) | 10:00 |
| lucasagomes | o/ | 10:00 |
| *** amotoki has quit IRC | 10:05 | |
| *** sayantani01 has joined #openstack-ironic | 10:09 | |
| *** sayantani01 has quit IRC | 10:14 | |
| *** amotoki has joined #openstack-ironic | 10:15 | |
| openstackgerrit | Andrey Shestakov proposed openstack/bifrost: Allow enroll-dynamic on remote host https://review.openstack.org/411299 | 10:17 |
| openstackgerrit | Andrey Shestakov proposed openstack/bifrost: Allow deploy-dynamic on remote host https://review.openstack.org/411304 | 10:17 |
| openstackgerrit | Andrey Shestakov proposed openstack/bifrost: Add support of collecting logs from baremetal nodes https://review.openstack.org/415503 | 10:17 |
| *** awiddersheim has joined #openstack-ironic | 10:19 | |
| *** milan has joined #openstack-ironic | 10:20 | |
| *** mgoddard has joined #openstack-ironic | 10:21 | |
| *** awiddersheim has quit IRC | 10:24 | |
| *** dtantsur|afk is now known as dtantsur | 10:26 | |
| dtantsur | morning Ironic | 10:26 |
| dtantsur | morning vsaienk0, lucasagomes, aarefiev, vdrok, pas-ha | 10:26 |
| vdrok | Morning dtantsur o/ | 10:27 |
| lucasagomes | dtantsur, o/ | 10:28 |
| *** sambetts|afk is now known as sambetts | 10:28 | |
| sambetts | Morning all o/ | 10:28 |
| dtantsur | hey sambetts | 10:28 |
| sambetts | o/ dtantsur | 10:29 |
| lucasagomes | hey sambetts | 10:29 |
| sambetts | dtantsur: just looking at: https://review.openstack.org/#/c/417871/ | 10:29 |
| patchbot | patch 417871 - ironic-python-agent - Remove compatibility with old bash-based introspec... | 10:29 |
| vdrok | Morning sambetts ! | 10:29 |
| sambetts | hey lucasagomes vdrok | 10:29 |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/bifrost: Remove local actions for ssh keys https://review.openstack.org/426188 | 10:30 |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/bifrost: Refresh Ansible facts after installing packages https://review.openstack.org/426702 | 10:30 |
| aarefiev | morning vsaienk0, dtantsur, sambetts | 10:30 |
| *** Nisha_Agarwal has quit IRC | 10:32 | |
| *** amotoki has quit IRC | 10:34 | |
| sambetts | dtantsur: I can't decide if I'm being over cautious or not regarding the old inventory format and Ironic inspector. I guess ultimatly in the future we should have a ramdisk version check in the processing hooks, then we can have a check for it when we get the JSON, and processing hooks that are too old for the loaded ramdisk could either error or ignore the request. thinking about it this way | 10:38 |
| sambetts | makes be wonder if doing this means we should be cutting a major ramdisk version because we're breaking compatiblity, (are we cutting a major IPA version for Ocata?) | 10:38 |
| dtantsur | sambetts, we can bump major version whenever we think it's right. I'd also prefer we define our backward compatibility promise on this JSON before we try to stick to it. | 10:40 |
| dtantsur | I think we even say somewhere that this blob post-processing is not subject to backward compatibility | 10:40 |
| sambetts | dtantsur: I don't nessesaryly mind the backwards compatiblity thing as such, I just want to make sure people know to update their plugins before updating their version of IPA | 10:41 |
| dtantsur | well, we have release notes.. | 10:42 |
| sambetts | I don't think people are very cautious about looking for breaking changes in release notes without a major version | 10:43 |
| sambetts | thats what a major version indicates | 10:43 |
| *** yuanying_ has joined #openstack-ironic | 10:48 | |
| dtantsur | they should be, but yeah, we can bump a major version | 10:50 |
| dtantsur | jroll, ^^^ | 10:50 |
| *** yuanying has quit IRC | 10:51 | |
| milan | morning Ironic! :) | 10:52 |
| milan | morning aarefiev, dtantsur, lucasagomes, sambetts, vdrok and all! :) | 10:52 |
| dtantsur | morning milan | 10:52 |
| karthiks | morning Milan | 10:52 |
| vdrok | morning milan | 10:53 |
| milan | morning karthiks! :) | 10:53 |
| mat128 | morning all, milan | 10:53 |
| milan | morning mat128! :) | 10:53 |
| sambetts | dtantsur: cool, we'll see what jroll thinks, but other than that I think I'm +2A on that patch, having thought about it more, dealing with backward compat in inspector stuff will be unnesssary, as long as people know what they're doing before they update their agent ramdisk | 10:53 |
| sambetts | and I think the version will help that | 10:53 |
| jroll | morning | 10:56 |
| jroll | dtantsur: can you s/other/upgrades/ on that release note? | 10:56 |
| jroll | well maybe not | 10:56 |
| jroll | there's no discrete action to take | 10:56 |
| *** sayantani01 has joined #openstack-ironic | 11:03 | |
| *** sayantani01 has quit IRC | 11:08 | |
| dtantsur | yeah | 11:11 |
| *** mrtenio has quit IRC | 11:14 | |
| *** Nisha_Agarwal has joined #openstack-ironic | 11:15 | |
| *** udesale has quit IRC | 11:25 | |
| *** yuanying_ has quit IRC | 11:25 | |
| *** dims has joined #openstack-ironic | 11:26 | |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/bifrost: Allow to create cloud.yaml with many clouds https://review.openstack.org/426717 | 11:27 |
| sambetts | jroll: we currently have no way to push out warnings for deprecation of an option through node validate right? | 11:28 |
| *** nmathew has quit IRC | 11:34 | |
| openstackgerrit | Aparna proposed openstack/proliantutils: Add HPSUM firmware update support https://review.openstack.org/422004 | 11:37 |
| *** Nisha_Agarwal has quit IRC | 11:40 | |
| vdrok | morning jroll | 11:45 |
| *** vsaienko has joined #openstack-ironic | 11:50 | |
| *** jkilpatr has joined #openstack-ironic | 11:57 | |
| *** sayantani01 has joined #openstack-ironic | 11:57 | |
| *** sayantani01 has quit IRC | 12:02 | |
| dtantsur | sambetts, right. there was a spec by yuriyz, dunno how it ended up | 12:02 |
| *** jaganathan has joined #openstack-ironic | 12:06 | |
| *** jcoufal has joined #openstack-ironic | 12:10 | |
| sambetts | dtantsur: I think I remember that, I don't know what happened to it either | 12:10 |
| *** vsaienko has quit IRC | 12:13 | |
| dtantsur | sambetts, https://review.openstack.org/#/c/279070/ | 12:13 |
| patchbot | patch 279070 - ironic-specs - Driver configuration warnings | 12:13 |
| dtantsur | I'd like to put it on agenda for Pike fwiw | 12:13 |
| *** hshiina has joined #openstack-ironic | 12:15 | |
| sambetts | dtantsur: I think we could build it into this proposal https://bugs.launchpad.net/ironic/+bug/1655633 | 12:18 |
| openstack | Launchpad bug 1655633 in Ironic "[RFE] Enhance /drivers/properties API to have machine readable detail view " [Wishlist,New] | 12:18 |
| sambetts | dtantsur: I propose to change up the driver properties to be more like oslo config opts | 12:19 |
| *** foexle has joined #openstack-ironic | 12:19 | |
| *** awiddersheim has joined #openstack-ironic | 12:21 | |
| *** rbartal has quit IRC | 12:24 | |
| *** awiddersheim has quit IRC | 12:26 | |
| dtantsur | oh, cool, /me will check | 12:27 |
| dtantsur | sambetts, have a spec on it already? our UI folks will cry of happiness | 12:28 |
| sambetts | dtantsur: still writing it ATM, similarly we wanted it to improve the UI experience | 12:29 |
| dtantsur | sure, I'll add it to my personal priority list when it's up | 12:29 |
| sambetts | :) coolios | 12:29 |
| *** alexpilotti has quit IRC | 12:30 | |
| openstackgerrit | Aparna proposed openstack/proliantutils: Add return code for hpsum firmware update https://review.openstack.org/423856 | 12:30 |
| *** alexpilotti has joined #openstack-ironic | 12:30 | |
| *** alexpilotti has quit IRC | 12:35 | |
| openstackgerrit | Miles Gould proposed openstack/ironic: Allow setting UEFI boot devices in DRAC driver https://review.openstack.org/420107 | 12:38 |
| jaganathan | Hi, please let me know the steps to connect ironic node during introspection to see ironic python agent log | 12:38 |
| amarao | Hello everyone. I found something strange with multi-switch configuration for servers with more than one NIC: https://bugs.launchpad.net/ironic/+bug/1660324 | 12:40 |
| openstack | Launchpad bug 1660324 in ironic (Ubuntu) "There is no way to specify which port use for which network" [Undecided,New] | 12:40 |
| amarao | How does ironic (nova?) decide which ironic port is assigned to which nova interface? | 12:40 |
| milan | aarefiev, hey, if you got time: https://review.openstack.org/#/c/375045/ you already +2 it before and was +2 recently by dtantsur ;) thx! | 12:40 |
| patchbot | patch 375045 - ironic-inspector-specs - Endpoint to list introspection statuses | 12:40 |
| dtantsur | amarao, I think it's nova that decides that, and IIRC it's quite random. I may be wrong on that tho | 12:41 |
| milan | sambetts, hey, you might want to check the spec too ;) ^^ | 12:41 |
| vsaienk0 | dtantsur, jroll: http://logs.openstack.org/65/426665/3/experimental/gate-ironic-dsvm-standalone-ubuntu-xenial-nv/58bd374/console.html#_2017-01-30_12_14_45_094917 | 12:42 |
| amarao | dtantsur. Thanks. This is really bad if switch1 and switch2 are serving different types of networks (internet and local, f.e.). | 12:42 |
| dtantsur | vsaienk0, nice! will check the patch ~ soon | 12:42 |
| sambetts | amarao: Our current multi tenant implmentation only allows for one neutron physnet, and expects both switches to have access to that | 12:42 |
| milan | guys, also qq about https://review.openstack.org/#/c/421285/7/ironic_inspector/db.py@74 < uuid vs id: might exposing look-up attributes over API make sense to you one day? | 12:44 |
| patchbot | patch 421285 - ironic-inspector - Find a node by multiple attributes | 12:44 |
| milan | aarefiev, sambetts ^ | 12:44 |
| dtantsur | sambetts, does https://bugs.launchpad.net/ironic/+bug/1660324 sound like an RFE? it may be part of what dsneddon_afk was working on for tripleo | 12:44 |
| openstack | Launchpad bug 1660324 in ironic (Ubuntu) "There is no way to specify which port use for which network" [Undecided,New] | 12:44 |
| amarao | sambetts... Except for this randomness, is other places where this 'one' is defined? It can boot with two tenant networks (I just tested that) and it works fine except for randomness. | 12:44 |
| vsaienk0 | sambetts: I think amarao is looking for neutron router networks support, am I right? | 12:44 |
| sambetts | vsaienk0: yup | 12:44 |
| dtantsur | yep | 12:44 |
| amarao | Not exactly. I have --router:external for both networks. | 12:45 |
| *** lucasagomes is now known as lucas-hungry | 12:45 | |
| aarefiev | milan: I think we have rfe about exposing look up attributes | 12:45 |
| dtantsur | amarao, we have a guy looking into a similar or the same thing, dsneddon_afk. he's on US west coast though, and you seem to be from east EU... | 12:45 |
| amarao | I just want to fix somehow which ironic port use for which instance interface. | 12:45 |
| milan | aarefiev, thx me looks | 12:46 |
| *** ccamacho is now known as ccamacho|lunch | 12:46 | |
| amarao | I'll try now to use neutron port-create/nova interface-attach interface to bind them in proper order. | 12:46 |
| sambetts | amarao: we don't support post-deploy interface attach at the moment | 12:46 |
| amarao | sambetts, oh, thanks. | 12:47 |
| sambetts | vsaienk0: I have been thinking about this for a little while but I think we might need to have a "physnet" field on the Ironic port to allow for placement of neutron ports on to the right NIC | 12:48 |
| sambetts | vsaienk0: this would allow for both placement and validation to ensure that a neutron port will work with our Ironic node | 12:49 |
| milan | aarefiev, I can't see it, mind posting a link? | 12:49 |
| dtantsur | milan, https://bugs.launchpad.net/ironic-inspector/+bug/1525231 ? | 12:50 |
| openstack | Launchpad bug 1525231 in Ironic Inspector "[RFE] API for dynamic lookup attributes " [Wishlist,Confirmed] | 12:50 |
| sambetts | vsaienk0: becuase right now we have no way to prevent someone attaching a neutron port that is physically unreachable | 12:50 |
| amarao | sambetts, yes, it sounds reasonable. | 12:50 |
| milan | dtantsur, thx, was looking into specs | 12:50 |
| *** sayantani01 has joined #openstack-ironic | 12:51 | |
| vsaienk0 | sambetts: I think it is exactly what neutron routed networks provide. For example user have two neutron networks A and B. Network A is associated with subnet sub_A and network B is associated with sub_B. We should be able to specify subnets (not networks) that are available for specific port or node | 12:52 |
| sambetts | vsaienk0: I also wonder what would happen if you had two physnets each supporting vlans 0 - 100, our baremetal node is on physnet 1, you get a port vlan 50 on physnet 2, would neutron bind vlan 50 on the wrong physnet if it just talked to the switch we had in our local link? | 12:53 |
| amarao | sambetts, with generic_switch it just goes to specific switch and configure it. No any smart logic. | 12:54 |
| amarao | generic_switch+ml2+neutron just ignores 'physnet' part right now. | 12:54 |
| sambetts | yeah... thats worrying ... we need to associate switches to physnets somehow... | 12:55 |
| amarao | *Except for HDCP, which is separate thing. I use ConfigDrive for instances, so I didn't dig too deep. | 12:55 |
| sambetts | vsaienk0: ^ | 12:56 |
| amarao | I saw old brocade code, it uses physical segment of network to choose switch. | 12:56 |
| *** sayantani01 has quit IRC | 12:56 | |
| amarao | And it has map physnet<->switch in own config. | 12:56 |
| *** dtantsur is now known as dtantsur|brb | 12:56 | |
| amarao | I think best way to deal with 'multiswitch' configuration is to put physnet to local_link_connection. It makes perfect sense. | 12:58 |
| vsaienk0 | sambetts: yeah, looks like neutron allows to have overlapping segmentation ID on different phys nets | 13:01 |
| vsaienk0 | sambetts: mapping subnet to switch should help here | 13:01 |
| sambetts | vsaienk0: we currently have a (pretty major) sec bug here though, which is the networking-gs will bind a port into the wrong physnet which might result in a port being bound into someone elses network | 13:02 |
| *** bcornec has quit IRC | 13:02 | |
| sambetts | jroll: ^^ | 13:02 |
| jroll | :| | 13:04 |
| vsaienk0 | sambetts: it may occur only when administrator uses overlapping vlans for different physnets, and looks fear for all ML2 drivers | 13:05 |
| *** dprince has joined #openstack-ironic | 13:06 | |
| jroll | so this is just a neutron / networking-generic-switch problem, right? not something we can fix in ironic? | 13:06 |
| amarao | It's never occur to me that it's possible to use overlapping vlans on different physnets. And I'm not sure if this would work as expected in ovs of-rules (in openvswitch agent). Same vlans may share traffic. | 13:06 |
| amarao | I think vlan overlapping in physnets as secondary issue. Just note it in documentation as "don't do it". | 13:07 |
| vsaienk0 | sambetts: also this may happen only when both networks are allowed on the same switch, which is impossible how administrator will guarantee that both networks with the same VLAN are isolated? | 13:07 |
| sambetts | amarao: ovs has internal VLANs which it uses to separate traffic which it should then output to a different bridge for each NIC | 13:07 |
| sambetts | vsaienk0: two physnets each supporting vlans 0 - 100, our baremetal node is on physnet 1, you get a port vlan 50 on physnet 2, would neutron bind vlan 50 on the wrong physnet 1 but some else might already own physnet 1 vlan 50 | 13:08 |
| vsaienk0 | sambetts: officially we support only VLAN network, and we pick network segmentation ID not internal segmentation ID value. | 13:08 |
| *** vikrant has quit IRC | 13:09 | |
| vsaienk0 | sambetts: ah... looks like an real issue | 13:10 |
| amarao | I want to try to create server with port-id instead of net-id. Is this make any sense? (Create ports before booting baremetal) | 13:10 |
| sambetts | amarao: that should work, but nic -> port mapping is still arbitrary :( | 13:11 |
| *** athomas has quit IRC | 13:11 | |
| openstackgerrit | Naohiro Tamura proposed openstack/ironic: Update the existing APIs due to adding get_supported_power_states https://review.openstack.org/382194 | 13:11 |
| amarao | sambetts :( | 13:11 |
| sambetts | jroll: ideally I think we need two fixes, we need to sort the VIF to Port mapping on the Ironic side to ensure that we put Vifs on the right ports for the physnet they are attached too, and we need to ensure that networking-gs doesn't bind ports onto switches on the wrong physnet | 13:12 |
| sambetts | jroll: unfortunatly I don't think there is a way to query neutron to find out which physnet a ironic ports switch is in... which means duplicating config... or we just leave it down to the point we bind the port to fail, but that would result in random deployment failures depending on the order of the mapping of the vif to ports | 13:15 |
| *** trown|outtypewww is now known as trown | 13:16 | |
| TheJulia | Good morning | 13:17 |
| amarao | hi | 13:17 |
| jroll | sambetts: hrm, that first one is a bug we've had for years, right | 13:20 |
| *** athomas has joined #openstack-ironic | 13:20 | |
| jroll | sambetts: which we're getting... closer to fixing, I think | 13:20 |
| jroll | but isn't really a drop everything and fix it this week bug | 13:20 |
| sambetts | jroll: to some degree yes, it just gets worse now we're using tenant networks because they are allocated to physnets by neutron not by an admin | 13:21 |
| jroll | oh, neutron network != physnet | 13:21 |
| jroll | sigh | 13:21 |
| sambetts | jroll: no, physnet == physical L2 layer | 13:22 |
| jroll | is physnet a thing in neutron, though, that represents an L2 network? | 13:22 |
| sambetts | jroll: yeah, you configure it in neutron like vlan_ranges= physnet1:100:500,physnet2:100:500 | 13:23 |
| sambetts | jroll: physnet1 and physnet2 represent destict physical networks | 13:23 |
| sambetts | jroll: vlan 100 in physnet1 != vlan 100 in physnet2 | 13:24 |
| jroll | right | 13:24 |
| vsaienk0 | sambetts: shouldn't we store what physnets is available on the switch in neutron ml2 config and just use this information during port binding by Neutron ML2? | 13:24 |
| jroll | so the issue is a port that is supposed to be on physnet1 could land on a switch on physnet 2, where $vlan belongs to someone else?: | 13:24 |
| jroll | sambetts: ^ | 13:24 |
| sambetts | jroll: yup | 13:24 |
| jroll | ok | 13:25 |
| jroll | is there not some mapping from port:physnet or network:physnet? | 13:25 |
| jroll | or is it the switch:physnet mapping that's missing? | 13:26 |
| sambetts | switch:physnet is missing | 13:26 |
| vsaienk0 | sambetts, jroll also we need to implement port status notification in Ironic, we completely ignoring so not able to check if any issue happened during port binding | 13:26 |
| *** bfournie has quit IRC | 13:26 | |
| jroll | one thing at a time......... | 13:26 |
| openstackgerrit | Aparna proposed openstack/ironic: Doc for sanitize erase support in iLO drivers https://review.openstack.org/408441 | 13:26 |
| *** hshiina has quit IRC | 13:27 | |
| jroll | sambetts: iirc switches don't exist in neutron core, so it seems like it's a thing that ml2 plugins need to handle, right? | 13:27 |
| jroll | or ml2 plugins need to block it | 13:27 |
| jroll | and we need to do our best to map it correctly | 13:27 |
| sambetts | vsaienk0, jroll: ensuring that neutron side would reject the bind request for the wrong switch + physnet would prevent the sec issue, but would also result in random failure if the vif -> port mapping is wrong | 13:28 |
| openstackgerrit | Merged openstack/ironic: [Devstack] Set DEFAULT_IMAGE_NAME variable https://review.openstack.org/425599 | 13:28 |
| vsaienk0 | jroll even if ML2 plugin will reject binding, we do not check it on Ironic side | 13:28 |
| sambetts | so we need both I guess | 13:28 |
| vsaienk0 | + to sambetts: it will lead to random failures | 13:28 |
| jroll | which would likely need the switch:physnet api, OR we would need to have something like port.local_link_info['physnets'] = ['foo', 'bar'], ya? | 13:29 |
| jroll | and that would allow us to determine if a neutron port is being bound to an ironic port that will work correctly | 13:29 |
| *** awiddersheim has joined #openstack-ironic | 13:29 | |
| sambetts | jroll: I think physnet in local_link is the only solution we can implment, I don't know if any way to query it right now | 13:30 |
| sambetts | +1 | 13:30 |
| vsaienk0 | sambetts: why to store it in local_link_info and not in the ML2 config? | 13:30 |
| jroll | sambetts: yeah, we'd have to convince neutron to do the former | 13:30 |
| sambetts | vsaienk0: we need to do both | 13:30 |
| jroll | really sucks to have to put this in local_link_info :/ | 13:31 |
| sambetts | the ml2 driver need to do it to prevent bad binding, and Ironic needs to do it to prevent bad mapping | 13:31 |
| *** aparnav has quit IRC | 13:31 | |
| vsaienk0 | sambetts: why both? we reject binding on ML2 side, and only need to have notification from Neutron that port wasn't bound | 13:31 |
| sambetts | vsaienk0: and then deployment fails even if the Ironic node has a NIC in that physnet | 13:32 |
| sambetts | vsaienk0: but it was just mapped wrogn | 13:32 |
| sambetts | jroll: alternative is to progressivly try to bind a vif to each Ironic port until we get a sucessful binding | 13:32 |
| jroll | sambetts: mehhhhh | 13:32 |
| sambetts | jroll: quite | 13:32 |
| jroll | I'd rather we had APIs for physical things in neutron proper | 13:33 |
| vsaienk0 | sambetts: yeah, we still need to know physnet to port mapping in ironic to pick correct NIC for a VIF | 13:34 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Add Ironic standalone tests https://review.openstack.org/423556 | 13:34 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: DNM/DNR: Temp patch to verify ironic standalone tests https://review.openstack.org/426665 | 13:34 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: [Devstack] Add stack user to libvirt group https://review.openstack.org/426664 | 13:34 |
| amarao | vsaienk0: If binding failed it does not report errors on port binding, just park port as 'binding_failed'. When I had debugged generic_switch I often had client server booting into provision network due to this 'binding_failed' | 13:35 |
| sambetts | amarao: woah?! it shouldn't boot into the provsioing network regardless of tenant binding failed, the vlan should be removed from the switch when the provisioning nic is deleted right?! | 13:36 |
| sambetts | vsaienk0: ^ | 13:36 |
| vsaienk0 | amarao: ironic do not check neutron port status binding at the moment, so it will continue deploy even with unbound ports | 13:37 |
| vsaienk0 | amarao the same is for tenant network | 13:37 |
| amarao | vsaienk0 It sounds as CVE to me. | 13:38 |
| vsaienk0 | amarao what exactly that ironic do not check Neutron port status? | 13:39 |
| *** sayantani01 has joined #openstack-ironic | 13:39 | |
| amarao | Steps to reproduce: | 13:40 |
| amarao | 1. Boot tenant instance. It will goes to provision network, get image. | 13:40 |
| amarao | 2. Ironic asks neutron to swap port to different vlan. If this operation failed (binding_failed) than instance booting with client image and provision vlan on port (inside switch). | 13:40 |
| amarao | *then | 13:41 |
| sambetts | amarao: we delete the neutron port for the provisioning vlan, neutron should clean up that port's binding on the switch | 13:41 |
| amarao | sambetts: And if this operation fails? | 13:42 |
| vsaienk0 | sambetts: if delete failed, Ironic will not check this | 13:42 |
| *** srobert has joined #openstack-ironic | 13:42 | |
| amarao | When I debugged generic_switch I saw this many times. Typo in 'removal code' - and tenant has access to provision network. | 13:42 |
| *** lucas-hungry is now known as lucasagomes | 13:42 | |
| amarao | * And it can be not just typo, but lagging ssh on switch, temporal network failure, etc. | 13:43 |
| *** sdake has joined #openstack-ironic | 13:43 | |
| sambetts | *face palm* | 13:43 |
| vsaienk0 | sambetts, amarao this problem is not only fear to generic-switch, Ironic do not check port status so we can't predict if operation completed successfully | 13:43 |
| sambetts | ML2 driver should ensure that even on network outtage or anything like that, that the port is cleaned up | 13:44 |
| amarao | Nova has special option for those cases: vif_plugging_is_fatal = True | 13:44 |
| *** sayantani01 has quit IRC | 13:44 | |
| amarao | sambetts If neutron-server has outage to switch, ML2 has no physical means to do this. | 13:45 |
| amarao | (BoolOpt) Fail instance boot if vif plugging fails | 13:45 |
| amarao | I think same should be for 'fail to delete/unplug from provision network'. | 13:45 |
| vsaienk0 | sambetts, jroll what do you think to reconsider priority for neutron port status notifications? | 13:46 |
| jroll | vsaienk0: I've wanted to do it for some time, but there's no way we can do it in ocata | 13:47 |
| jroll | vsaienk0: but yes we should talk about it as a pike thing | 13:48 |
| sambetts | So as I see it we have several issues we've discussed so far: | 13:48 |
| sambetts | * Ironic physnet mapping | 13:48 |
| sambetts | * networking-gs needs physnet mapping | 13:49 |
| vsaienk0 | jroll, the main concern was where to put the notificator code in neutron tree or not. But neutron team doesn't want to put it in the tree as they provide ML2 mechanism | 13:49 |
| sambetts | * Ironic not checking if port binding/unbinding is successful | 13:49 |
| sambetts | vsaienk0: Neutron where happy to put it in their tree if it was similar code wise to the nova one | 13:50 |
| sambetts | s/where/were | 13:50 |
| vsaienk0 | sambetts: networking-gs mapping is the lowest priority as it will not work without notifications | 13:50 |
| amarao | Can ironic just check port status prior to performing any driver operations (like pxe boot, or power changes)? | 13:50 |
| *** baoli has joined #openstack-ironic | 13:50 | |
| *** ppiela has joined #openstack-ironic | 13:51 | |
| sambetts | vsaienk0: networking-gs should implmented it anyway because even if the binding_fails and the deploy finishes the node won't be connected to the wrong network | 13:51 |
| vsaienk0 | sambetts: we still need to fix issue with ports in flat network, they always remain bound, and requires a separate ML2 | 13:52 |
| vsaienk0 | s/bound/unbound | 13:52 |
| vsaienk0 | sambetts: so once we have a separate ML2 for fake flat port binding why not keep notificator code there? | 13:53 |
| sambetts | vsaienk0: notifiers and ml2 drivers are very different things | 13:53 |
| vsaienk0 | sambetts: it is possible to implement notifier on top of ML2 | 13:54 |
| *** sasha2 has joined #openstack-ironic | 13:54 | |
| vsaienk0 | sambetts: some time ago I've did PoC on top on networking-gs https://review.openstack.org/#/c/357780 | 13:54 |
| patchbot | patch 357780 - networking-generic-switch - WIP: Enable ironic notifications | 13:54 |
| sambetts | vsaienk0: sure, but they don't need to be the same thing | 13:54 |
| sambetts | vsaienk0: and neutron wants us to implemented it like the https://github.com/openstack/neutron/blob/b9d0a5b885282f170952b102a8f69f3112df9406/neutron/notifiers/nova.py | 13:55 |
| vsaienk0 | sambetts: yes. Ok I will reupload https://review.openstack.org/#/c/343684/ so we can start implementing ironic side, which is the bigest part. Where to keep notificator code is not so important for now | 13:55 |
| patchbot | patch 343684 - ironic-specs - Add synchronize-events-with-neutron spec. | 13:55 |
| jroll | sambetts | vsaienk0: Neutron where happy to put it in their tree if it was similar code wise to the nova one <- +1, that's what I recall | 13:56 |
| sambetts | +1 | 13:57 |
| *** sdake has quit IRC | 13:58 | |
| *** karthiks has quit IRC | 13:59 | |
| vsaienk0 | jroll, sambetts: but we can't enable notifications for both flat and multitenancy networks, until we fix unbound state for ports in flat network, which requires separate ML2 https://review.openstack.org/#/c/339129/ | 14:00 |
| patchbot | patch 339129 - neutron - Add BaremetalMechanismDriver driver. | 14:00 |
| *** bfournie has joined #openstack-ironic | 14:00 | |
| vsaienk0 | unless we decide to support 'partial' notifications | 14:00 |
| *** jschwarz|lunch is now known as jschwarz | 14:01 | |
| *** dtantsur|brb is now known as dtantsur | 14:01 | |
| *** kjw3 has quit IRC | 14:04 | |
| openstackgerrit | edan david proposed openstack/ironic-inspector: Adding InfiniBand Support https://review.openstack.org/264257 | 14:04 |
| *** Goneri has joined #openstack-ironic | 14:06 | |
| *** rloo has joined #openstack-ironic | 14:07 | |
| *** sdake has joined #openstack-ironic | 14:10 | |
| amarao | I have one question: When Ironic creates console via shellinabox, is it protected by token or password? I see a plain link: http://192.168.26.61:11623/, which is funny, but it not very multitenant. | 14:12 |
| *** clenimar has quit IRC | 14:12 | |
| *** gcb has joined #openstack-ironic | 14:13 | |
| openstackgerrit | Merged openstack/ironic-inspector: Switch to decorators.idempotent_id https://review.openstack.org/426099 | 14:13 |
| *** sdake has quit IRC | 14:14 | |
| *** clenimar has joined #openstack-ironic | 14:16 | |
| *** alexpilotti has joined #openstack-ironic | 14:16 | |
| *** sdake has joined #openstack-ironic | 14:19 | |
| *** jheroux has joined #openstack-ironic | 14:19 | |
| vsaienk0 | jroll, dtantsur please review https://review.openstack.org/#/c/426700/ with changes standalone tests passing CI http://logs.openstack.org/65/426665/4/experimental/gate-ironic-dsvm-standalone-ubuntu-xenial-nv/f08ae81/console.html#_2017-01-30_14_03_31_298726 | 14:19 |
| patchbot | patch 426700 - openstack-infra/project-config - Update experimental ironic-standalone job | 14:19 |
| *** alexpilotti has quit IRC | 14:20 | |
| *** ccamacho|lunch is now known as ccamacho | 14:22 | |
| *** sdake has quit IRC | 14:23 | |
| *** trown is now known as trown|brb | 14:24 | |
| *** trown|brb is now known as trown | 14:25 | |
| *** dyasny has joined #openstack-ironic | 14:25 | |
| *** dyasny has quit IRC | 14:25 | |
| *** dyasny has joined #openstack-ironic | 14:25 | |
| *** alexpilotti has joined #openstack-ironic | 14:26 | |
| openstackgerrit | Merged openstack/ironic-inspector: Allow diskless nodes introspection https://review.openstack.org/326620 | 14:28 |
| *** Nisha_Agarwal has joined #openstack-ironic | 14:30 | |
| *** sdake has joined #openstack-ironic | 14:31 | |
| *** sdake has quit IRC | 14:31 | |
| vsaienk0 | dtantsur: just resolved you comments https://review.openstack.org/#/c/426700/2 | 14:32 |
| patchbot | patch 426700 - openstack-infra/project-config - Update experimental ironic-standalone job | 14:32 |
| *** sayantani01 has joined #openstack-ironic | 14:33 | |
| dtantsur | thnx, will check | 14:34 |
| *** athomas has quit IRC | 14:34 | |
| dtantsur | milan, mind reviewing https://review.openstack.org/#/c/425641/ please? it fixes an issue with bulk introspection and virtualbmc | 14:34 |
| patchbot | patch 425641 - ironic-inspector - Deprecate introspection_delay_drivers option and m... | 14:34 |
| milan | dtantsur, sure thing | 14:35 |
| milan | dtantsur, thx for pointing | 14:35 |
| *** alexpilotti has quit IRC | 14:37 | |
| *** sayantani01 has quit IRC | 14:38 | |
| *** dyasny has quit IRC | 14:39 | |
| *** dyasny has joined #openstack-ironic | 14:39 | |
| openstackgerrit | Grzegorz Grasza (xek) proposed openstack/ironic: Add support for DBDeadlock handling https://review.openstack.org/394413 | 14:40 |
| *** karthiks has joined #openstack-ironic | 14:46 | |
| *** sdake has joined #openstack-ironic | 14:46 | |
| *** athomas has joined #openstack-ironic | 14:46 | |
| *** rcernin has quit IRC | 14:48 | |
| openstackgerrit | Merged openstack/ironic: Add Inject NMI to api-ref https://review.openstack.org/423941 | 14:49 |
| milan | dtantsur, +2 | 14:50 |
| dtantsur | thnx | 14:51 |
| *** alexpilotti has joined #openstack-ironic | 14:51 | |
| *** sdake_ has joined #openstack-ironic | 14:54 | |
| vdrok | sambetts: there is a small issue in the first of the 'fix subcontrollers collisions' patch. I'm +2 if you fix it :) | 14:55 |
| *** udesale has joined #openstack-ironic | 14:57 | |
| sambetts | vdrok: cool, I saw your comments, not sure what to do about the error code, seems weird we've made it inconsistent :/ | 14:57 |
| *** sdake has quit IRC | 14:57 | |
| openstackgerrit | Bob Fournier proposed openstack/ironic-inspector: Add plugin to process basic management LLDP TLVs https://review.openstack.org/406496 | 14:58 |
| vdrok | sambetts: yup. I think about just leaving things as is now, and refactor it later, when we make them all consistent | 14:58 |
| *** causten has joined #openstack-ironic | 14:58 | |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Forbid passing portgroup_uuid to POST /ports in old API https://review.openstack.org/404839 | 14:58 |
| *** dyasny has quit IRC | 14:59 | |
| *** dyasny has joined #openstack-ironic | 14:59 | |
| *** rcernin has joined #openstack-ironic | 15:00 | |
| *** crushil has joined #openstack-ironic | 15:01 | |
| *** causten has quit IRC | 15:05 | |
| *** jaganathan has quit IRC | 15:07 | |
| *** rloo has quit IRC | 15:09 | |
| *** rloo has joined #openstack-ironic | 15:09 | |
| *** dyasny has quit IRC | 15:11 | |
| *** dyasny has joined #openstack-ironic | 15:13 | |
| jroll | rloo: morning, can we figure out https://review.openstack.org/#/c/417970/24/ironic/tests/unit/conductor/test_manager.py@3863 before the meeting maybe? :) | 15:14 |
| patchbot | patch 417970 - ironic - Add dynamic driver functionality to REST API | 15:14 |
| rloo | jroll: hi, morning. I haven't yet looked, give me 5 min or so to finish something up. | 15:15 |
| jroll | rloo: ok, no worries, just want to get it going | 15:15 |
| rloo | jroll: def. didn't realize it was 'stuck' (probably cuz of me.) just getting internal status report done first... | 15:16 |
| jroll | rloo: I responded to you with a question, I think the current code may be okay, but we maybe should just remove that test | 15:16 |
| jroll | no rush though | 15:16 |
| *** baoli has quit IRC | 15:18 | |
| *** dyasny has quit IRC | 15:19 | |
| *** baoli has joined #openstack-ironic | 15:19 | |
| *** vinsh has joined #openstack-ironic | 15:20 | |
| *** alexpilotti has quit IRC | 15:20 | |
| *** rpioso has joined #openstack-ironic | 15:22 | |
| *** jaosorior has joined #openstack-ironic | 15:23 | |
| jroll | vdrok: turns out you were right on the default validation thing, I've added that to my todo list | 15:23 |
| milan | morning jroll, rloo! :) | 15:23 |
| *** udesale has quit IRC | 15:24 | |
| jroll | dtantsur: sambetts: mind reviewing this (updated) RFE: https://bugs.launchpad.net/ironic/+bug/1609622 | 15:25 |
| openstack | Launchpad bug 1609622 in Ironic "RFE: Fix iLO drivers to get local_gb for all storages" [Wishlist,In progress] - Assigned to Nisha Agarwal (agarwalnisha1980) | 15:25 |
| jroll | seems like it should be okay to me | 15:25 |
| jroll | odd that we need snmp to do that thing, but ya know, hardware be weird | 15:25 |
| dtantsur | jroll, I hope "SNMP inspection will be invoked only if user asks for it." is not correct (seems fixed in other places) | 15:26 |
| jroll | dtantsur: I read that as must have driver_info[snmp_inspection] = True | 15:27 |
| jroll | which seems fine to me | 15:27 |
| dtantsur | this is where I'm -1 to it... | 15:27 |
| jroll | oh? | 15:27 |
| dtantsur | ETOOMANYOPTIONS | 15:27 |
| *** jaosorior has quit IRC | 15:27 | |
| dtantsur | one more thing for operators to figure out to make their cloud behave | 15:27 |
| jroll | surely some folks turn off snmp in their DC | 15:27 |
| jroll | but yeah | 15:27 |
| jroll | maybe that can be detected instead | 15:27 |
| dtantsur | we can try and fail. anyway, it's only done if we don't get local_gb, right? | 15:28 |
| rloo | hi milan | 15:28 |
| jroll | dtantsur: unclear, seems maybe both are done (as one way gets some disks, the other gets other disks) | 15:28 |
| jroll | but yeah I see what you're saying now | 15:28 |
| dtantsur | jroll, well, it should be made clear IMO | 15:28 |
| dtantsur | fwiw we already have drivers using snmp when their vendor interface cannot do something | 15:29 |
| * jroll now wonders why auth_user, etc, aren't snmp_auth_user | 15:29 | |
| dtantsur | jroll, this was my last comment. should be e.g. snmp_username, etc | 15:30 |
| jroll | dtantsur: yep | 15:30 |
| jroll | thanks | 15:30 |
| * jroll does the comments | 15:31 | |
| *** jaosorior has joined #openstack-ironic | 15:32 | |
| *** rajinir has joined #openstack-ironic | 15:33 | |
| *** sayantani01 has joined #openstack-ironic | 15:34 | |
| rloo | jroll: i looked at the comments in 417970. seems too early in the week for me to answer :) | 15:34 |
| NobodyCam | Good Morning Ironicers | 15:35 |
| jroll | rloo: :( | 15:35 |
| dtantsur | morning NobodyCam, rloo | 15:35 |
| rloo | jroll: i think the first question which might be the question you asked in the patch. what is the no-X interfaces meant to do | 15:35 |
| sergek | Morning NobodyCam , rloo | 15:35 |
| rloo | hey, good morning NobodyCam, sergek, dtantsur | 15:35 |
| NobodyCam | Morning dtantsur rloo serge :) | 15:35 |
| JayF | gm everyone o/ | 15:35 |
| NobodyCam | morning JayF :) | 15:35 |
| rloo | jroll: if it is really 'impossible' for an operator to set the default_X_interface (or is it the enabled_X_interface) to None, will that be documented? cuz it ain't clear. | 15:36 |
| dtantsur | morning JayF | 15:36 |
| mariojv | good morning | 15:36 |
| JayF | https://review.openstack.org/#/c/426647/ easy docs review if someone has a second | 15:36 |
| patchbot | patch 426647 - ironic - Update the outdated link in user-guide | 15:36 |
| dtantsur | rloo, both are possible. what is not possible is to end up with a valid node with one of interfaces being None. | 15:36 |
| * dtantsur is not sure if his comment is relevant | 15:37 | |
| jroll | ^^ | 15:37 |
| jroll | rloo: CONF.default_foo_interface=None means "use the first supported interface that is also enabled" | 15:37 |
| NobodyCam | morning mariojv | 15:37 |
| NobodyCam | morning jroll | 15:37 |
| jroll | heya NobodyCam | 15:37 |
| mariojv | o/ | 15:37 |
| NobodyCam | :) | 15:37 |
| rloo | jroll, dtantsur: i looked at the test. i was trying to set enabled_raid_interface = None, and I got the NoValidDefaultForInterface exception. | 15:38 |
| dtantsur | rloo, this is right, yes | 15:38 |
| rloo | jroll, dtantsur: so does that mean, we cannot set enabled_raid_interface to None? | 15:38 |
| dtantsur | why not? | 15:38 |
| dtantsur | ah, enabled - yes | 15:38 |
| dtantsur | you have to leave no-raid there. | 15:38 |
| dtantsur | default one can be perfectly set to None though | 15:39 |
| patchbot | Error: 'supybot.one' is not a valid configuration variable. | 15:39 |
| dtantsur | WUT? | 15:39 |
| jroll | lol | 15:39 |
| dtantsur | default username | 15:39 |
| patchbot | Error: 'supybot.username' is not a valid configuration variable. | 15:39 |
| dtantsur | hmm, I wonder if I can find some use of this non-sense :) | 15:39 |
| dtantsur | anyway | 15:39 |
| rloo | dtantsur, jroll: I just looked at the description for the enabled_X_interfaces. No where does it say that it cannot be None. | 15:40 |
| dtantsur | we agreed that a node without a valid interface should not exist | 15:40 |
| jroll | rloo: then we should fix that | 15:40 |
| dtantsur | rloo, we can update it, yeah. do we say that enabled_drivers cannot be None? | 15:40 |
| rloo | dtantsur: no we don't. | 15:40 |
| *** yolanda has quit IRC | 15:41 | |
| JayF | default ; drop table supybot; | 15:41 |
| patchbot | Error: 'supybot.;' is not a valid configuration variable. | 15:41 |
| dtantsur | LOL | 15:41 |
| dtantsur | rloo, yeah, we assume it's obvious. maybe it's no longer with all this driver composition stuff. | 15:41 |
| rloo | dtantsur, jroll: so I will assume that we will update the enabled_X to ay they cannot be None. Would like to make sure we also code for that. Just thinking about it, I wonder if we/code should allow None and replace with no-X interface. On the other hand, if we code so None is not allowed, why would we even need the no-X interfaces? | 15:42 |
| openstackgerrit | Sam Betts proposed openstack/ironic: Rename and redocument the Cisco UCSM driver https://review.openstack.org/400673 | 15:42 |
| dtantsur | rloo, I think we've had a loooong discussion on accepting None.. tl;dr None means "calculated default" | 15:42 |
| dtantsur | on nodes | 15:42 |
| rloo | dtantsur: for the default_x_interface, i know. did we also discuss for enabled_x_interfaces? | 15:43 |
| rloo | dtantsur: /me looks at spec... | 15:43 |
| dtantsur | I'm not sure why you see it as problem, to be honest | 15:43 |
| dtantsur | there may be quite a few configuration options where we don't accept None as value | 15:43 |
| rloo | dtantsur: cuz i, as dummy user, tried setting it to None and got totally confused as to how things were meant to work | 15:43 |
| dtantsur | as a user, what did you expect to see by setting enable_x_interfaces to None? | 15:44 |
| jroll | so, as usual, we need to write good docs :) | 15:44 |
| JayF | Is there something we can do in the help text to prevent a user from thinking None is valid? | 15:44 |
| * JayF == jroll , this seems like a docs issue | 15:44 | |
| rloo | dtantsur: that it would 'turn off' that interface. eg, the hardware (or node's whatever) doesn't support it | 15:44 |
| dtantsur | rloo, we don't have a notion of turning off interfaces.. it may make sense for RAID, but no whatsoever for e.g. power | 15:45 |
| *** rbudden has joined #openstack-ironic | 15:45 | |
| rloo | dtantsur: what about storage interface (the future)? | 15:45 |
| dtantsur | rloo, it has a noop implementation, just like network one. it cannot be None still | 15:45 |
| * jroll needs an early lunch today, will be back before the meeting | 15:46 | |
| openstackgerrit | Merged openstack/ironic: Update the outdated link in user-guide https://review.openstack.org/426647 | 15:46 |
| rloo | dtantsur: i'm not arguing that we should allow None. I am just saying that there is something wrong now, and we need to fix/address it. | 15:46 |
| dtantsur | we used to have things like .inspect = None in classic drivers. but it ended up confusing, and also it's harder to code around (you need task.driver.inspect is not None everywhere) | 15:46 |
| dtantsur | I'm in the documentation camp, just trying to verify our assumptions | 15:46 |
| rloo | dtantsur: so what do you suggest? | 15:46 |
| jroll | dtantsur: fwiw, this started with line 3855 here: https://review.openstack.org/#/c/417970/24/ironic/tests/unit/conductor/test_manager.py@3858 | 15:47 |
| patchbot | patch 417970 - ironic - Add dynamic driver functionality to REST API | 15:47 |
| jroll | which in classic driver world tests the case where driver.raid == None | 15:47 |
| rloo | jroll: are you leaving now? should we resume discussion after the meeting? | 15:47 |
| *** yolanda has joined #openstack-ironic | 15:47 | |
| jroll | so the question is, should no-raid fail validation, and we test that? or is this test invalid for dynamic drivers | 15:47 |
| jroll | rloo: I'm flexible | 15:47 |
| dtantsur | rloo, documenting that enabled_xxx_interface should have at least one value, with no-x being the way to disable it. | 15:48 |
| *** zackf has joined #openstack-ironic | 15:48 | |
| dtantsur | jroll, rloo, the test looks completely valid to me | 15:48 |
| rloo | dtantsur: ok, that is fine with me. But we also should code that if someone sets enabled_X_interface to None, that the conductor doesn't start up, right? | 15:49 |
| jroll | dtantsur: does this look valid? https://review.openstack.org/#/c/417970/24/ironic/drivers/modules/noop.py | 15:49 |
| patchbot | patch 417970 - ironic - Add dynamic driver functionality to REST API | 15:49 |
| *** sayantani01 has quit IRC | 15:49 | |
| jroll | rloo: yes, that's on my todo list already (thanks to this discussion) | 15:49 |
| rloo | jroll: ok thx. | 15:49 |
| *** sayantani01 has joined #openstack-ironic | 15:50 | |
| jroll | rloo: specifically, validate we have a good default interface for each hardware type, during conductor startup | 15:50 |
| dtantsur | jroll, no, because of signature mismatch. but the idea is valid | 15:50 |
| openstackgerrit | Sam Betts proposed openstack/ironic: Reorganise Cisco UCSM driver code to match its name https://review.openstack.org/426814 | 15:50 |
| rloo | jroll, dtantsur: so as far as the NoX interface classes go. Are they meant to fail for *any* method call? | 15:50 |
| jroll | dtantsur: heh, good point, thanks | 15:50 |
| JayF | rloo: not fail, but noop | 15:51 |
| dtantsur | rloo, yes | 15:51 |
| vdrok | morning rloo NobodyCam and JayF | 15:51 |
| dtantsur | JayF, this is not about noop, this is about NoRaid and family | 15:51 |
| dtantsur | the difference is whether a user invoke them or ironic itself | 15:51 |
| JayF | dtantsur: no-raid implements every method in the interface as a noop | 15:51 |
| dtantsur | I think we have it in the spec | 15:51 |
| JayF | dtantsur: which is different than "fail"ing the call, at least in my mind | 15:51 |
| NobodyCam | Morning vdrok :) | 15:51 |
| jroll | JayF: it does not | 15:52 |
| vdrok | should we also add no-* to all drivers as supported interfaces? | 15:52 |
| jroll | JayF: https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/noop.py#L65 | 15:52 |
| dtantsur | vdrok, we do it, but I'd not enforce it in code | 15:52 |
| vdrok | dtantsur: yeah, I meant enforcing | 15:52 |
| dtantsur | as we no longer know which interfaces are "optional" in the world of the driver composition | 15:53 |
| JayF | jroll: oooh. okay. Does that mean something like rescue needs to have a no-rescue interface, and we have to handle UnsupportedDriverExtension? | 15:53 |
| rloo | jroll, dtantsur (and everyone else interested, yay): for the NoX interface classes, we want .get_properties() to fail too? (I didn't test, but I think it will pass) | 15:53 |
| jroll | JayF: possibly, not all interfaces have one (like deploy), though now that I check we already have a NoRescue class | 15:54 |
| JayF | that's a hell of a lot more complicated :( | 15:54 |
| jroll | rloo: I don't think so, because e.g. using no-raid will mean one can't get driver properties from the API (because that call would fail) | 15:54 |
| JayF | mariojv: ^ read scrollback, tl;dr we have to handle the no-rescue case, which raises UnsupportedDriverExtension | 15:55 |
| dtantsur | rloo, I think it's going to screw up the whole get properties API, unless we start handling it.. | 15:55 |
| milan | sambetts, qq: do you plan reviewing https://review.openstack.org/#/c/396147/ | 15:55 |
| patchbot | patch 396147 - ironic-inspector-specs - Spec for retrieving NUMA node information | 15:55 |
| jroll | JayF: let me put it this way, should an operator be able to disable rescue? | 15:55 |
| milan | ? | 15:55 |
| JayF | jroll: yeah, I think so? | 15:55 |
| rloo | jroll, dtantsur: oh ... <whatever word you want>. We need to think/document clearly what the No* classes should do or not do... | 15:55 |
| jroll | JayF: yeah, then we need to handle a no-rescue thing | 15:56 |
| mariojv | i think we already do | 15:56 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: [Devstack] Download both disk and uec images https://review.openstack.org/420692 | 15:56 |
| dtantsur | rloo, we have a lot of docs to write indeed | 15:56 |
| mariojv | https://review.openstack.org/#/c/350831/26/ironic/conf/default.py default=['no-rescue'] | 15:57 |
| patchbot | patch 350831 - ironic - Add API and conductor methods for [un]rescue | 15:57 |
| rloo | dtantsur, jroll: yes to the docs. and i don't think we should merge the comp stuff w/o the docs or at least a rough draft of docs, so it is clear (and acceptable) to folks, how this is going to work, etc. | 15:57 |
| jroll | JayF: actually, you should be able to just let it bubble up, that seems like a legit error to return from the API :) | 15:57 |
| sambetts | milan: honestly didn't even know it existed, /me is a bit swamped atm | 15:57 |
| mariojv | JayF: https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/noop.py#L47 | 15:57 |
| dtantsur | rloo, we've long decided to not block stuff on docs | 15:58 |
| jroll | rloo: I disagree, I think we should press on, we're too far gone to put the brakes on this now IMO | 15:58 |
| dtantsur | because the only thing that takes longer than writing docs is reviewing docs | 15:58 |
| jroll | I will have docs up this week | 15:58 |
| milan | sambetts, OK cool | 15:58 |
| JayF | mariojv: well, jroll just clarified the part I was confused about -- we don't need to handle that exception, bubbling it up is the right thing | 15:58 |
| rloo | dtantsur: we did? (not to block stuff on docs?) | 15:58 |
| mariojv | cool, didn't read full scrollback | 15:58 |
| dtantsur | rloo, that's why we don't demand docs in the same patch as features | 15:58 |
| rloo | dtantsur: yes, we don't demand docs in the same patch. that is fine. but in some cases (like with rolling upgrades) and other features where things are not clear by just looking at code/spec, i think it is reasonable to ask for doc patches that have that info (regardless of whether they land or not, i think there should be patches that exist) | 16:00 |
| rloo | dtantsur: cuz where else is this information going to be captured BEFORE the code gets merged? | 16:01 |
| *** dtantsur is now known as dtantsur|mtg | 16:02 | |
| jroll | rloo: would a list of specific things that must be documented help alleviate your concerns? | 16:02 |
| rloo | jroll: yes, for me, anyway. | 16:02 |
| rloo | jroll: i guess if you put it in that patch, and we assume no one else is going to ask the same questions w/o seeing that list... | 16:03 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: WIP: docs for dynamic drivers https://review.openstack.org/426819 | 16:03 |
| jroll | rloo: okay, mind adding a comment with things you have in mind there? ^^ | 16:03 |
| rloo | jroll: whoa, making me do more work? :) | 16:04 |
| jroll | rloo: what is unclear to you is different than what is unclear to me, so I need your help here | 16:04 |
| rloo | jroll: ha ha. | 16:04 |
| jroll | rloo: seriously :) | 16:04 |
| jroll | rloo: just anything you think is non-obvious, you don't need to list everything we need to document :) | 16:05 |
| rloo | jroll: yeah, i know. will do. | 16:05 |
| jroll | thank you | 16:05 |
| jroll | I'll rev that API patch with a couple minor fixes after lunch/meeting | 16:05 |
| jroll | bbl | 16:05 |
| jroll | thanks for a good caht btw :) | 16:05 |
| jroll | chat* | 16:05 |
| rloo | jroll: :) | 16:05 |
| *** sdake_ has quit IRC | 16:06 | |
| openstackgerrit | Sam Betts proposed openstack/ironic: Reorganise Cisco UCSM driver code to match its name https://review.openstack.org/426814 | 16:06 |
| openstackgerrit | Sam Betts proposed openstack/ironic: Rename and redocument the Cisco UCSM driver https://review.openstack.org/400673 | 16:06 |
| openstackgerrit | Merged openstack/ironic-inspector: Deprecate introspection_delay_drivers option and make it no-op https://review.openstack.org/425641 | 16:07 |
| openstackgerrit | yolanda.robla proposed openstack/ironic-specs: Ironic deployment steps https://review.openstack.org/412523 | 16:07 |
| *** sdake has joined #openstack-ironic | 16:08 | |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Refactor waiters in our tempest plugin https://review.openstack.org/410684 | 16:13 |
| *** aNuposic has joined #openstack-ironic | 16:18 | |
| *** baoli has quit IRC | 16:19 | |
| *** dsneddon_afk is now known as dsneddon | 16:19 | |
| lucasagomes | folks, random question: Is someone working on the node-get-clean-steps API/CLI ? | 16:20 |
| lucasagomes | I don't see it in our API/CLI yet ?(ref: https://specs.openstack.org/openstack/ironic-specs/specs/5.0/manual-cleaning.html#ironic-node-get-clean-steps) | 16:20 |
| rloo | lucasagomes: oh, that might have been me. sec... | 16:21 |
| rloo | lucasagomes: you mean this? https://review.openstack.org/#/c/278566/ | 16:22 |
| patchbot | patch 278566 - ironic - Add 'GET /nodes/<node>/cleaning/steps' API request | 16:22 |
| lucasagomes | rloo, that's it! Thanks | 16:22 |
| rloo | lucasagomes: that fell between the cracks | 16:22 |
| rloo | lucasagomes: did you need/want it in ocata? | 16:22 |
| rloo | lucasagomes: i can go back to it. | 16:22 |
| lucasagomes | rloo, yeah, just asking cause the QE in red hat is testing the manual cleaning steps and I was looking around how we can discover the cleaning steps via the api | 16:23 |
| lucasagomes | but I couldn't find in the API/CLI | 16:23 |
| lucasagomes | rloo, it's ok, if you want I can take over as well | 16:23 |
| rloo | lucasagomes: oh, i'd rather work on that code than review :) | 16:23 |
| lucasagomes | rloo, do you have time to fix it ? | 16:23 |
| lucasagomes | rloo, fair enough :D | 16:24 |
| rloo | lucasagomes: yeah, see ^^ :) | 16:24 |
| lucasagomes | rloo, and pike should be fine for it as well | 16:24 |
| lucasagomes | thanks rloo ! | 16:24 |
| *** rcernin has quit IRC | 16:24 | |
| rloo | lucasagomes: oh darn. ok then, guess i'll continue looking at driver comp first :) | 16:24 |
| openstackgerrit | Merged openstack/python-ironicclient: Use same variable for --[no-]maintenance https://review.openstack.org/425889 | 16:25 |
| lucasagomes | rloo, cool np! | 16:26 |
| *** baoli has joined #openstack-ironic | 16:36 | |
| *** baoli has quit IRC | 16:37 | |
| *** e0ne has quit IRC | 16:37 | |
| *** rbudden has quit IRC | 16:39 | |
| *** aNuposic has quit IRC | 16:41 | |
| rloo | ding dong, 15 minutes to meeting. please update status reports if you haven't already. the management thanks you in advance. | 16:45 |
| *** mat128_ has joined #openstack-ironic | 16:45 | |
| openstackgerrit | Mario Villaplana proposed openstack/ironic: Agent driver rescue implementation https://review.openstack.org/400437 | 16:46 |
| openstackgerrit | Mario Villaplana proposed openstack/ironic: Add API and conductor methods for [un]rescue https://review.openstack.org/350831 | 16:46 |
| mariojv | just a rebase ^ | 16:46 |
| *** Syed__ has joined #openstack-ironic | 16:46 | |
| *** baoli has joined #openstack-ironic | 16:47 | |
| JayF | rloo: jroll: Should rescue mode still be in status reports, given it's bumped to Pike? I think yes? | 16:47 |
| rloo | JayF: yup | 16:47 |
| rloo | JayF: we'll update what should be a subteam, *after* the ptg/new priorities are decided. | 16:48 |
| rloo | JayF: i guess we were overly optimistic wrt ocata priorities. | 16:48 |
| JayF | rloo: eh, I prefer to think of it as, estimating how long stuff takes can be extremely difficult | 16:48 |
| JayF | mariojv: what's the link to that image of rescue working in devstack? was going to put it in the status update | 16:49 |
| rloo | JayF: that too! :) | 16:49 |
| *** sdake_ has joined #openstack-ironic | 16:49 | |
| openstackgerrit | Chris Krelle proposed openstack/bifrost: DNM: remove pxe_amt driver from pxe enable drivers list https://review.openstack.org/425871 | 16:49 |
| jroll | I like to think of cycle priorities as the priorities in which things get done, not a promise they will get done. there will always be work to do, here's the order in which we want to do it | 16:49 |
| mariojv | JayF: 1 sec | 16:51 |
| mariojv | JayF: pm'ed | 16:51 |
| mariojv | i'm ashamed of my terminal font, lol | 16:51 |
| *** rama_y has joined #openstack-ironic | 16:51 | |
| mariojv | JayF: but go ahead and put it in the update | 16:51 |
| mariojv | JayF: rebase is done btw | 16:52 |
| *** jkilpatr has quit IRC | 16:52 | |
| *** sdake has quit IRC | 16:53 | |
| rloo | jroll: agree, although maybe we shouldn't be too optimistic. am worried it causes folks to feel a bit let down. | 16:53 |
| JayF | mariojv: if you can check that status update make sure it looks sane to you? | 16:53 |
| openstackgerrit | Markos Chandras proposed openstack/bifrost: roles: bifrost-create-vm-nodes: Allow overriding the disk caching method https://review.openstack.org/426844 | 16:53 |
| mariojv | looking | 16:53 |
| *** amarao has quit IRC | 16:54 | |
| jlvillal | vdrok: Did you have any time to look at that grenade + multi-tenant failure that we talked about on Thursday? | 16:54 |
| mariojv | JayF: LGTM. | 16:54 |
| jroll | rloo: dunno, I think we got a ton of work done the last two cycles (and I hear that from other devs too) | 16:54 |
| rloo | jroll: i dunno either. i was just guessing/thinking that. would need to get confirmation about whether it is good or not to list so many things. for me, it was fine. | 16:55 |
| mariojv | i think it's fine as long as the ordering's clear | 16:55 |
| jroll | rloo: well, it seems to be working, let's keep doing it unless someone says it's unhelpful :) | 16:55 |
| *** stendulker has joined #openstack-ironic | 16:55 | |
| *** hshiina has joined #openstack-ironic | 16:56 | |
| rloo | jroll: true. that's what retrospectives are for, so if no feedback... :) | 16:56 |
| vdrok | jlvillal: ugh, I completely forgot about that :( | 16:56 |
| vdrok | will look now | 16:56 |
| vsaienk0 | jlvillal, good morning! could you please add to your review queue https://review.openstack.org/#/c/415881/ I've resolved your comments, thanks! | 16:56 |
| patchbot | patch 415881 - ironic - Validate port info before assume we may use it | 16:56 |
| jroll | rloo: :) | 16:56 |
| *** bnemec has quit IRC | 16:56 | |
| openstackgerrit | Markos Chandras proposed openstack/bifrost: roles: bifrost-create-vm-nodes: Allow overriding the disk caching method https://review.openstack.org/426844 | 16:56 |
| vgadiraj | morning all o/ | 16:58 |
| *** tesseract has quit IRC | 16:58 | |
| jroll | meeting in #openstack-meeting-3 in 1 minute and change | 16:58 |
| *** baoli has quit IRC | 16:59 | |
| *** bnemec has joined #openstack-ironic | 17:01 | |
| *** stendulker_ has joined #openstack-ironic | 17:02 | |
| mariojv | JayF: should we propose a rescue mode ptg session? | 17:02 |
| JayF | mariojv: I don't think it's needed | 17:02 |
| vgadiraj | vdrok, vsaienk0: I am trying to figure out the grenade + multi-tenant failures as well with jlvillal. I am guessing that network connection to the node is lost after grenade.sh completes, and am wondering if this is due to Neutron restarting and the vlan tags changing?http://logs.openstack.org/49/422149/7/experimental/gate-grenade-dsvm-ironic-multitenant-ubuntu-xenial-nv/d416cdb/logs/new/ | 17:03 |
| jlvillal | vgadiraj: There is a meeting going on now. Likely limited replies right now | 17:03 |
| vdrok | vgadiraj: right, might be. I also see these http://logs.openstack.org/49/422149/5/experimental/gate-grenade-dsvm-ironic-multitenant-ubuntu-xenial-nv/74c9ed9/logs/new/screen-n-cpu.txt.gz#_2017-01-22_17_26_43_606 | 17:04 |
| *** stendulker_ has quit IRC | 17:05 | |
| *** fellypefca has joined #openstack-ironic | 17:05 | |
| *** stendulker_ has joined #openstack-ironic | 17:05 | |
| *** fellypefca has quit IRC | 17:05 | |
| *** sdake has joined #openstack-ironic | 17:06 | |
| *** stendulker has quit IRC | 17:06 | |
| *** rama_y has quit IRC | 17:06 | |
| *** mgould|afk is now known as mgould | 17:06 | |
| *** dtantsur|mtg is now known as dtantsur | 17:07 | |
| *** gcb has quit IRC | 17:07 | |
| *** stendulker_ has quit IRC | 17:07 | |
| *** stendulker_ has joined #openstack-ironic | 17:07 | |
| *** jkilpatr has joined #openstack-ironic | 17:08 | |
| *** sdake_ has quit IRC | 17:09 | |
| *** rbudden has joined #openstack-ironic | 17:09 | |
| *** aNuposic has joined #openstack-ironic | 17:12 | |
| vdrok | vgadiraj: also vsaienk0 pointed that networking-generic-switch seems to be empty for some reason http://logs.openstack.org/49/422149/5/experimental/gate-grenade-dsvm-ironic-multitenant-ubuntu-xenial-nv/74c9ed9/logs/new/screen-q-svc.txt.gz#_2017-01-22_17_23_05_849 | 17:13 |
| vdrok | which causes all the errors in q-svc | 17:14 |
| *** mtanino has joined #openstack-ironic | 17:14 | |
| *** baoli has joined #openstack-ironic | 17:15 | |
| vgadiraj | vdrok: gotcha, this is helpful :) | 17:18 |
| vdrok | vgadiraj: whoops, networking-generic-switch config is empty, that's what I meant :) | 17:19 |
| vdrok | np | 17:19 |
| openstackgerrit | Aparna proposed openstack/ironic: Support `update_firmware` to hpsum based firmware update https://review.openstack.org/422572 | 17:27 |
| *** hshiina has quit IRC | 17:30 | |
| rloo | jroll: you know what to do to unblock https://review.openstack.org/#/c/417970/24/ironic/tests/unit/conductor/test_manager.py@3863? | 17:30 |
| patchbot | patch 417970 - ironic - Add dynamic driver functionality to REST API | 17:30 |
| openstackgerrit | Aparna proposed openstack/proliantutils: Add HPSUM firmware update support https://review.openstack.org/422004 | 17:32 |
| jroll | rloo: yep, I think that patch may actually be ready to land, other than there's a signature match here: https://review.openstack.org/#/c/417970/24/ironic/drivers/modules/noop.py | 17:32 |
| patchbot | patch 417970 - ironic - Add dynamic driver functionality to REST API | 17:32 |
| jroll | s/match/mismatch | 17:32 |
| rloo | jroll: ok, if you want that test to fail when calling validate_raid_config? | 17:33 |
| jroll | rloo: yeah, I think that's the right thing to do | 17:33 |
| rloo | jroll: then we should add a comment to that test, i'll comment there. | 17:34 |
| dtantsur | milan, I've posted a few minor reviews against inspector, could you please go over them? https://review.openstack.org/#/q/project:openstack/ironic-inspector+status:open+owner:%22Dmitry+Tantsur%22+NOT+label:Code-Review%253C%253D-1 | 17:34 |
| vdrok | jroll: also, I guess it needs adding no-* to fake hardware? | 17:34 |
| jroll | rloo: the test is already doing the right thing, right? | 17:34 |
| vdrok | jroll: otherwise, it should not be possible to specify unsupported raid interface to node. in both create node and update node it's checked iirc | 17:35 |
| *** rbudden has quit IRC | 17:35 | |
| jroll | vdrok: it's kind of a hack, because I'm skipping the validation. I'll probably just change that test to use manual-management | 17:35 |
| vdrok | or using the hardware type that has no-raid as supported raid interface? | 17:35 |
| vdrok | aha | 17:35 |
| rloo | jroll: sure, it just needs to be clear what the 'right thing' is. I commented. cuz 'not supported' isn't clear. | 17:38 |
| *** rbudden has joined #openstack-ironic | 17:39 | |
| jroll | rloo: I don't see what isn't clear about 'not supported' for that | 17:40 |
| *** amarao has joined #openstack-ironic | 17:41 | |
| rloo | jroll: in classic, not supported == interface = None. | 17:41 |
| *** rbudden has quit IRC | 17:41 | |
| jroll | yep | 17:41 |
| rloo | jroll: if we assume people know that in dynamic, not supported == interface=NoInterface, then I guess it is ok. | 17:41 |
| jroll | rloo: I mean, what would clear it up, a comment or? | 17:42 |
| dtantsur | milan, also updated inspector items in the PTG etherpad, please check if it makes sense to you | 17:42 |
| rloo | jroll: if one just looks at the two tests there, one for classic and the one you wrote for dynamic, it is assymetrical. | 17:42 |
| rloo | jroll: a comment would. that's what i mentioned in the patch. | 17:42 |
| rloo | jroll: or a different test name. i also commented about that :) | 17:42 |
| jroll | rloo: okay, I haven't seen your comment | 17:42 |
| rloo | jroll: no worries, it is hot off the press | 17:42 |
| *** crushil has quit IRC | 17:43 | |
| openstackgerrit | Aparna proposed openstack/ironic: Support `update_firmware` to hpsum based firmware update https://review.openstack.org/422572 | 17:43 |
| dtantsur | vdrok, keep in mind that FakeHardware is hardcoded to be compatible with everything (ugly hacks ftw) | 17:43 |
| vdrok | dtantsur: ahh, right, I recall it now :) | 17:44 |
| openstackgerrit | Aparna proposed openstack/ironic: Support `update_firmware` to hpsum based firmware update https://review.openstack.org/422572 | 17:45 |
| * jroll bbiab | 17:46 | |
| milan | dtantsur, thx, will check | 17:46 |
| *** lucasagomes is now known as lucas-afk | 17:46 | |
| dtantsur | sambetts, https://review.openstack.org/#/c/417871/ has 4x +2 now. What's the final conclusion? :) | 17:47 |
| patchbot | patch 417871 - ironic-python-agent - Remove compatibility with old bash-based introspec... | 17:47 |
| *** stendulker_ has quit IRC | 17:47 | |
| JayF | dtantsur: one of those is sam | 17:48 |
| JayF | dtantsur: so I presume it should be safe to land | 17:48 |
| dtantsur | hopefully, I just dunno why he didn't do +A | 17:48 |
| openstackgerrit | Chris Krelle proposed openstack/bifrost: Remove pxe_amt driver from pxe enable drivers list https://review.openstack.org/425871 | 17:50 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: WIP: Allow to VIF detail fields https://review.openstack.org/421544 | 17:53 |
| *** rama_y has joined #openstack-ironic | 17:55 | |
| *** mgoddard has quit IRC | 17:57 | |
| *** baoli has quit IRC | 17:58 | |
| *** baoli has joined #openstack-ironic | 17:59 | |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: [Devstack] Download both disk and uec images https://review.openstack.org/420692 | 17:59 |
| sambetts | dtantsur: didn't +A to give people a chance to saw my message before we pushed it through | 18:00 |
| sambetts | :) | 18:00 |
| dtantsur | ah, I see :) makes sense | 18:01 |
| dtantsur | will call it a day now, see you | 18:02 |
| *** harlowja has joined #openstack-ironic | 18:02 | |
| *** ralonsoh has quit IRC | 18:02 | |
| openstackgerrit | Oksana Voshchana proposed openstack/ironic: WIP: TEST GRENADE DEV https://review.openstack.org/397126 | 18:02 |
| sambetts | o/ dtantsur | 18:03 |
| *** dtantsur is now known as dtantsur|afk | 18:03 | |
| *** derekh has quit IRC | 18:03 | |
| *** dyasny has joined #openstack-ironic | 18:04 | |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: [Devstack] Download both disk and uec images https://review.openstack.org/420692 | 18:04 |
| *** absubram has joined #openstack-ironic | 18:04 | |
| *** mgould is now known as mgould|afk | 18:07 | |
| *** e0ne has joined #openstack-ironic | 18:12 | |
| vdrok | good night dtantsur|afk | 18:15 |
| *** athomas has quit IRC | 18:15 | |
| *** ijw has joined #openstack-ironic | 18:17 | |
| *** alexpilotti has joined #openstack-ironic | 18:18 | |
| *** ijw has quit IRC | 18:19 | |
| *** ijw has joined #openstack-ironic | 18:19 | |
| *** alexpilotti has quit IRC | 18:22 | |
| *** crushil has joined #openstack-ironic | 18:24 | |
| *** Nisha_Agarwal has quit IRC | 18:26 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Allow using resource classes https://review.openstack.org/407651 | 18:28 |
| *** crushil has quit IRC | 18:30 | |
| *** crushil has joined #openstack-ironic | 18:31 | |
| *** amarao has quit IRC | 18:33 | |
| *** jkilpatr has quit IRC | 18:33 | |
| *** Sukhdev has joined #openstack-ironic | 18:36 | |
| *** milan has quit IRC | 18:38 | |
| *** crushil has quit IRC | 18:38 | |
| *** sayantani01 has quit IRC | 18:42 | |
| *** mgoddard has joined #openstack-ironic | 18:44 | |
| *** jkilpatr has joined #openstack-ironic | 18:45 | |
| *** foexle has quit IRC | 18:46 | |
| *** rama_y has quit IRC | 18:46 | |
| *** e0ne has quit IRC | 18:51 | |
| *** e0ne has joined #openstack-ironic | 18:55 | |
| *** sayantani01 has joined #openstack-ironic | 19:04 | |
| openstackgerrit | Chris Krelle proposed openstack/bifrost: Remove pxe_amt driver from pxe enable drivers list https://review.openstack.org/425871 | 19:17 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Treat unknown fields as subcontrollers https://review.openstack.org/426257 | 19:21 |
| vdrok | good night everyone! | 19:23 |
| rloo | bye vdrok! | 19:24 |
| vgadiraj | good night vdrok! | 19:28 |
| *** jaosorior is now known as jaosorior_lunch | 19:28 | |
| *** sdake has quit IRC | 19:32 | |
| *** rpioso has quit IRC | 19:33 | |
| *** mgoddard has quit IRC | 19:34 | |
| *** jtomasek_ has joined #openstack-ironic | 19:34 | |
| *** rbudden has joined #openstack-ironic | 19:42 | |
| *** sdake has joined #openstack-ironic | 19:48 | |
| *** jheroux has quit IRC | 20:00 | |
| *** dyasny has quit IRC | 20:02 | |
| *** ChubYann has joined #openstack-ironic | 20:06 | |
| *** crushil has joined #openstack-ironic | 20:07 | |
| mrda | Morning Ironic | 20:10 |
| *** e0ne has quit IRC | 20:12 | |
| openstackgerrit | Ramamani Yeleswarapu proposed openstack/ironic: Enable OSProfiler support in Ironic https://review.openstack.org/358142 | 20:13 |
| *** vsaienko has joined #openstack-ironic | 20:14 | |
| *** ijw has quit IRC | 20:16 | |
| *** rloo_ has joined #openstack-ironic | 20:16 | |
| *** e0ne has joined #openstack-ironic | 20:16 | |
| *** ijw has joined #openstack-ironic | 20:16 | |
| *** dyasny has joined #openstack-ironic | 20:18 | |
| *** ElCoyote_ has joined #openstack-ironic | 20:18 | |
| *** rloo has quit IRC | 20:18 | |
| vsaienko | jroll with concurrency 4 it took near 30 min to run deploy tests for pxe|agent ipmitool + partitioned|wholedisk + manual cleaning for agent|pxe http://logs.openstack.org/65/426665/4/experimental/gate-ironic-dsvm-standalone-ubuntu-xenial-nv/00e220b/console.html#_2017-01-30_19_52_22_041681 | 20:19 |
| *** alexpilotti has joined #openstack-ironic | 20:21 | |
| mariojv | JayF: here's a bug for the logging thing we discussed the other day https://bugs.launchpad.net/ironic/+bug/1660432 | 20:21 |
| openstack | Launchpad bug 1660432 in Ironic "Oslo logging doesn't work during IPA functional tests" [Undecided,New] - Assigned to Mario Villaplana (mario-villaplana-j) | 20:21 |
| *** jtomasek_ has quit IRC | 20:22 | |
| jroll | vsaienko: that sounds like a good thing, yes? :) | 20:23 |
| *** rloo_ has quit IRC | 20:23 | |
| jroll | morning mrda | 20:23 |
| *** baoli has quit IRC | 20:25 | |
| *** sambetts is now known as sambetts|afk | 20:25 | |
| *** rpioso has joined #openstack-ironic | 20:25 | |
| *** alexpilotti has quit IRC | 20:26 | |
| * sambetts|afk must have been up late a wild mrda has appeared! | 20:26 | |
| *** baoli has joined #openstack-ironic | 20:26 | |
| mrda | lol | 20:26 |
| mrda | sambetts|afk: or I'm up earlier than usual :) | 20:26 |
| *** rloo has joined #openstack-ironic | 20:26 | |
| *** baoli has quit IRC | 20:27 | |
| vsaienko | jroll: yup, looks like we are pretty close in decreasing number of CI jobs | 20:28 |
| JayF | vsaienko: I was actually going to ask about one of those related patches | 20:29 |
| *** dyasny has quit IRC | 20:29 | |
| JayF | vsaienko: https://review.openstack.org/#/c/420692/8 seems like something we shouldn't land yet, imo, as just downloading the image doesn't get it put in glance, right? | 20:29 |
| patchbot | patch 420692 - ironic - [Devstack] Download both disk and uec images | 20:29 |
| JayF | vsaienko: so what's the value of that code without any further patches? | 20:29 |
| jroll | vsaienko: awesome, nice work :) | 20:30 |
| *** dsneddon is now known as dsneddon_afk | 20:30 | |
| *** dyasny has joined #openstack-ironic | 20:32 | |
| *** sayantani01 has quit IRC | 20:32 | |
| JayF | we're OK with changes like https://review.openstack.org/#/c/316149 that are breaking in the API (because we created APIs we didn't intend)? | 20:34 |
| patchbot | patch 316149 - ironic - Fix invalid chassis/nodes/chassis_uuid API url | 20:34 |
| JayF | It is a breaking change for someone (mis)using our API, which is why it made me hrm it wasn't behind a versino change | 20:34 |
| vsaienko | JayF previously when download logic was inside devstack we download both partitioned and wholedisk images, but after we moved it into ironic https://github.com/openstack-dev/devstack/commit/d89b175321ac293454ad15caaee13c0ae46b0bd6 for partitioned jobs we start downloading only partitioned image | 20:36 |
| *** baoli has joined #openstack-ironic | 20:36 | |
| vsaienko | JayF as partitioned image is default in devstack https://github.com/openstack-dev/devstack/blob/d89b175321ac293454ad15caaee13c0ae46b0bd6/stackrc#L655 | 20:36 |
| JayF | vsaienko: OH! | 20:36 |
| JayF | vsaienko: gotcha, I think I get it now | 20:36 |
| JayF | vsaienko: the second image still doesn't /do/ anything | 20:37 |
| JayF | vsaienko: but not downloading it was a breaking change | 20:37 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Improve conductor driver validation at startup https://review.openstack.org/426213 | 20:37 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Add dynamic interfaces fields to nodes API https://review.openstack.org/424720 | 20:37 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Add dynamic driver functionality to REST API https://review.openstack.org/417970 | 20:37 |
| vsaienko | right, while we didn't use it at the CI, we still had 2 images | 20:37 |
| vsaienko | JayF yes https://review.openstack.org/#/c/420692 restore original behavior | 20:37 |
| patchbot | patch 420692 - ironic - [Devstack] Download both disk and uec images | 20:37 |
| jroll | rloo: okay, let's see if we've got it this time | 20:38 |
| jroll | still more to come, ofc, but wanted to get folks started on this one | 20:38 |
| JayF | vsaienko: awesome, just landed it | 20:38 |
| rloo | thx jroll. we'll see :) | 20:39 |
| vsaienko | thanks JayF! Once you have a time please add to your review queue https://review.openstack.org/#/c/423556 | 20:39 |
| patchbot | patch 423556 - ironic - Add Ironic standalone tests | 20:39 |
| * JayF sees ironic_tempest_plugin and groans | 20:40 | |
| JayF | but you got it open in my tab so you win anyway :P | 20:40 |
| vsaienko | :) | 20:43 |
| JayF | vsaienko: -1 with comment. mainly there's nothing, commit, bug, anything, describing a higher level goal that's being accomplished | 20:44 |
| JayF | vsaienko: so it's impossible for me to review to see if it achieves the goal if the goal isn't written out ;) | 20:44 |
| *** rbudden has quit IRC | 20:57 | |
| *** sayantani01 has joined #openstack-ironic | 20:57 | |
| vsaienko | JayF the goal is to decrease a number of CI jobs, combine different agents/partiton schemes boot options etc... The approach we using at the moment 1 job per agent is not scalable. I will check if the bug was opened or create a new one if not. | 20:58 |
| JayF | vsaienko: thanks, that's why I assumed, but folks looking at git log don't have the benefit of temporal context :) | 20:58 |
| JayF | *what | 20:58 |
| *** vsaienko has quit IRC | 21:05 | |
| *** sdake has quit IRC | 21:06 | |
| *** dsneddon has joined #openstack-ironic | 21:06 | |
| *** mat128_ has quit IRC | 21:06 | |
| *** sdake has joined #openstack-ironic | 21:07 | |
| *** vsaienko has joined #openstack-ironic | 21:10 | |
| *** e0ne has quit IRC | 21:10 | |
| *** crushil has quit IRC | 21:12 | |
| rloo | JayF: wrt your question about 316149, good question, i don't know the answer, but commented. i think f we can't decide, it is worth discussing at the monday meeting. | 21:12 |
| JayF | rloo: I don't really have an opinion one way or the other, just sorta seems like something that should be decided instead of just quietly done | 21:13 |
| rloo | JayF: ++ | 21:13 |
| *** jtomasek_ has joined #openstack-ironic | 21:20 | |
| *** jaosorior_lunch is now known as jaosorior | 21:21 | |
| *** crushil has joined #openstack-ironic | 21:21 | |
| *** Goneri has quit IRC | 21:28 | |
| *** sayantani01 has quit IRC | 21:29 | |
| *** rbudden has joined #openstack-ironic | 21:32 | |
| *** sayantani01 has joined #openstack-ironic | 21:39 | |
| * mat128 out | 21:41 | |
| *** baoli has quit IRC | 21:46 | |
| *** jtomasek_ has quit IRC | 21:47 | |
| *** sayantani01 has quit IRC | 21:51 | |
| *** ccamacho has quit IRC | 21:52 | |
| *** vsaienko has quit IRC | 21:53 | |
| *** sayantani01 has joined #openstack-ironic | 21:57 | |
| *** baoli has joined #openstack-ironic | 21:58 | |
| openstackgerrit | Merged openstack/ironic-python-agent: Remove compatibility with old bash-based introspection ramdisk https://review.openstack.org/417871 | 21:58 |
| *** trown is now known as trown|outtypewww | 22:02 | |
| *** foexle has joined #openstack-ironic | 22:09 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Improve conductor driver validation at startup https://review.openstack.org/426213 | 22:13 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Add dynamic interfaces fields to nodes API https://review.openstack.org/424720 | 22:13 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Add dynamic driver functionality to REST API https://review.openstack.org/417970 | 22:13 |
| * jroll probably out for the night, see y'all | 22:13 | |
| *** crushil has quit IRC | 22:20 | |
| *** dprince has quit IRC | 22:20 | |
| *** jcoufal has quit IRC | 22:21 | |
| *** crushil has joined #openstack-ironic | 22:24 | |
| mariojv | night jroll | 22:27 |
| *** foexle has quit IRC | 22:27 | |
| *** mgoddard has joined #openstack-ironic | 22:27 | |
| openstackgerrit | Merged openstack/ironic: Devstack: Create a "no ansi" logfile for the baremetal console logs https://review.openstack.org/425457 | 22:31 |
| *** crushil has quit IRC | 22:31 | |
| *** srobert_ has joined #openstack-ironic | 22:33 | |
| openstackgerrit | Merged openstack/ironic: [Devstack] Download both disk and uec images https://review.openstack.org/420692 | 22:35 |
| *** rama_y has joined #openstack-ironic | 22:35 | |
| *** srobert has quit IRC | 22:36 | |
| *** srobert_ has quit IRC | 22:37 | |
| *** baoli has quit IRC | 22:40 | |
| *** dsneddon has quit IRC | 22:43 | |
| openstackgerrit | Annie Lezil proposed openstack/ironic-python-agent: [WIP] Collect NIC name given by BIOS https://review.openstack.org/396406 | 22:47 |
| openstackgerrit | Annie Lezil proposed openstack/ironic-python-agent: [WIP] Collect NIC name given by BIOS https://review.openstack.org/396406 | 22:53 |
| *** jkilpatr has quit IRC | 22:54 | |
| openstackgerrit | John L. Villalovos proposed openstack/ironic: WIP/DNM: Testing multinode job https://review.openstack.org/422149 | 22:55 |
| *** mgoddard has quit IRC | 22:59 | |
| *** bfournie has quit IRC | 23:04 | |
| openstackgerrit | John L. Villalovos proposed openstack/ironic: WIP/DNM: Testing serial console logs https://review.openstack.org/426965 | 23:09 |
| *** sdake has quit IRC | 23:09 | |
| *** jrist has quit IRC | 23:13 | |
| *** sayantani01 has quit IRC | 23:27 | |
| *** jrist has joined #openstack-ironic | 23:29 | |
| *** rama_y_ has joined #openstack-ironic | 23:30 | |
| *** ElCoyote_ has left #openstack-ironic | 23:37 | |
| *** ElCoyote_ has joined #openstack-ironic | 23:37 | |
| *** bcornec has joined #openstack-ironic | 23:37 | |
| *** pmannidi has joined #openstack-ironic | 23:39 | |
| *** wrouesnel has joined #openstack-ironic | 23:47 | |
| *** sdake has joined #openstack-ironic | 23:50 | |
| *** baoli has joined #openstack-ironic | 23:55 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!