| *** ayoung has joined #openstack-ironic | 00:03 | |
| *** piet has quit IRC | 00:12 | |
| openstackgerrit | Stephanie Miller proposed openstack/bifrost: Unify testing scripts https://review.openstack.org/297424 | 00:12 |
|---|---|---|
| *** Sukhdev has joined #openstack-ironic | 00:15 | |
| *** sambetts|afk has quit IRC | 00:22 | |
| *** rloo has quit IRC | 00:23 | |
| *** alex_xu has quit IRC | 00:26 | |
| *** hoangcx has joined #openstack-ironic | 00:26 | |
| *** alex_xu has joined #openstack-ironic | 00:27 | |
| *** yarkot1 has joined #openstack-ironic | 00:29 | |
| *** yarkot1 has quit IRC | 00:30 | |
| *** sambetts has joined #openstack-ironic | 00:31 | |
| *** akrish9 has joined #openstack-ironic | 00:31 | |
| *** yarkot1 has joined #openstack-ironic | 00:34 | |
| *** akrish9 has quit IRC | 00:44 | |
| *** yarkot1 has quit IRC | 00:44 | |
| *** akrish9 has joined #openstack-ironic | 00:45 | |
| openstackgerrit | Gonéri Le Bouder proposed openstack/ironic: ipxe: retry on failure https://review.openstack.org/306196 | 00:59 |
| *** ptx0 has joined #openstack-ironic | 01:00 | |
| *** ptx0 has joined #openstack-ironic | 01:00 | |
| *** akrish9 has quit IRC | 01:02 | |
| *** akrish9 has joined #openstack-ironic | 01:02 | |
| *** akrish9 has quit IRC | 01:04 | |
| *** akrish9 has joined #openstack-ironic | 01:06 | |
| *** hoangcx_ has joined #openstack-ironic | 01:08 | |
| *** baoli has joined #openstack-ironic | 01:15 | |
| *** amotoki_ has quit IRC | 01:16 | |
| *** Guest17323 has quit IRC | 01:20 | |
| *** Sukhdev has quit IRC | 01:25 | |
| *** yarkot1 has joined #openstack-ironic | 01:33 | |
| *** baoli has quit IRC | 01:35 | |
| *** piet has joined #openstack-ironic | 01:35 | |
| *** baoli has joined #openstack-ironic | 01:35 | |
| *** mtanino has quit IRC | 01:36 | |
| *** suro-patz has joined #openstack-ironic | 01:47 | |
| *** Marga__ has joined #openstack-ironic | 01:48 | |
| *** baoli has quit IRC | 01:48 | |
| *** baoli has joined #openstack-ironic | 01:48 | |
| *** suro-patz1 has joined #openstack-ironic | 01:50 | |
| *** Marga_ has quit IRC | 01:51 | |
| *** suro-patz has quit IRC | 01:51 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/bifrost: Updated from global requirements https://review.openstack.org/305167 | 01:52 |
| *** Marga__ has quit IRC | 01:52 | |
| openstackgerrit | Tan Lin proposed openstack/ironic: Typo https://review.openstack.org/306209 | 01:54 |
| *** Haomeng has joined #openstack-ironic | 02:00 | |
| *** cameron_C has joined #openstack-ironic | 02:01 | |
| *** alex_xu has quit IRC | 02:02 | |
| *** baoli has quit IRC | 02:07 | |
| *** suro-patz1 has quit IRC | 02:08 | |
| *** alex_xu has joined #openstack-ironic | 02:09 | |
| *** suro-patz has joined #openstack-ironic | 02:09 | |
| *** chlong has joined #openstack-ironic | 02:20 | |
| *** amotoki has joined #openstack-ironic | 02:25 | |
| *** kwazar has quit IRC | 02:26 | |
| *** thrash|g0ne has quit IRC | 02:26 | |
| *** kwazar has joined #openstack-ironic | 02:26 | |
| *** Haomeng|2 has joined #openstack-ironic | 02:26 | |
| *** Haomeng has quit IRC | 02:26 | |
| *** suro-patz has quit IRC | 02:31 | |
| *** hoangcx has quit IRC | 02:42 | |
| *** yuanying has quit IRC | 02:48 | |
| *** e0ne has joined #openstack-ironic | 02:51 | |
| *** Haomeng|2 has quit IRC | 02:52 | |
| *** Haomeng|2 has joined #openstack-ironic | 02:53 | |
| *** amotoki has quit IRC | 02:55 | |
| *** e0ne has quit IRC | 02:56 | |
| *** amotoki has joined #openstack-ironic | 02:59 | |
| *** e0ne has joined #openstack-ironic | 03:00 | |
| *** hoangcx_ has quit IRC | 03:04 | |
| *** piet has quit IRC | 03:07 | |
| *** e0ne has quit IRC | 03:08 | |
| *** baoli has joined #openstack-ironic | 03:10 | |
| *** e0ne has joined #openstack-ironic | 03:11 | |
| *** e0ne has quit IRC | 03:13 | |
| *** e0ne has joined #openstack-ironic | 03:17 | |
| *** harlowja_at_home has joined #openstack-ironic | 03:18 | |
| *** thrash has joined #openstack-ironic | 03:22 | |
| *** thrash has quit IRC | 03:22 | |
| *** thrash has joined #openstack-ironic | 03:22 | |
| *** Haomeng|2 has quit IRC | 03:22 | |
| *** ijw_ has joined #openstack-ironic | 03:23 | |
| *** pleia2_ has joined #openstack-ironic | 03:27 | |
| *** izaakk_ has joined #openstack-ironic | 03:27 | |
| *** Mr_T has quit IRC | 03:28 | |
| *** akrish9 has quit IRC | 03:28 | |
| *** piet has joined #openstack-ironic | 03:29 | |
| *** akrish9 has joined #openstack-ironic | 03:29 | |
| *** harlowja_at_home has quit IRC | 03:30 | |
| *** sambetts_ has joined #openstack-ironic | 03:31 | |
| *** baoli has quit IRC | 03:32 | |
| *** sambetts has quit IRC | 03:32 | |
| *** ijw has quit IRC | 03:32 | |
| *** xavierr has quit IRC | 03:32 | |
| *** dhellmann has quit IRC | 03:32 | |
| *** izaakk has quit IRC | 03:32 | |
| *** pleia2 has quit IRC | 03:32 | |
| *** izaakk_ is now known as izaakk | 03:32 | |
| *** Sukhdev has joined #openstack-ironic | 03:33 | |
| *** xavierr has joined #openstack-ironic | 03:38 | |
| *** Mr_T has joined #openstack-ironic | 03:45 | |
| *** pleia2_ is now known as pleia2 | 03:46 | |
| *** Haomeng has joined #openstack-ironic | 03:46 | |
| *** yuanying has joined #openstack-ironic | 03:47 | |
| *** e0ne has quit IRC | 04:00 | |
| *** vmud213 has joined #openstack-ironic | 04:01 | |
| *** e0ne has joined #openstack-ironic | 04:04 | |
| *** dhellmann has joined #openstack-ironic | 04:05 | |
| openstackgerrit | Debayan Ray proposed openstack/proliantutils: Add proper error message for extraction failure https://review.openstack.org/305331 | 04:13 |
| *** vishwanathj_away has quit IRC | 04:15 | |
| *** Marga_ has joined #openstack-ironic | 04:16 | |
| *** Marga_ has quit IRC | 04:17 | |
| *** Marga_ has joined #openstack-ironic | 04:18 | |
| *** e0ne has quit IRC | 04:22 | |
| *** e0ne has joined #openstack-ironic | 04:26 | |
| *** ijw_ has quit IRC | 04:29 | |
| *** e0ne has quit IRC | 04:30 | |
| *** piet has quit IRC | 04:38 | |
| *** piet has joined #openstack-ironic | 04:39 | |
| *** chlong has quit IRC | 04:48 | |
| *** harlowja_at_home has joined #openstack-ironic | 04:50 | |
| *** vinaym213 has joined #openstack-ironic | 04:53 | |
| *** vishwanathj has joined #openstack-ironic | 04:57 | |
| *** piet has quit IRC | 04:57 | |
| *** vmud213 has quit IRC | 04:57 | |
| *** vishwanathj is now known as vishwanathj_zzz | 04:57 | |
| *** harlowja_at_home has quit IRC | 05:00 | |
| *** chlong has joined #openstack-ironic | 05:01 | |
| *** alex_xu_ has joined #openstack-ironic | 05:10 | |
| *** rcernin has joined #openstack-ironic | 05:11 | |
| *** alex_xu has quit IRC | 05:12 | |
| *** piet has joined #openstack-ironic | 05:14 | |
| *** causten_ has joined #openstack-ironic | 05:24 | |
| *** smoriya has quit IRC | 05:26 | |
| *** amotoki_ has joined #openstack-ironic | 05:26 | |
| *** smoriya_afk has joined #openstack-ironic | 05:26 | |
| *** mat128 has quit IRC | 05:26 | |
| *** mkoderer__ has quit IRC | 05:26 | |
| *** smoriya_afk is now known as smoriya | 05:26 | |
| *** amotoki has quit IRC | 05:27 | |
| *** xavierr has quit IRC | 05:28 | |
| *** ChrisAusten has quit IRC | 05:28 | |
| *** jlvillal has quit IRC | 05:28 | |
| *** dkehn has quit IRC | 05:28 | |
| *** bcornec has quit IRC | 05:28 | |
| *** bcornec has joined #openstack-ironic | 05:29 | |
| *** xavierr has joined #openstack-ironic | 05:30 | |
| *** Marga_ has quit IRC | 05:33 | |
| *** daemontool has joined #openstack-ironic | 05:33 | |
| *** jlvillal has joined #openstack-ironic | 05:34 | |
| *** mkoderer__ has joined #openstack-ironic | 05:35 | |
| *** piet has quit IRC | 05:35 | |
| *** mat128 has joined #openstack-ironic | 05:36 | |
| *** Marga_ has joined #openstack-ironic | 05:38 | |
| *** davidlenwell has quit IRC | 05:44 | |
| *** davidlenwell has joined #openstack-ironic | 05:46 | |
| *** dkehn has joined #openstack-ironic | 05:53 | |
| openstackgerrit | Debayan Ray proposed openstack/ironic: Add note on prerequisite of 'rpm' file extraction https://review.openstack.org/306247 | 05:53 |
| *** hoangcx has joined #openstack-ironic | 05:58 | |
| *** Sukhdev has quit IRC | 05:59 | |
| *** chlong has quit IRC | 06:02 | |
| *** amotoki_ has quit IRC | 06:07 | |
| *** Fdaisuke has quit IRC | 06:19 | |
| *** zhenguo_ has quit IRC | 06:20 | |
| *** tesseract has joined #openstack-ironic | 06:29 | |
| *** tesseract is now known as Guest63229 | 06:29 | |
| *** divya has joined #openstack-ironic | 06:43 | |
| *** Marga__ has joined #openstack-ironic | 06:55 | |
| *** irf has joined #openstack-ironic | 06:56 | |
| irf | Hello folks .. i am back | 06:56 |
| irf | i have verified my hardware and it has Intel core I5 CPU | 06:57 |
| irf | In the boot order , i have UEFI option available | 06:57 |
| *** jlvillal has quit IRC | 06:58 | |
| irf | motherboard does not have IPMI controller , i have confirmed that so cant use IPMI driver | 06:58 |
| irf | can someone suggest , which driver i can use ? | 06:59 |
| irf | i am reffering to this doc : http://docs.openstack.org/developer/ironic/deploy/drivers.html | 06:59 |
| *** davidlenwell_ has joined #openstack-ironic | 07:01 | |
| *** mkoderer__ has quit IRC | 07:03 | |
| irf | plz suggest | 07:03 |
| *** ohamada has joined #openstack-ironic | 07:03 | |
| *** davidlenwell has quit IRC | 07:04 | |
| *** Marga_ has quit IRC | 07:04 | |
| *** mat128 has quit IRC | 07:05 | |
| *** mkoderer__ has joined #openstack-ironic | 07:09 | |
| *** jlvillal has joined #openstack-ironic | 07:11 | |
| *** mat128 has joined #openstack-ironic | 07:12 | |
| Haomeng | irf: depends on how can we control your server poweron/off, is there api interface to be called, or are you use PDU to control your server's power? | 07:13 |
| Haomeng | irf: is your machine pc or server? | 07:14 |
| Haomeng | irf: almost x86 server has ipmi card | 07:14 |
| *** mkovacik has joined #openstack-ironic | 07:15 | |
| irf | i have checked that .. | 07:15 |
| irf | on my motherboard, i dont have IPMI controller | 07:15 |
| irf | mine is just desktop PC | 07:16 |
| Haomeng | irf: if it is pc, you can check if it supports AMT | 07:16 |
| irf | ok | 07:17 |
| *** ifarkas has joined #openstack-ironic | 07:17 | |
| irf | let me read this : http://docs.openstack.org/developer/ironic/drivers/amt.html | 07:17 |
| Haomeng | irf: and PC supports wol - http://docs.openstack.org/developer/ironic/drivers/wol.html | 07:19 |
| *** achanda has joined #openstack-ironic | 07:20 | |
| *** athomas has joined #openstack-ironic | 07:22 | |
| *** divya has quit IRC | 07:22 | |
| *** amotoki has joined #openstack-ironic | 07:25 | |
| *** mgould has joined #openstack-ironic | 07:25 | |
| *** athomas has quit IRC | 07:27 | |
| mgould | morning Ironic! | 07:28 |
| *** pcaruana has joined #openstack-ironic | 07:31 | |
| *** athomas has joined #openstack-ironic | 07:31 | |
| *** karimb has joined #openstack-ironic | 07:47 | |
| *** e0ne has joined #openstack-ironic | 07:58 | |
| pas-ha | morning mgould and all | 07:58 |
| *** jtomasek_ has joined #openstack-ironic | 08:00 | |
| *** jcoufal has joined #openstack-ironic | 08:00 | |
| *** achanda has quit IRC | 08:06 | |
| *** yuriyz has quit IRC | 08:06 | |
| *** openstackgerrit has quit IRC | 08:08 | |
| *** yuikotakada has joined #openstack-ironic | 08:08 | |
| *** yuriyz has joined #openstack-ironic | 08:10 | |
| *** stendulker has joined #openstack-ironic | 08:10 | |
| *** mbound has joined #openstack-ironic | 08:11 | |
| *** b_bezak has joined #openstack-ironic | 08:15 | |
| *** alex_xu_ has quit IRC | 08:16 | |
| *** kromanenko has joined #openstack-ironic | 08:18 | |
| *** b_bezak has left #openstack-ironic | 08:18 | |
| *** openstackgerrit has joined #openstack-ironic | 08:18 | |
| *** alex_xu has joined #openstack-ironic | 08:23 | |
| openstackgerrit | Sofiia Andriichenko proposed openstack/python-ironicclient: Add test for 'port-show --field' command https://review.openstack.org/304008 | 08:28 |
| *** jistr has joined #openstack-ironic | 08:30 | |
| *** e0ne has quit IRC | 08:43 | |
| *** ppiela has quit IRC | 08:43 | |
| *** ppiela has joined #openstack-ironic | 08:44 | |
| *** dims_ has quit IRC | 08:46 | |
| *** dims has joined #openstack-ironic | 08:49 | |
| *** e0ne has joined #openstack-ironic | 08:53 | |
| *** dtantsur|afk is now known as dtantsur | 08:58 | |
| dtantsur | Morning Ironic | 08:58 |
| -openstackstatus- NOTICE: Gerrit is going to be restarted because is not processing new changes | 09:05 | |
| dtantsur | sambetts_, ifarkas, aarefiev, hi folks, please review https://review.openstack.org/305960 as soon as gerrit is back up | 09:06 |
| dtantsur | it's a pretty bad bug, I want to have it fixed as soon as possible | 09:06 |
| *** alex_xu has quit IRC | 09:07 | |
| *** alex_xu has joined #openstack-ironic | 09:09 | |
| sambetts_ | Morning all | 09:10 |
| *** sambetts_ is now known as sambetts | 09:11 | |
| vdrok | morning ironic, Haomeng dtantsur and sambetts ! | 09:17 |
| dtantsur | o/ | 09:17 |
| *** amotoki has quit IRC | 09:18 | |
| *** yuikotakada has quit IRC | 09:19 | |
| -openstackstatus- NOTICE: No jobs are being processed by gerrit and zuul . We are working to solve the problem, please be aware that no changes have been sent to the queue in the last hour, so you will need to recheck jobs for that period. | 09:20 | |
| *** ChanServ changes topic to "No jobs are being processed by gerrit and zuul . We are working to solve the problem, please be aware that no changes have been sent to the queue in the last hour, so you will need to recheck jobs for that period." | 09:20 | |
| sambetts | Hi dtantsur, vdrok, nice to see Gerrit and Zuul continuing with their Friday tradition | 09:22 |
| dtantsur | wonderful..... | 09:23 |
| *** alex_xu_ has joined #openstack-ironic | 09:25 | |
| *** alex_xu has quit IRC | 09:26 | |
| *** divya has joined #openstack-ironic | 09:29 | |
| divya | ironic node-set-provision-state --config-drive /tmp/$NAME $NAME active | 09:30 |
| divya | command to deploy image using ipmi agent? | 09:31 |
| *** manfredp has joined #openstack-ironic | 09:32 | |
| *** manfredp_ has joined #openstack-ironic | 09:33 | |
| *** manfredp has quit IRC | 09:36 | |
| *** ndipanov has joined #openstack-ironic | 09:40 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-specs: [WIP] Driver composition reform spec https://review.openstack.org/188370 | 09:42 |
| irf | ok thanks | 09:47 |
| *** alex_xu_ has quit IRC | 09:47 | |
| *** ChanServ changes topic to "Bare Metal Provisioning | Status: http://bit.ly/ironic-whiteboard | Docs: http://docs.openstack.org/developer/ironic/ | Bugs: https://bugs.launchpad.net/ironic" | 09:48 | |
| -openstackstatus- NOTICE: Zuul and gerrit are working normally now. Please recheck any jobs that may have been affected by this failure. | 09:48 | |
| *** alex_xu has joined #openstack-ironic | 09:49 | |
| stendulker | dtantsur: Hi, have a minute... wanted to discuss your comment for spec https://review.openstack.org/#/c/230274/7/specs/approved/secure-baremetal-deploy-in-uefi.rst | 09:50 |
| *** amotoki has joined #openstack-ironic | 09:55 | |
| openstackgerrit | vinay kumar muddu proposed openstack/ironic-python-agent: Wait for all interfaces to be up before node lookup https://review.openstack.org/303503 | 09:56 |
| dtantsur | stendulker, yep. even though it's better to leave a reason on the spec itself | 09:58 |
| stendulker | dtantsur: Basically I did not understand your comment :( | 09:58 |
| dtantsur | stendulker, my comment is "why" | 09:59 |
| dtantsur | you introduce an option, which does not seem required to me. you don't explain why you do it. | 09:59 |
| stendulker | dtantsur: I do not find any other option to enable this feature... | 10:00 |
| divya | 2016-04-15 15:28:14.297 ^[[01;31mERROR ironic.conductor.manager [^[[01;36mreq-1c001fa8-4a52-4a30-8a2e-ff86cafdef49 ^[[00;36mNone None^[[01;31m] ^[[01;35m^[[01;31mError while preparing to deploy to node e2fb844c-1af6-47ae-9ab7-5b651262afd9: No VIFs found for node e2fb844c-1af6-47ae-9ab7-5b651262afd9 when attempting to update DHCP BOOT options. | 10:00 |
| divya | Getting this error while deploying bare metal node | 10:00 |
| dtantsur | stendulker, why do you need to *disable* it? | 10:00 |
| dtantsur | also sorry, running for lunch, see you in a few | 10:00 |
| *** dtantsur is now known as dtantsur|lunch | 10:00 | |
| stendulker | dtantsur: you mean all the deploys should use secure boot? | 10:01 |
| stendulker | dtantsur|lunch: Will discuss this later... | 10:01 |
| *** nathan-h has joined #openstack-ironic | 10:04 | |
| *** amotoki has quit IRC | 10:05 | |
| *** yolanda has quit IRC | 10:12 | |
| *** alex_xu has quit IRC | 10:14 | |
| *** alex_xu has joined #openstack-ironic | 10:16 | |
| *** ndipanov has quit IRC | 10:22 | |
| *** ndipanov has joined #openstack-ironic | 10:24 | |
| *** daemontool has quit IRC | 10:28 | |
| *** divya_ has joined #openstack-ironic | 10:32 | |
| divya_ | hi Ironics | 10:38 |
| *** chopmann has joined #openstack-ironic | 10:40 | |
| *** chopmann has quit IRC | 10:41 | |
| openstackgerrit | Tan Lin proposed openstack/ironic: [PoC]Ironic live upgrade framework - Make service compatible https://review.openstack.org/306357 | 10:45 |
| openstackgerrit | Tan Lin proposed openstack/ironic: [PoC] Refactor configdrive into a new field https://review.openstack.org/306358 | 10:45 |
| *** vinaym213 has quit IRC | 10:52 | |
| *** jtomasek_ has quit IRC | 10:56 | |
| *** stendulker has quit IRC | 11:04 | |
| *** daemontool has joined #openstack-ironic | 11:06 | |
| *** dtantsur|lunch is now known as dtantsur | 11:09 | |
| *** irf has quit IRC | 11:10 | |
| *** irf has joined #openstack-ironic | 11:10 | |
| irf | hi | 11:11 |
| irf | do i need to use both the drivers ? | 11:11 |
| irf | pxe_wol and agent_wol? | 11:11 |
| *** electrofelix has joined #openstack-ironic | 11:11 | |
| sambetts | irf: One node can only be using one driver at a time | 11:12 |
| *** hoangcx has quit IRC | 11:12 | |
| sambetts | the pxe_ and agent_ define different methods that ironic can do the deployment, pxe_ will use iscsi to write the image to the node, agent_ uses the Ironic Python agent's http API to transfer the image to the node | 11:13 |
| *** jtomasek_ has joined #openstack-ironic | 11:14 | |
| aarefiev | dtantsur: nice catch, will you add release note in follow | 11:15 |
| *** divya_ has quit IRC | 11:15 | |
| dtantsur | aarefiev, oh, a release note... I should probably add it in this patch, as a plan to backport it | 11:15 |
| dtantsur | will do, if you guys can quickly re-review | 11:15 |
| sambetts | dtantsur: sure :) | 11:15 |
| *** divya has quit IRC | 11:16 | |
| irf | thanks | 11:17 |
| dtantsur | hmm, sambetts, aarefiev, wdyt about testing rerunning introspection in the discovery job (it's a bit less busy)? | 11:17 |
| *** jtomasek_ has quit IRC | 11:18 | |
| *** cppforlife_ has quit IRC | 11:19 | |
| *** yonglihe has quit IRC | 11:19 | |
| *** sergek has quit IRC | 11:19 | |
| *** ifarkas has quit IRC | 11:19 | |
| *** jtomasek_ has joined #openstack-ironic | 11:20 | |
| *** cppforlife_ has joined #openstack-ironic | 11:20 | |
| *** divya has joined #openstack-ironic | 11:20 | |
| *** yonglihe has joined #openstack-ironic | 11:20 | |
| *** sergek has joined #openstack-ironic | 11:20 | |
| divya | Hi Sam | 11:21 |
| *** clif_h has quit IRC | 11:21 | |
| sambetts | dtantsur: I ultimatly don't mind, although I'd like for us to maybe try to add a separate job for it for ease of debugging, so we can be like ok, main inspection passed, discovery failed, but reinstpection passed etc, maybe that'll be easier once we've got tempest going ;/ | 11:21 |
| sambetts | Hi | 11:21 |
| divya | Failed to prepare to deploy. Error: No VIFs found for node cb24e0b9 | | | -f19b-4c27-b287-58d6133973ff when attempting to update DHCP BOOT | | 11:21 |
| sambetts | divya: How did you try to deploy? | 11:22 |
| dtantsur | sambetts, fine, lets wait for tempest, should be there soon :) | 11:22 |
| *** xek has joined #openstack-ironic | 11:22 | |
| divya | | last_error | Failed to prepare to deploy. Error: No VIFs found for node cb24e0b9-f19b-4c27-b287-58d6133973ff when attempting to update DHCP BOOT options. | 11:22 |
| aarefiev | yeah I can do it in tempest | 11:22 |
| dtantsur | great | 11:22 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-inspector: Make sure to clean the blacklist cache when disabling the firewall https://review.openstack.org/305960 | 11:23 |
| dtantsur | sambetts, aarefiev, ^^^ | 11:23 |
| dtantsur | with the release note | 11:23 |
| divya | i used this command to deploy | 11:23 |
| divya | #ironic node-set-provision-state --config-drive /tmp/$NAME $NAME active | 11:23 |
| TheJulia | NobodyCam: Sorry was asleep | 11:23 |
| TheJulia | also, good morning everyone! | 11:23 |
| sambetts | divya: Are you using Ironic in OpenStack? | 11:23 |
| divya | yes | 11:23 |
| *** clif_h has joined #openstack-ironic | 11:23 | |
| sambetts | then you don't use ironic to deploy a node, you use nova | 11:24 |
| jroll | good morning TheJulia, sambetts, dtantsur, aarefiev, any other friends I have missed :) | 11:24 |
| jroll | and happy friday \o/ | 11:24 |
| sambetts | o/ jroll | 11:24 |
| dtantsur | morning jroll, \o/ | 11:24 |
| sambetts | TGIF | 11:24 |
| TheJulia | good morning jroll | 11:24 |
| divya | is this steps required before deploying node. https://github.com/jriguera/ansible-ironic-standalone/wiki/Cloud-Init-and-Config-Drive | 11:24 |
| dtantsur | jroll, gate fix for stable/mitaka: https://review.openstack.org/#/c/306323/ | 11:25 |
| divya | i used this doc to deploy bare metal node https://github.com/jriguera/ansible-ironic-standalone/wiki/Using-Ironic-client-with-agent_ipmitool-driver | 11:25 |
| aarefiev | morning jroll | 11:25 |
| sambetts | divya: its in the URL, thats for Ironic standalone | 11:26 |
| sambetts | divya: I'd suggest taking a look through this: http://docs.openstack.org/developer/ironic/dev/dev-quickstart.html | 11:26 |
| jroll | dtantsur: +A, do we need that on liberty as well? | 11:28 |
| dtantsur | jroll, likely, not sure. I got hit by it in mitaka right now | 11:28 |
| jroll | dtantsur: okay, I'll jfdi | 11:28 |
| irf | hi | 11:28 |
| sambetts | jroll: I would expect it to affect all branches that have that script | 11:28 |
| irf | i am not able to delete the port using port-delete command | 11:29 |
| irf | ironic port-create -n da7e87a0-230a-4ba6-bd11-0f44e652d69f -a 00:00:00:00:00:88 | 11:29 |
| jroll | merge conflict :| | 11:29 |
| dtantsur | jroll, ah yeah, that's why I didn't propose it | 11:29 |
| irf | it says mac address already in use | 11:30 |
| irf | and gives this error | 11:30 |
| irf | v | 11:30 |
| irf | No valid host was found. Reason: No conductor service registered which supports driver pxe_ssh. ( | 11:30 |
| irf | i am not able to delete the port | 11:30 |
| sambetts | irf: It sounds like you have a node enrolled that has the driver pxe_ssh, but you've removed support for pxe_ssh from your conductor | 11:31 |
| jroll | dtantsur: https://review.openstack.org/306369 | 11:32 |
| sambetts | irf: you need to add pxe_ssh to the enabled_drivers, then restart the conductor, then delete the node or update it to use a different driver , then you can remove the pxe_ssh from enabled drivers again | 11:32 |
| irf | ohh ok | 11:32 |
| *** jtomasek_ has quit IRC | 11:33 | |
| *** dprince has joined #openstack-ironic | 11:33 | |
| *** Marga__ has quit IRC | 11:34 | |
| sambetts | dtantsur: I've got a slight english nit with your release note to make it read better, do you have time to push a quick fix?, if not I'm happy to push it through | 11:35 |
| sambetts | jroll: woops... should have refrshed that page before commenting :-p | 11:38 |
| irf | done | 11:38 |
| irf | thanks | 11:38 |
| sambetts | irf: :) | 11:39 |
| jroll | sambetts: ? | 11:39 |
| *** baoli has joined #openstack-ironic | 11:40 | |
| *** baoli_ has joined #openstack-ironic | 11:41 | |
| jroll | oh I see it, heh | 11:42 |
| *** dencaval has quit IRC | 11:43 | |
| sambetts | jroll: Sorry about the lack of context on that patch I meant to create a bug for it then got distracted | 11:45 |
| jroll | sambetts: nah it's all good | 11:45 |
| *** baoli has quit IRC | 11:45 | |
| *** dims has quit IRC | 11:46 | |
| openstackgerrit | Merged openstack/bifrost: Updated from global requirements https://review.openstack.org/305167 | 11:51 |
| *** openstack has quit IRC | 11:51 | |
| *** openstack has joined #openstack-ironic | 11:53 | |
| dtantsur | sambetts, sure, gimme 5 mins | 11:55 |
| *** dims has joined #openstack-ironic | 11:56 | |
| openstackgerrit | Sofiia Andriichenko proposed openstack/python-ironicclient: Add test for chassis-node-list https://review.openstack.org/299482 | 11:57 |
| *** athomas has quit IRC | 11:58 | |
| *** athomas has joined #openstack-ironic | 11:58 | |
| openstackgerrit | Mathieu Mitchell proposed openstack/ironic-specs: Container-based display drivers https://review.openstack.org/306078 | 11:59 |
| mat128 | good morning ironic! | 12:02 |
| sambetts | o/ mat128 | 12:02 |
| mat128 | \o | 12:02 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-inspector: Make sure to clean the blacklist cache when disabling the firewall https://review.openstack.org/305960 | 12:03 |
| dtantsur | sambetts, aarefiev, one more update :) ^^^ | 12:04 |
| sambetts | dtantsur: :) thanks | 12:04 |
| vdrok | morning TheJulia jroll and mat128 | 12:04 |
| dtantsur | sambetts, have a minute chatting about driver composition? | 12:05 |
| dtantsur | sambetts, tl;dr I don't believe that "CIMC power with IPMI management" is something we should support. and I don't think it's something we've ever supported to be honest | 12:05 |
| sambetts | dtantsur: Sure, I don't know why we shouldn't support that | 12:06 |
| dtantsur | sambetts, I'd rather ask why should we? like, getting the worst of both worlds? :) | 12:06 |
| dtantsur | I see people often coding drivers in assumption that power == management, and it's a sane thing to assume, provided that the same interface is usually used | 12:07 |
| dtantsur | jroll, you might want to follow our discussion ^^^ | 12:07 |
| jroll | ooooo | 12:08 |
| sambetts | that was just an example, I just don't want to limit the flexibilty that we could offer | 12:08 |
| * jroll listens | 12:08 | |
| mat128 | dtantsur: I'm with sambetts on this. If operators could chose all drivers independently CMIC power with IPMI management would be possible, wouldn't it? | 12:09 |
| sambetts | jroll: context is my comment: https://review.openstack.org/#/c/188370/ | 12:09 |
| dtantsur | sambetts, it's fine, but it complicates the matter. I mean, I can make all the drivers pluggable, but who will guarantee that all possible permutations work? | 12:09 |
| dtantsur | mat128, sambetts, also, who would support that? Sam, as a Cisco representative, will you work on supporting this combination? establish CI for it? | 12:09 |
| sambetts | dtantsur: thats what the comcept of the hardware_type is for | 12:10 |
| mat128 | Nova has this concept of supported / untested and warns if you are using something untested: https://github.com/openstack/nova/blob/master/nova/virt/libvirt/driver.py#L618 | 12:10 |
| sambetts | dtantsur: I think thats what I address in the last part, its down to the vendors to gurantee support for their hardware type | 12:10 |
| dtantsur | sambetts, my question is about human resources and vendor buy-in. if no vendors will declare official support for this, what's the use in complicating everything? | 12:11 |
| dtantsur | sambetts, will Cisco do that? I mean support CIMC + IPMI, IPMI + CIMC, etc? | 12:11 |
| mat128 | dtantsur: if we have a compatibility map, how will we manage 3rd party drivers, especially, how will they hook into the map if they require? | 12:12 |
| dtantsur | my point here is that supporting such things is a burden on vendors | 12:12 |
| dtantsur | mat128, entry points are extensible by 3rd parties. that's why we don't hardcode things in python code any more | 12:13 |
| dtantsur | I'd like vendors to stand up and say "yes, we want to support this" before we go into complicating the spec even more | 12:14 |
| dtantsur | hence my question: is it something planned by someone here from vendors side? | 12:15 |
| jroll | dtantsur: fwiw, I actually imagined getting rid of node.driver as well | 12:17 |
| jroll | but didn't think about the mixing of power/mgmt much | 12:17 |
| dtantsur | jroll, that's a separate topic, kind of | 12:17 |
| jroll | kind of :) | 12:17 |
| sambetts | I don't really get how this would be any different from a vendors perspective than the current proposed solution except that the current solution doesn't define supported purmutations | 12:17 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Fix API node name updates https://review.openstack.org/300983 | 12:17 |
| jroll | rather, I think I agree with sambetts in general about how this should work, however I also agree I'm not sure we should allow mixing eg. cimc/ipmi | 12:18 |
| jroll | but if the vendor *does* want to support that, fine | 12:18 |
| jroll | I just don't understand why a user would desire that | 12:18 |
| dtantsur | jroll++ for why a user would desire that | 12:18 |
| sambetts | if we don't allow them to be customised whats the point in having them as separate interfaces? | 12:19 |
| dtantsur | sambetts, good question :) | 12:20 |
| jroll | right, I think we should build the system to enable it | 12:21 |
| sambetts | I think there are cases when you can only provide power e.g. AMT/WOL, and no mgnt, but someone might have some custome vendor mgnt solution, so then they create a WOL + <some custom thing> and that wouldn't be possible if we lock them together | 12:21 |
| jroll | but very critically ask "what's the use case" when someone does something crazy in-tree | 12:22 |
| mat128 | for an example of operator config, we have a power driver that does IPMI and falls back to SNMP/PDU (Fallback Power). We use this power driver alongside it's equivalent management driver, FallbackManagement. | 12:22 |
| mat128 | if you had a list of supported options, I would be SOL because the list wouldn't contain any of my stuff | 12:22 |
| dtantsur | sambetts, yeah, this example is more interesting | 12:22 |
| jroll | ++ | 12:23 |
| dtantsur | hmm, thanks folks, this makes much more sense at a second glance | 12:23 |
| sambetts | then that vendor can add a new hardware_type, and add a new mgnt interface, then just has to add his hardware type to both the WOL interface and his mgnt interface | 12:23 |
| dtantsur | now, do we want to get rid of driver/hardware_type completely? | 12:23 |
| dtantsur | sambetts, why do we need a hardware_type then? | 12:23 |
| jroll | idk about the whole hardware_type thing, but I don't have a better proposal | 12:24 |
| sambetts | well how to define supported/unsupport interfaces | 12:24 |
| jroll | AIUI, it's meant to say "hardware of this type can use these drivers" | 12:24 |
| dtantsur | sambetts, how you do this? (in python code is NOT an answer) | 12:24 |
| mat128 | if all drivers respect the interface, why wouldn't it be possible to mix anything? | 12:24 |
| jroll | ^ | 12:24 |
| *** trown|outtypewww is now known as trown | 12:24 | |
| jroll | that's my next question, you'd need that to be pluggable | 12:24 |
| jroll | for out of tree things | 12:24 |
| sambetts | so originally I started by defining a Hardware type class that defined what interfaces were valid to use with that hardware | 12:25 |
| dtantsur | sambetts, no hardcoding in Python please :) you will totally exclude 3rd parties | 12:26 |
| sambetts | they would be stevedored | 12:26 |
| sambetts | but the probelm I ran into with that was that often you want to support all interfaces, except a few, e.g. CIMC works with agent, iscsi and maybe others in the future, but it'll never support ILO Virutal media | 12:27 |
| *** b_bezak_ has joined #openstack-ironic | 12:27 | |
| dtantsur | sambetts++ | 12:27 |
| dtantsur | so | 12:27 |
| sambetts | thats why I wanted to move the relationship the other way around | 12:27 |
| mat128 | "This spec proposes redefining a driver as power + management + passthru interfaces with other interfaces being changeable by an operator." < if you do this, and provide new fields for deploy, console, etc. I guess your goal is met | 12:28 |
| dtantsur | we need an entry point like: ironic.hardware-types.cimc.boot listing all available boot interfaces. then 3rd parties can add to it | 12:28 |
| dtantsur | OR | 12:28 |
| dtantsur | we can leave it up to the configuration file like I have right now in the proposal | 12:28 |
| dtantsur | (then potentially operators can screw themselves) | 12:28 |
| sambetts | I see it more like: ironic.power.interface.ipmipower.supported_hardware_types: .... | 12:29 |
| sambetts | and have a ironic.hardwre_types one too | 12:29 |
| sambetts | ironic.interface.power.ipmitool.support_hardware_types: ... | 12:30 |
| sambetts | thats better ^ | 12:30 |
| dtantsur | yeah. maybe even s/support_// | 12:30 |
| *** ndipanov has quit IRC | 12:31 | |
| dtantsur | now a tricky question for you folks: what defines the vendor passthru? | 12:31 |
| dtantsur | sambetts, mat128 ^^ | 12:32 |
| jroll | oh my | 12:32 |
| mat128 | sambetts, dtantsur: ok different idea (bear with me for a sec): hardware_types specify all fields (like the current driver), but they can also be defined individually. All hardware_types should be CI tested by vendor, but if you mix individual options you're on your own. | 12:32 |
| sambetts | its should be just another interface | 12:32 |
| dtantsur | sambetts, in 99% cases it depends on management/power | 12:32 |
| * jroll thinks we should kill vendor passthru with fire | 12:32 | |
| dtantsur | jroll, vendors will cry | 12:32 |
| dtantsur | mat128, I don't quite get how it's different from what sam proposes | 12:33 |
| * sambetts thinks this wouldn't be a probelm if deployment didn't rely on it | 12:33 | |
| dtantsur | well... | 12:33 |
| dtantsur | sambetts, if you mix drac vendor with ipmi mgmt, it won't work | 12:33 |
| jroll | dtantsur: I don't think there's many useful passthrus in tree | 12:33 |
| jroll | right? | 12:33 |
| mat128 | you dont need to provide any list of supported X, you just provide "bundles" that are known to work | 12:34 |
| mat128 | like we do right now | 12:34 |
| dtantsur | jroll, drac bios settings iirc; something in ipmi drivers | 12:34 |
| dtantsur | sambetts, or maybe it will... | 12:34 |
| jroll | dtantsur: they should move to cleaning :) | 12:34 |
| *** keedya has joined #openstack-ironic | 12:34 | |
| *** irf has quit IRC | 12:34 | |
| jroll | anyway, we can't kill it for now | 12:34 |
| jroll | and remember we mix in multiple passthru classes | 12:34 |
| dtantsur | oh yeah, moar fun | 12:35 |
| jroll | maybe passthrus=agent,ipmitool | 12:35 |
| *** Goneri has joined #openstack-ironic | 12:35 | |
| dtantsur | jroll, lets kill the agent passthru finally. or rather, make it generic. I can file an RFE | 12:35 |
| *** b_bezak_ has left #openstack-ironic | 12:35 | |
| sambetts | a passthru can have supported_hardware_types too, and also there is no reason why a particular interface couldn't add something in their validate call to say hey I don't work with your current purmutation of interfaces | 12:36 |
| dtantsur | fair | 12:36 |
| mat128 | sambetts: does that mean we don't need any supported* in stevedore? | 12:37 |
| dtantsur | mat128, why? | 12:38 |
| sambetts | mat128: you still need that so that you can customise the support_hardware_types | 12:38 |
| sambetts | for each interface | 12:38 |
| *** irf has joined #openstack-ironic | 12:38 | |
| mat128 | hmm, I fail to see why we should limit the operators in any way | 12:38 |
| jroll | dtantsur: +1, good point | 12:39 |
| sambetts | mat128: we're not? | 12:39 |
| mat128 | we define hardware types that have supported interfaces (deploy/mgmt/power/etc.) | 12:39 |
| sambetts | no the other way around | 12:39 |
| mat128 | ok yes, the other way around | 12:39 |
| mat128 | what are we going to do with this information? | 12:40 |
| mat128 | raise on unsupported setups? | 12:40 |
| mat128 | warn? | 12:40 |
| *** manfredp_ has quit IRC | 12:40 | |
| sambetts | raise, mat128: https://review.openstack.org/#/c/188370/ my full explanation is here | 12:40 |
| mat128 | '(mat128) hmm, I fail to see why we should limit the operators in any way' | 12:40 |
| openstackgerrit | Gonéri Le Bouder proposed openstack/ironic: ipxe: retry on failure https://review.openstack.org/306196 | 12:41 |
| mat128 | so we limit them in their choices, don't we? | 12:41 |
| dtantsur | jroll, https://bugs.launchpad.net/ironic/+bug/1570841 wanna a spec for that? | 12:42 |
| openstack | Launchpad bug 1570841 in Ironic "[RFE] Promote agent vendor passthru to core API" [Wishlist,Confirmed] - Assigned to Dmitry Tantsur (divius) | 12:42 |
| sambetts | as a vendor I want to define my support for Ironic, so I don't get emails e.g. I've configured this interface with this interface and it doesn't work | 12:42 |
| mat128 | so it's more of declaring what is supported | 12:42 |
| jroll | dtantsur: I don't, but the "this is an api change and will have deprecation things" makes me think yes, a quick one | 12:42 |
| dtantsur | jroll, sure | 12:43 |
| jroll | dtantsur: more about how we do the deprecation period and such | 12:43 |
| mat128 | which goes back to my nova example of 'you're on your own now': https://github.com/openstack/nova/blob/master/nova/virt/libvirt/driver.py#L618 | 12:43 |
| *** keedya has quit IRC | 12:43 | |
| mat128 | I feel this would be much simpler if we just allowed independent power/deploy/mgmt/console/etc. drivers | 12:43 |
| sambetts | I think we could have a hardware_type which is ANYTHING, and just allow any interface permutation if that hardware_type is being used | 12:43 |
| dtantsur | sambetts, aka "fake" :) | 12:43 |
| sambetts | dtantsur: exactly | 12:43 |
| sambetts | but if you declare your hardware as my hardware, then I'm going to tell you what I support | 12:44 |
| mat128 | so ``node.hardware_type`` is the new ``node.driver`` | 12:44 |
| *** ndipanov has joined #openstack-ironic | 12:46 | |
| sambetts | not exactly, node.driver is a fixed relationship between a set of interfaces, hardware_type lets us define support for a wide range on interfaces that can all be mixed and matched, and if it validates you shoule be able to guarentee it works | 12:46 |
| mat128 | good point, yes | 12:47 |
| mat128 | ok then, I think it makes sense | 12:47 |
| mat128 | so I can stevedore configure a new type that allows whatever permutations I want, it's just that when I call you up and tell you this, you can tell me that I'm on my own :) | 12:47 |
| sambetts | exactly | 12:48 |
| mat128 | good stuff, sorry for disturbing the peace on this beautiful morning | 12:48 |
| openstackgerrit | Julia Kreger proposed openstack/ironic: Add config option for ATA erase fallback in agent https://review.openstack.org/302819 | 12:48 |
| sambetts | thats all good, its a work in progress idea, so brian storming like this is the best thing for it | 12:48 |
| sambetts | brain * | 12:48 |
| sambetts | man it must be Friday | 12:48 |
| TheJulia | That it is :) | 12:49 |
| jroll | http://www.reactiongifs.us/wp-content/uploads/2014/02/jumping_high_five_family_guy.gif | 12:50 |
| jroll | tried to find brian storming through a door but no luck :( | 12:50 |
| sambetts | hahahaha | 12:50 |
| sambetts | Morning TheJulia | 12:50 |
| *** mgoddard__ has joined #openstack-ironic | 12:51 | |
| TheJulia | jroll: I think that means you just have to spend the day watching family guy :) | 12:51 |
| *** alaski is now known as lascii | 12:52 | |
| jroll | TheJulia: lol | 12:52 |
| vdrok | jroll: about that nova rebuild bug, wdyt about just adding if recreate before this line - https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L2786 | 12:52 |
| vdrok | the comment says about that, so we could just set scheduled_node to instance.node | 12:53 |
| vdrok | then migration won't be created | 12:53 |
| *** ndipanov has quit IRC | 12:53 | |
| *** mgoddard_ has quit IRC | 12:54 | |
| jroll | vdrok: idk, we probably want to hit this block: https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L2771 | 12:54 |
| jroll | no sense in making more conditionals, just pass the node in if it's a real rebuild call | 12:55 |
| vdrok | jroll: hm, I'm not a nova expert, but I thought that this rebuild claim creates migration in nova db | 12:55 |
| vdrok | and after that, after instance deletion update_available_resource complains about this migration, telling that instance is deleted | 12:56 |
| vdrok | so maybe nopclaim is better? | 12:56 |
| *** irf has quit IRC | 12:56 | |
| jroll | vdrok: rebuild does rebuild_instance on the same host, migrate/evacuate does rebuild_instance on a different host | 12:56 |
| vdrok | yep, I get that | 12:57 |
| jroll | idk about migration records, if they're meant to be created for rebuild or not | 12:57 |
| vdrok | well, they are, I'll paste a log | 12:57 |
| jroll | but really, the problem here is that the rebuild api doesn't take instance.node into account | 12:57 |
| jroll | idk | 12:57 |
| jroll | seems like you should ask nova folks :) | 12:57 |
| jroll | I mean for normal rebuilds, not ironic rebuilds | 12:57 |
| vdrok | jroll: http://paste.openstack.org/show/494215/ | 12:59 |
| vdrok | here is what happens after instance deletion | 12:59 |
| vdrok | jroll: yep, will ask around in nova channel, thanks! :) | 12:59 |
| vdrok | brb | 12:59 |
| jroll | yeah np | 13:00 |
| *** derekh has joined #openstack-ironic | 13:03 | |
| *** ndipanov has joined #openstack-ironic | 13:14 | |
| *** cdearborn has joined #openstack-ironic | 13:17 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-specs: Promote agent vendor passthru to core API https://review.openstack.org/306418 | 13:19 |
| dtantsur | jroll, any quick comments on ^^^? | 13:19 |
| * dtantsur didn't finish the RPC part yet, but wants some sanity check | 13:19 | |
| * jroll reads | 13:19 | |
| openstackgerrit | Merged openstack/ironic-inspector: Make sure to clean the blacklist cache when disabling the firewall https://review.openstack.org/305960 | 13:22 |
| *** jjohnson2_ has joined #openstack-ironic | 13:22 | |
| * sambetts wonders if we should look into creating an API extension framework similar to that provided in neutron, so that the agent deploy interface could define the lookup and heartbeat API end points | 13:23 | |
| dtantsur | sambetts, that's how we use vendor passthru now, I guess :) | 13:24 |
| jroll | dtantsur: done, good stuff, just a few things | 13:24 |
| jroll | no pls no extensions | 13:24 |
| sambetts | I guess its sort of like vendor passthru, except defined by the interfaces themselves not a separate thing | 13:24 |
| dtantsur | jroll, thanks, updating soon | 13:25 |
| jroll | cool | 13:25 |
| jroll | oh I forgot one | 13:25 |
| jroll | there | 13:26 |
| dtantsur | oh, this is a good one :) | 13:26 |
| sambetts | I guess lookup and heartbeat are specfic to use IPA, but not specific to a particuar deploy process, so unless where going to support people creating their own ramdisk with its own API then we don't need to make this extensible | 13:27 |
| sambetts | big +1 from me | 13:28 |
| dtantsur | sambetts, even 3rd party ramdisk will probably have to follow the same logic: lookup, heartbeat, then do whatever they want | 13:28 |
| sambetts | yeah, I was just thinking that | 13:28 |
| sambetts | tbh I'd really like it if the ramdisks didn't talk to the Ironic API | 13:29 |
| sambetts | seems like a layer break to me | 13:29 |
| dtantsur | sambetts, well, it's not mandatory to call these API.. and not mandatory to implement in the deploy interface | 13:30 |
| jroll | the other option is have them register via zookeeper/etc | 13:30 |
| dtantsur | hmm, yeah. are we talking about DLM on the summit? | 13:30 |
| *** piet has joined #openstack-ironic | 13:31 | |
| jroll | I think so | 13:32 |
| sambetts | Once the bash ramdisk is killed with fire, the deploy interface should be more or less transparent to the ramdisk being used right? | 13:33 |
| jroll | define transparent :) | 13:35 |
| sambetts | if we define a standard interface for the ramdisks to talk e.g. lookup/heartbeat, then the deploy interfaces should have a standard API for functions that are called when one of those events happen | 13:35 |
| jroll | dtantsur: oh, we aren't discussing it, we more or less have a plan | 13:36 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-specs: Promote agent vendor passthru to core API https://review.openstack.org/306418 | 13:36 |
| dtantsur | update ^^^ | 13:36 |
| dtantsur | jroll, aha, remembered | 13:36 |
| jroll | sambetts: yeah, sounds right | 13:36 |
| jroll | this spec reminds me, going to try to finish search api spec today | 13:38 |
| dtantsur | hehe :) so many specs, this is the 3rd from me in 2 days | 13:38 |
| jroll | nice, let's land them all next week :) | 13:38 |
| sambetts | Do you think I need to write a spec for the remove *_ssh drivers RFE? | 13:38 |
| dtantsur | speaking of which, sambetts, please check https://review.openstack.org/305864 when you have some time | 13:38 |
| jroll | sambetts: link? | 13:39 |
| sambetts | jroll: https://bugs.launchpad.net/ironic/+bug/1570301 | 13:39 |
| openstack | Launchpad bug 1570301 in Ironic "[RFE] Deprecate and remove the *_ssh drivers" [Wishlist,Confirmed] | 13:39 |
| jroll | hmm | 13:40 |
| jroll | sambetts: can you lay out the todo list for that, then I think we could approve it | 13:41 |
| jroll | with timelines for deprecation, preferably | 13:41 |
| jroll | (e.g. add vbmc to ironic project) | 13:41 |
| *** piet has quit IRC | 13:41 | |
| sambetts | in a full spec or in the RFE? | 13:41 |
| dtantsur | yay for vmbc in ironic :) | 13:41 |
| jroll | sambetts: in the rfe should be fine | 13:42 |
| sambetts | Cool :D sure I'll try to update that | 13:42 |
| jroll | thanks, feel free to ping me | 13:42 |
| jroll | I want to go through open rfe's this afternoon | 13:42 |
| *** piet has joined #openstack-ironic | 13:42 | |
| sambetts | will do, I'm not sure what lucasagomes's plan is for vbmc | 13:42 |
| sambetts | lucasagomes: you here? | 13:42 |
| dtantsur | sambetts, he's on travel today | 13:43 |
| dtantsur | sambetts, but moving vbmc to ironic was on his plans :) | 13:43 |
| dtantsur | jroll, a concerning thing: http://tarballs.openstack.org/ironic-python-agent/coreos/files/ shows the newest images from Apr 5 | 13:43 |
| jroll | dtantsur: https://review.openstack.org/#/c/305988/ fixes | 13:44 |
| jroll | but being a pain in the gate for some reason :/ | 13:44 |
| dtantsur | damn | 13:44 |
| jroll | looks like this round is debian mirrors :| | 13:45 |
| jroll | and timeouts | 13:45 |
| jroll | urgh | 13:45 |
| dtantsur | ..... | 13:46 |
| sambetts | sigh... | 13:46 |
| sambetts | dtantsur: yeah, not sure what the timeline is on it though :/ | 13:47 |
| jroll | we could get it done next week easy | 13:47 |
| dtantsur | yep, not so hard. just wait for Lucas to come back from Brno | 13:47 |
| sambetts | awesome :D | 13:47 |
| *** [1]cdearborn has joined #openstack-ironic | 13:49 | |
| *** piet has quit IRC | 13:51 | |
| *** piet has joined #openstack-ironic | 13:51 | |
| sambetts | hmmm, deprecate for removal in OpenStack Ocata or P..? | 13:51 |
| sambetts | removal in Ocata leaves them deprecated for the whole of Newton | 13:52 |
| *** jayc__ has joined #openstack-ironic | 13:52 | |
| jroll | otaca imo | 13:55 |
| jroll | ocata* | 13:56 |
| *** Nisha_away has joined #openstack-ironic | 13:57 | |
| sambetts | cool, thats what I was thinking :) | 13:57 |
| dtantsur | ++ for ocata | 13:58 |
| *** daemontool_ has joined #openstack-ironic | 13:58 | |
| dtantsur | people should not use them in production (but they do) | 13:58 |
| *** ametts has joined #openstack-ironic | 13:59 | |
| jroll | O_o | 13:59 |
| *** cdearborn has quit IRC | 14:00 | |
| sambetts | jroll: Something like https://bugs.launchpad.net/ironic/+bug/1570301 | 14:00 |
| openstack | Launchpad bug 1570301 in Ironic "[RFE] Deprecate and remove the *_ssh drivers" [Wishlist,Confirmed] | 14:00 |
| *** dims has quit IRC | 14:01 | |
| *** daemontool has quit IRC | 14:01 | |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Add unit tests for ClusteredComputeManager https://review.openstack.org/306445 | 14:03 |
| jroll | sambetts: lgtm, dtantsur are you good with that as-is? | 14:03 |
| jroll | I added a couple words | 14:03 |
| sambetts | jroll: we've already got IPMItool driver gate tests running and voting by the Ocata cycle | 14:04 |
| *** dims has joined #openstack-ironic | 14:04 | |
| jroll | sambetts: yeah, but not for all cases | 14:04 |
| sambetts | oh? | 14:04 |
| *** mgould has quit IRC | 14:04 | |
| dtantsur | jroll, sambetts, lgtm | 14:05 |
| jroll | we just have it for pxe_ipa I believe | 14:05 |
| *** yolanda has joined #openstack-ironic | 14:05 | |
| sambetts | jroll: we have separate tests for *_ipmitool | 14:05 |
| sambetts | https://review.openstack.org/#/c/305868/ | 14:06 |
| sambetts | for both pxe and agent | 14:06 |
| jroll | oh we do have agent, nice | 14:06 |
| sambetts | we'd need them for tinyipa and postgres ones I guess | 14:06 |
| sambetts | too | 14:06 |
| jroll | ya | 14:06 |
| jroll | I also no longer see an ipxe test? | 14:07 |
| sambetts | pxe_ipa is ipxe | 14:07 |
| jroll | aha, cool | 14:07 |
| jroll | so yeah, just need to make sure we have the same coverage | 14:07 |
| sambetts | either dtantsur or lucasagomes merged them | 14:07 |
| *** dansmith is now known as superdan | 14:07 | |
| sambetts | ++ | 14:07 |
| jroll | but we can just switch them over in ocata | 14:07 |
| dtantsur | yep, pxe_ipa uses ipxe | 14:08 |
| jroll | no need to run duplicate jobs for all of newton | 14:08 |
| jroll | at any rate, rfe is now approved | 14:08 |
| sambetts | cool :D | 14:08 |
| jroll | thanks for that | 14:08 |
| dtantsur | awesome | 14:09 |
| *** divya has quit IRC | 14:09 | |
| * sambetts did it so that he doesn't have deal with confused people any more ;) | 14:09 | |
| *** mtanino has joined #openstack-ironic | 14:14 | |
| NobodyCam | good morning Ironicers...and of course... | 14:16 |
| NobodyCam | TGIF!!! | 14:16 |
| sambetts | hey hey NobodyCam | 14:16 |
| NobodyCam | morning sambetts :) | 14:16 |
| *** maurosr has quit IRC | 14:19 | |
| *** [1]cdearborn has quit IRC | 14:20 | |
| *** maurosr has joined #openstack-ironic | 14:20 | |
| jroll | morning NobodyCam \o | 14:21 |
| *** [1]cdearborn has joined #openstack-ironic | 14:22 | |
| *** jayc__ has quit IRC | 14:22 | |
| *** jayc_ has joined #openstack-ironic | 14:23 | |
| NobodyCam | morning jroll how goes the house hunt? | 14:23 |
| *** cdearborn has joined #openstack-ironic | 14:23 | |
| *** Nisha_away has quit IRC | 14:23 | |
| jroll | NobodyCam: good, we're signing on one sunday unless a couple more we see tonight are amazing :) | 14:24 |
| NobodyCam | jroll: awesome :) and congratz :) | 14:24 |
| jroll | thanks! | 14:25 |
| TheJulia | wow, congrats | 14:25 |
| *** jayc_ is now known as jayc | 14:25 | |
| jroll | super excited, has a fenced in yard for the dog and less than a mile walk to a small little downtown area | 14:25 |
| jroll | very american dream-y | 14:25 |
| *** e0ne_ has joined #openstack-ironic | 14:25 | |
| *** e0ne has quit IRC | 14:26 | |
| NobodyCam | anyone seeing timeout in gate checks https://review.openstack.org/#/c/272658 has had several | 14:26 |
| jroll | yep, another patch is being problematic too | 14:26 |
| jroll | haven't dug into if something changed | 14:26 |
| *** vishwanathj_zzz has quit IRC | 14:27 | |
| *** vishwanathj_zzz has joined #openstack-ironic | 14:27 | |
| NobodyCam | ya I haven't dug in to much: http://logs.openstack.org/58/272658/38/check/gate-tempest-dsvm-ironic-pxe_ipa/bcf088f/console.html#_2016-04-15_07_13_40_438 | 14:28 |
| *** vishwanathj_zzz has quit IRC | 14:28 | |
| NobodyCam | but looks like the old standard timeout at least on the surface | 14:28 |
| *** vishwanathj_zzz has joined #openstack-ironic | 14:29 | |
| *** achanda has joined #openstack-ironic | 14:30 | |
| jroll | yep | 14:30 |
| dtantsur | morning NobodyCam | 14:31 |
| NobodyCam | morning dtantsur ... happy Friday :) | 14:31 |
| *** achanda has quit IRC | 14:32 | |
| *** mkovacik has quit IRC | 14:33 | |
| *** mkovacik has joined #openstack-ironic | 14:33 | |
| dtantsur | jroll, stable backport please: https://review.openstack.org/#/c/305916/ | 14:33 |
| *** athomas has quit IRC | 14:34 | |
| dtantsur | sambetts, aarefiev, and 2 backports for one of you guys: https://review.openstack.org/306421 and https://review.openstack.org/306429 :) | 14:34 |
| jroll | dtantsur: done | 14:36 |
| sambetts | dtantsur: also done :-P | 14:36 |
| sambetts | dtantsur: I forgot we backported that firewall stuff into liberty | 14:36 |
| *** [1]cdearborn has quit IRC | 14:37 | |
| dtantsur | we did | 14:38 |
| dtantsur | now we have a fix for the fix :) | 14:38 |
| sambetts | yay :D | 14:38 |
| *** cdearborn has quit IRC | 14:40 | |
| NobodyCam | jroll: https://review.openstack.org/#/c/304578 has 2 +2's wanted a IPA expert to hit the button.. | 14:40 |
| *** cdearborn has joined #openstack-ironic | 14:40 | |
| *** electrofelix has quit IRC | 14:43 | |
| * NobodyCam grabs more coffee | 14:44 | |
| *** mgoddard has joined #openstack-ironic | 14:46 | |
| *** mgoddard__ has quit IRC | 14:46 | |
| *** mkovacik has quit IRC | 14:49 | |
| *** mkovacik has joined #openstack-ironic | 14:50 | |
| *** electrofelix has joined #openstack-ironic | 14:51 | |
| *** vishwanathj_zzz is now known as vishwanathj | 14:51 | |
| *** daemontool__ has joined #openstack-ironic | 14:54 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-inspector: Store ramdisk logs on all processing failures, not only reported by the ramdisk https://review.openstack.org/300011 | 14:55 |
| *** mbound has quit IRC | 14:55 | |
| *** daemontool_ has quit IRC | 14:56 | |
| *** ptx0 has quit IRC | 14:57 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-specs: Promote agent vendor passthru to core API https://review.openstack.org/306418 | 14:58 |
| dtantsur | yuriyz, ^^^ | 14:58 |
| NobodyCam | so ya seeing timeouts on more patches, /me suspects something bigger is afoot. :( | 14:58 |
| sambetts | :'( | 14:58 |
| vdrok | morning NobodyCam | 15:01 |
| vdrok | yep, something strange happens - http://logs.openstack.org/83/300983/10/check/gate-tempest-dsvm-ironic-pxe_ssh/fae7ec2/console.html#_2016-04-15_12_58_40_684 | 15:01 |
| vdrok | 12:45:48,323 and the next log line is 12:58:17,745 | 15:01 |
| NobodyCam | morning vdrok | 15:01 |
| sambetts | jlvillal: Btw I've had our recheck command adapted to make sure you can do a single Cisco recheck or a full recheck | 15:03 |
| *** mbound has joined #openstack-ironic | 15:06 | |
| *** pcaruana has quit IRC | 15:06 | |
| jroll | NobodyCam: I'll let JayF have that one given sambetts comments | 15:08 |
| *** superdan is now known as _-_ | 15:08 | |
| *** _-_ is now known as superdan | 15:09 | |
| *** MattMan has quit IRC | 15:10 | |
| *** MattMan has joined #openstack-ironic | 15:10 | |
| *** kozhukalov has quit IRC | 15:11 | |
| *** serverascode has quit IRC | 15:11 | |
| *** JoshNang has quit IRC | 15:11 | |
| *** ChrisMead has quit IRC | 15:11 | |
| *** vdrok has quit IRC | 15:11 | |
| *** boris-42 has quit IRC | 15:11 | |
| *** devananda has quit IRC | 15:12 | |
| *** jmccrory has quit IRC | 15:12 | |
| *** fyxim_ has quit IRC | 15:12 | |
| *** JayF has quit IRC | 15:12 | |
| *** odyssey4me has quit IRC | 15:13 | |
| *** anush has quit IRC | 15:13 | |
| *** lekha has quit IRC | 15:13 | |
| *** [1]cdearborn has joined #openstack-ironic | 15:16 | |
| *** mgoddard_ has joined #openstack-ironic | 15:16 | |
| *** daemontool_ has joined #openstack-ironic | 15:16 | |
| *** ppiela_ has joined #openstack-ironic | 15:16 | |
| *** odyssey4me has joined #openstack-ironic | 15:17 | |
| *** electrofelix has quit IRC | 15:18 | |
| *** maurosr has quit IRC | 15:18 | |
| *** anush has joined #openstack-ironic | 15:18 | |
| *** MattMan__ has joined #openstack-ironic | 15:18 | |
| *** JayF has joined #openstack-ironic | 15:19 | |
| *** JayF has joined #openstack-ironic | 15:19 | |
| *** sambetts_ has joined #openstack-ironic | 15:19 | |
| *** boris-42 has joined #openstack-ironic | 15:19 | |
| *** mat128_ has joined #openstack-ironic | 15:19 | |
| *** sturivnyi_ has joined #openstack-ironic | 15:20 | |
| *** MattMan has quit IRC | 15:20 | |
| *** daemontool__ has quit IRC | 15:20 | |
| *** mgoddard has quit IRC | 15:20 | |
| *** jayc has quit IRC | 15:20 | |
| *** vdrok_ has joined #openstack-ironic | 15:20 | |
| jroll | JayF: icymi in the netsplit, your eyes are appreciated on https://review.openstack.org/#/c/304578 | 15:20 |
| *** tonyb_ has joined #openstack-ironic | 15:20 | |
| *** pleia2_ has joined #openstack-ironic | 15:21 | |
| *** fyxim_ has joined #openstack-ironic | 15:23 | |
| *** baoli has joined #openstack-ironic | 15:24 | |
| *** davidlenwell has joined #openstack-ironic | 15:24 | |
| *** jlvillal_ has joined #openstack-ironic | 15:24 | |
| *** vdrok has joined #openstack-ironic | 15:24 | |
| *** vdrok_ has left #openstack-ironic | 15:25 | |
| *** maurosr has joined #openstack-ironic | 15:25 | |
| *** ChrisMead has joined #openstack-ironic | 15:25 | |
| *** e0ne_ has quit IRC | 15:25 | |
| *** dims has quit IRC | 15:25 | |
| *** ametts has quit IRC | 15:25 | |
| *** ndipanov has quit IRC | 15:25 | |
| *** derekh has quit IRC | 15:25 | |
| *** baoli_ has quit IRC | 15:25 | |
| *** dprince has quit IRC | 15:25 | |
| *** ppiela has quit IRC | 15:25 | |
| *** mat128 has quit IRC | 15:25 | |
| *** jlvillal has quit IRC | 15:25 | |
| *** ohamada has quit IRC | 15:25 | |
| *** davidlenwell_ has quit IRC | 15:25 | |
| *** xavierr has quit IRC | 15:25 | |
| *** dhellmann has quit IRC | 15:25 | |
| *** sambetts has quit IRC | 15:25 | |
| *** pleia2 has quit IRC | 15:25 | |
| *** cameron_C has quit IRC | 15:25 | |
| *** kbyrne has quit IRC | 15:25 | |
| *** hamzy has quit IRC | 15:25 | |
| *** tonyb has quit IRC | 15:25 | |
| *** EmilienM has quit IRC | 15:25 | |
| *** sturivnyi has quit IRC | 15:25 | |
| *** pleia2_ is now known as pleia2 | 15:26 | |
| *** johnthetubaguy has quit IRC | 15:26 | |
| *** xavierr has joined #openstack-ironic | 15:26 | |
| *** dims has joined #openstack-ironic | 15:26 | |
| *** e0ne has joined #openstack-ironic | 15:26 | |
| openstackgerrit | Anton Arefiev proposed openstack/ironic-inspector: [WIP] Tempest: add basic test https://review.openstack.org/293926 | 15:26 |
| JayF | jroll: looking | 15:26 |
| *** dhellmann has joined #openstack-ironic | 15:26 | |
| *** amotoki has joined #openstack-ironic | 15:27 | |
| *** jlvillal_ is now known as jlvillal | 15:27 | |
| *** cdearborn has quit IRC | 15:27 | |
| *** ptx0 has joined #openstack-ironic | 15:27 | |
| JayF | jroll: mat128_ added a conductor side check that complains loudly if zeroize=false and iterations=0 | 15:27 |
| *** jlvillal is now known as Guest81062 | 15:27 | |
| JayF | jroll: IMO we don't need to scream in IPA logs as well, which nobody would see anyway | 15:28 |
| *** EmilienM has joined #openstack-ironic | 15:28 | |
| *** dprince has joined #openstack-ironic | 15:28 | |
| *** greghaynes has quit IRC | 15:28 | |
| *** mkovacik has quit IRC | 15:29 | |
| openstackgerrit | Kyrylo Romanenko proposed openstack/python-ironicclient: Add tests for baremetal subcommands in openstackclient https://review.openstack.org/301687 | 15:29 |
| *** ametts has joined #openstack-ironic | 15:29 | |
| *** ndipanov has joined #openstack-ironic | 15:30 | |
| *** ohamada has joined #openstack-ironic | 15:30 | |
| *** jmccrory has joined #openstack-ironic | 15:30 | |
| *** jistr has quit IRC | 15:30 | |
| *** rcernin has quit IRC | 15:30 | |
| *** derekh has joined #openstack-ironic | 15:30 | |
| *** hamzy has joined #openstack-ironic | 15:31 | |
| JayF | sambetts_: https://review.openstack.org/#/c/304578 I just +2'd this with a comment re: your concern, can you have another look? If you feel OK with that I'll aldn it | 15:31 |
| JayF | land | 15:31 |
| *** Guest63229 has quit IRC | 15:31 | |
| *** kbyrne has joined #openstack-ironic | 15:32 | |
| *** jayc has joined #openstack-ironic | 15:33 | |
| *** devananda has joined #openstack-ironic | 15:33 | |
| *** JoshNang has joined #openstack-ironic | 15:33 | |
| dtantsur | have a nice weekend everyone! | 15:34 |
| *** dtantsur is now known as dtantsur|afk | 15:35 | |
| NobodyCam | have a great weekend dtantsur|afk | 15:35 |
| vdrok | dtantsur|afk: you too! good night | 15:35 |
| *** cameron_C has joined #openstack-ironic | 15:36 | |
| *** [1]cdearborn has quit IRC | 15:36 | |
| *** lekha has joined #openstack-ironic | 15:37 | |
| *** [1]cdearborn has joined #openstack-ironic | 15:38 | |
| openstackgerrit | Kyrylo Romanenko proposed openstack/python-ironicclient: Add tests for openstackclient plugin https://review.openstack.org/301687 | 15:40 |
| *** jistr has joined #openstack-ironic | 15:42 | |
| *** johnthetubaguy has joined #openstack-ironic | 15:46 | |
| *** kozhukalov has joined #openstack-ironic | 15:47 | |
| *** greghaynes has joined #openstack-ironic | 15:47 | |
| *** vishwana_ has joined #openstack-ironic | 15:47 | |
| *** serverascode has joined #openstack-ironic | 15:47 | |
| *** vishwana_ is now known as vishwanathj_ | 15:48 | |
| *** vishwanathj has quit IRC | 15:50 | |
| *** rama_y has joined #openstack-ironic | 15:53 | |
| *** jistr is now known as jistr|afk | 16:00 | |
| *** mgoddard__ has joined #openstack-ironic | 16:01 | |
| *** mgoddard_ has quit IRC | 16:04 | |
| sambetts_ | JayF: Make sense to me, I was looking for a coresponding patch that might do it, but the bug report seems to not have updated with patch info :( | 16:07 |
| sambetts_ | o/ dtantsur|afk | 16:07 |
| JayF | sambetts_: yeah, something has been ... weird since they moved the gerrit server | 16:07 |
| JayF | sambetts_: like emails coming in late, bugs getting updated later than usual | 16:07 |
| *** sambetts_ is now known as sambetts | 16:07 | |
| *** jcoufal has quit IRC | 16:08 | |
| JayF | sambetts_: you wanna flip your vote then I'll alnd it? | 16:08 |
| sambetts | JayF: done :) | 16:08 |
| devananda | morning, all | 16:09 |
| JayF | gm | 16:09 |
| sambetts | hey devananda | 16:09 |
| *** jayc has quit IRC | 16:10 | |
| *** serverascode has quit IRC | 16:13 | |
| *** gabriel-bezerra has quit IRC | 16:13 | |
| *** serverascode has joined #openstack-ironic | 16:14 | |
| *** gabriel-bezerra has joined #openstack-ironic | 16:15 | |
| *** yarkot has quit IRC | 16:15 | |
| *** anush has quit IRC | 16:15 | |
| *** anush has joined #openstack-ironic | 16:16 | |
| *** sergek has quit IRC | 16:17 | |
| *** mbound has quit IRC | 16:17 | |
| *** yonglihe has quit IRC | 16:17 | |
| *** yarkot1 has quit IRC | 16:17 | |
| *** akrish9 has quit IRC | 16:17 | |
| *** agordeev has quit IRC | 16:17 | |
| vdrok | morning devananda and JayF | 16:19 |
| JayF | gm | 16:19 |
| *** agordeev has joined #openstack-ironic | 16:20 | |
| *** sergek has joined #openstack-ironic | 16:20 | |
| NobodyCam | morning devananda | 16:20 |
| *** yonglihe has joined #openstack-ironic | 16:20 | |
| vdrok | devananda: wdyt about https://review.openstack.org/306010 if you have some time? | 16:20 |
| NobodyCam | morning JayF | 16:20 |
| *** Sukhdev has joined #openstack-ironic | 16:23 | |
| *** amotoki has quit IRC | 16:24 | |
| *** jayc has joined #openstack-ironic | 16:24 | |
| *** achanda has joined #openstack-ironic | 16:26 | |
| *** kromanenko has quit IRC | 16:27 | |
| devananda | vdrok: interesting. what is the goal // what are you trying to fix? | 16:30 |
| *** achanda has quit IRC | 16:30 | |
| vdrok | devananda: after I do nova rebuild the first time, instance.node is being replaced by the first ironic node associated with nova compute, and subsequent rebuilds fail | 16:32 |
| *** jayc_ has joined #openstack-ironic | 16:32 | |
| devananda | oh | 16:33 |
| *** jayc_ has quit IRC | 16:33 | |
| devananda | ah, because of this? compute_node = self._get_compute_info(context, self.host) | 16:33 |
| vdrok | yep | 16:33 |
| devananda | ugh. yea, that's definitely not righ | 16:34 |
| *** ohamada has quit IRC | 16:34 | |
| devananda | it is assuming that self.host has only one hypervisor_hostname | 16:34 |
| vdrok | scheduled_node is passed in only in case of evacuate | 16:34 |
| devananda | which is not true for the nova.virt.ironic driver | 16:34 |
| vdrok | yeah | 16:34 |
| *** jayc has quit IRC | 16:34 | |
| vdrok | devananda: another fix I thought about it this one - https://review.openstack.org/#/c/306011/ | 16:34 |
| vdrok | but if done that way, migration is created in nova, which then creates troubles in update_available_resource | 16:35 |
| *** suro-patz has joined #openstack-ironic | 16:36 | |
| *** cfarquhar has quit IRC | 16:37 | |
| devananda | vdrok: that would also prevent an admin from using nova's force-hosts capabilities | 16:37 |
| devananda | actually, question | 16:37 |
| devananda | the passed "instance" parameter -- this is the current instance, which is to be rebuilt, isn't it? | 16:37 |
| vdrok | yes | 16:37 |
| devananda | you're forcing it to be rebuilt on the same ironic Node | 16:37 |
| devananda | which prevents evacuation or rebuilding on a different host from working -- even when that is desired | 16:38 |
| devananda | in this one, rebuild onto a new Node would work, because scheduled_node would not be None | 16:38 |
| devananda | https://review.openstack.org/#/c/306010/2/nova/compute/manager.py | 16:38 |
| devananda | so yea, that's better | 16:38 |
| vdrok | yep, thanks, will add tests and more details there | 16:39 |
| devananda | vdrok: also, reading the doc block on that method, I think you want it to be "if migration:" | 16:41 |
| *** cfarquhar has joined #openstack-ironic | 16:43 | |
| *** cfarquhar has quit IRC | 16:43 | |
| *** cfarquhar has joined #openstack-ironic | 16:43 | |
| *** harlowja has quit IRC | 16:44 | |
| *** harlowja has joined #openstack-ironic | 16:45 | |
| vdrok | devananda: I thought they are equivalent in our case? looking at compute.api.evacuate, migration is created there and recreate set to true | 16:46 |
| JayF | Is there some way to assert nothing happened at all? | 16:46 |
| JayF | Trying to write a test to confirm that nothing happens, and it's surprisingly hard | 16:47 |
| JayF | https://review.openstack.org/#/c/263842/12/ironic/drivers/modules/pxe.py line 521 | 16:47 |
| JayF | trying to test if CONF.pxe.manage_ramdisk_boot is false | 16:47 |
| JayF | but it literally just noops the whole method | 16:47 |
| JayF | is there some kind of "assert this did nothing whatsoever" trick I don't know about, lol | 16:47 |
| JayF | I guess I could just negative-check against the other path, that's what I'll do unless there is a better suggestion here | 16:48 |
| sambetts | JayF: assrt that dhcp_options_for_instance isn't called, but try_set_boot_device is called? | 16:48 |
| JayF | this is for clean_up_ramdisk | 16:48 |
| JayF | where literally I do nothing if it's set false | 16:49 |
| sambetts | assert _get_deploy_image_info is never called? | 16:49 |
| JayF | yeah, that's what I'm going to do | 16:49 |
| devananda | JayF: assert call count is 0 ? | 16:49 |
| devananda | yea | 16:49 |
| JayF | copy the test for the other settings, then assert all those calls didn't happen | 16:49 |
| *** trown is now known as trown|lunch | 16:51 | |
| JayF | I just kinda wish there was a assertNoop() or similar, but I guess that'd be kinda hard to write :) | 16:51 |
| *** jayc has joined #openstack-ironic | 16:52 | |
| sambetts | JayF: you'd have to assert that the stack frame didn't change or something *mind melt* | 16:53 |
| JayF | yeah, exactly | 16:53 |
| sambetts | devananda: I've got to head off but I'd like to hear your thoughts on the idea I left in the comments for the Driver comp reform spec https://review.openstack.org/#/c/188370/ when you get a moment, it sort of changes up the way we were thinking about it | 16:54 |
| *** anush has quit IRC | 16:55 | |
| *** anush has joined #openstack-ironic | 16:55 | |
| *** sambetts is now known as sambetts|afk | 16:56 | |
| *** Sukhdev has quit IRC | 16:57 | |
| *** Sukhdev has joined #openstack-ironic | 16:57 | |
| *** derekh has quit IRC | 16:58 | |
| *** piet has quit IRC | 16:59 | |
| *** jistr|afk is now known as jistr | 17:00 | |
| *** piet has joined #openstack-ironic | 17:00 | |
| Guest81062 | JayF, Do you have a link to actual tests to this point? | 17:03 |
| *** achanda has joined #openstack-ironic | 17:04 | |
| JayF | Guest81062: might wanna /nick :) | 17:04 |
| Guest81062 | Ah. darn | 17:04 |
| JayF | Guest81062: I fixed them, fixed them so well they're failing now | 17:04 |
| *** Guest81062 is now known as jlvillal | 17:04 | |
| JayF | so I have to find out why it's failing, lol | 17:04 |
| *** Sukhdev has quit IRC | 17:04 | |
| jlvillal | JayF, Okay back to being jlvillal | 17:05 |
| JayF | like the test is correct now; it's showing my code may be incorrect | 17:05 |
| JayF | so digging into that | 17:05 |
| JayF | although not that test in particular, another test bug ruby found | 17:05 |
| jlvillal | JayF, Okay. Happy to help if I can. Though in an all-day meeting which have to pay attention to for the most part. | 17:06 |
| JayF | I think I'm good for now, except for the fact my code is broken, lol | 17:06 |
| JayF | I think I'm going to simplify it, instead of trying to print the deprecation once per driver, just going to print it once overall | 17:07 |
| jlvillal | JayF, :) Okay. | 17:07 |
| *** achanda has quit IRC | 17:08 | |
| *** mgoddard__ has quit IRC | 17:10 | |
| *** mgoddard has joined #openstack-ironic | 17:10 | |
| devananda | sambetts|afk: will take a look, thanks | 17:11 |
| sambetts|afk | devananda: thanks | 17:12 |
| sambetts|afk | jlvillal: cisco-ironic-recheck is our new recheck command | 17:13 |
| * sambetts|afk is really AFK now :D | 17:13 | |
| *** yuanying_ has joined #openstack-ironic | 17:18 | |
| *** vdrok has quit IRC | 17:18 | |
| *** mat128_ has quit IRC | 17:18 | |
| *** nathan-h has quit IRC | 17:19 | |
| *** suro-patz1 has joined #openstack-ironic | 17:20 | |
| *** fyxim_ has quit IRC | 17:20 | |
| *** kbyrne has quit IRC | 17:21 | |
| *** daemontool__ has joined #openstack-ironic | 17:22 | |
| *** jmccrory has quit IRC | 17:22 | |
| *** pleia2 has quit IRC | 17:22 | |
| *** mat128 has joined #openstack-ironic | 17:23 | |
| *** yonglihe_ has joined #openstack-ironic | 17:23 | |
| *** boris-42 has quit IRC | 17:24 | |
| *** pleia2 has joined #openstack-ironic | 17:24 | |
| *** akrish9 has joined #openstack-ironic | 17:26 | |
| *** piet has quit IRC | 17:26 | |
| *** anush has quit IRC | 17:26 | |
| *** jayc has quit IRC | 17:26 | |
| *** cfarquhar has quit IRC | 17:26 | |
| *** suro-patz has quit IRC | 17:26 | |
| *** yonglihe has quit IRC | 17:26 | |
| *** sergek has quit IRC | 17:26 | |
| *** agordeev has quit IRC | 17:26 | |
| *** serverascode has quit IRC | 17:26 | |
| *** yonglihe_ is now known as yonglihe | 17:27 | |
| *** gabriel-bezerra has quit IRC | 17:27 | |
| *** hamzy has quit IRC | 17:27 | |
| *** ndipanov has quit IRC | 17:27 | |
| *** ametts has quit IRC | 17:27 | |
| *** EmilienM has quit IRC | 17:27 | |
| *** dhellmann has quit IRC | 17:27 | |
| *** ChrisMead has quit IRC | 17:27 | |
| *** sambetts|afk has quit IRC | 17:27 | |
| *** ppiela_ has quit IRC | 17:27 | |
| *** daemontool_ has quit IRC | 17:27 | |
| *** Goneri has quit IRC | 17:27 | |
| *** cppforlife_ has quit IRC | 17:27 | |
| *** alex_xu has quit IRC | 17:27 | |
| *** openstackgerrit has quit IRC | 17:27 | |
| *** bcornec has quit IRC | 17:27 | |
| *** smoriya has quit IRC | 17:27 | |
| *** yuanying has quit IRC | 17:27 | |
| *** Haomeng has quit IRC | 17:27 | |
| *** thrash has quit IRC | 17:27 | |
| *** EmilienM has joined #openstack-ironic | 17:29 | |
| *** bcornec has joined #openstack-ironic | 17:29 | |
| *** maurosr has quit IRC | 17:31 | |
| *** e0ne has quit IRC | 17:31 | |
| *** JayF has quit IRC | 17:31 | |
| *** sambetts has joined #openstack-ironic | 17:32 | |
| *** smoriya_afk has joined #openstack-ironic | 17:33 | |
| *** smoriya_afk is now known as smoriya | 17:34 | |
| *** ndipanov has joined #openstack-ironic | 17:34 | |
| *** b_bezak has joined #openstack-ironic | 17:34 | |
| *** david-lyle has quit IRC | 17:34 | |
| *** david-lyle has joined #openstack-ironic | 17:34 | |
| *** jayc has joined #openstack-ironic | 17:36 | |
| *** ppiela has joined #openstack-ironic | 17:36 | |
| *** jmccrory has joined #openstack-ironic | 17:37 | |
| *** maurosr has joined #openstack-ironic | 17:37 | |
| *** yarkot has joined #openstack-ironic | 17:37 | |
| *** b_bezak has quit IRC | 17:38 | |
| *** cfarquhar has joined #openstack-ironic | 17:39 | |
| *** cfarquhar has quit IRC | 17:39 | |
| *** cfarquhar has joined #openstack-ironic | 17:39 | |
| devananda | sambetts: comments posted | 17:39 |
| *** kbyrne has joined #openstack-ironic | 17:40 | |
| devananda | I like your idea, but am suggesting a different approach | 17:40 |
| *** anush has joined #openstack-ironic | 17:40 | |
| *** daemontool__ has quit IRC | 17:40 | |
| *** ChrisMead has joined #openstack-ironic | 17:40 | |
| *** alex_xu has joined #openstack-ironic | 17:41 | |
| *** sergek has joined #openstack-ironic | 17:42 | |
| *** ndipanov has quit IRC | 17:44 | |
| *** ametts has joined #openstack-ironic | 17:44 | |
| *** dhellmann has joined #openstack-ironic | 17:44 | |
| *** fyxim_ has joined #openstack-ironic | 17:44 | |
| *** JayF has joined #openstack-ironic | 17:45 | |
| *** JayF has joined #openstack-ironic | 17:45 | |
| mat128 | JayF: mock objects always have a .called, you could assert that this is false | 17:45 |
| *** Goneri has joined #openstack-ironic | 17:45 | |
| *** dhellmann_ has joined #openstack-ironic | 17:46 | |
| JayF | mat128: yeah; I alreayd pushed up a new patch with the test fixes | 17:47 |
| *** dhellmann_ has quit IRC | 17:47 | |
| JayF | just think it got eaten by the netsplit | 17:47 |
| *** openstackgerrit has joined #openstack-ironic | 17:48 | |
| *** agordeev has joined #openstack-ironic | 17:49 | |
| *** dhellmann has quit IRC | 17:50 | |
| *** trown|lunch is now known as trown | 17:50 | |
| *** hamzy has joined #openstack-ironic | 17:50 | |
| *** Sukhdev has joined #openstack-ironic | 17:51 | |
| *** serverascode has joined #openstack-ironic | 17:51 | |
| *** yarkot1 has joined #openstack-ironic | 17:51 | |
| *** jistr has quit IRC | 17:51 | |
| *** boris-42 has joined #openstack-ironic | 17:53 | |
| *** sergek has quit IRC | 17:54 | |
| *** agordeev has quit IRC | 17:54 | |
| *** dhellmann has joined #openstack-ironic | 17:55 | |
| *** yonglihe has quit IRC | 17:55 | |
| *** sergek has joined #openstack-ironic | 17:56 | |
| *** dhellmann has quit IRC | 17:57 | |
| *** raginbajin has quit IRC | 17:57 | |
| openstackgerrit | Gonéri Le Bouder proposed openstack/ironic: iscsi: wipe the disk before deployment https://review.openstack.org/303605 | 17:57 |
| *** dhellmann has joined #openstack-ironic | 17:57 | |
| *** ndipanov has joined #openstack-ironic | 17:58 | |
| jroll | dtantsur|afk: forgot to tell you, we can't have multiple release liaisons in newton, so I'll have to ack inspector releases | 17:58 |
| jroll | http://lists.openstack.org/pipermail/openstack-dev/2016-April/092459.html | 17:58 |
| *** agordeev has joined #openstack-ironic | 17:58 | |
| jroll | speaking of releases, I think I'd like to release some things next week before summit, to get us in the habit of frequent releases | 17:58 |
| *** vdrok has joined #openstack-ironic | 17:59 | |
| jroll | just throwing that out there :) | 17:59 |
| *** cppforlife_ has joined #openstack-ironic | 17:59 | |
| *** yonglihe has joined #openstack-ironic | 18:00 | |
| *** ravig_ has joined #openstack-ironic | 18:00 | |
| *** jayc has quit IRC | 18:01 | |
| *** causten__ has joined #openstack-ironic | 18:01 | |
| *** raginbajin has joined #openstack-ironic | 18:02 | |
| ravig_ | NobodyCam, TheJulia: Can we discuss on this spec: https://review.openstack.org/#/c/294995/ | 18:03 |
| *** Marga_ has joined #openstack-ironic | 18:03 | |
| *** vishwanathj has joined #openstack-ironic | 18:03 | |
| NobodyCam | ravig_: I'm on a conference call ATM. | 18:03 |
| devananda | jroll: ++ to frequent point releases | 18:04 |
| ravig_ | NobodyCam: Can you please let me know, whenever you are free | 18:04 |
| *** alex_xu has quit IRC | 18:04 | |
| *** causten_ has quit IRC | 18:04 | |
| devananda | jroll: also, I'd like to chat about client version stuff when you have time. perhaps at the summit, perhaps before | 18:05 |
| *** gabriel-bezerra has joined #openstack-ironic | 18:05 | |
| *** alex_xu has joined #openstack-ironic | 18:07 | |
| *** vishwanathj_ has quit IRC | 18:07 | |
| vdrok | good night everyone! | 18:08 |
| NobodyCam | night vdrok | 18:08 |
| *** Goneri has quit IRC | 18:09 | |
| TheJulia | ravig_: NobodyCam and I should be off the phone in about 45 minutes | 18:13 |
| ravig_ | TheJulia: Cool thanks :) | 18:14 |
| *** robcresswell has quit IRC | 18:14 | |
| *** robcresswell has joined #openstack-ironic | 18:15 | |
| *** e0ne has joined #openstack-ironic | 18:19 | |
| *** thrash has joined #openstack-ironic | 18:23 | |
| *** thrash has quit IRC | 18:23 | |
| *** thrash has joined #openstack-ironic | 18:23 | |
| JayF | https://github.com/openstack/nova/blob/master/nova/virt/ironic/driver.py#L301 is there a reason we don't check for node_resources_unavailable() before node_resources_used() in the nova driver/ | 18:24 |
| JayF | I think it'd resolve a bug we're seeing downstream | 18:24 |
| JayF | https://bugs.launchpad.net/nova/+bug/1503453 bug to be specific | 18:24 |
| openstack | Launchpad bug 1503453 in OpenStack Compute (nova) "unavailable ironic nodes being scheduled to" [Medium,Triaged] | 18:24 |
| JayF | jroll: ^ wdyt | 18:24 |
| JayF | Madasi: ^ that's your bug iirc | 18:25 |
| *** sinval has joined #openstack-ironic | 18:26 | |
| Madasi | that's a good question | 18:29 |
| Madasi | nothing jumps out at me as obvious why that wouldn't work | 18:29 |
| Madasi | hmm, the code has changed enough that the reference from the bug isn't right anymore. Should have linked to a specific commit instead of master I suppose | 18:32 |
| Madasi | that, or it's been so long since I looked at that code that is doesn't look familiar anymore. Not sure which | 18:34 |
| *** Sukhdev has quit IRC | 18:37 | |
| Madasi | JayF: I think swapping the order might catch cases where the orphaned node is also in maintenance, but not others | 18:38 |
| Madasi | heh, I actually considered that | 18:38 |
| Madasi | https://bugs.launchpad.net/nova/+bug/1503453/comments/1 | 18:38 |
| openstack | Launchpad bug 1503453 in OpenStack Compute (nova) "unavailable ironic nodes being scheduled to" [Medium,Triaged] | 18:38 |
| JayF | Madasi: what's the harm in -$total_ram / 0 | 18:39 |
| JayF | Madasi: does it actually try to provision it back to 0? | 18:40 |
| Madasi | I think the concern was that swapping the order might also cause https://bugs.launchpad.net/bugs/1502177 | 18:40 |
| openstack | Launchpad bug 1502177 in OpenStack Compute (nova) "Existing Ironic instances report negative available RAM for the node after upgrade " [High,Fix released] - Assigned to Jim Rollenhagen (jim-rollenhagen) | 18:40 |
| JayF | lame :( | 18:40 |
| JayF | jcook: ^ | 18:40 |
| Madasi | not following where the / 0 comes in | 18:41 |
| jcook | Nova won't see an instance on the node if it's an orphaned instance uuid | 18:42 |
| openstackgerrit | Andre Aranha proposed openstack/python-oneviewclient: Extend python-oneviewclient to Uplink Set https://review.openstack.org/295981 | 18:43 |
| Madasi | right, and that causes nova to "fix" the reported resources available, and try to schedule to that node, which ironic will fail b/c it sees an instance already there | 18:43 |
| anteaya | does anyone know if any ironic summit sessions with be discussing third party ci? | 18:44 |
| anteaya | I didn't see any on the ironic schedule | 18:44 |
| jcook | Madasi: it'll set the available and used? I thought only the available | 18:44 |
| jcook | Madasi: oh then it'll say it's using more than available and thusly total negative? | 18:44 |
| openstackgerrit | Andre Aranha proposed openstack/python-oneviewclient: Extended client to modify Connections https://review.openstack.org/302365 | 18:45 |
| jcook | Madasi: wouldn't it be more appropriate to set available to min of memory_mb and size of server? | 18:46 |
| TheJulia | ravig_: So, brains are now available, whats up? | 18:46 |
| anteaya | I'm trying to sort out my schedule and at present only nova seems to have a session where they plan on discussing third party ci and i just don't believe it | 18:46 |
| anteaya | that they would be the only project | 18:46 |
| NobodyCam | Yes /me is also back | 18:47 |
| Madasi | jcook: I think the usage gets changed here - https://github.com/openstack/nova/blob/a3da334656bac94967af58d1b9a95dbb26ebf44b/nova/compute/resource_tracker.py#L823-L824 | 18:48 |
| Madasi | but it's been quite a while, I'd have to sit down and try to wrap my head around the code again to give good answers | 18:48 |
| jcook | Madasi: memory is set to 0 for error states? | 18:49 |
| ravig_ | TheJulia: I have a question regarding the https://review.openstack.org/#/c/294995/ : Regarding the scenario 3, where we are booting from rbd, are we planning to have rbdroot parameter in kernel parameters in pxefile | 18:49 |
| jcook | Madasi: neutron_failed == no schedule, neutron_failed with instance uuid == schedule; neutron_failed is preventing build alone by, I'm guessing, setting memory to 0 | 18:50 |
| *** Goneri has joined #openstack-ironic | 18:50 | |
| TheJulia | ravig_: That was how I anticipated it working | 18:50 |
| TheJulia | ravig_: since the only other way I think that could work would be to assemble ramdisk to bootstrap the machine into using a value burned into the ramdisk | 18:51 |
| Madasi | jcook: by nova or by ironic? ironic does it here for error https://github.com/openstack/nova/blob/master/nova/virt/ironic/driver.py#L181 | 18:51 |
| jcook | Madasi: also just set this to min(0, self.compute_node.memory_md - self.compute_node.memory_mb_used): https://github.com/openstack/nova/blob/a3da334656bac94967af58d1b9a95dbb26ebf44b/nova/compute/resource_tracker.py#L845 | 18:51 |
| jcook | Madasi: what you just linked | 18:52 |
| Madasi | good luck getting nova to go along with that change :) | 18:52 |
| Madasi | and i'm not positive it fixes both cases | 18:52 |
| Madasi | would have to think about it more | 18:52 |
| jcook | either that or set self.compute_node.memory_mb_used to max(self.compute_node.memory_mb_used, self.compute_node.memory_mb) | 18:53 |
| Madasi | wouldn't that always show all the ram used? | 18:54 |
| ravig_ | Since we rely it on kernel parameter, wouldn't that cause problem - recently we faced a problem, when we are using RHEL with syslinux as bootloader after the update, it stopped updating few parameters | 18:55 |
| jcook | sorry, that should be a min | 18:55 |
| anteaya | NobodyCam: do you know of any ironic design summit sessions that will be talking about third party ci? | 18:55 |
| Madasi | I suspect that would break several things around reserved memory for hypervisor overhead in nova | 18:56 |
| *** Goneri has quit IRC | 18:56 | |
| Madasi | but might be wrong | 18:56 |
| *** ametts has quit IRC | 18:56 | |
| jcook | Madasi: what does memory look like for error state? | 18:56 |
| jcook | available = 0 used = 0? | 18:56 |
| *** dprince has quit IRC | 18:57 | |
| TheJulia | ravig_: So, the downside of pxe booting is that the kernel/initial ramdisk may go out of sync what is on the actual "disk" as time goes on, could that be the cause of what you experienced? I'm only thinking that we wouldn't be loading another bootloader really, we would be loading the intended kernel and ramdisk, and wouldn't be able to try and read the ceph volume to load the bootloader that m | 18:57 |
| ravig_ | TheJulia: Instead can we update the bootdisk name in ipxe file. Do you think it makes sense - But this may cause one ipxe file for each node... | 18:57 |
| TheJulia | ay be on the volume | 18:57 |
| jcook | Madasi: https://github.com/openstack/nova/blob/master/nova/virt/ironic/driver.py#L313 memory_mb is definitely 0 | 18:57 |
| TheJulia | ravig_: there would be an ipxe file per noe anyway | 18:57 |
| TheJulia | s/noe/node/ | 18:57 |
| jcook | not sure what used is | 18:57 |
| NobodyCam | anteaya: not specially on that topic. here is the planning ether pad: https://etherpad.openstack.org/p/ironic-newton-summit | 18:58 |
| anteaya | NobodyCam: thank you | 18:58 |
| Madasi | if there is an instance, we will return all the memory as used, regardless of error state | 18:58 |
| jcook | Madasi: so that negative RAM scenario is true right now for errored nodes? | 18:59 |
| Madasi | if there isn't an instance, then we look at maintenance, power state, and provision state and report 0 present for the node | 18:59 |
| anteaya | NobodyCam: apparently it is an item on the Gate session | 19:00 |
| Madasi | jcook: no, was only true when we switched over from looking at node resources to instance_info resources, and only for pre-existing builds | 19:00 |
| NobodyCam | ah yes that does have a section fot 3rd party test | 19:01 |
| Madasi | but the concern was the swapping to order of checks from instance and then error/maintenance state to the other way around could introduce the same effect | 19:01 |
| Madasi | where we report 0 resources present for the node, the nova sees the instance and subtracts what it is using from the present amount, causing -instance size resources available | 19:01 |
| anteaya | NobodyCam: thank you | 19:01 |
| anteaya | I figured you would have something, I just didn't want to find it at the last minute | 19:02 |
| NobodyCam | anteaya: see the linked ehterpad in 3.1.6.3 | 19:02 |
| JayF | Madasi: I wonder if we could get the 0 resources available special-cased in the generic RT code | 19:02 |
| JayF | Madasi: that seems like an easier change to get in than changing the driver api signatures | 19:02 |
| Madasi | JayF: how? by having nova check ironic states? | 19:03 |
| jcook | Madasi, JayF: well, I think then putting a min in here might be the right way to handle it: https://github.com/openstack/nova/blob/a3da334656bac94967af58d1b9a95dbb26ebf44b/nova/compute/resource_tracker.py#L845 | 19:03 |
| JayF | yep | 19:03 |
| anteaya | NobodyCam: yeah that was from Tokyo | 19:03 |
| anteaya | NobodyCam: thanks | 19:03 |
| JayF | Madasi: jcook ^ just suggested what I was suggesting, in a more understandable way | 19:03 |
| JayF | Madasi: tl;dr: if self.compute_node.memory_mb == 0; don't erase that stuff, dammit! | 19:03 |
| JayF | :) | 19:03 |
| JayF | Madasi: or making _used min(0) | 19:04 |
| ravig_ | TheJulia: Let me talk through the scenario. Please correct me if I am wrong - iPXE loads kernel and initramdisk with some parameters related to rbdroot and rbd volume will be booted from sanboot or sanhook | 19:04 |
| Madasi | wouldn't a min always return 0 or negative? | 19:04 |
| Madasi | I'm not following your logic here | 19:04 |
| jcook | min(0, memory - used) | 19:04 |
| jcook | erm max | 19:04 |
| TheJulia | ravig_: ipxe does not know about rdb volumes, so it wouldn't use sanboot | 19:04 |
| JayF | make 0 the minimum that value can ever be set to | 19:05 |
| JayF | which I guess means using max like jcook said | 19:05 |
| jcook | yeah, w/e that code is lol | 19:05 |
| JayF | that code is like, the source of 50% of our scheduling errors | 19:05 |
| Madasi | yeah, that makes more sense, essentially just setting 0 as the floor | 19:05 |
| jcook | yes | 19:05 |
| JayF | I mean, lets try to get that change in then? | 19:05 |
| jcook | +1 | 19:05 |
| Madasi | but that assumes nobody else is using that for something special, but who knows | 19:05 |
| Madasi | that seems like it would be silly | 19:05 |
| JayF | I mean, Madasi, let code review handle that | 19:05 |
| jcook | subtracting and getting a negative number seems wrong in this case, and if it's being used like that it's a hack | 19:06 |
| JayF | I mean, does someone wanna push that up? I'll give jcook or Madasi the chance since you all had the idea / wrote the original bug | 19:06 |
| Madasi | but not sure that setting 0 as a min fixes the orphan problem | 19:06 |
| JayF | Madasi: you swap the checks AND force that as a min | 19:06 |
| Madasi | ah, ok | 19:06 |
| jcook | it doesn't fix the orphan it fixes scheduling to an orphan | 19:06 |
| Madasi | so a cross project code fix :) | 19:07 |
| *** Sukhdev has joined #openstack-ironic | 19:07 | |
| JayF | no | 19:07 |
| JayF | the ironic driver is in nova | 19:07 |
| ravig_ | TheJulia: So, thats what I was wondering too, so in a way, will it be inside the OS's bootloader which is on rbd volume? | 19:07 |
| Madasi | oh, true | 19:07 |
| JayF | so the whole patch would be in nova | 19:07 |
| Madasi | forgot that for a min | 19:07 |
| TheJulia | ravig_: It loads the kernel into ram along with the ramdisk and initiates boot of those two items, the command line parameter that gets passed along to the kernel is read by the ramdisk, and it's initscripts, if they support it, would then mount the rdb volume as root | 19:07 |
| JayF | Madasi: giving you the chance to write the upstream patch, I'll help with review comments if you can get it up | 19:07 |
| TheJulia | ravig_:So, the RDB volume's bootloader, if it has one, cannot be used, only the ramdisk can bridge the gap for instance, if that makes sense | 19:08 |
| Madasi | Would like to, but I won't have bandwidth for that for a while | 19:09 |
| *** lennyb has quit IRC | 19:09 | |
| jcook | JayF: I'll make the change | 19:09 |
| *** afaranha has quit IRC | 19:10 | |
| TheJulia | ravig_: sorry, RBD | 19:10 |
| JayF | jcook: awesome; I'll offer you the same deal, if you want I can help fixup after review comments and all | 19:11 |
| jcook | JayF: cool, I'll try to knock that out right mewo | 19:12 |
| jcook | meow* | 19:12 |
| TheJulia | speaking of cats... *looks down to an angry cat in her lap* | 19:13 |
| *** mgoddard has quit IRC | 19:14 | |
| *** keedya has joined #openstack-ironic | 19:15 | |
| *** rama_y has quit IRC | 19:15 | |
| *** sinval has quit IRC | 19:15 | |
| *** aginwala has joined #openstack-ironic | 19:17 | |
| *** anush has quit IRC | 19:18 | |
| *** akrish9 is now known as anush | 19:18 | |
| *** Marga__ has joined #openstack-ironic | 19:19 | |
| *** aginwala_ has joined #openstack-ironic | 19:19 | |
| *** suro-patz has joined #openstack-ironic | 19:20 | |
| *** serverascode_ has joined #openstack-ironic | 19:22 | |
| *** bcornec1 has joined #openstack-ironic | 19:22 | |
| *** serverascode has quit IRC | 19:22 | |
| *** e0ne has quit IRC | 19:22 | |
| *** ChrisMead has quit IRC | 19:22 | |
| *** yarkot has quit IRC | 19:22 | |
| *** maurosr has quit IRC | 19:22 | |
| *** bcornec has quit IRC | 19:22 | |
| *** suro-patz1 has quit IRC | 19:22 | |
| *** maurosr has joined #openstack-ironic | 19:22 | |
| *** Marga_ has quit IRC | 19:22 | |
| *** aginwala has quit IRC | 19:22 | |
| *** aginwala_ has quit IRC | 19:23 | |
| *** maurosr has quit IRC | 19:23 | |
| *** maurosr has joined #openstack-ironic | 19:23 | |
| *** serverascode_ is now known as serverascode | 19:23 | |
| *** e0ne has joined #openstack-ironic | 19:24 | |
| *** piet has joined #openstack-ironic | 19:24 | |
| *** xavierr has quit IRC | 19:25 | |
| *** aginwala has joined #openstack-ironic | 19:27 | |
| *** raildo is now known as raildo-afk | 19:27 | |
| *** ChrisMead has joined #openstack-ironic | 19:31 | |
| *** e0ne has quit IRC | 19:31 | |
| jcook | JayF: is there an existing bug somewhere for this in launchpad? | 19:32 |
| JayF | jcook: I linked it above | 19:32 |
| JayF | 1503453 | 19:33 |
| JayF | https://bugs.launchpad.net/nova/+bug/1503453 bug to be specific | 19:33 |
| openstack | Launchpad bug 1503453 in OpenStack Compute (nova) "unavailable ironic nodes being scheduled to" [Medium,Triaged] | 19:33 |
| JayF | there's the full url | 19:33 |
| *** ravig_ has quit IRC | 19:33 | |
| * JayF afk for a couple of hours | 19:34 | |
| jcook | got it thanks | 19:35 |
| *** Sukhdev has quit IRC | 19:36 | |
| *** ametts has joined #openstack-ironic | 19:42 | |
| *** achanda has joined #openstack-ironic | 19:42 | |
| *** baoli has quit IRC | 19:45 | |
| *** aginwala has quit IRC | 19:48 | |
| *** piet has quit IRC | 19:50 | |
| NobodyCam | vdrok: are you by chance look'n to that time out issue ? | 19:52 |
| *** rloo has joined #openstack-ironic | 19:52 | |
| *** aginwala has joined #openstack-ironic | 19:53 | |
| *** raildo-afk is now known as raildo | 19:55 | |
| *** Goneri has joined #openstack-ironic | 19:55 | |
| *** aginwala has quit IRC | 20:01 | |
| *** cfarquhar has quit IRC | 20:03 | |
| *** aginwala has joined #openstack-ironic | 20:04 | |
| *** aginwala has quit IRC | 20:04 | |
| *** aginwala has joined #openstack-ironic | 20:05 | |
| *** achanda has quit IRC | 20:05 | |
| *** cfarquhar has joined #openstack-ironic | 20:06 | |
| *** cfarquhar has quit IRC | 20:06 | |
| *** cfarquhar has joined #openstack-ironic | 20:06 | |
| *** baoli has joined #openstack-ironic | 20:07 | |
| *** vishwana_ has joined #openstack-ironic | 20:09 | |
| *** vishwanathj has quit IRC | 20:11 | |
| *** raildo is now known as raildo-afk | 20:11 | |
| *** suro-patz has quit IRC | 20:12 | |
| *** raildo-afk is now known as raildo | 20:13 | |
| *** vishwanathj has joined #openstack-ironic | 20:15 | |
| *** vishwanathj has quit IRC | 20:15 | |
| *** vishwanathj has joined #openstack-ironic | 20:16 | |
| *** vishwanathj has quit IRC | 20:16 | |
| *** vishwanathj has joined #openstack-ironic | 20:17 | |
| *** vishwana_ has quit IRC | 20:20 | |
| *** Nisha_away has joined #openstack-ironic | 20:20 | |
| *** aginwala has quit IRC | 20:20 | |
| *** aginwala has joined #openstack-ironic | 20:21 | |
| *** ChrisMead has quit IRC | 20:21 | |
| *** aginwala has quit IRC | 20:22 | |
| jroll | jcook: JayF: sounds like y'all got your question sorted-ish? | 20:23 |
| *** aginwala has joined #openstack-ironic | 20:23 | |
| jroll | I only skimmed scrollback | 20:23 |
| jcook | jroll: well, I think there is an existing bug | 20:24 |
| *** Nisha_away has quit IRC | 20:24 | |
| *** ChrisMead has joined #openstack-ironic | 20:24 | |
| *** trown is now known as trown|outtypewww | 20:24 | |
| jroll | jcook: sure, I get that | 20:25 |
| jcook | jroll: if you maintenance nodes with instances on them, you'll have memory = 0 and used = all | 20:25 |
| jcook | jroll: no, I mean another bug in this code | 20:25 |
| *** aginwala has quit IRC | 20:25 | |
| jroll | jcook: does that cause problems or is just incorrect> | 20:26 |
| jroll | ? | 20:26 |
| jcook | I *think* you could end up with negative memory just like https://bugs.launchpad.net/nova/+bug/1502177 | 20:26 |
| openstack | Launchpad bug 1502177 in OpenStack Compute (nova) "Existing Ironic instances report negative available RAM for the node after upgrade " [High,Fix released] - Assigned to Jim Rollenhagen (jim-rollenhagen) | 20:26 |
| *** aginwala has joined #openstack-ironic | 20:27 | |
| jcook | It looks like the resource tracker takes the total memory and then subtracts every instance from it (+ overheads which for ironic should generally be 0, I would think) | 20:27 |
| jroll | oooo | 20:27 |
| jroll | and if your total for all nodes is negative it gets real problematic | 20:27 |
| jcook | I was trying to avoid doing that with the fix for scheduling to orphan instances, but I think it's already possible | 20:27 |
| jroll | so the original question is what if we swap the used/unavailable check | 20:27 |
| jroll | which makes sense at a glance | 20:28 |
| jroll | yeah, sounds like it | 20:28 |
| jcook | well, I just droped the 'el' from 'elif', but yeah | 20:28 |
| jcook | dropped* | 20:28 |
| jroll | mmm | 20:28 |
| *** mkovacik has joined #openstack-ironic | 20:28 | |
| jcook | so, I'm still looking at it | 20:28 |
| jroll | cool | 20:29 |
| jroll | this is easy to test :) | 20:29 |
| *** Goneri has quit IRC | 20:29 | |
| jcook | I'm not positive where total memory is added up, but it looks like it's pulled from db | 20:29 |
| *** ijw has joined #openstack-ironic | 20:29 | |
| *** achanda has joined #openstack-ironic | 20:29 | |
| NobodyCam | so fyi looking into the timeout issue I'm seeing things come to a strange end here https://github.com/openstack/ironic/blob/master/ironic_tempest_plugin/tests/scenario/test_baremetal_basic_ops.py#L78 at least in this test run: http://logs.openstack.org/58/272658/38/check/gate-tempest-dsvm-ironic-pxe_ssh/9462e59/console.html#_2016-04-15_15_17_49_950 | 20:30 |
| jroll | jcook: well, I'm not sure if a regular deployment of nova+ironic would be broken here, but in a cells configuration, it would drop at the cells level when it sees that cell has no capacity | 20:31 |
| jroll | NobodyCam: oooooo | 20:32 |
| jroll | vdrok mentinoed something similar earlier | 20:32 |
| jcook | in ironic the compute memory == the memory of every node under that compute? | 20:33 |
| jroll | NobodyCam: so it's somewhere here: https://github.com/openstack/ironic/blob/master/ironic_tempest_plugin/tests/scenario/test_baremetal_basic_ops.py#L122-L126 | 20:33 |
| jroll | jcook: erm, "it depends" | 20:33 |
| *** aginwala has quit IRC | 20:34 | |
| vdrok | NobodyCam: nope I have not looked yet, but it seems to be something during SSH command run/initialization | 20:34 |
| jcook | the sum of the reported by every node* | 20:34 |
| *** alex_xu has quit IRC | 20:34 | |
| jroll | jcook: so a resource tracker object is actually created for every (host, node) combination (where host == compute service, node == "compute node" e.g. ironic node) | 20:34 |
| *** achanda has quit IRC | 20:34 | |
| jroll | jcook: and there's a row in the compute_nodes table for every (host, node) combination | 20:34 |
| jcook | o_o | 20:34 |
| jroll | which are iterated over during scheduling | 20:35 |
| jroll | yeah dude, nova be cray | 20:35 |
| jcook | that's super helpful | 20:35 |
| *** ijw has quit IRC | 20:35 | |
| jcook | and answers a huge '?' I had | 20:35 |
| jroll | heh, yeah it was enlightening when I learned that | 20:35 |
| *** suro-patz has joined #openstack-ironic | 20:35 | |
| jcook | cool, so this makes this waaaaaaay easier to fix | 20:35 |
| *** alex_xu has joined #openstack-ironic | 20:36 | |
| jroll | jcook: just remember in cells, it does evaluate the sum of all nodes for choosing a cell to schedule to | 20:36 |
| *** mkovacik_ has joined #openstack-ironic | 20:36 | |
| jcook | jroll: resource tracker runs at global level too? | 20:37 |
| *** mkovacik has quit IRC | 20:37 | |
| jcook | jroll: or just in the scheduler? | 20:37 |
| jcook | erm | 20:37 |
| jcook | nova-cells global not scheduler | 20:38 |
| jroll | jcook: nope, only in computes. RT is what populates compute_nodes table. scheduler and cells services just read that table | 20:38 |
| jcook | so nova cells does a sum(memory) - sum(used)? | 20:38 |
| jcook | I'm guessing in the filters | 20:38 |
| jroll | so child cells service reads that table for totals, and passes up a dict of {ram_for_flavor: number_of_slots, (continue for each flavor)} | 20:39 |
| jroll | and global cells just uses number of slots for the requested flavor in each cell, to decide which cell | 20:39 |
| jcook | I'm familiar with that bit | 20:40 |
| jroll | https://github.com/openstack/nova/blob/master/nova/cells/state.py#L209 | 20:40 |
| jroll | right | 20:40 |
| jcook | me trying to think how that is calculated for ironic | 20:40 |
| jroll | here's the code that sums things: https://github.com/openstack/nova/blob/master/nova/cells/state.py#L256 | 20:40 |
| jroll | that iterates over each row in compute_nodes | 20:41 |
| jcook | ah | 20:41 |
| jcook | so *maybe* that needs a change | 20:41 |
| *** baoli has quit IRC | 20:41 | |
| jroll | I guess I'm lost as to what bug you're fixing at this point :P | 20:41 |
| jcook | so if a node has an instance it reports all used | 20:42 |
| jcook | and if it's in a "bad" state i returns none available | 20:42 |
| jcook | 0 - used = negative number | 20:42 |
| jcook | Ironic uses bad state to prevent scheduling to it | 20:43 |
| NobodyCam | vdrok: see mine and jroll's comments just above | 20:43 |
| jroll | jcook: right, so you want to make it such that it reports x total, x used, 0 available? | 20:44 |
| jcook | ultimately I want it not to schedule orphans :P | 20:44 |
| jroll | aha | 20:44 |
| jroll | orphan being ironic thinks it has an instance, but it does not? | 20:44 |
| jcook | yeah | 20:45 |
| jroll | right, so this is the really fun part | 20:45 |
| jcook | reservations | 20:45 |
| jcook | ? | 20:45 |
| jroll | if we tell the resource tracker that the resources are used | 20:45 |
| jroll | but there is no instance in nova related to that node | 20:46 |
| jcook | it "fixes" it | 20:46 |
| jcook | yeah.. | 20:46 |
| jroll | it will correct us and say there are ... yeah | 20:46 |
| jroll | there is a virt driver method to help correct this | 20:46 |
| jroll | I forget what it's called | 20:47 |
| jroll | but it does not take node into account, only host | 20:47 |
| jroll | and so is completely useless to us | 20:47 |
| jroll | and nova is not interested in adding node to this | 20:47 |
| jcook | yeah, I heard a bit about that >_> | 20:48 |
| jroll | Madasi originally tracked that down | 20:48 |
| jcook | I *think* I *might* know how to fix this **maybe** | 20:48 |
| jcook | without changing apis etc | 20:48 |
| jroll | I'd be very curious to see that :) | 20:48 |
| jcook | we'll see :) | 20:48 |
| *** suro-patz1 has joined #openstack-ironic | 20:49 | |
| *** jjohnson2_ has quit IRC | 20:49 | |
| *** suro-patz has quit IRC | 20:49 | |
| *** jjohnson2_ has joined #openstack-ironic | 20:50 | |
| Madasi | jcook: are you saying you think maintenanced nodes w/ instance could have negative ram reported now, or after swapping that if/elif logic? | 20:55 |
| Madasi | they shouldn't now | 20:55 |
| *** Sukhdev has joined #openstack-ironic | 20:56 | |
| NobodyCam | ah ha | 20:56 |
| NobodyCam | jroll: vdrok: TheJulia: http://paste.openstack.org/show/WGdfWq34V3KQFtvLOXIh/ <- this could be what's going on | 20:57 |
| *** ametts has quit IRC | 20:58 | |
| NobodyCam | this is the line we are failing on: https://github.com/openstack/ironic/blob/master/ironic_tempest_plugin/tests/scenario/test_baremetal_basic_ops.py#L78 | 20:58 |
| *** lascii is now known as alaski | 21:01 | |
| jroll | NobodyCam: why would that make it hang, though? are we sure client.exec_command uses os.system under the hood? | 21:01 |
| NobodyCam | I am not sure was just testing things | 21:02 |
| jroll | gotcha | 21:02 |
| jroll | I'm going to poke my head in -qa as well | 21:03 |
| *** ijw has joined #openstack-ironic | 21:03 | |
| NobodyCam | :) | 21:04 |
| *** jjohnson2_ has quit IRC | 21:06 | |
| jcook | Madasi: when I started convo, I thought now; but yeah after changing that logic | 21:11 |
| Madasi | ah, yeah | 21:11 |
| jcook | Madasi: I'm gonna ponder while I push my little one on the swing. I think I know how to fix. | 21:12 |
| Madasi | that was the thing that made just swapping the order not an easy fix. The floor to the available ram should mitigate that though | 21:12 |
| jcook | yeah, I think that might be sufficient, wanna make sure though | 21:12 |
| Madasi | but I'm not sure if there are any other consequences from marking 0 ram on nodes with instances that we don't know about yet | 21:12 |
| Madasi | but yeah, it's certainly farther along to a fix than I got | 21:13 |
| jcook | we'll see | 21:13 |
| Madasi | and doesn't take an api change, which is great | 21:13 |
| jcook | yeah +1 to that | 21:13 |
| * JayF glad to have a manager that can be nerd-sniped into fixing longstanding bugs | 21:13 | |
| Madasi | JayF: definitely a feature | 21:14 |
| JayF | reminder: if you're going to the summit and want to be a part of group hangouts for ironic, etc, please add your info here: https://etherpad.openstack.org/p/ironic-summit-arrivals | 21:15 |
| Madasi | If I wasn't so busy standing up new servers, I'd have loved to run with it, but it's better to get it done than wait around for me to get free time | 21:15 |
| Madasi | Going to have to miss the mini pub crawl on Sunday, a bit bummed about that | 21:15 |
| JayF | I basically get in at 1am monday mornign / sunday night | 21:16 |
| JayF | but in exchange I got a nonstop flight both ways | 21:16 |
| JayF | #worthit | 21:16 |
| Madasi | may have to make time to hit up austin eastciders on my own time | 21:16 |
| Madasi | nonstops are certainly nice | 21:16 |
| Madasi | i must agree | 21:16 |
| Madasi | hmm, should maybe make a trip to Meridian Hive Meadery too. Too many good places I wouldn't mind hitting up | 21:17 |
| Madasi | which is ironic, since I didn't go to them while I lived there :) | 21:18 |
| JayF | no, that's not ironic | 21:18 |
| JayF | ironic is software for provisioning bare metal nodes as part of an openstack cloud | 21:18 |
| JayF | or a great song without any irony (see, it's ironic?) from the 90s | 21:18 |
| *** raildo is now known as raildo-afk | 21:18 | |
| JayF | jcook: paranoid about spam much? lol | 21:20 |
| jcook | yes | 21:20 |
| jcook | :P | 21:20 |
| jroll | it blows my mind that people using gmail are worried about spam :P | 21:21 |
| jroll | filters are pretty good these days | 21:21 |
| jroll | NobodyCam: so turns out that create_timestamp() thing calls sync and umount | 21:22 |
| jroll | which is probably what's hanging | 21:22 |
| jroll | but disk io looks okay on that particular run, unless there's a neighbor crushing the disks | 21:22 |
| openstackgerrit | Chris Krelle proposed openstack/ironic: Test fix for tests https://review.openstack.org/306634 | 21:22 |
| NobodyCam | jroll: ahh | 21:22 |
| jroll | I need to head out for now, have a good weekend y'all | 21:30 |
| NobodyCam | night jroll | 21:31 |
| Madasi | o/ | 21:32 |
| *** lenny has joined #openstack-ironic | 21:32 | |
| devananda | g'night, jroll o/ | 21:32 |
| *** Sukhdev has quit IRC | 21:52 | |
| *** piet has joined #openstack-ironic | 21:53 | |
| *** raginbajin has quit IRC | 22:00 | |
| *** vishwanathj has quit IRC | 22:01 | |
| *** raginbajin has joined #openstack-ironic | 22:05 | |
| *** jjohnson2_ has joined #openstack-ironic | 22:06 | |
| *** dhellmann has left #openstack-ironic | 22:07 | |
| *** alex_xu has quit IRC | 22:13 | |
| *** alex_xu has joined #openstack-ironic | 22:15 | |
| *** KennethWilke has quit IRC | 22:16 | |
| *** karimb has quit IRC | 22:16 | |
| *** KennethWilke has joined #openstack-ironic | 22:17 | |
| *** odyssey4me has quit IRC | 22:22 | |
| *** rloo has quit IRC | 22:25 | |
| *** odyssey4me has joined #openstack-ironic | 22:27 | |
| *** piet has quit IRC | 22:30 | |
| *** baoli has joined #openstack-ironic | 22:30 | |
| *** jjohnson2_ has quit IRC | 22:41 | |
| *** baoli has quit IRC | 22:48 | |
| *** suro-patz1 has quit IRC | 22:49 | |
| jlvillal | NobodyCam: What kind of bug are you seeing in the test? | 22:52 |
| jlvillal | Seems like exec_command should already be returning a string. | 22:52 |
| NobodyCam | hey hey jlvillal | 22:52 |
| * jlvillal is just being curious | 22:53 | |
| NobodyCam | ya I think jroll found the real answer | 22:53 |
| jlvillal | Okay good :) | 22:53 |
| NobodyCam | see his comment at 2:22 pst | 22:53 |
| jlvillal | NobodyCam: Thanks | 22:53 |
| *** Sukhdev has joined #openstack-ironic | 22:55 | |
| *** suro-patz has joined #openstack-ironic | 23:04 | |
| *** ppiela has quit IRC | 23:12 | |
| *** [1]cdearborn has quit IRC | 23:14 | |
| *** Sukhdev has quit IRC | 23:25 | |
| *** davideag_ has quit IRC | 23:31 | |
| *** tonyb_ is now known as tonyb | 23:33 | |
| *** piet has joined #openstack-ironic | 23:36 | |
| *** piet has quit IRC | 23:52 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
