| jlvillal | Damn JayF I got caught by the Gerrit restart again :( | 00:01 |
|---|---|---|
| *** moshele has joined #openstack-ironic | 00:02 | |
| JayF | yep, it's always how you know it's getting close to time :) | 00:02 |
| *** mtanino has quit IRC | 00:11 | |
| *** baoli has joined #openstack-ironic | 00:33 | |
| *** hoangcx has joined #openstack-ironic | 00:40 | |
| *** hoangcx2 has quit IRC | 00:43 | |
| *** dprince has quit IRC | 00:44 | |
| *** baoli has quit IRC | 00:45 | |
| *** joprovost has joined #openstack-ironic | 00:55 | |
| devananda | gah. the problem with my keystone policy patch was s/role/rule/ ... | 00:57 |
| *** sdake has quit IRC | 00:57 | |
| devananda | took me far longer than it should have to spot that | 00:58 |
| *** joprovost1 has joined #openstack-ironic | 00:58 | |
| *** sdake has joined #openstack-ironic | 00:59 | |
| *** joprovost has quit IRC | 01:00 | |
| *** joprovost1 is now known as joprovost | 01:00 | |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Add keystone policy support to Ironic https://review.openstack.org/325599 | 01:00 |
| *** phuongnh has joined #openstack-ironic | 01:01 | |
| *** Nisha_away has joined #openstack-ironic | 01:04 | |
| *** Nisha_away has quit IRC | 01:08 | |
| *** Sukhdev has quit IRC | 01:12 | |
| openstackgerrit | Alexander Bashmakov proposed openstack/ironic: Rewrite devstack/tools/ironic/scripts/create-node.sh in Python. https://review.openstack.org/346832 | 01:13 |
| *** rbudden has quit IRC | 01:23 | |
| *** adu has joined #openstack-ironic | 01:25 | |
| *** wajdi has joined #openstack-ironic | 01:29 | |
| *** wajdi_ has joined #openstack-ironic | 01:31 | |
| *** tangchen has quit IRC | 01:33 | |
| *** wajdi has quit IRC | 01:34 | |
| *** tangchen has joined #openstack-ironic | 01:39 | |
| openstackgerrit | Yuiko Takada Mori proposed openstack/ironic-specs: Fix Nova-compatible Serial Console's Other deployer impact https://review.openstack.org/335378 | 01:43 |
| *** sdake has quit IRC | 01:44 | |
| *** baoli has joined #openstack-ironic | 01:49 | |
| openstackgerrit | zhangjl proposed openstack/ironic: Installation Guide doc code is out of date https://review.openstack.org/350340 | 02:03 |
| openstackgerrit | zhangjl proposed openstack/ironic: Installation Guide doc code is out of date https://review.openstack.org/350340 | 02:07 |
| openstackgerrit | Zhenguo Niu proposed openstack/ironic-inspector: Check whether action value is string before calling format() https://review.openstack.org/349392 | 02:09 |
| *** joprovost has quit IRC | 02:21 | |
| *** moshele has quit IRC | 02:22 | |
| openstackgerrit | xiexs proposed openstack/virtualbmc: Adds power_soft support to vbmc https://review.openstack.org/347367 | 02:38 |
| openstackgerrit | xiexs proposed openstack/virtualbmc: Adds power diag support to vbmc https://review.openstack.org/347368 | 02:38 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Update documentation for keystone policy support https://review.openstack.org/345556 | 02:44 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Mask configdrive contents in API responses https://review.openstack.org/326768 | 02:44 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Add keystone policy support to Ironic https://review.openstack.org/325599 | 02:44 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: [WIP] Hide driver[_internal]_info from observer role https://review.openstack.org/345076 | 02:44 |
| *** yuanying has quit IRC | 02:48 | |
| *** adu has quit IRC | 02:56 | |
| *** Nisha_away has joined #openstack-ironic | 02:58 | |
| *** Nisha_away is now known as Nisha1 | 02:58 | |
| *** adu has joined #openstack-ironic | 02:58 | |
| *** rama_y has quit IRC | 03:04 | |
| *** bvandewa_ has joined #openstack-ironic | 03:05 | |
| *** bvandewa_ has quit IRC | 03:05 | |
| *** bvandewa_ has joined #openstack-ironic | 03:06 | |
| *** bvandewa_ has quit IRC | 03:07 | |
| *** bvandewa has quit IRC | 03:07 | |
| *** adu has quit IRC | 03:18 | |
| *** adu has joined #openstack-ironic | 03:21 | |
| *** adu has quit IRC | 03:22 | |
| openstackgerrit | Zhenguo Niu proposed openstack/ironic: Clean nodes stuck in CLEANING state when ir-cond restarts https://review.openstack.org/349971 | 03:25 |
| *** Nakato has quit IRC | 03:27 | |
| *** Nakato has joined #openstack-ironic | 03:32 | |
| openstackgerrit | xiexs proposed openstack/virtualbmc: Adds power_soft support to vbmc https://review.openstack.org/347367 | 03:39 |
| *** yuanying has joined #openstack-ironic | 03:47 | |
| *** ayoung has quit IRC | 03:48 | |
| *** yuanying has quit IRC | 03:49 | |
| *** yuanying has joined #openstack-ironic | 03:49 | |
| *** bvandewa has joined #openstack-ironic | 03:49 | |
| openstackgerrit | Tang Chen proposed openstack/ironic: Split API root unit and functional tests https://review.openstack.org/348965 | 03:50 |
| openstackgerrit | Tang Chen proposed openstack/ironic: Prepare for functional tests https://review.openstack.org/349421 | 03:50 |
| openstackgerrit | Tang Chen proposed openstack/ironic: Move api driver tests to functional dir https://review.openstack.org/350365 | 03:50 |
| *** sdake has joined #openstack-ironic | 03:50 | |
| *** bvandewa has quit IRC | 03:51 | |
| *** bvandewa has joined #openstack-ironic | 03:52 | |
| *** sdake_ has joined #openstack-ironic | 03:54 | |
| openstackgerrit | xiexs proposed openstack/virtualbmc: Adds power diag support to vbmc https://review.openstack.org/347368 | 03:57 |
| *** sdake has quit IRC | 03:58 | |
| *** vmud213 has joined #openstack-ironic | 04:00 | |
| *** appu103_ has joined #openstack-ironic | 04:03 | |
| *** appu103_ is now known as aparnav | 04:03 | |
| *** sdake_ has quit IRC | 04:05 | |
| openstackgerrit | xiexs proposed openstack/pyghmi: Supports cold reset in bmc https://review.openstack.org/350368 | 04:05 |
| *** links has joined #openstack-ironic | 04:10 | |
| *** bvandewa has quit IRC | 04:10 | |
| *** Nisha1 has quit IRC | 04:13 | |
| *** Sukhdev has joined #openstack-ironic | 04:13 | |
| *** amotoki has joined #openstack-ironic | 04:20 | |
| *** vmud213 has quit IRC | 04:46 | |
| *** vmud213 has joined #openstack-ironic | 04:52 | |
| openstackgerrit | Zhenguo Niu proposed openstack/ironic: Remove unneeded 'reserved' filter in cleaning stuck nodes https://review.openstack.org/350377 | 04:55 |
| *** Nisha_away has joined #openstack-ironic | 05:10 | |
| openstackgerrit | Aparna proposed openstack/proliantutils: Modify error message of InvalidInputExecption in hpssa https://review.openstack.org/338894 | 05:11 |
| *** e0ne has joined #openstack-ironic | 05:12 | |
| *** e0ne has quit IRC | 05:15 | |
| *** mbound has joined #openstack-ironic | 05:16 | |
| openstackgerrit | Merged openstack/ironic: Timing metrics for agent deploy classes https://review.openstack.org/350243 | 05:17 |
| *** mbound has quit IRC | 05:21 | |
| *** rcernin has joined #openstack-ironic | 05:28 | |
| openstackgerrit | Aparna proposed openstack/proliantutils: Fix 'raid_config' validation error message https://review.openstack.org/344196 | 05:30 |
| *** rama_y has joined #openstack-ironic | 05:42 | |
| *** ChubYann has quit IRC | 05:42 | |
| *** fragatina has joined #openstack-ironic | 05:45 | |
| *** hshiina has joined #openstack-ironic | 05:49 | |
| *** fragatina has quit IRC | 05:51 | |
| *** fragatina has joined #openstack-ironic | 05:52 | |
| *** Sukhdev has quit IRC | 05:55 | |
| *** rbartal has joined #openstack-ironic | 05:56 | |
| openstackgerrit | Aparna proposed openstack/proliantutils: Modify error message of InvalidInputException in hpssa https://review.openstack.org/338894 | 05:58 |
| *** rama_y has quit IRC | 06:00 | |
| *** mjura has joined #openstack-ironic | 06:04 | |
| *** sdake has joined #openstack-ironic | 06:07 | |
| *** sdake has joined #openstack-ironic | 06:07 | |
| *** tiendc has joined #openstack-ironic | 06:07 | |
| *** Nisha_away has quit IRC | 06:11 | |
| *** mjura has quit IRC | 06:12 | |
| *** mjura has joined #openstack-ironic | 06:13 | |
| *** arturas has joined #openstack-ironic | 06:16 | |
| *** arturas has quit IRC | 06:18 | |
| *** arturas has joined #openstack-ironic | 06:19 | |
| *** sdake has quit IRC | 06:23 | |
| openstackgerrit | zhangjl proposed openstack/ironic: Installation Guide doc code is out of date https://review.openstack.org/350340 | 06:34 |
| *** milan has joined #openstack-ironic | 06:36 | |
| *** fumie has quit IRC | 06:40 | |
| *** tesseract- has joined #openstack-ironic | 06:45 | |
| *** arturas has quit IRC | 06:47 | |
| *** lok has joined #openstack-ironic | 06:53 | |
| *** Lord_Phoenix has joined #openstack-ironic | 06:54 | |
| *** moshele has joined #openstack-ironic | 06:57 | |
| *** tangchen has quit IRC | 07:00 | |
| *** tangchen_ has quit IRC | 07:00 | |
| *** aparnav has quit IRC | 07:03 | |
| *** aparnav has joined #openstack-ironic | 07:04 | |
| openstackgerrit | lokesh s proposed openstack/ironic: Adding follow-up patch for audit middleware https://review.openstack.org/350399 | 07:04 |
| *** e0ne has joined #openstack-ironic | 07:16 | |
| *** ohamada has joined #openstack-ironic | 07:19 | |
| *** uggla_ has joined #openstack-ironic | 07:20 | |
| *** pcaruana has joined #openstack-ironic | 07:21 | |
| uggla_ | Hello | 07:21 |
| *** links has quit IRC | 07:23 | |
| *** rbartal has quit IRC | 07:25 | |
| *** ekarlso has quit IRC | 07:29 | |
| *** tiendc has quit IRC | 07:30 | |
| *** jpich has joined #openstack-ironic | 07:35 | |
| openstackgerrit | Aparna proposed openstack/proliantutils: Modify error message of InvalidInputException in hpssa https://review.openstack.org/338894 | 07:36 |
| *** links has joined #openstack-ironic | 07:36 | |
| *** ekarlso has joined #openstack-ironic | 07:37 | |
| *** rbartal has joined #openstack-ironic | 07:38 | |
| openstackgerrit | kesper proposed openstack/proliantutils: Adds constructor for InvalidInputError https://review.openstack.org/328904 | 07:39 |
| openstackgerrit | kesper proposed openstack/proliantutils: Adds constructor for InvalidInputError https://review.openstack.org/328904 | 07:44 |
| *** Goneri has joined #openstack-ironic | 07:54 | |
| *** Goneri has quit IRC | 07:56 | |
| *** Goneri has joined #openstack-ironic | 07:56 | |
| *** daemontool has joined #openstack-ironic | 07:57 | |
| *** fxpester has joined #openstack-ironic | 07:57 | |
| *** jistr is now known as jistr|training | 07:59 | |
| *** zzzeek has quit IRC | 08:00 | |
| *** amotoki has quit IRC | 08:00 | |
| *** bvandewa has joined #openstack-ironic | 08:02 | |
| *** zzzeek has joined #openstack-ironic | 08:02 | |
| *** amotoki has joined #openstack-ironic | 08:04 | |
| *** amotoki has quit IRC | 08:05 | |
| alineb | morning all | 08:15 |
| openstackgerrit | Merged openstack/ironic-inspector: Updated from global requirements https://review.openstack.org/348625 | 08:15 |
| aarefiev | morning! | 08:16 |
| *** amotoki has joined #openstack-ironic | 08:17 | |
| *** amotoki_ has joined #openstack-ironic | 08:18 | |
| *** vmud213 has quit IRC | 08:19 | |
| lucasagomes | alineb, aarefiev morning | 08:20 |
| openstackgerrit | Zhenguo Niu proposed openstack/ironic: Clean nodes stuck in DELETING state when ir-cond restarts https://review.openstack.org/350439 | 08:20 |
| aarefiev | morning lucasagomes | 08:21 |
| *** amotoki has quit IRC | 08:21 | |
| *** vmud213 has joined #openstack-ironic | 08:23 | |
| alineb | morning lucasagomes, aarefiev | 08:25 |
| aarefiev | o/ alineb | 08:26 |
| *** baoli has quit IRC | 08:27 | |
| openstackgerrit | Cornelio Hopmann proposed openstack/ironic-specs: Partial-Bug: #1583065 [RFE] Support network switches provisioning https://review.openstack.org/338215 | 08:28 |
| openstack | bug 1583065 in Ironic "[RFE] Support network switches provisioning" [Wishlist,Incomplete] https://launchpad.net/bugs/1583065 - Assigned to Cornelio Hopmann (chopmann) | 08:28 |
| *** amotoki_ has quit IRC | 08:29 | |
| moshele | lucasagomes: moring | 08:37 |
| moshele | lucasagomes: I have a question regarding the multi tenant feature | 08:37 |
| vsaienk0 | morning moshele, aarefiev, dtantsur, lucasagomes, vdrok, jroll and all Ironic'ers | 08:40 |
| moshele | lucasagomes: I was able to make it work on InfiniBand we my patches https://review.openstack.org/#/c/264263/ and https://review.openstack.org/#/c/264257/ | 08:40 |
| moshele | lucasagomes: but I had to change the https://review.openstack.org/#/c/339129/ Add BaremetalMechanismDriver driver. to support vlan network | 08:41 |
| moshele | morning: vsaienk0 | 08:41 |
| moshele | vsaienk0: regarding your BaremetalMechanismDriver driver. t https://review.openstack.org/#/c/339129/ why is it only for flat? | 08:42 |
| *** electrofelix has joined #openstack-ironic | 08:43 | |
| *** amotoki has joined #openstack-ironic | 08:45 | |
| *** priteau has joined #openstack-ironic | 08:48 | |
| *** jistr|training is now known as jistr | 08:49 | |
| openstackgerrit | kesper proposed openstack/proliantutils: Adds constructor for InvalidInputError https://review.openstack.org/328904 | 08:52 |
| *** amotoki has quit IRC | 08:53 | |
| openstackgerrit | Merged openstack/python-ironic-inspector-client: Updated from global requirements https://review.openstack.org/348672 | 08:59 |
| *** hshiina has quit IRC | 09:00 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ironic: Updated from global requirements https://review.openstack.org/350097 | 09:01 |
| *** amotoki has joined #openstack-ironic | 09:02 | |
| *** mgould|afk is now known as mgould | 09:02 | |
| mgould | morning Ironic! | 09:02 |
| *** amitkqed has quit IRC | 09:04 | |
| *** amitkqed has joined #openstack-ironic | 09:04 | |
| vdrok | morning alineb aarefiev lucasagomes moshele vsaienk0 and mgould ! :) | 09:05 |
| *** bvandewa has quit IRC | 09:06 | |
| moshele | vdrok: good morning | 09:06 |
| moshele | vdrok: can you please review the InfiniBand patches https://review.openstack.org/#/c/264257/ and https://review.openstack.org/#/c/264263/ | 09:07 |
| vdrok | moshele: yup, will do | 09:07 |
| moshele | vdrok: I tested with the multi tenant feature and it working | 09:08 |
| vdrok | cool :) | 09:08 |
| moshele | vdrok: I just update the BaremetalMechanismDriver to support vlan https://review.openstack.org/#/c/339129/ | 09:08 |
| moshele | vdrok: do you know why it is only for flat? | 09:09 |
| mgould | morning vdrok alineb aarefiev lucasagomes moshele vsaienk0 | 09:09 |
| Romanenko_K | morning! | 09:09 |
| aarefiev | morning vsaienk0 vdrok moshele mgould Romanenko_K | 09:10 |
| mgould | morning Romanenko_K | 09:10 |
| sambetts|afk | Morning all | 09:10 |
| *** sambetts|afk is now known as sambetts | 09:11 | |
| mgould | morning sambetts | 09:11 |
| vdrok | morning Romanenko_K and sambetts | 09:11 |
| sambetts | o/ mgould | 09:11 |
| vsaienk0 | moshele: what do you mean "just update the BaremetalMechanismDriver to support vlan"? | 09:11 |
| vdrok | moshele: maybe because of this - https://github.com/openstack/networking-generic-switch/blob/master/networking_generic_switch/generic_switch_mech.py ? | 09:11 |
| moshele | vdrok: ok, but shouldn't this be done by BaremetalMechanismDriver , In infiniband we don't configure switches | 09:13 |
| vsaienk0 | moshele: BaremetalMechanismDriver is a noop driver that will provide fake port binding in flat network and will contain notification support. It shouldn't work with vlans | 09:13 |
| moshele | vdrok: I was thinking that BaremetalMechanismDriver bind the barmetal port and you have other drivers that configure the switches | 09:14 |
| vdrok | moshele: vsaienk0 is an expert here :) | 09:14 |
| sambetts | the other vendor switch mech drivers should support the baremetal vnic type to do binding on their switches | 09:15 |
| vsaienk0 | moshele: once port is bound, neutron will not call other drivers to bound the port. So de facto only one driver can bound a port. | 09:15 |
| moshele | vsaienk0, sambetts: I see so I need to update our mlnx_infiniband to bind barmetal ports | 09:16 |
| sambetts | Yup | 09:16 |
| moshele | ok cool :) | 09:17 |
| vsaienk0 | yes, seems to be that is correct place | 09:17 |
| openstackgerrit | Zhenguo Niu proposed openstack/ironic-inspector: Check whether action value is string before calling format() https://review.openstack.org/349392 | 09:17 |
| moshele | sambetts: please review my inifinband patches when you have time https://review.openstack.org/#/c/264263/ https://review.openstack.org/#/c/264257/ , thanks :) | 09:18 |
| moshele | sambetts: and thanks for your comments it was really helpful | 09:19 |
| sambetts | :) | 09:20 |
| vmud213 | morning all | 09:21 |
| *** dtantsur|afk is now known as dtantsur | 09:24 | |
| dtantsur | Morning Ironic | 09:24 |
| sambetts | o/ dtantsur | 09:24 |
| *** tangchen has joined #openstack-ironic | 09:24 | |
| mgould | morning dtantsur | 09:29 |
| yibo | dtantsur: Would you have a look at https://bugs.launchpad.net/ironic/+bug/1582964? RFE needs be approved before according patch can pass review. I'm not sure what I should do now. Thanks. | 09:34 |
| openstack | Launchpad bug 1582964 in Ironic "[RFE] Support multi arch deployments" [Wishlist,In progress] - Assigned to Yibo Cai (cyb70289) | 09:34 |
| dtantsur | yibo, let's wait for jroll to come and discuss | 09:35 |
| yibo | dtantsur: okay, thanks | 09:35 |
| vdrok | morning vmud213 and dtantsur | 09:36 |
| vmud213 | morning vdrok | 09:39 |
| *** electrofelix has quit IRC | 09:40 | |
| *** electrofelix has joined #openstack-ironic | 09:40 | |
| aarefiev | morning dtantsur sambetts | 09:42 |
| *** evgenyl has quit IRC | 09:42 | |
| milan | morning Ironic! :) | 09:43 |
| pas-ha | morning milan aarefiev vdrok and all :) | 09:44 |
| aarefiev | morning pas-ha milan | 09:44 |
| *** evgenyl has joined #openstack-ironic | 09:44 | |
| milan | morning aarefiev dtantsur mgould moshele pas-ha sambetts vdrok vmud213 vsaienk0 yibo | 09:45 |
| milan | :) | 09:45 |
| milan | and morning all! :) | 09:45 |
| dtantsur | lucasagomes and others: wanna some huge patch reviewing? ;) https://review.openstack.org/#/c/236982/ brings us possibility to use keystone v3.. | 09:47 |
| dtantsur | (thanks pas-ha for pushing it forward) | 09:47 |
| openstackgerrit | Anton Arefiev proposed openstack/python-ironicclient: Fail with more meaningful error while creating client https://review.openstack.org/344791 | 09:47 |
| milan | guys, I've added new column into db in inspector, now getting this error in unit tests: http://paste.openstack.org/show/547734/ I'm missing some step somewhere but failing to locate it | 09:47 |
| pas-ha | milan: have you added a corresponding DB migration? | 09:49 |
| milan | pas-ha, I did, but I've modified it later manually | 09:49 |
| milan | could it get out of sync somehow? | 09:49 |
| dtantsur | probably it did, hard to tell without looking at the code | 09:50 |
| pas-ha | I'd think that somewhere in basic unittest setup there is a code that brings the db to the current state | 09:50 |
| milan | dtantsur, pas-ha I guess I'll push it with this failure so you could take look at it | 09:51 |
| * milan needs to fix pep8 now ;) | 09:51 | |
| vmud213 | morning milan | 09:51 |
| milan | vmud213, o/ :) | 09:51 |
| dtantsur | milan, that would be handy | 09:51 |
| dtantsur | milan, also sometimes the gate has different opinions on things | 09:51 |
| milan | heh, OK let's see | 09:51 |
| vmud213 | vsaienk0: I am just planning to start a devstack, to test the recently merged patches related to networking | 09:51 |
| vmud213 | What configurational changes may i require? | 09:52 |
| * dtantsur will get lunch and start answering emails and customer escalations....... | 09:52 | |
| vmud213 | I hope all the patches are merged in the upstream. | 09:52 |
| lucasagomes | dtantsur, added to the list | 09:54 |
| lucasagomes | moshele, sorry, did you get the answer already? | 09:54 |
| lucasagomes | vdrok, mgould, everyone else: morning :-) | 09:55 |
| milan | lucasagomes, morning! :) | 09:55 |
| moshele | lucasagomes: yes, thanks :) | 09:55 |
| *** Jaison has joined #openstack-ironic | 09:58 | |
| *** links has quit IRC | 09:59 | |
| openstackgerrit | Yuiko Takada Mori proposed openstack/ironic-specs: Fix Nova-compatible Serial Console's Other deployer impact https://review.openstack.org/335378 | 09:59 |
| openstackgerrit | Anton Arefiev proposed openstack/ironic-inspector: Tempest: increase ironic sync timeout https://review.openstack.org/350499 | 10:05 |
| *** Goneri has quit IRC | 10:08 | |
| *** fragatina has quit IRC | 10:09 | |
| *** amit213 has quit IRC | 10:10 | |
| *** amit213 has joined #openstack-ironic | 10:12 | |
| openstackgerrit | Aline Bousquet proposed openstack/ironic: Add metrics for the ipmitool driver https://review.openstack.org/350502 | 10:13 |
| *** serverascode has quit IRC | 10:15 | |
| *** serverascode has joined #openstack-ironic | 10:16 | |
| openstackgerrit | Milan Kováčik proposed openstack/ironic-inspector: Introducing node introspection state management https://review.openstack.org/348943 | 10:17 |
| milan | dtantsur, ^ ;) | 10:20 |
| openstackgerrit | Merged openstack/ironic-inspector: Use OSC in exercise.sh https://review.openstack.org/344616 | 10:21 |
| dtantsur | okie, will check today, unless somebody does it earlier | 10:23 |
| *** Goneri has joined #openstack-ironic | 10:24 | |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic: Extend the "configuring ironic-api behind mod_wsgi" guide https://review.openstack.org/350507 | 10:25 |
| *** hoangcx has quit IRC | 10:27 | |
| *** vmud213 has quit IRC | 10:27 | |
| *** vmud213 has joined #openstack-ironic | 10:28 | |
| *** amotoki has quit IRC | 10:30 | |
| *** lok has quit IRC | 10:35 | |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic: Extend the "configuring ironic-api behind mod_wsgi" guide https://review.openstack.org/350507 | 10:35 |
| openstackgerrit | Kyrylo Romanenko proposed openstack/python-ironicclient: Add baremetal port list command to OSC plugin https://review.openstack.org/346722 | 10:36 |
| *** amotoki has joined #openstack-ironic | 10:39 | |
| lucasagomes | dtantsur, https://bugs.launchpad.net/ironic/+bug/1608252 works for me, I think we can mark it as invalid | 10:39 |
| openstack | Launchpad bug 1608252 in Ironic "Ironic API deployed with Apache: Exception registering nodes: Timed out waiting for a reply to message ID" [High,Confirmed] - Assigned to Lucas Alvares Gomes (lucasagomes) | 10:39 |
| lucasagomes | for the ironic case | 10:39 |
| lucasagomes | I suspect the puppet-ironic modules are not configuring it correctly and that's why it's failing for tripleo | 10:40 |
| dtantsur | lucasagomes, yeah, unless we get an environment (not just logs) with a reproducer, we'll have to close it | 10:40 |
| dtantsur | EmilienM, ^^^ | 10:40 |
| *** Lord_Phoenix has quit IRC | 10:40 | |
| phuongnh | morning Ironic'er | 10:40 |
| mgould | morning phuongnh | 10:40 |
| *** amotoki has quit IRC | 10:41 | |
| *** joprovost has joined #openstack-ironic | 10:41 | |
| phuongnh | I try to install ironic based on all-in-one Openstack deployed by devstack | 10:42 |
| phuongnh | I update [ironic] setting in nova.conf and restart nova compute service | 10:43 |
| phuongnh | and I got an error | 10:43 |
| phuongnh | http://pasteboard.co/pLGl832o.jpg | 10:44 |
| lucasagomes | phuongnh, what did you change? There's not much in that section AFAICT, only credentials and endpoints to | 10:45 |
| phuongnh | I checked ironic service and it is running normaly, where should I check in next step? | 10:45 |
| phuongnh | only small changed as guided: http://pasteboard.co/49Jpwqfuv.jpg | 10:47 |
| mgould | phuongnh: is the nova service running normally? | 10:47 |
| * milan relocates | 10:48 | |
| *** milan has quit IRC | 10:48 | |
| jroll | phuongnh: that's a 500 error from ironic, so there must be a traceback in ironic-api logs somewhere | 10:48 |
| *** Jaison has quit IRC | 10:49 | |
| jroll | dtantsur: yibo: what do you want to discuss about that? the "_by_arch" thing isn't my favorite but if dmitry and lucas are good with it that's fine with me | 10:49 |
| jroll | all cores: let's get this chain done so we can land the networking code in nova https://review.openstack.org/#/q/status:open+project:openstack/ironic+branch:master+topic:bug/1608511 | 10:50 |
| jroll | and good morning :) | 10:50 |
| dtantsur | I haven't checked it yet (probably) | 10:50 |
| dtantsur | morning jroll, lemme finish a thing here and I'll review it | 10:50 |
| phuongnh | mgould, yes, they is running normaly except n-cpu | 10:51 |
| jroll | dtantsur: thanks | 10:52 |
| mgould | phuongnh: I wasn't thinking straight, you should do what jroll suggested instead :-) | 10:52 |
| phuongnh | jroll, I am checking the log | 10:53 |
| *** links has joined #openstack-ironic | 10:53 | |
| phuongnh | mgould, thanks | 10:53 |
| vdrok | morning jroll | 10:54 |
| jroll | hey vdrok :) | 10:54 |
| phuongnh | jroll, there is no error in ironic-api log http://pasteboard.co/49U43NIej.jpg | 10:57 |
| jroll | hmmmm | 10:58 |
| jroll | is this behind apache? | 10:58 |
| phuongnh | yes, I use apache and it is started as normal | 10:58 |
| jroll | ah, maybe check apache error log as well | 10:58 |
| phuongnh | jroll, yes, I will do it | 10:59 |
| phuongnh | jroll, yes, you are correct, I found ironic-error.log file | 11:03 |
| phuongnh | jroll, http://paste.openstack.org/show/547742/ | 11:03 |
| openstackgerrit | Sam Betts proposed openstack/python-ironicclient: Fix uuid to UUID in expected error messages https://review.openstack.org/350522 | 11:12 |
| sambetts | dtantsur, jroll: ironicclient gate is currently broken, we need this ^ to unblock it | 11:13 |
| openstackgerrit | Debayan Ray proposed openstack/proliantutils: Adds constructor for InvalidInputError https://review.openstack.org/328904 | 11:14 |
| lucasagomes | phuongnh, IOError: [Errno 13] Permission denied: '/var/log/ironic/app.wsgi.log' | 11:16 |
| lucasagomes | there's a bunch of errors there | 11:16 |
| phuongnh | lucasagomes, the file '/var/log/ironic/app.wsgi.log' is not found | 11:19 |
| lucasagomes | phuongnh, you may want to check ur /etc/httpd/conf.d/ironic config | 11:19 |
| openstackgerrit | Debayan Ray proposed openstack/proliantutils: Adds constructor for InvalidInputError https://review.openstack.org/328904 | 11:19 |
| lucasagomes | (replace httpd with apache2 for debian systems) | 11:19 |
| lucasagomes | phuongnh, that's where you set the log files path for apache + mod_wsgi | 11:20 |
| * lucasagomes will go grab some bites, be right back | 11:21 | |
| *** lucasagomes is now known as lucas-hungry | 11:21 | |
| dtantsur | jroll, sambetts, hmm, https://review.openstack.org/#/c/350006/ needs a release note. I'm fine with following up with it, but wdyt? | 11:22 |
| phuongnh | lucasagomes, are you talking about this file? /etc/apache2/sites-available/ironic.conf | 11:23 |
| dtantsur | jroll, sambetts, or maybe it does not, provided that we just landed this new field... | 11:25 |
| sambetts | dtantsur: I'm looking at the existing release notes, and they don't mention if its nullable or not, so I think it might be ok http://docs.openstack.org/releasenotes/ironic/current-series.html#new-features | 11:29 |
| dtantsur | sambetts, yeah. but there is a patch adding a migration, where I think a release note would be handy | 11:30 |
| phuongnh | lucasagemes, this is my file http://paste.openstack.org/show/547743/ | 11:30 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Operator documentation for multitenancy https://review.openstack.org/228496 | 11:30 |
| sambetts | dtantsur: hmm, good point, I not sure what the release not would say, any thoughts? | 11:31 |
| sambetts | s/not/note* | 11:31 |
| sambetts | perhaps we should modify the existing one to mention that the default will be populated if not specified | 11:32 |
| dtantsur | sambetts, actually there is such patch, but for some reason it's separate. I vote for merging these two together | 11:33 |
| dtantsur | jroll, ^^ | 11:33 |
| mat128 | good morning Ironic! | 11:34 |
| dtantsur | morning mat128 | 11:34 |
| sambetts | Hi mat128 | 11:34 |
| jroll | dtantsur: I can squash if that's helpful | 11:36 |
| dtantsur | yes please | 11:36 |
| jroll | k, one sec | 11:36 |
| dtantsur | there are enough people here to fast-track it :) | 11:37 |
| vdrok | morning mat128 | 11:37 |
| TheJulia | good morning | 11:38 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Add a data migration to fill node.network_interface https://review.openstack.org/349528 | 11:39 |
| vdrok | and morning TheJulia | 11:39 |
| jroll | dtantsur: ^^ | 11:39 |
| dtantsur | thanks, +2 | 11:39 |
| dtantsur | morning TheJulia | 11:39 |
| mat128 | morning TheJulia | 11:40 |
| jroll | TheJulia: mornin' | 11:41 |
| jroll | thanks dmitry | 11:41 |
| jroll | so once that lands, we need a requirements bump (already in gerrit), and then the nova thing can go \o/ | 11:43 |
| *** livelace has joined #openstack-ironic | 11:44 | |
| *** hshiina has joined #openstack-ironic | 11:45 | |
| *** dprince has joined #openstack-ironic | 11:45 | |
| sambetts | \o/ | 11:46 |
| TheJulia | woot | 11:46 |
| *** bfournie has quit IRC | 11:52 | |
| *** Lord_Phoenix has joined #openstack-ironic | 11:54 | |
| moshele | hi does ironic can send notification when changing a node state (for example for manage -> inspection) | 11:57 |
| jroll | moshele: it's being worked on, but not yet | 11:58 |
| jroll | moshele: I guess power state is first, but provision state should be a quick follow after that: https://review.openstack.org/#/q/topic:bug/1526408 | 11:59 |
| *** narasimha_SV has joined #openstack-ironic | 11:59 | |
| moshele | jroll: cool thanks | 11:59 |
| narasimha_SV | http://paste.openstack.org/show/547746/ | 11:59 |
| narasimha_SV | ConfigFileValueError: Value for option scheduler_host_manager is not valid: Valid values are [host_manager, ironic_host_manager], but found 'nova.scheduler.ironic_host_manager.IronicHostManager' | 12:00 |
| narasimha_SV | nova api is unable to start because of this error | 12:00 |
| jroll | narasimha_SV: that error tells you exactly what to do | 12:00 |
| narasimha_SV | but as per mitaka configurations this is the same | 12:00 |
| narasimha_SV | the value is the same | 12:01 |
| jroll | right, nova broke that | 12:01 |
| narasimha_SV | jroll: Do I need to change this ironic_host_manager ???? | 12:01 |
| narasimha_SV | yes | 12:01 |
| jroll | yes | 12:01 |
| narasimha_SV | ok thank you | 12:01 |
| jroll | np | 12:01 |
| * jroll updates docs | 12:01 | |
| jroll | narasimha_SV: it might be ironic_host_manager.IronicHostManager, I'm not sure | 12:02 |
| * jroll tests | 12:02 | |
| *** phuongnh has quit IRC | 12:03 | |
| narasimha_SV | ok let me check | 12:04 |
| jroll | sorry, it is just scheduler_host_manager = ironic_host_manager | 12:04 |
| *** milan has joined #openstack-ironic | 12:05 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Update nova host_manager config docs https://review.openstack.org/350542 | 12:09 |
| jroll | easy peasy review there | 12:10 |
| dtantsur | jroll, don't we need a class name in ^^^? | 12:10 |
| EmilienM | lucas-hungry, dtantsur: so you suspect a bad config. Have you looked at Ironic config in TripleO before? | 12:10 |
| sambetts | dtantsur: its a stevedore name now | 12:11 |
| dtantsur | sambetts, oh, got it | 12:11 |
| *** adu has joined #openstack-ironic | 12:11 | |
| jroll | dtantsur: apparently not, see e.g. http://logs.openstack.org/06/350006/2/check/gate-tempest-dsvm-ironic-ipa-wholedisk-agent_ssh-tinyipa/a3ca84b/logs/etc/nova/nova.conf.txt.gz | 12:11 |
| jroll | oh commit message is wrong though | 12:11 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Update nova scheduler_host_manager config docs https://review.openstack.org/350542 | 12:11 |
| jroll | thought it was just host_manager | 12:11 |
| *** ccamacho is now known as ccamacho|lunch | 12:11 | |
| TheJulia | so silly question, when did that change occur because our install guide is occassionally referenced by people with older versions of code? | 12:12 |
| narasimha_SV | jroll: thank you it worked | 12:12 |
| openstackgerrit | Merged openstack/proliantutils: Fix 'raid_config' validation error message https://review.openstack.org/344196 | 12:12 |
| jroll | narasimha_SV: no problem | 12:12 |
| jroll | TheJulia: this cycle | 12:12 |
| sambetts | TheJulia, dtantsur, jroll: https://github.com/openstack/nova/commit/158c6d64c2da48ec5fb3382eb64cd5c5e9c5c2d9 | 12:12 |
| jroll | TheJulia: fwiw, we should be leading people to e.g. http://docs.openstack.org/developer/ironic/mitaka/deploy/install-guide.html | 12:13 |
| TheJulia | sambetts: I have the exact same link on my screen | 12:13 |
| dtantsur | sambetts, thanks, got it | 12:13 |
| TheJulia | jroll: true, I'm more thinking the times people have come in with Kilo or.. now I guess liberty code and tried to follow the instructions :) | 12:13 |
| jroll | yeah | 12:13 |
| jroll | fwiw I think ironic_host_manager.IronicHostManager will still work | 12:14 |
| jroll | oh no, it won't | 12:14 |
| jroll | because choices in the opt | 12:14 |
| * jroll fixes | 12:14 | |
| lazy_prince | jroll: regarding https://review.hpcloud.net/#/c/120942/, Say I have 200 nodes registered in Ironic in one by one using ironic cli, How much time would it take for the resource tracker to run and by when hash-ring would have stabilized.. | 12:14 |
| TheJulia | lazy_prince: uhhh.... I doubt jroll can see an internal review | 12:15 |
| jroll | probably not :) | 12:15 |
| TheJulia | unless there is something going on I'm not aware of :) | 12:15 |
| * dtantsur clicks just of curiosity | 12:15 | |
| jroll | lazy_prince: RT runs every 5 minutes by default, and is adjustable | 12:15 |
| jroll | lazy_prince: adding a node shouldn't change the ring very much, that's the "consistent" part of "consistent hashing" | 12:16 |
| TheJulia | lazy_prince: time to run would truly be a cpu/io/architecture layout result | 12:16 |
| lazy_prince | aha.. sorry.. didn't realize I was using the internal link.. | 12:17 |
| TheJulia | lazy_prince: your question was good enough for people to figure out what you were talking about ;) | 12:17 |
| TheJulia | even without the link | 12:17 |
| lazy_prince | so anyways.. another thing: if I bring up a new compute proxy host, it would have to wait for 5 mins (based on default config) to get hash ring updated.. is that right understanding..? | 12:18 |
| *** MattMan has quit IRC | 12:19 | |
| *** MattMan has joined #openstack-ironic | 12:19 | |
| jroll | the new host would see the new hash ring immediately, the other hosts would take *up to* 5 minutes to see it | 12:19 |
| TheJulia | lazy_prince: and as it stands the resource tracker has to collect data anyway before a node can be scheduled, so from a standpoint of eventually consistent, I think it works | 12:20 |
| lazy_prince | okay.. I get it now.. | 12:20 |
| jroll | TheJulia: oh, I lied, that host_manager behavior was deprecated in mitaka and removed in newton | 12:21 |
| TheJulia | jroll: c'est la vie :) | 12:22 |
| jroll | I went to add a comment to the compat handling and poof | 12:22 |
| *** adu has quit IRC | 12:23 | |
| *** baoli has joined #openstack-ironic | 12:26 | |
| *** bfournie has joined #openstack-ironic | 12:26 | |
| *** baoli_ has joined #openstack-ironic | 12:27 | |
| openstackgerrit | Cornelio Hopmann proposed openstack/ironic-specs: Partial-Bug: #1583065 [RFE] Support network switches provisioning https://review.openstack.org/338215 | 12:28 |
| openstack | bug 1583065 in Ironic "[RFE] Support network switches provisioning" [Wishlist,Incomplete] https://launchpad.net/bugs/1583065 - Assigned to Cornelio Hopmann (chopmann) | 12:28 |
| *** baoli has quit IRC | 12:30 | |
| *** hoangcx has joined #openstack-ironic | 12:31 | |
| *** afaranha has quit IRC | 12:32 | |
| *** afaranha has joined #openstack-ironic | 12:32 | |
| *** joprovost has quit IRC | 12:34 | |
| *** daemontool_ has joined #openstack-ironic | 12:35 | |
| *** sdake has joined #openstack-ironic | 12:37 | |
| *** adu has joined #openstack-ironic | 12:37 | |
| *** daemontool has quit IRC | 12:37 | |
| jroll | nice | 12:41 |
| *** lucas-hungry is now known as lucasagomes | 12:43 | |
| lucasagomes | EmilienM, sorry the delay was having lunch... I haven't looked at the tripleo one | 12:43 |
| EmilienM | ah :) | 12:43 |
| lucasagomes | EmilienM, I will take a look, I just saw ur reply to the email | 12:43 |
| EmilienM | so how you can say it comes from puppet-ironic? ;) | 12:43 |
| EmilienM | lucasagomes: check your emails, I replied to the thread | 12:44 |
| EmilienM | good | 12:44 |
| lucasagomes | EmilienM, I suspected, cause I've manually set it up according to our documentation and it worked | 12:44 |
| lucasagomes | where tripleo uses puppet to configure it, so I suspected it was a misconfig in puppet somewhere | 12:44 |
| EmilienM | lucasagomes: it works in devstack (tm) | 12:44 |
| lucasagomes | :D | 12:44 |
| lucasagomes | always | 12:44 |
| jroll | I wonder if it's something about the number of threads? | 12:45 |
| jroll | that's the main difference I see | 12:45 |
| jroll | tripleo has 2, devstack has 10 | 12:45 |
| lucasagomes | yeah, I'm rebuilding that env cause I've trashed it :-( | 12:45 |
| jroll | TheJulia: do you have any suggestions for making the reno better here? https://review.openstack.org/#/c/349528 | 12:48 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-specs: Boot management for in-band inspection https://review.openstack.org/305864 | 12:53 |
| dtantsur | folks, an easy spec for review ^^^ | 12:53 |
| dtantsur | does not touch anything beyond ironic-inspector inspection interface | 12:53 |
| *** guolidong has joined #openstack-ironic | 13:00 | |
| openstackgerrit | Merged openstack/ironic: Move default network_interface logic in node object https://review.openstack.org/350006 | 13:02 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Add nova scheduler_host_subset_size option to docs https://review.openstack.org/350569 | 13:03 |
| openstackgerrit | Cornelio Hopmann proposed openstack/ironic-specs: [RFE] Support network switches provisioning https://review.openstack.org/350570 | 13:03 |
| *** guolidong is now known as Sylvester | 13:04 | |
| openstackgerrit | Shivanand Tendulker proposed openstack/ironic: Fix iLO drivers inconsistent boot mode default value https://review.openstack.org/345231 | 13:07 |
| openstackgerrit | Merged openstack/pyghmi: Supports cold reset in bmc https://review.openstack.org/350368 | 13:08 |
| *** jcoufal has joined #openstack-ironic | 13:12 | |
| *** ccamacho|lunch is now known as ccamacho | 13:17 | |
| openstackgerrit | Kyrylo Romanenko proposed openstack/python-ironicclient: Fix 'uuid' to 'UUID' in negative chassis-create test https://review.openstack.org/350578 | 13:17 |
| Romanenko_K | now gate-ironicclient-dsvm-functional fails due to change of exception strings | 13:18 |
| sambetts | Romanenko_K: Please see my comment, there is already a patch to fix it | 13:19 |
| Romanenko_K | oh, good! | 13:19 |
| jroll | sambetts: where's that patch? | 13:19 |
| jroll | I can't find it now, saw it earlier | 13:19 |
| sambetts | jroll: https://review.openstack.org/#/c/350522/ | 13:19 |
| *** rbudden has joined #openstack-ironic | 13:19 | |
| jroll | thanks | 13:19 |
| jroll | oh, +W yay | 13:19 |
| * lucasagomes approved the keystone session patch | 13:20 | |
| sambetts | jroll: yup just waiting on jenkins :) | 13:20 |
| jroll | lucasagomes: \o/ | 13:20 |
| sambetts | lucasagomes: woop | 13:20 |
| TheJulia | jroll: possibly split apart the note on line 30/31 from the sentence on line 32 sinec they are separate items in my mind. The sticking point for me is how does an operator truly know what the setting should be when we highlight what they should make sure something is right | 13:20 |
| *** adu has quit IRC | 13:20 | |
| milan | guys, I've got this error: http://logs.openstack.org/43/348943/2/check/gate-ironic-inspector-python27-db-ubuntu-xenial/7d54a8c/console.html#_2016-08-03_12_39_06_476253 in inspector's patch: https://review.openstack.org/#/c/348943/2 which I can't figure out; I've added a db column and a migration for it (which I later updated manually) help appreciated! thx! | 13:21 |
| jroll | TheJulia: as in, switch them around or make it a separate note? | 13:21 |
| *** wlfightup has quit IRC | 13:22 | |
| TheJulia | jroll: kind of, or sub-item bullet points. I guess the conundrum is someone who only looks at upgrading every six months might not have any context into what they should expect for the configuration without knowing where to go to gain all that knowledge | 13:23 |
| TheJulia | jroll: where as we can look at it and go "Oh, okay!" | 13:23 |
| TheJulia | jroll: am I making sense? | 13:23 |
| jroll | TheJulia: kinda | 13:23 |
| jroll | TheJulia: the logic to figure it out is in the same note, though | 13:24 |
| TheJulia | jroll: it is almost like we're inviting someone upgrading to possibly set the wrong thing when we should say "Realistically, you likely won't need to modify this unless a standard is required in the deployment for all new nodes being added." | 13:24 |
| jroll | TheJulia: I almost feel like the migration should only do flat/noop | 13:25 |
| jroll | idk | 13:25 |
| jroll | but CD and all | 13:25 |
| TheJulia | yeah.. I kind of feel the same, but I can see the case _if_ someone has done all the work ahead of time or is just upgrading a greenfield env | 13:26 |
| dtantsur | milan, commented | 13:26 |
| milan | dtantsur, thx, /me looks | 13:26 |
| *** joprovost has joined #openstack-ironic | 13:26 | |
| jroll | TheJulia: well, if they're on 6.0.0, they still couldn't have set the network_interface field | 13:26 |
| *** fxpester has quit IRC | 13:26 | |
| jroll | so any upgrade should be coming from flat/noop, unless CD | 13:26 |
| *** vmud213 has quit IRC | 13:27 | |
| TheJulia | Yeah, for readability, maybe just re-ordering or squashing 30/31 into line 25 | 13:27 |
| jroll | oh, and this only migrates NULLs | 13:27 |
| TheJulia | yup | 13:27 |
| jroll | TheJulia: maybe I just add "if you haven't set default_network_interface, everything should go as expected" | 13:27 |
| TheJulia | works for me, as is it works for me, I just worry I guess | 13:28 |
| *** fxpester has joined #openstack-ironic | 13:28 | |
| *** aparnav has quit IRC | 13:28 | |
| *** fragatina has joined #openstack-ironic | 13:28 | |
| *** joprovost1 has joined #openstack-ironic | 13:29 | |
| milan | dtantsur +1 :) works now | 13:30 |
| *** joprovost has quit IRC | 13:30 | |
| *** joprovost1 is now known as joprovost | 13:30 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Add a data migration to fill node.network_interface https://review.openstack.org/349528 | 13:31 |
| jroll | TheJulia: see how that looks | 13:31 |
| openstackgerrit | Merged openstack/ironic-inspector: Check whether action value is string before calling format() https://review.openstack.org/349392 | 13:31 |
| sambetts | milan: qq about that patch, why create utils.enum over just a dict or a set? | 13:31 |
| milan | basically just shorter in code: States.finishing | 13:32 |
| milan | vs | 13:32 |
| milan | States['finishing'] | 13:32 |
| *** rbudden has quit IRC | 13:32 | |
| milan | sambetts, other option would be to use class States(object): ..... | 13:32 |
| milan | but it doesn't have the .attributes()/attribute() magic | 13:33 |
| milan | sambetts, also py3.x has a native Enum type | 13:33 |
| milan | this is just a quick hack O:-) | 13:34 |
| sambetts | huh, I din't know about the attributes/attribute magic, do yuo have a link to somewhere thats documented? | 13:35 |
| * milan found it on stack overflow looks again | 13:36 | |
| dtantsur | sambetts, have you seen https://bugs.launchpad.net/bugs/1606231 ? | 13:36 |
| openstack | Launchpad bug 1606231 in Ironic "vif_port_id of ironic port is not updating after neutron port-delete" [Undecided,In progress] - Assigned to Anton Kremenetsky (anton-kremenetsky) | 13:36 |
| dtantsur | I wonder if it's related to your attach/detach work | 13:36 |
| sambetts | hmmm interesting, without listneing for some trigger from neutron I don't know how we can do that | 13:37 |
| jroll | O_o | 13:37 |
| jroll | interesting | 13:37 |
| *** hoangcx has quit IRC | 13:37 | |
| xavierr | Good morning Ironic | 13:37 |
| sambetts | I think this is related to the ML conversation we had a week or so back about adding a hooks into neutron so we receive notifications on port changes | 13:38 |
| jroll | sambetts: it says the notification goes through to the virt driver | 13:38 |
| sambetts | well that depends if its the teannt port or not | 13:38 |
| jroll | but we don't implement detach | 13:38 |
| milan | sambetts, http://stackoverflow.com/questions/36932/how-can-i-represent-an-enum-in-python https://docs.python.org/3/library/enum.html (py3.x is of course a fine-tuned thing ;) | 13:38 |
| jroll | sambetts: well, this is from nova user perspective, so it would be the tenant port | 13:39 |
| *** hj-hpe has joined #openstack-ironic | 13:39 | |
| sambetts | jroll: in that case it'd work for the multi-tenant driver but not for the flat driver | 13:39 |
| jroll | sambetts: why not for the flat driver? | 13:40 |
| jroll | flat driver uses the tenant port for provisioning | 13:40 |
| sambetts | jroll: because even if the port is deleted it isn't actually detached from the network | 13:40 |
| sambetts | jroll: the only thing that changes is that its IP address is deallocatd | 13:40 |
| jroll | anyway, this bug just tells me that yes we should implement attach/detach | 13:40 |
| *** thiagop has joined #openstack-ironic | 13:41 | |
| jroll | sure | 13:41 |
| sambetts | yeah but I think we should move attach/detach functionality into the network drviers, and have nova pass it through to Ironic in all cases, and then the network driver can decide if it should do something or not | 13:43 |
| jroll | I agree | 13:43 |
| sambetts | also I wonder if we should work out if we can prevent neutron port deletion if its a baremetal flat port | 13:43 |
| sambetts | unless it comes from nova | 13:43 |
| jroll | why would we need to prevent that? | 13:44 |
| sambetts | hmm, perhaps prevent is the wrong word, but maybe in the flat case we can make port deletion error because it can't be unbound in the flat case | 13:45 |
| jroll | yeah, maybe | 13:45 |
| sambetts | so if the instance is up its bound and unless you delete the instance it can't be unbound | 13:45 |
| thiagop | Morning Ironic | 13:46 |
| sambetts | I think there is some sort of logic like that in nova for VMs right? for if it fails to delete the OVS/linuxbridge ports or something? | 13:46 |
| sambetts | thiagop: congrats on the scholarship | 13:47 |
| thiagop | hey, thanks sambetts | 13:47 |
| milan | morning thiagop, xavierr :) | 13:50 |
| *** fragatina has quit IRC | 13:51 | |
| milan | sambetts, do you think the inspector state patch makes sense esp. how the state is synced with db? | 13:52 |
| NobodyCam | Morning ironic'ers | 13:53 |
| thiagop | morning NobodyCam | 13:53 |
| jlvillal | Good morning Ironic'ers :) | 13:53 |
| *** afaranha has quit IRC | 13:53 | |
| NobodyCam | Morning thiagop, jlvillal :) | 13:53 |
| jlvillal | :) | 13:54 |
| milan | sambetts, some kind of a -2 vs -1 in an early stage of this patch ;) | 13:54 |
| milan | jlvillal, NobodyCam morning! :) | 13:54 |
| NobodyCam | Morning milan ;) | 13:55 |
| jroll | heyo NobodyCam, jlvillal, thiagop | 13:56 |
| thiagop | morning jlvillal | 13:56 |
| openstackgerrit | Merged openstack/python-ironicclient: Fix uuid to UUID in expected error messages https://review.openstack.org/350522 | 13:56 |
| jlvillal | Thanks all :) | 13:57 |
| NobodyCam | Morning jroll ;) | 13:57 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-specs: Boot management for in-band inspection https://review.openstack.org/305864 | 13:57 |
| dtantsur | morning thiagop, jlvillal, NobodyCam | 13:57 |
| NobodyCam | Morning dtantsur :) | 13:58 |
| *** fragatina has joined #openstack-ironic | 14:00 | |
| *** krot_sickleave is now known as krotscheck | 14:00 | |
| *** anush_ has joined #openstack-ironic | 14:00 | |
| jroll | sambetts: this scares me http://logs.openstack.org/20/349520/5/gate/gate-ironic-python35-db/a515e8e/console.html#_2016-08-03_11_36_55_251703 | 14:00 |
| jroll | sambetts: I feel like we've introduced a race/flakiness in that test | 14:01 |
| jroll | that patch didn't change that object https://review.openstack.org/#/c/349520/ | 14:01 |
| dtantsur | ouch | 14:01 |
| dtantsur | maybe it started depending on the overridden defaults? | 14:02 |
| sambetts | ^ this | 14:02 |
| jroll | right | 14:02 |
| sambetts | it hashed the object, so depending on the default set the object will be slightly different :/ so I wonder if we need to add a defined config/reset the object for that test | 14:03 |
| jroll | yeah, that's what I'm doing now | 14:03 |
| *** wajdi_ has quit IRC | 14:04 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Clear overrides before testing object versions https://review.openstack.org/350600 | 14:06 |
| jroll | I think that should do it | 14:06 |
| jroll | I'm just concerned these will keep popping up on various tests | 14:06 |
| *** joprovost1 has joined #openstack-ironic | 14:07 | |
| sambetts | jroll: I'm not sure that will do it, because the objects.node will need reloading | 14:07 |
| sambetts | to rebuild the class before hashing unless it does that somewhere else | 14:07 |
| jroll | mmm | 14:07 |
| *** joprovost has quit IRC | 14:08 | |
| *** joprovost1 is now known as joprovost | 14:08 | |
| sambetts | what we might be able to do is remove the default= thats in the fields= dict, and rely on the constructor setting it | 14:08 |
| sambetts | because that would remove anything set on the class | 14:08 |
| jroll | yeah, that too | 14:08 |
| vdrok | sambetts: iirc when I tried this locally, after setting default in the node constructor, it did change in node.Node.fields['network_interface'].default | 14:10 |
| sambetts | woah?! an instance of an object shouldn't change the class right? | 14:11 |
| sambetts | unless I'm missing something :/ | 14:11 |
| jroll | unless the fields are singletons | 14:11 |
| jroll | O_O | 14:12 |
| * sambetts voms | 14:12 | |
| *** ayoung has joined #openstack-ironic | 14:12 | |
| *** narasimha_SV has quit IRC | 14:13 | |
| sambetts | actually I guess unless on class create it takes a deep copy of the fields dict, every instanec of the node object has the same dict with the same objects in it | 14:14 |
| * jroll bugs dan | 14:14 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Remove default from network_interface field https://review.openstack.org/350600 | 14:14 |
| NobodyCam | morning vdrok :) | 14:15 |
| vdrok | morning NobodyCam ! | 14:15 |
| jroll | sambetts: it is shared :| | 14:15 |
| sambetts | yeah :/ | 14:16 |
| jroll | we dun messed up | 14:16 |
| *** Lord_Phoenix has quit IRC | 14:16 | |
| sambetts | jroll: sigh, why does something that works so well in production because it only loaded once at start, have to suck sooooo much in test | 14:17 |
| sambetts | s | 14:17 |
| jroll | sambetts: well, we just established it doesn't work in prod, right? | 14:17 |
| jroll | because changing one node's default changes the class | 14:17 |
| jroll | we should probably revert that patch | 14:18 |
| sambetts | jroll: but our defaults won't change after conductor start | 14:18 |
| *** amotoki has joined #openstack-ironic | 14:18 | |
| sambetts | API start | 14:18 |
| jroll | sambetts: they will, at each Node.__init__ | 14:18 |
| vdrok | yeah, we don't have sighup thing yet? | 14:18 |
| sambetts | the _default_network_interface function will always return a consistant result | 14:18 |
| jroll | oh wait, this is all about config | 14:18 |
| vdrok | yup | 14:18 |
| jroll | ok | 14:18 |
| *** joprovost1 has joined #openstack-ironic | 14:18 | |
| sambetts | thats why we could actually do it fine on class build | 14:19 |
| jroll | so I guess 350600 isn't that bad | 14:19 |
| sambetts | but because tests are derp | 14:19 |
| jroll | :| | 14:19 |
| vdrok | maybe we can still workaround that thing here? https://github.com/openstack/oslo.versionedobjects/blob/master/oslo_versionedobjects/fixture.py#L268 | 14:19 |
| jroll | although idk if 350600 fixes the thing | 14:19 |
| *** joprovost has quit IRC | 14:19 | |
| *** joprovost1 is now known as joprovost | 14:19 | |
| sambetts | jroll: I don't think it will if the defaults are shared | 14:20 |
| jroll | it's still a valid cleanup, so maybe we land it and see | 14:20 |
| jroll | yeah | 14:20 |
| vdrok | jroll: I think it's still possible to hit this with parallel run | 14:20 |
| sambetts | yeah | 14:20 |
| jroll | ugh | 14:20 |
| jroll | I need to step away for a few, sambetts/vdrok mind seeing if you can figure this out? | 14:21 |
| vdrok | sure | 14:21 |
| sambetts | sure | 14:21 |
| jroll | thanks! | 14:21 |
| sambetts | vdrok: do you think replacing the network_interface with a fresh StringField object in the constructor would work? | 14:22 |
| *** afaranha has joined #openstack-ironic | 14:22 | |
| vdrok | sambetts: the hash will still change I think | 14:22 |
| sambetts | so doing self.fields['network_interface'] = object_fields.StringField( | 14:23 |
| sambetts | nullable=False, default=_default_network_interface()), | 14:23 |
| sambetts | would still modify all the objects? | 14:23 |
| sambetts | including the class? | 14:23 |
| *** links has quit IRC | 14:23 | |
| vdrok | lemme try to be sure | 14:23 |
| sambetts | it would probably be nicer to do that anyway to avoid changing a "private" attribute of the StringField too | 14:24 |
| *** ppiela has quit IRC | 14:26 | |
| sambetts | milan: I have a question about the state property function in node info, is it really a race condition if we don't find a node with the uuid we're looking for? Is the race that the node has been deleted without this node info object knowing? | 14:27 |
| *** wajdi has joined #openstack-ironic | 14:27 | |
| vdrok | sambetts: nope, it still changes | 14:28 |
| *** ppiela has joined #openstack-ironic | 14:28 | |
| sambetts | I guess the dict becomes shared then :( it could be pssoble to do self.fields = copy(Node.fields) then update the network_interface, but I feel like we're getting more and more hacky | 14:30 |
| vdrok | so the thing I was thinking about is to mock the node fields in object registry | 14:31 |
| vdrok | in ovo | 14:31 |
| sambetts | I did think about whether or not we could/should use mock here | 14:32 |
| vdrok | yeah, if the only problem is the unittest then we can do that, if not, idk the good way :) | 14:32 |
| *** rbartal has quit IRC | 14:33 | |
| sambetts | I think it is just the unittests, because in a real world scenario _default_network_interface always returns a consistant output, because it won't change unless you change the config | 14:33 |
| milan | sambetts, it's when the combination of uuid,state changes | 14:34 |
| milan | so someone else updated the record | 14:34 |
| jroll | sambetts: what if we just don't set a default, and when we init the node object, set network_interface if it's none | 14:34 |
| sambetts | milan: you aren't filtering on state=self.state though | 14:35 |
| milan | sambetts, I thought I was :D | 14:35 |
| * milan checks | 14:35 | |
| milan | sambetts, +1 I'm not | 14:36 |
| * jroll actually steps away now | 14:36 | |
| sambetts | jroll: that won't update it if someone updates the node object with network_interface = None :( | 14:36 |
| dtantsur | milan, https://github.com/openstack/requirements/blob/master/global-requirements.txt#L39 | 14:36 |
| jroll | sambetts: and/or make it a proprty? | 14:36 |
| milan | dtantsur, cool, thx! | 14:37 |
| * milan will use that instead in next commit | 14:37 | |
| jroll | sambetts: I feel like we should bug dan about the best way to do this, or go back to my original change | 14:37 |
| sambetts | jroll: the o.ov does all the magic they do by making it a property of the node, but I don't know if you can override properties set on the parent | 14:37 |
| jroll | sambetts: don't see why not | 14:38 |
| jroll | oh, I see | 14:38 |
| jroll | probably best to bug dan s or something :P | 14:39 |
| * jroll really really steps away | 14:39 | |
| *** mtanino has joined #openstack-ironic | 14:40 | |
| *** michauds has joined #openstack-ironic | 14:40 | |
| *** amotoki has quit IRC | 14:40 | |
| vdrok | sambetts: it seems that fields defined in obj_extra_fields are not hashed | 14:43 |
| thiagop | If some core have a review slot to +A https://review.openstack.org/#/c/286192/, I'll deeply appreciate. It has 2 +2s already | 14:43 |
| *** amotoki has joined #openstack-ironic | 14:45 | |
| milan | dtantsur, class MyE(enum.Enum): a = 1 ; MyE.a == 1 # ~> False O_o | 14:46 |
| dtantsur | I suspect it's intentional... | 14:47 |
| milan | MyE.a.value == 1 works | 14:47 |
| milan | so code would be States.error.value :-/ | 14:47 |
| *** fragatina has quit IRC | 14:49 | |
| *** fragatina has joined #openstack-ironic | 14:49 | |
| sambetts | vdrok: does that mean we can put network_interface into a different place and then it won't be considered in the hash? | 14:50 |
| vdrok | sambetts: I'm looking at the code now, it seems to be the case, but we'll have to do all the property magic by ourselves | 14:51 |
| apughazh | Good morning everyone, my patch needs a workflow +1 could someone give it a look? | 14:51 |
| apughazh | https://review.openstack.org/#/c/346894/ | 14:51 |
| vdrok | sambetts: because in the metaclass properties are created only for fields attribute - https://github.com/openstack/oslo.versionedobjects/blob/master/oslo_versionedobjects/base.py#L52 | 14:52 |
| sambetts | vdrok: I have an idea! We can subclass the StringField type so that it can take a function as the default | 14:53 |
| openstackgerrit | Kyrylo Romanenko proposed openstack/python-ironicclient: Add openstack baremetal driver commands https://review.openstack.org/350050 | 14:54 |
| sambetts | vdrok: then the node class would be consistant because the functions object id wouldn't change | 14:54 |
| sambetts | vdrok: shall I put up a ptch | 14:54 |
| sambetts | ? | 14:54 |
| vdrok | sambetts: hm, that might work, if there is nothing tied to the type of the default in the base class | 14:55 |
| sambetts | vdrok: nothing I can see, and anything that is we can override | 14:55 |
| vdrok | sambetts: we'll also have to define our own string type that handles functions in the default then | 14:58 |
| vdrok | because of this - https://github.com/openstack/oslo.versionedobjects/blob/master/oslo_versionedobjects/fields.py#L169 | 14:58 |
| sambetts | vdrok: thats the exact thing I'm overriding | 14:59 |
| vdrok | yeah, right, that should work :) | 14:59 |
| *** fragatina has quit IRC | 15:02 | |
| sambetts | vdrok: just testing it now | 15:02 |
| jroll | woo | 15:04 |
| *** amotoki has quit IRC | 15:06 | |
| jroll | apughazh: left a question inline | 15:06 |
| *** mjura has quit IRC | 15:06 | |
| jroll | looks fine, but consistency and all | 15:07 |
| *** PollyZ has joined #openstack-ironic | 15:07 | |
| lucasagomes | strange error: http://logs.openstack.org/82/236982/49/gate/gate-ironic-python34-db/d18eecb/console.html#_2016-08-03_13_24_14_975719 | 15:09 |
| lucasagomes | this failed for py34 in gate, but passed on 27 and 35 | 15:09 |
| sambetts | lucasagomes: working on it now | 15:09 |
| lucasagomes | maybe something merged in between and changed it!? | 15:09 |
| lucasagomes | sambetts, oh, right, what's the problem? | 15:09 |
| vdrok | lucasagomes: that's a race | 15:09 |
| *** rama_y has joined #openstack-ironic | 15:09 | |
| lucasagomes | ew :-( | 15:09 |
| *** Sylvester has quit IRC | 15:10 | |
| *** dprince has quit IRC | 15:10 | |
| * jroll files a bug for visibility | 15:10 | |
| vdrok | because we might change network_interface field default and the hash of the node class will change | 15:10 |
| jroll | sambetts: here's the bug for your commit message https://bugs.launchpad.net/ironic/+bug/1609455 | 15:12 |
| openstack | Launchpad bug 1609455 in Ironic "flaky unit test for object hashes" [Critical,Confirmed] | 15:12 |
| sambetts | jroll: thanks :) | 15:12 |
| apughazh | jroll: Let me double check on that for you | 15:13 |
| *** rpioso has joined #openstack-ironic | 15:16 | |
| pas-ha | got -2 from Jenkins on keystoneauth patch. py34 unitests fail for objects fingerprints mismatch http://logs.openstack.org/82/236982/49/gate/gate-ironic-python34-db/d18eecb/console.html#_2016-08-03_13_24_14_976095 - should I reverify or I try rebasing to master? | 15:16 |
| pas-ha | btw lucasagomes thanks a lot :) | 15:17 |
| lucasagomes | pas-ha, sambetts is looking into it | 15:17 |
| lucasagomes | so probably don't recheck yet | 15:17 |
| pas-ha | ok, cool | 15:17 |
| pas-ha | won't | 15:17 |
| * jroll wonders if we should revert in the meantime | 15:19 | |
| *** priteau has quit IRC | 15:19 | |
| vdrok | jroll: if we revert, it might be easier to go with your first approach I think | 15:20 |
| jroll | vdrok: yeah, but it's painful from a dev perspective | 15:20 |
| *** amotoki has joined #openstack-ironic | 15:22 | |
| dtantsur | I'd wait for a fix, landing a revert will bring us fun with merge conflicts, etc | 15:23 |
| *** hoangcx has joined #openstack-ironic | 15:24 | |
| openstackgerrit | Alexander Bashmakov proposed openstack/ironic: Rewrite devstack/tools/ironic/scripts/create-node.sh in Python. https://review.openstack.org/346832 | 15:31 |
| sambetts | jroll: py35 tox env makes me sad everytime I try to run tox through without any params, I have no idea why but it won't even install the virtualenv on my VM | 15:35 |
| JayF | sambetts: you may not have python 3.5 binary on your vm? | 15:36 |
| sambetts | JayF: I think its something to do with missing C libraries | 15:36 |
| JayF | weird; if you pastebin a thing I'll take a look and try to help | 15:36 |
| sambetts | JayF: I need to dig into it, but it just dumps the worlds biggest trace into my terminal :-P | 15:36 |
| jroll | maybe python3.5-dev or something? | 15:37 |
| sambetts | something to do with pillow and libimaging I think | 15:38 |
| sambetts | at first glance | 15:38 |
| sambetts | :/ | 15:38 |
| jroll | hrm | 15:38 |
| jroll | idk, I only run 2.7 locally and let the gate handle the rest :P | 15:38 |
| dtantsur | ditto here :) | 15:38 |
| jroll | er, s/gate/CI/ | 15:39 |
| sambetts | hehe, I just like the py27 followed by pep8 thing, maybe I'll alias tox to tox -epy27 pep8 | 15:39 |
| * dtantsur is learning German, so when he sees a sentence starting with "er" he automatically translates it as "he" >_< | 15:39 | |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Operator documentation for multitenancy https://review.openstack.org/228496 | 15:39 |
| sambetts | dtantsur: oh nice! any particular reason or just for fun? | 15:40 |
| dtantsur | sambetts, largely for fun.. with a vague and uncertain idea to move there eventually | 15:40 |
| jlvillal | Does anyone know off the top of their head. When doing a gate job and devstack. Does it install the packages in test-requirements.txt? | 15:40 |
| vdrok | dtantsur: then study swiss german :) I've heard it's very different | 15:41 |
| jlvillal | Just wondering if code that only runs for a devstack run should have the requirement listed in requirements.txt or test-requirements.txt | 15:41 |
| dtantsur | jlvillal, it does iirc. at least used to | 15:41 |
| JayF | jlvillal: how could something be required for tempest tests in devstack but not be required for "normal" runs? | 15:41 |
| jroll | eh, I don't think it does | 15:41 |
| sambetts | only if it runs the tox env | 15:41 |
| jlvillal | JayF: It is python code in devstack/.... | 15:41 |
| JayF | oh poop | 15:42 |
| JayF | maybe that's a good reason to have left that script as bash... | 15:42 |
| * JayF not sure why we even had a bug about making it python when all of devstack is bash | 15:42 | |
| JayF | I don't wanna add something to ironic requirements if it's only needed by devstack though. | 15:42 |
| * jlvillal looks around and wonders if anyone knows he filed the RFE bug... | 15:43 | |
| dtantsur | https://github.com/openstack-dev/devstack/blob/master/inc/python#L165-L176 | 15:43 |
| jlvillal | dtantsur: Woot! thanks :) | 15:43 |
| dtantsur | so I suspect it does, it might make sense to verify by looking at some existing logs | 15:43 |
| jroll | what does it need to add to requirements? | 15:45 |
| *** pece has joined #openstack-ironic | 15:45 | |
| * jroll clicks the review | 15:45 | |
| dtantsur | I don't think there is any reason to write so much code in bash, but I don't know the details ofc | 15:46 |
| jroll | dtantsur: you were the one that asked for this, iirc :) | 15:46 |
| dtantsur | well, it's likely :) | 15:46 |
| JayF | dtantsur: Devstack is full of bash :) | 15:46 |
| jlvillal | JayF: Don't bash devstack now ;) | 15:47 |
| dtantsur | I'd not be proud of it... | 15:47 |
| dtantsur | also I remember someone here disliking DIB for this reason ;) | 15:47 |
| * jroll wishes we were spending time on higher priorities tbh | 15:47 | |
| JayF | jlvillal: it has a tough shell | 15:47 |
| jlvillal | heh | 15:47 |
| jroll | grooooan | 15:48 |
| jlvillal | JayF: So are you officially a Pacific NW resident now? | 15:48 |
| JayF | jlvillal: since ~7/5 | 15:48 |
| JayF | jroll: tired from having bourne the weight of all my puns again? | 15:48 |
| dtantsur | jroll, do we have a list of priority patches just awaiting reviews? | 15:48 |
| jlvillal | JayF: Sweet. Welcome to gray sky winters! | 15:48 |
| JayF | jlvillal: I don't mind so much :) | 15:49 |
| jroll | dtantsur: trello? it doesn't say if things are waiting for reviews or updates, though | 15:49 |
| jlvillal | JayF: I know some people who use the "happy lights" during the winter. | 15:49 |
| dtantsur | yeah, I've opened https://review.openstack.org/#/q/status:open+branch:master+topic:bug/1526403 and most of stuff is red :( | 15:50 |
| jroll | dtantsur: let me put it this way, I wish people were reviewing the agent APIs instead of this :) | 15:50 |
| dtantsur | ++ | 15:50 |
| dtantsur | btw we could use some non-inspector reviews on the inspector LLDP patch https://review.openstack.org/#/c/321082/ | 15:50 |
| jroll | sure | 15:50 |
| * jroll does it | 15:50 | |
| * JayF notes metrics is a nice to have prio https://review.openstack.org/#/c/350317/ https://review.openstack.org/#/c/301919/ | 15:50 | |
| dtantsur | my agent API patch conflicts with both metrics and keystone policy >_< | 15:51 |
| * dtantsur reviews the former | 15:51 | |
| jroll | oh my, https://review.openstack.org/#/c/321082/15/ironic_inspector/plugins/local_link_connection.py | 15:52 |
| dtantsur | blame sambetts for everything there :D | 15:52 |
| JayF | dtantsur: should we put some of this LLDP parsing code in ironic_lib and share between inspector and IPA? | 15:52 |
| JayF | dtantsur: b/c IPA already does some of this | 15:52 |
| dtantsur | JayF, I don't think IPA does the same, we're basing on the results from IPA. but will defer it to sambetts | 15:53 |
| sambetts | JayF: IPA does none of this | 15:53 |
| *** moshele has quit IRC | 15:53 | |
| sambetts | JayF: IPA does not processing of the LLDP packet passed dviding it into types and values | 15:53 |
| JayF | sambetts: https://github.com/openstack/ironic-python-agent/blob/master/ironic_python_agent/netutils.py | 15:53 |
| *** PollyZ has quit IRC | 15:53 | |
| *** Sukhdev has joined #openstack-ironic | 15:54 | |
| sambetts | JayF: I use that code to get this information | 15:54 |
| JayF | sambetts: basically IPA is doing a D- job, and that inspector code does an A+ job? | 15:54 |
| dtantsur | JayF, do I get it right that you'll kill me with fire if I ask to rebase metrics against the agent IPA patch? | 15:55 |
| lazy_prince | yeah.. so we code move the code to ironic-lib and let IPA also do A+ job.. | 15:55 |
| sambetts | JayF: we don't want any processing on the IPA side because it depends on vendor, and rebuilding ramdisks is hard and loading plugins in inspector is easy | 15:55 |
| sambetts | lazy_prince: ^ | 15:55 |
| JayF | dtantsur: honestly, I'd really, really rather just land my patch :( I think it's less difficult to review and less likely to have churn | 15:55 |
| JayF | dtantsur: but I bet you feel the same way the other way so *shruG* | 15:56 |
| dtantsur | sigh... | 15:56 |
| JayF | dtantsur: yeah, exactly | 15:56 |
| openstackgerrit | Akilan Pughazhendi proposed openstack/ironic: Updating dev quickstart to include compatiblity for newest distros https://review.openstack.org/346894 | 15:56 |
| * JayF been trying to get that reviewed and landed for weeks | 15:56 | |
| dtantsur | JayF, anyway, the change LGTM, let the fastest win :) | 15:57 |
| JayF | dtantsur: I think the A+ case is policy and metrics lands this morning, agent api rebased and lands tomorrow | 15:58 |
| dtantsur | this is not impossible :) | 15:58 |
| *** PollyZ has joined #openstack-ironic | 15:58 | |
| dtantsur | JayF, an idea for improvement: by default derive the metrics name from the callable name | 15:59 |
| dtantsur | would save a lot of typing | 15:59 |
| JayF | dtantsur: is there a way to do that *without* modifying the metrics library? | 16:00 |
| JayF | dtantsur: __name__ won't work in the decorator aiui | 16:00 |
| dtantsur | JayF, wait, why? | 16:00 |
| JayF | I'm pretty sure it'll evaluate outside the function it's wrapping | 16:01 |
| JayF | like you'd need __name__ inside the method, but you're passing in __name__ outside the method | 16:01 |
| *** mgoddard_ has joined #openstack-ironic | 16:01 | |
| dtantsur | JayF, I may be misunderstanding something, but I meant something like http://paste.openstack.org/show/547806/ | 16:01 |
| JayF | dtantsur: that would require modifying the library and getting another ironic-lib release out :( | 16:02 |
| JayF | dtantsur: I agree it'd be an improvement to default to that if nothing is passed in | 16:02 |
| dtantsur | sure, I don't plan on blocking due to it. just an idea for the future | 16:02 |
| JayF | dtantsur: but not sure it's worth the time to jump through all the hoops today | 16:02 |
| dtantsur | no, let's get the whole thing in, then refactor | 16:03 |
| JayF | dtantsur: btw part of why I want that agent metrics piece in is I'd like to land all the inside-ipa metrics before the cycle ends too :( | 16:03 |
| dtantsur | fair enough | 16:03 |
| jroll | getting an ironic-lib release out is easy, fwiw, and we'll be doing it in 3 weeks no matter what | 16:04 |
| JayF | sambetts: I grok what you meant about the LLDP thing now | 16:04 |
| dtantsur | JayF, do you plan on adding metrics to https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/agent_base_vendor.py as well? without it https://review.openstack.org/#/c/350317/ looks incomplete | 16:05 |
| sambetts | JayF: awsome :) | 16:05 |
| *** mgoddard has quit IRC | 16:05 | |
| JayF | dtantsur: yeah; if you want it in the same patch I can jfdi right now | 16:05 |
| jroll | sambetts: did you figure out the objects thing? | 16:06 |
| *** jistr is now known as jistr|biab | 16:06 | |
| JayF | dtantsur: it's just sorta deps all the way down :) if I follow all the dep trees up and down I end up with one giant hairy metrics patch | 16:06 |
| JayF | lol | 16:06 |
| sambetts | jroll: yup, I have it working locally, just making it better then I'll push the fix :D | 16:07 |
| JayF | dtantsur: but I landed "agent deploy driver metrics" before without that, so I should add it | 16:07 |
| jroll | sambetts: awesome, ty | 16:07 |
| * jroll anxious to get networking landed in nova | 16:07 | |
| dtantsur | JayF, I'd prefer you add it | 16:08 |
| JayF | dtantsur: already halfway there, lol | 16:08 |
| dtantsur | cool! | 16:08 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ironic: Updated from global requirements https://review.openstack.org/350097 | 16:09 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ironic-inspector: Updated from global requirements https://review.openstack.org/350669 | 16:09 |
| *** bvandewa has joined #openstack-ironic | 16:09 | |
| *** bvandewa has quit IRC | 16:09 | |
| dtantsur | I have to bail out already, sorry. I'll check whatever remains of the metrics patches tomorrow morning | 16:11 |
| jroll | dtantsur: before you go | 16:13 |
| openstackgerrit | Jay Faulkner proposed openstack/ironic: Timing metrics: pxe boot and iscsi deploy driver https://review.openstack.org/350317 | 16:13 |
| JayF | dtantsur: ^ | 16:13 |
| jroll | dtantsur: did we bikeshed these agent endpoints properly in the spec? e.g. /v1/nodes/uuid/heartbeat vs /v1/heartbeat? | 16:13 |
| dtantsur | no, I think devananda was against the former | 16:13 |
| jroll | so we did bikeshed about it :) | 16:13 |
| jroll | thanks | 16:14 |
| *** bvandewa has joined #openstack-ironic | 16:14 | |
| JayF | this violates REST principles. "heartbeat" is not a member of the nodes/<UUID>/ resource, and is not returned from a call to "GET /v1/nodes/<UUID/" so therefor is not discoverable | 16:14 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/python-ironicclient: Updated from global requirements https://review.openstack.org/348673 | 16:14 |
| JayF | ^ from deva, patch set 7 on the spec | 16:14 |
| JayF | looks properly bikeshedded to me | 16:15 |
| JayF | lol | 16:15 |
| dtantsur | JayF, +2 and I have to run for real :) | 16:15 |
| dtantsur | see you all tomorrow | 16:15 |
| JayF | :D | 16:15 |
| *** dtantsur is now known as dtantsur|afk | 16:15 | |
| openstackgerrit | Sam Betts proposed openstack/ironic: Enable StringFields to take functions as defaults https://review.openstack.org/350679 | 16:15 |
| sambetts | jroll, vdrok ^^ | 16:16 |
| jroll | sambetts: yay | 16:16 |
| sambetts | hopefully we shouldn't have to use this code much, and really it should be deprecated when we deprecate the dhcp providers I think | 16:17 |
| *** Goneri has quit IRC | 16:18 | |
| sambetts | although I guess we'll still have the default_network_interface config option to contend with | 16:18 |
| *** PollyZ has quit IRC | 16:18 | |
| * JayF only +1s that one | 16:18 | |
| vdrok | sambetts: hmm, that's not what I thought | 16:18 |
| jroll | sambetts: I wonder if that should be a different class | 16:19 |
| sambetts | jroll: to make it more explict? | 16:19 |
| jroll | sambetts: yeah, but maybe not necessary | 16:20 |
| sambetts | vdrok: what were you thinking? | 16:20 |
| vdrok | sambetts: why not overload String coerce? and only __repr__ in StringField? that way we would avoid overriding private _null method https://github.com/openstack/oslo.versionedobjects/blob/master/oslo_versionedobjects/fields.py#L254 | 16:20 |
| sambetts | vdrok: because _null uses self._default directly, and we can't change that without changing all instances of the Node object | 16:21 |
| *** jistr|biab is now known as jistr | 16:22 | |
| vdrok | sambetts: yep, but we don't need to change that I think - https://github.com/openstack/oslo.versionedobjects/blob/master/oslo_versionedobjects/fields.py#L169 | 16:23 |
| *** hshiina has quit IRC | 16:23 | |
| vdrok | if self._type here will be our new String, that accepts function as a value argument to coerce | 16:24 |
| vdrok | we wouldn't need to change it | 16:24 |
| vdrok | self._default != UnspecifiedDefault will still be true for a function | 16:25 |
| sambetts | so you want a new String FieldType ? | 16:25 |
| vdrok | yup | 16:26 |
| jroll | accepting a function as a value in coerce() would allow someone to set a function though (e.g. node.network_interface = some_func) | 16:26 |
| jroll | which isn't great | 16:26 |
| jroll | but probably okay | 16:26 |
| vdrok | dunno if it is much better, but still we're not overrding private methods | 16:26 |
| *** ccamacho is now known as ccamacho|out | 16:26 | |
| jroll | yeah, maybe I do like overriding coerce() instead | 16:26 |
| jroll | class StringThatAcceptsAFunctionType() :P | 16:27 |
| vdrok | :) | 16:27 |
| *** dansmith has quit IRC | 16:27 | |
| sambetts | we'd still need to override the __repr__ on the field, but I guess thats ok | 16:27 |
| vdrok | yep | 16:27 |
| vdrok | __repr__ is a standard thing so is OK | 16:28 |
| *** dansmith has joined #openstack-ironic | 16:28 | |
| *** dansmith is now known as Guest20988 | 16:28 | |
| jroll | I do think the upstream o.vo patch in the future should change _null | 16:28 |
| jroll | and then we can drop our subclass | 16:28 |
| *** ohamada has quit IRC | 16:29 | |
| *** Guest20988 is now known as dansmith | 16:29 | |
| *** dprince has joined #openstack-ironic | 16:33 | |
| sambetts | jroll, vdrok: can we think of a better name that StringFieldThatAcceptsAFunction? | 16:34 |
| sambetts | than* | 16:34 |
| jroll | lol | 16:34 |
| sambetts | I guess its explict ;) | 16:34 |
| jroll | StringAcceptsCallableField? | 16:34 |
| jroll | with a giant note about it being a hack | 16:34 |
| vdrok | yeah, whatever we call it it just needs to mention in the docstring DO NOT USE DO NOT INHERIT :) | 16:35 |
| sambetts | heh yes :-P | 16:35 |
| sambetts | tbh it feels less of a hack now I'm not overriding privates ;-P | 16:36 |
| devananda | morning, all | 16:37 |
| sambetts | hey devananda | 16:37 |
| jroll | morning devananda | 16:37 |
| *** rcernin has quit IRC | 16:38 | |
| *** syed_ has joined #openstack-ironic | 16:38 | |
| vdrok | devananda: morning | 16:39 |
| * devananda dives back into the policy work before other distractions find him | 16:39 | |
| *** Sukhdev has quit IRC | 16:40 | |
| *** vmud213 has joined #openstack-ironic | 16:41 | |
| *** pece has quit IRC | 16:44 | |
| *** e0ne has quit IRC | 16:44 | |
| lucasagomes | folks, I'm calling it a day | 16:45 |
| lucasagomes | have a great evening all | 16:45 |
| sambetts | o/ lucasagomes | 16:45 |
| lucasagomes | :-) | 16:45 |
| *** lucasagomes is now known as lucas-dinner | 16:45 | |
| *** hoangcx has quit IRC | 16:46 | |
| *** vmud213 has quit IRC | 16:49 | |
| *** vmud213 has joined #openstack-ironic | 16:49 | |
| *** krotscheck is now known as kro_focused | 16:53 | |
| openstackgerrit | Kyrylo Romanenko proposed openstack/python-ironicclient: Add openstack baremetal driver commands https://review.openstack.org/350050 | 16:54 |
| *** afaranha_ has joined #openstack-ironic | 16:57 | |
| openstackgerrit | Sam Betts proposed openstack/ironic: Create a custom StringField that can process functions https://review.openstack.org/350679 | 16:58 |
| sambetts | jroll, vdrok ^^ | 16:58 |
| *** afaranha has quit IRC | 16:58 | |
| *** pcaruana has quit IRC | 17:00 | |
| vdrok | sambetts: thx, looks good! will wait for the gate | 17:00 |
| openstackgerrit | Alexander Bashmakov proposed openstack/ironic: Rewrite devstack/tools/ironic/scripts/create-node.sh in Python. https://review.openstack.org/346832 | 17:00 |
| jroll | sambetts: couple things there | 17:01 |
| openstackgerrit | Stephanie Miller proposed openstack/python-ironicclient: Add volume_connector support https://review.openstack.org/214786 | 17:03 |
| *** tesseract- has quit IRC | 17:03 | |
| sambetts | jroll: updating now | 17:04 |
| jroll | thanks | 17:04 |
| openstackgerrit | Stephanie Miller proposed openstack/python-ironicclient: Add volume_connector support https://review.openstack.org/214786 | 17:05 |
| vdrok | jroll: hm, didn't know super with staticmethod is possible :) | 17:06 |
| openstackgerrit | Sam Betts proposed openstack/ironic: Create a custom StringField that can process functions https://review.openstack.org/350679 | 17:06 |
| sambetts | vdrok: apprently so ^ | 17:06 |
| *** openstackgerrit_ has joined #openstack-ironic | 17:06 | |
| vdrok | yep | 17:06 |
| openstackgerrit | Stephanie Miller proposed openstack/python-ironicclient: Add volume_target support https://review.openstack.org/285233 | 17:07 |
| *** openstackgerrit_ has quit IRC | 17:08 | |
| *** Romanenko_K has quit IRC | 17:08 | |
| jroll | anything is possible in python :P | 17:08 |
| thiagop | jroll: zombo.com | 17:08 |
| thiagop | "you can do anything, anything at all" | 17:08 |
| *** Romanenko_K has joined #openstack-ironic | 17:08 | |
| jroll | sambetts: see the unit test comment? | 17:09 |
| sambetts | jroll: woops, nope :-P I went straght to fields.py | 17:09 |
| jroll | heh | 17:10 |
| *** mgould is now known as mgould|afk | 17:11 | |
| *** mgoddard_ has quit IRC | 17:12 | |
| *** mgoddard has joined #openstack-ironic | 17:12 | |
| *** vmud213 has quit IRC | 17:13 | |
| *** Romanenko_K has quit IRC | 17:15 | |
| *** Romanenko_K has joined #openstack-ironic | 17:17 | |
| *** PollyZ has joined #openstack-ironic | 17:17 | |
| *** moshele has joined #openstack-ironic | 17:19 | |
| openstackgerrit | Stephanie Miller proposed openstack/python-ironicclient: Add volume_target support https://review.openstack.org/285233 | 17:19 |
| *** e0ne has joined #openstack-ironic | 17:27 | |
| *** e0ne has quit IRC | 17:28 | |
| * jlvillal eats his popcorn and reads: http://paste.openstack.org/show/545744/ | 17:33 | |
| openstackgerrit | Akilan Pughazhendi proposed openstack/ironic: Updated Dev quickstart for viewing doc changes https://review.openstack.org/350705 | 17:37 |
| *** daemontool_ has quit IRC | 17:38 | |
| mat128 | jlvillal: was that on the ML? | 17:43 |
| openstackgerrit | Milan Kováčik proposed openstack/ironic-inspector: Introducing node introspection state management https://review.openstack.org/348943 | 17:43 |
| *** PollyZ has quit IRC | 17:43 | |
| *** e0ne has joined #openstack-ironic | 17:43 | |
| jlvillal | mat128: It doesn't appear so. Or if it was it was a link to the paste. Though I didn't see it | 17:43 |
| JayF | https://review.openstack.org/#/c/350317 easy review if someone wants to land a thing | 17:43 |
| *** narasimha_SV has joined #openstack-ironic | 17:44 | |
| JayF | https://review.openstack.org/#/c/350502 similar one here (not mine though, thanks alineb ) | 17:44 |
| narasimha_SV | http://paste.openstack.org/show/547819/ | 17:45 |
| narasimha_SV | while executing ironic-dbsync --config-file /etc/ironic/ironic.conf create_schema | 17:45 |
| narasimha_SV | getitng the above ERROR | 17:45 |
| *** moshele has quit IRC | 17:46 | |
| mat128 | narasimha_SV: all I can think of is reserved keyword? | 17:46 |
| *** bvandewa has quit IRC | 17:47 | |
| jroll | what version of mysql? | 17:47 |
| *** mat128 is now known as mat128|afk | 17:47 | |
| narasimha_SV | jroll: mysql version - 5.6.28 | 17:50 |
| openstackgerrit | Sam Betts proposed openstack/ironic: Create a custom StringField that can process functions https://review.openstack.org/350679 | 17:51 |
| sambetts | jroll: just wanted to cover as many bases as possible ^ | 17:51 |
| jroll | sambetts: awesome | 17:52 |
| *** milan has quit IRC | 17:53 | |
| * sambetts needs a beer now | 17:53 | |
| devananda | jroll: 'online' does not appear in the reserved keywords list: https://dev.mysql.com/doc/refman/5.7/en/keywords.htmlKeywords | 17:54 |
| jroll | devananda: agree | 17:54 |
| jroll | I'm not sure why that'd break | 17:54 |
| narasimha_SV | I am using kolla to deploy ironic. | 17:56 |
| narasimha_SV | sometime node_tags and portgroup tables are not getting created | 17:57 |
| narasimha_SV | and sometimes getting this issue | 17:57 |
| JayF | This happens during every devstack build, doesn't it? | 17:57 |
| devananda | that sounds like a bug in kolla? | 17:57 |
| JayF | like I'm wondering what could be different about kolla vs the normal devstack install to cause this failure | 17:57 |
| jroll | sambetts: +2, someone mind hitting the button on 350679? | 17:57 |
| jroll | yeah agree, we run this on every CI run | 17:57 |
| *** bvandewa has joined #openstack-ironic | 17:57 | |
| narasimha_SV | there no much difference apart from kolla deploys on top of docker containers | 17:58 |
| narasimha_SV | both use the git code to build the components | 17:58 |
| *** harlowja has quit IRC | 17:58 | |
| narasimha_SV | http://paste.openstack.org/show/547827/ | 17:59 |
| narasimha_SV | this time tables got created but node_tags and portgroup are missing | 17:59 |
| *** e0ne has quit IRC | 18:00 | |
| *** harlowja has joined #openstack-ironic | 18:00 | |
| jroll | yeah, that's certainly odd | 18:00 |
| devananda | narasimha_SV: is kolla pulling from master, or installing from packages? | 18:01 |
| jroll | I can't find this code in kolla either | 18:01 |
| narasimha_SV | devananda: yes | 18:01 |
| * sambetts is calling it night | 18:02 | |
| devananda | vdrok: your last comment on https://review.openstack.org/#/c/236982/49 -- the issue is being fixed. what do you mean? | 18:02 |
| jlvillal | sambetts: Ciao! | 18:02 |
| sambetts | cya y'all tomorrow o/ | 18:02 |
| *** sambetts is now known as sambetts|afk | 18:02 | |
| vdrok | devananda: https://review.openstack.org/350679 | 18:02 |
| vdrok | we have a race in unittests right now, that^ is the fix | 18:03 |
| devananda | ahh. ty | 18:04 |
| *** PollyZ has joined #openstack-ironic | 18:06 | |
| devananda | vdrok: afaict, once those two patches land, we should be able to switch ironic devstack plugin to keystone v3. WDYT? | 18:06 |
| vdrok | devananda: yep, that should work now | 18:07 |
| vdrok | after those I mean :) | 18:07 |
| devananda | it passed in my local tests after just applying the first patch | 18:07 |
| devananda | testing now with the devstack patch, too | 18:07 |
| devananda | awesome | 18:07 |
| *** Romanenko_K has quit IRC | 18:12 | |
| devananda | vdrok: any issue if I rebase 236982 on current master? | 18:12 |
| devananda | that'll make rebasing my policy patches on it a bit cleaner | 18:12 |
| vdrok | devananda: sure, I guess it won't hurt | 18:13 |
| openstackgerrit | Merged openstack/ironic: Pass agent metrics config via conductor https://review.openstack.org/301919 | 18:19 |
| JayF | \o/ | 18:20 |
| *** jpich has quit IRC | 18:22 | |
| *** electrofelix has quit IRC | 18:22 | |
| openstackgerrit | Jay Faulkner proposed openstack/ironic-python-agent: Add metrics support to IPA https://review.openstack.org/119981 | 18:25 |
| * vdrok leaves | 18:25 | |
| vdrok | good night everyone! | 18:25 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: [WIP] add test to ensure policy is always authorized https://review.openstack.org/350177 | 18:26 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Add keystone policy support to Ironic https://review.openstack.org/325599 | 18:26 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Update devstack plugin with new auth options https://review.openstack.org/297312 | 18:26 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Migrate to using keystoneauth Sessions https://review.openstack.org/236982 | 18:26 |
| JayF | devananda: ^ isn't that migrate to using keystoneauth sessions going to break backwards compat for existing configs for talking to neutron/glance? | 18:29 |
| JayF | hmm. I guess the stuff in ironic/common/keystone.py does it | 18:31 |
| devananda | JayF: it passed grenade (tm) | 18:32 |
| devananda | :) | 18:32 |
| JayF | Hrh | 18:32 |
| JayF | Heh | 18:32 |
| JayF | I guess that should be somewhat comforting | 18:33 |
| JayF | I just know reading this that I'll be real careful(tm) when deploying this to my environment | 18:33 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Test that network_interface is explicitly set on POST/PATCH https://review.openstack.org/349520 | 18:34 |
| devananda | yea - it's a significant change to configs for services auth. I'd expect you to be careful. | 18:34 |
| devananda | but it's also a really good improvement | 18:34 |
| * jroll approves that, rebased on top of the flaky test fix thing | 18:34 | |
| JayF | I'll take your word for it w/r/t it being better :P. I know it's what we're supposed to do, at least | 18:35 |
| devananda | JayF: it's better for everyone who actually uses keystone :P | 18:35 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Add a data migration to fill node.network_interface https://review.openstack.org/349528 | 18:35 |
| devananda | jroll: ++ | 18:35 |
| * TheJulia looks at the session stuff, looks at what she wrote for a cinder interface, and climbs back into the code | 18:35 | |
| jroll | approved that second one as well | 18:35 |
| jroll | there's nothing without a +W in the nova networking dep chain | 18:36 |
| devananda | TheJulia: did you see the keystoneauth patch? that may either conflict with - or make easier - your cinder integrations | 18:36 |
| *** amotoki has quit IRC | 18:37 | |
| TheJulia | devananda: somewhat cleaner... I think, I'd have to swap the ksa session tokens stuffs back into my brain | 18:38 |
| *** ayoung has quit IRC | 18:38 | |
| *** ccamacho|out is now known as ccamacho | 18:39 | |
| TheJulia | devananda: https://review.openstack.org/#/c/236982/50/ironic/common/neutron.py@44 sticks out as a possible thing to dig into further because I think there was a desire to move to auth_plugin = None as the param, but I'm not aware of anyone using cinder in noauth mode right now... and the auth changes in ksa are a blur in my brain right now. | 18:44 |
| TheJulia | err, s/cinder/neutron/ | 18:45 |
| *** ametts has joined #openstack-ironic | 18:45 | |
| * TheJulia goes back to cinder code before brain goes even more off track | 18:45 | |
| JayF | devananda: do you want to add-back the change to keystone v3 as a part of your policy patch? | 18:45 |
| JayF | devananda: since it's rebased on the pieces that should add v3 compat, right? | 18:46 |
| jroll | I feel like that is actually a variable | 18:48 |
| jroll | or should be one | 18:48 |
| jroll | there's #ENABLE_IDENTITY_V2 | 18:51 |
| jroll | s/#/$/ | 18:51 |
| JayF | I can't find the piece I was looking for | 18:51 |
| JayF | where in a previous patchset it was changed | 18:51 |
| JayF | aha, found it, and it is set to v3 | 18:51 |
| JayF | that's what I get for looking at patchset over patchset diffs | 18:51 |
| JayF | https://review.openstack.org/#/c/325599/24/ironic/tests/unit/api/base.py ln 48 is what I was talking about jroll | 18:52 |
| jroll | ah, idk why that would matter | 18:52 |
| jroll | there's a devstack variable for it too, that was changed | 18:52 |
| JayF | https://review.openstack.org/#/c/325599/24/devstack/lib/ironic | 18:53 |
| JayF | where? | 18:53 |
| jroll | JayF: previous patch https://review.openstack.org/#/c/325599/20/devstack/lib/ironic | 18:53 |
| JayF | aha, yes, that's what I was looking for | 18:53 |
| JayF | for some reason, 20..24 didn't show that | 18:54 |
| JayF | I'm wondering if we want to bump that back now that it's based on the keystoneauth stuff | 18:54 |
| jroll | well, devstack is supposed to work with v2 and v3 | 18:55 |
| JayF | ah, the keystone auth change completely does away with that setting I believe | 18:55 |
| jroll | I'm not sure if one or the other is always present | 18:55 |
| jroll | I guess we should move to v3 asap though | 18:55 |
| openstackgerrit | Namrata proposed openstack/ironic: Add guru meditation report for ironic https://review.openstack.org/350736 | 18:55 |
| JayF | https://review.openstack.org/#/c/297312/26/devstack/lib/ironic | 18:56 |
| JayF | completely removes that option | 18:56 |
| JayF | and I think represents a change to devstack running by default in ks v3 | 18:56 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ironic: Updated from global requirements https://review.openstack.org/350097 | 18:56 |
| jroll | oh, neat | 18:56 |
| jroll | I feel like keystoneauth1 will figure out what's available and do the right thing | 18:57 |
| JayF | exactly | 18:57 |
| JayF | that's the piece I was missing before | 18:57 |
| JayF | I have +1s/+2s across those patches now | 18:58 |
| JayF | didn't +2 the big keystoneauth sessions patch b/c I don't understand it well | 18:58 |
| JayF | but if you wanna go through them now I think they can all land | 18:58 |
| jroll | I thought we landed keystoneauth this morning | 18:58 |
| JayF | it failed unit tests | 18:59 |
| jroll | rebase, whomp | 18:59 |
| JayF | and deva rebased it | 18:59 |
| JayF | I have the devstack stuff already +2A so when you land that it'll alnd the devstack stuff too | 18:59 |
| JayF | we should get the policy stuff all landing too | 18:59 |
| JayF | so dtantsur|afk finally gets out of rebase hell | 18:59 |
| jroll | yar | 18:59 |
| JayF | man, that keystone patch from 10/19/15 | 19:00 |
| JayF | that's a long time comin | 19:01 |
| pas-ha | yep, 9 months. almost like having a baby :) | 19:05 |
| jroll | pas-ha: hey, so if the version isn't specified in the endpoint, keystoneauth1 will discover if it should use v2 or v3 right? | 19:07 |
| * JayF brb lunching | 19:08 | |
| pas-ha | jroll: yes. there is an api version discovery baked in, depending also on what options are in config, e.g. when v3 will be used only when domain stuff is specified | 19:09 |
| jroll | right, cool | 19:10 |
| jroll | thanks | 19:10 |
| jlvillal | jroll, The project-config patch: https://review.openstack.org/#/c/350703/ from today's Ironic QA meeting | 19:12 |
| jlvillal | FYI | 19:12 |
| jroll | jlvillal: thanks | 19:12 |
| pas-ha | having moved to v3, we could start thinking about hiding the agent-related endpoints from other users. e.g. create a separate ironic domain and have a user there that by roles and policies can only access the lookup and heartbeat. than make a v3 trust for it so it can auth itself, and pass that via kernel params inside. | 19:13 |
| devananda | JayF: actually the explicit v3 bigs are not needed any longer | 19:13 |
| pas-ha | alrigt, getting pretty late here. hope to the stuff merged in the morning :) bye all | 19:14 |
| devananda | JayF: the move to ksauth makes ironic "just work" when I change my env vars to point to v3 | 19:14 |
| *** Sukhdev has joined #openstack-ironic | 19:15 | |
| jroll | pas-ha: I just +W'd it :) | 19:18 |
| jroll | g'night \o | 19:18 |
| clenimar | great job, pas-ha. glad to see this landing :) | 19:23 |
| openstackgerrit | Varun Gadiraju proposed openstack/ironic: Fix for "db type could not be determined" error message https://review.openstack.org/350302 | 19:23 |
| *** livelace has quit IRC | 19:30 | |
| *** sabeen has joined #openstack-ironic | 19:31 | |
| *** tesseract- has joined #openstack-ironic | 19:31 | |
| *** ametts has quit IRC | 19:32 | |
| *** clenimar has quit IRC | 19:34 | |
| *** jcoufal has quit IRC | 19:34 | |
| *** clenimar has joined #openstack-ironic | 19:35 | |
| *** tesseract- has quit IRC | 19:37 | |
| *** ametts has joined #openstack-ironic | 19:45 | |
| *** ayoung has joined #openstack-ironic | 19:45 | |
| JayF | jroll: shouldn't we fix that one typo in the keystoneauth stuff quickly? | 19:49 |
| JayF | jroll: it looks like that's broken enough to break deployments | 19:49 |
| jroll | JayF: the SERVICE thing? | 19:49 |
| JayF | yes | 19:49 |
| jroll | it's typo'd in both places | 19:49 |
| jroll | so it works | 19:49 |
| JayF | OH | 19:49 |
| JayF | lol | 19:49 |
| jroll | :D | 19:49 |
| jroll | yeah I'd be very curious why that passed CI if it wasn't in both | 19:50 |
| jroll | or pep8 even | 19:50 |
| *** Goneri has joined #openstack-ironic | 19:51 | |
| *** narasimha_SV has quit IRC | 19:52 | |
| *** dprince has quit IRC | 19:58 | |
| *** Goneri has quit IRC | 20:10 | |
| *** daemontool has joined #openstack-ironic | 20:10 | |
| *** joprovost has quit IRC | 20:12 | |
| openstackgerrit | Jay Faulkner proposed openstack/ironic-specs: Implement rescue mode https://review.openstack.org/171878 | 20:23 |
| openstackgerrit | Alexander Bashmakov proposed openstack/ironic: Rewrite devstack/tools/ironic/scripts/create-node.sh in Python. https://review.openstack.org/346832 | 20:31 |
| *** baoli_ has quit IRC | 20:39 | |
| *** baoli has joined #openstack-ironic | 20:40 | |
| *** daemontool has quit IRC | 20:48 | |
| *** Sukhdev has quit IRC | 20:52 | |
| *** ametts has quit IRC | 20:52 | |
| *** Sukhdev has joined #openstack-ironic | 20:53 | |
| *** sabeen has quit IRC | 21:03 | |
| openstackgerrit | Merged openstack/ironic: Devstack should use a prebuilt ramdisk by default https://review.openstack.org/349112 | 21:05 |
| openstackgerrit | Merged openstack/ironic: Updating dev quickstart to include compatiblity for newest distros https://review.openstack.org/346894 | 21:05 |
| *** michauds has quit IRC | 21:23 | |
| JayF | sambetts|afk: http://logs.openstack.org/17/350317/2/check/gate-grenade-dsvm-ironic/45801d7/logs/old/localrc.txt.gz our gate is still building ramdisks in all cases | 21:24 |
| *** rpioso has quit IRC | 21:26 | |
| *** baoli has quit IRC | 21:27 | |
| openstackgerrit | Merged openstack/ironic: Update nova scheduler_host_manager config docs https://review.openstack.org/350542 | 21:29 |
| JayF | jroll: we need to revert https://review.openstack.org/349112 | 21:31 |
| JayF | jroll: that's going to break IPA gate | 21:31 |
| JayF | jroll: and I'll put patches up that'll fix the other issue | 21:31 |
| JayF | (break meaning in the worst way possible: we'll just not be testing IPA) | 21:31 |
| *** moshele has joined #openstack-ironic | 21:32 | |
| openstackgerrit | Jay Faulkner proposed openstack/ironic: Revert "Devstack should use a prebuilt ramdisk by default" https://review.openstack.org/350789 | 21:33 |
| JayF | jroll: ^^^ | 21:33 |
| JayF | https://review.openstack.org/350791 is the fix for our gate issue | 21:35 |
| JayF | https://github.com/openstack-infra/devstack-gate/blob/master/devstack-vm-gate.sh#L354 being the relevant piece | 21:36 |
| devananda | JayF: wait - i thought the plan was devstack defaults to use prebuilt, then in the IPA gate, we set a flag to make it actually build it? | 21:36 |
| JayF | devananda: the logic in project-config is reversed | 21:36 |
| devananda | oh crap :( | 21:36 |
| JayF | devananda: and because devstack-gate only sets it if it's told NOT to build it | 21:37 |
| JayF | devananda: we end up in a case where we build the whole time | 21:37 |
| JayF | devananda: except now, that landing means we're in a case where we DON'T build the whole time | 21:37 |
| JayF | that project-config link I pasted fixes one half | 21:37 |
| JayF | and I'm about to push something to ds-gate which will set the var either way | 21:37 |
| *** narasimha_SV has joined #openstack-ironic | 21:38 | |
| openstackgerrit | Merged openstack/ironic: Extend the "configuring ironic-api behind mod_wsgi" guide https://review.openstack.org/350507 | 21:39 |
| JayF | https://review.openstack.org/350795 Explicitly set IRONIC_BUILD_DEPLOY_RAMDISK | 21:39 |
| JayF | is the devstack-gate change | 21:39 |
| narasimha_SV | http://paste.openstack.org/show/547860/ | 21:42 |
| narasimha_SV | getting this issue when ever I try to dync db for ironic | 21:42 |
| narasimha_SV | because of this node_tags and portgroups tables are not getting created | 21:42 |
| narasimha_SV | did anyone face this issue ? | 21:42 |
| JayF | narasimha_SV: as we said earlier, this works in devstack everytime our tests run, so there's something different about what kolla is doing | 21:42 |
| narasimha_SV | please help me in resolving this | 21:42 |
| JayF | narasimha_SV: figuring out that difference is key to troubleshooting this | 21:43 |
| *** sdake has quit IRC | 21:43 | |
| JayF | narasimha_SV: also you have not filed a bug against kolla and/or ironic about this yet either | 21:43 |
| narasimha_SV | JayF: create_schema is working in devstack ? | 21:47 |
| JayF | narasimha_SV: we have to create a db schema on every devstac run | 21:48 |
| narasimha_SV | actualyy i just see the code in kolla it is running ironic-dbsync upgrade | 21:49 |
| thiagop | narasimha_SV: have you tried to run this alter table command directly on mysql? | 21:50 |
| narasimha_SV | thiagop: ALTER TABLE conductors ADD COLUMN online BOOL; ERROR 1064 (42000): You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near 'online BOOL' at line 1 | 21:57 |
| JayF | devananda: https://github.com/openstack-infra/devstack-gate/blob/master/devstack-vm-gate-wrap.sh#L200 | 22:00 |
| * JayF throws hands up in the air | 22:00 | |
| thiagop | narasimha_SV: which version of mysql are you running? | 22:01 |
| JayF | devananda: jroll: Anyone else around who is a core: https://review.openstack.org/#/c/350789/ can we please land this ASAP before anything lands in IPA with invalid tests running/ | 22:05 |
| openstackgerrit | Jay Faulkner proposed openstack/ironic: Revert "Devstack should use a prebuilt ramdisk by default" https://review.openstack.org/350789 | 22:06 |
| *** jaybeale has joined #openstack-ironic | 22:07 | |
| *** wajdi_ has joined #openstack-ironic | 22:07 | |
| narasimha_SV | thiagop: mysql 5.6 | 22:08 |
| thiagop | narasimha_SV: the syntax seems legit | 22:08 |
| narasimha_SV | yes | 22:09 |
| thiagop | narasimha_SV: the table already exists, right? | 22:09 |
| *** moshele has quit IRC | 22:10 | |
| narasimha_SV | yes | 22:10 |
| narasimha_SV | mysql> desc conductors; +------------+--------------+------+-----+---------+----------------+ | Field | Type | Null | Key | Default | Extra | +------------+--------------+------+-----+---------+----------------+ | created_at | datetime | YES | | NULL | | | updated_at | datetime | YES | | NULL | | | id | int(11) | NO | PRI | NULL | 22:10 |
| *** wajdi has quit IRC | 22:11 | |
| narasimha_SV | http://paste.openstack.org/show/547862/ | 22:11 |
| narasimha_SV | http://paste.openstack.org/show/547863/ | 22:12 |
| narasimha_SV | tables created when database got sync happened | 22:12 |
| *** jaybeale has quit IRC | 22:12 | |
| *** wajdi_ has quit IRC | 22:13 | |
| jroll | JayF: lol, done | 22:17 |
| *** jaybeale has joined #openstack-ironic | 22:18 | |
| JayF | jroll: I just spent forever troubleshooting this stuff. So much misdirection /o\ | 22:18 |
| * jroll out for the night | 22:18 | |
| jroll | yeah, sorry :( | 22:18 |
| JayF | it's not your fault | 22:18 |
| mrda | Morning Ironic | 22:21 |
| thiagop | morning mrda | 22:28 |
| mrda | o/ | 22:28 |
| NobodyCam | morning mrda | 22:29 |
| NobodyCam | gah virtual box does not report memory with dmidecode! nice! | 22:30 |
| *** jaybeale has quit IRC | 22:35 | |
| *** bfournie has quit IRC | 22:38 | |
| *** sabeen has joined #openstack-ironic | 22:41 | |
| *** syed_ has quit IRC | 22:42 | |
| *** narasimha_SV has quit IRC | 22:44 | |
| *** thiagop has quit IRC | 22:51 | |
| * jlvillal is being asked by people if there is any Kubernetes integration work in progress for Ironic. | 22:55 | |
| devananda | jlvillal: integration in what sense? | 22:58 |
| jlvillal | devananda: Ironic on top of K8S is what I was asked about | 23:00 |
| jlvillal | They are running into an issue, but it might be a Nova thing. | 23:00 |
| jlvillal | quote: Nova-compute (with ironic driver) needs to be able to re-register hypervisor ID when it’s hostname changes. Right now each time nova-compute gets rescheduled, it registers all baremetals as new hypervisors. This happens in a K8S world when a POD gets rescheduled to a different host, so nova-compute changes its hostname. | 23:00 |
| JayF | jlvillal: that's very unlikely to be fixed | 23:02 |
| JayF | jlvillal: simply because it's a design thing in nova | 23:02 |
| JayF | jlvillal: you can set CONF.host on a node to something in order for that to be used instead of the hostname though | 23:02 |
| JayF | jlvillal: in a kubernetes world that's probably the proper workaround | 23:03 |
| jlvillal | JayF: Okay. Thanks. Yeah that is what they are doing. | 23:03 |
| *** sdake has joined #openstack-ironic | 23:03 | |
| JayF | but computes are id'd by their hostname, that's not even changing in the multi-compute stuff jroll is landing | 23:03 |
| jlvillal | Thanks! | 23:04 |
| *** PollyZ has quit IRC | 23:09 | |
| JayF | Uh, we appear to have tests failing right now for... strange reasons | 23:15 |
| JayF | http://logs.openstack.org/79/350679/4/check/gate-ironic-python27-db-ubuntu-xenial/e9f62eb/console.html#_2016-08-03_22_42_09_066090 | 23:15 |
| JayF | anyone seen that before? | 23:15 |
| JayF | hmm | 23:17 |
| JayF | that's not why it failed though | 23:17 |
| JayF | Because some passing jobs have those | 23:17 |
| jlvillal | JayF: On this change: https://review.openstack.org/#/c/350791/3/jenkins/jobs/ironic.yaml | 23:23 |
| jlvillal | JayF: Line 179 | 23:23 |
| jlvillal | Seems like it is swapping things, or am I confused??? | 23:23 |
| JayF | jlvillal: was supposed to have removed that part of the change and I didn't | 23:25 |
| devananda | jlvillal: JayF: actually that might be fixed, in a way, by jroll's multi-compute-host work | 23:25 |
| JayF | jlvillal: good -1 review, if you want credit for it be fast :) | 23:25 |
| jlvillal | I did do a -1 :) | 23:25 |
| devananda | the addition of a hash ring inside the ironic virt driver should allow it to recover from a hostname change | 23:25 |
| JayF | devananda: we had a long downstream chat about that, with the angle being 'how do we deploy it', and the answer is no | 23:25 |
| JayF | devananda: yes, but the leftover instances would still be "orphaned" without a similarly named compute host | 23:25 |
| devananda | ah - really? | 23:25 |
| JayF | devananda: and would not be able to have instance actions other than delete performed on them | 23:25 |
| JayF | yep | 23:25 |
| devananda | oh, right, inside nova | 23:26 |
| devananda | because of the assumption that if n-cpu died (or its POD died) then all the VMs it was managing are gone | 23:26 |
| devananda | (not true with ironic or vmware) | 23:26 |
| JayF | yep | 23:26 |
| * jlvillal listens and learns | 23:27 | |
| JayF | it's sorta a core nova assumption | 23:27 |
| JayF | which is why I presumed (and told jlvillal) that it's not likely to ever be fixed | 23:27 |
| devananda | jlvillal: the kolla project is trying to run ironic in containers, I believe. though they're also trying to use ironic to deploy container-OS's | 23:27 |
| jlvillal | devananda: Thanks | 23:27 |
| devananda | jlvillal: and CoreOS is working on a product that does much the same thing (runs all openstack services in containers) | 23:27 |
| devananda | though I don't know where they stand on support for ironic | 23:27 |
| JayF | I did suggest to jroll that we could've added support for an option to nova-compute as like a migration-hostname, i.e. an old hostname whose instances you should reassign to yourself | 23:27 |
| JayF | as a migration path | 23:27 |
| devananda | JayF: "evacuate" ;) | 23:27 |
| devananda | tell n-cpu to shed all the instances it "owns" by allowing other n-cpu to pick them up | 23:28 |
| JayF | but he said (and rightly, I believe) that it's mainly only helpful for folks who ran CCM (i.e., OnMetal, probably only us, lol) | 23:28 |
| JayF | and we basically said if you run CCM, and it breaks, you get two pieces | 23:28 |
| JayF | but implementing evacuate sounds even better | 23:28 |
| JayF | and good for an improvement to it in ocata | 23:28 |
| devananda | JayF: I agree with that approach for CCM - but I can also tell you you're not the only ones who have (tried to) run it | 23:28 |
| JayF | devananda: our deployment is basically going to be deploy new code (some later time), stop computes, remove CONF.host from configs, assign already-deployed instances manually in DB to one of the new nova-compute hostnames | 23:29 |
| devananda | anyway, we've digressed from the topic of k8s :) | 23:29 |
| JayF | well, the solution for k8s is clearly to set conf.host | 23:29 |
| JayF | and have an idea of "this is virtual-compute-1" regardless of hostname | 23:30 |
| JayF | because nova wants a static way to track it | 23:30 |
| JayF | and that's almost certainly more in line with k8s philosophy than calling nova evacuate everytime you cycle a container | 23:30 |
| devananda | JayF: I disagree | 23:30 |
| devananda | if you did that, then it would pin the number of (and configuration of!) k8s instances | 23:31 |
| devananda | until you kill of some of the running instances from ironic | 23:31 |
| devananda | because you would lose the ability to manage them, if you were not running a n-cpu with the same hostname that was used to launch the instance | 23:31 |
| JayF | devananda: I think we're talking different use cases | 23:32 |
| devananda | or to put another way, it means you would need to assert a fixed-size resource pool for your n-cpu workers | 23:32 |
| JayF | yes; I did have that assumption built into my statement | 23:32 |
| JayF | exactly | 23:32 |
| devananda | JayF: right, so, I think that runs counter to the spirit / philosophy of k8s | 23:33 |
| JayF | but you could add more n-cpu "workers", they would just not rebalance existing instances until those instances cycled | 23:33 |
| JayF | which isn't ideal but it better than anything we'd have today | 23:33 |
| devananda | JayF: you could add more, but you couldn't shrink | 23:33 |
| JayF | devananda: oooooh. | 23:33 |
| JayF | devananda: that's a real gotcha there I hadn't thought of | 23:34 |
| devananda | :P | 23:34 |
| JayF | devananda: that with the code as it sits to land in N, you can never shrink the number of n-cpu workers | 23:34 |
| JayF | jroll: ^ relevant to your interests above | 23:34 |
| devananda | JayF: indeed | 23:34 |
| JayF | (implied: without touching the database) | 23:35 |
| JayF | I still think it's better and should land | 23:35 |
| JayF | but we need to document that in bold flashing letters | 23:35 |
| JayF | until we make a way to evacuate in ocata | 23:35 |
| JayF | (basically it's at the point where it can land in nova-n as long as it doesn't grow, and I suspect implementing evacuate for it would be enough to make it lose the FFE) | 23:35 |
| JayF | I'm going to go put that comment on the rview | 23:36 |
| JayF | nice catch | 23:36 |
| devananda | I would still mark this feature as experimental'ish, fwiw | 23:37 |
| devananda | it's an improvement to be sure | 23:37 |
| JayF | if you read the release notes it is | 23:37 |
| devananda | :) | 23:37 |
| devananda | I haven't seen the patch in a couple days | 23:37 |
| JayF | I think part of the desire to get it in is that we're very eager to be the guinea pigs downstream | 23:37 |
| *** hoonetorg has quit IRC | 23:39 | |
| devananda | awesome | 23:41 |
| JayF | jroll: vdrok: sambetts|afk: https://review.openstack.org/#/c/350679 failed pep8, I'm pushing a patchset to fix teh pep8 failures, we should land this asap overnight or in the morning | 23:43 |
| *** hoonetorg has joined #openstack-ironic | 23:44 | |
| openstackgerrit | Jay Faulkner proposed openstack/ironic: Create a custom StringField that can process functions https://review.openstack.org/350679 | 23:44 |
| *** jrist has quit IRC | 23:49 | |
| *** sdake has quit IRC | 23:51 | |
| *** hoangcx has joined #openstack-ironic | 23:55 | |
| *** sdake has joined #openstack-ironic | 23:55 | |
| jlvillal | JayF: Seems reasonable to +2 and +A that, since it already happened before. Your changes are whitespace only. | 23:57 |
| jlvillal | But maybe others would disagree... | 23:57 |
| JayF | jlvillal: you're welcome to do that if you won't; as a personal rule I never land code I pushed the last change to | 23:57 |
| JayF | jlvillal: just sorta a CYA thing years in the industry taught me :D | 23:57 |
| jlvillal | heh | 23:57 |
| JayF | s/won't/want/ | 23:57 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
