| *** hoonetorg has quit IRC | 00:05 | |
| *** rbudden has quit IRC | 00:07 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/bifrost master: Updated from global requirements https://review.openstack.org/541944 | 00:07 |
|---|---|---|
| *** hshiina has joined #openstack-ironic | 00:14 | |
| jroll | TheJulia: congrats, condolences, and thank you :) | 00:23 |
| jroll | dtantsur|afk: thank you for your service, good work :D | 00:23 |
| *** crushil has joined #openstack-ironic | 00:24 | |
| TheJulia | jroll: <3 | 00:27 |
| rloo | huh? light-bulb goes on. Yay, all hail the new leader TheJulia!!! | 00:33 |
| rloo | and the first thing she does is go on PTO :) | 00:33 |
| *** charcol has joined #openstack-ironic | 00:37 | |
| TheJulia | lol | 00:39 |
| rloo | ok, i'm done for tonight. ciao. | 00:45 |
| *** rloo has quit IRC | 00:45 | |
| *** ijw has quit IRC | 00:53 | |
| *** crushil has quit IRC | 00:56 | |
| *** crushil has joined #openstack-ironic | 00:56 | |
| openstackgerrit | Merged openstack/ironic master: Don't try to lock for vif detach https://review.openstack.org/534441 | 01:03 |
| openstackgerrit | Merged openstack/ironic master: Remove unnecessary lines from sample local.conf https://review.openstack.org/541900 | 01:04 |
| openstackgerrit | Merged openstack/ironic master: Stop guessing mime types based on URLs https://review.openstack.org/541289 | 01:04 |
| *** trungnv has joined #openstack-ironic | 01:11 | |
| *** trungnv_ has joined #openstack-ironic | 01:11 | |
| openstackgerrit | Ruby Loo proposed openstack/ironic master: [reno] timeout parameter worked https://review.openstack.org/541961 | 01:11 |
| *** trungnv_ has quit IRC | 01:11 | |
| *** trungnv has quit IRC | 01:11 | |
| *** trungnv has joined #openstack-ironic | 01:12 | |
| *** rloo has joined #openstack-ironic | 01:14 | |
| rloo | dtantsur|afk, TheJulia: I just looked at hironori's comment in bug 1746849. Turns out it wasn't broken. we can either revert or leave the change and update the reno: https://review.openstack.org/#/c/541961/ | 01:15 |
| openstack | bug 1746849 in Ironic "conductor.utils.node_power_action() doesn't always use timeout" [Medium,Fix released] https://launchpad.net/bugs/1746849 - Assigned to Ruby Loo (rloo) | 01:15 |
| patchbot | patch 541961 - ironic - [reno] timeout parameter worked | 01:15 |
| rloo | dtantsur|afk, TheJulia: you can decide; will talk to you tomorrow :) | 01:16 |
| *** rloo has quit IRC | 01:17 | |
| *** jtomasek has quit IRC | 01:17 | |
| *** tiendc has joined #openstack-ironic | 01:23 | |
| openstackgerrit | melissaml proposed openstack/ironic-specs master: Modify grammatical errors https://review.openstack.org/541968 | 01:24 |
| openstackgerrit | melissaml proposed openstack/ironic master: Modify grammatical errors https://review.openstack.org/541969 | 01:26 |
| *** yangyapeng has quit IRC | 01:27 | |
| *** yangyapeng has joined #openstack-ironic | 01:27 | |
| *** mjturek has joined #openstack-ironic | 01:31 | |
| openstackgerrit | melissaml proposed openstack/ironic-inspector master: Modify grammatical errors https://review.openstack.org/541973 | 01:35 |
| openstackgerrit | Julia Kreger proposed openstack/ironic master: DNM Can we run a conductor on the subnode? https://review.openstack.org/535596 | 01:39 |
| * TheJulia thinks updating reno is easiest at the moment :\ | 01:41 | |
| * TheJulia will leave that call for dmitry tomorrow | 01:41 | |
| * TheJulia goes and has a beverage | 01:41 | |
| TheJulia | goodnight ironic! :) | 01:42 |
| *** hshiina2 has joined #openstack-ironic | 02:00 | |
| *** hshiina has quit IRC | 02:03 | |
| *** hshiina3 has joined #openstack-ironic | 02:03 | |
| *** hshiina2 has quit IRC | 02:06 | |
| *** hshiina3 is now known as hshiina | 02:08 | |
| *** Sukhdev has quit IRC | 02:27 | |
| *** hshiina2 has joined #openstack-ironic | 02:29 | |
| *** hshiina has quit IRC | 02:32 | |
| *** hshiina3 has joined #openstack-ironic | 02:37 | |
| *** hshiina2 has quit IRC | 02:41 | |
| *** hshiina3 is now known as hshiina | 02:52 | |
| openstackgerrit | Kaifeng Wang proposed openstack/ironic-inspector master: Introducing an api_workers option to inspector https://review.openstack.org/540722 | 02:56 |
| openstackgerrit | Merged openstack/ironic stable/pike: Do not pass credentials to the ramdisk on cleaning https://review.openstack.org/541703 | 03:01 |
| *** fragatina has quit IRC | 03:06 | |
| *** fragatina has joined #openstack-ironic | 03:06 | |
| *** fragatina has quit IRC | 03:06 | |
| *** fragatina has joined #openstack-ironic | 03:07 | |
| *** rbudden has joined #openstack-ironic | 03:11 | |
| *** trungnv has quit IRC | 03:14 | |
| *** trungnv has joined #openstack-ironic | 03:14 | |
| *** zshi has quit IRC | 03:16 | |
| *** zshi has joined #openstack-ironic | 03:16 | |
| *** jcoufal has quit IRC | 03:18 | |
| *** zshi has quit IRC | 03:26 | |
| *** zshi has joined #openstack-ironic | 03:26 | |
| absubram | hi all.. I followed the openstack docs and used disk-image-builder to build my own user images for centos and ubuntu.. is there a way to figure out the versions for my user images? | 03:31 |
| absubram | or will have to simply pxe boot my ironic nodes with these created images and then figure out the version from there? | 03:31 |
| openstackgerrit | Julia Kreger proposed openstack/ironic master: DNM Can we run a conductor on the subnode? https://review.openstack.org/535596 | 03:36 |
| *** mat128 has quit IRC | 03:39 | |
| *** absubram has quit IRC | 03:39 | |
| *** mat128 has joined #openstack-ironic | 03:41 | |
| openstackgerrit | ya.wang proposed openstack/ironic master: Add description for response erroe code https://review.openstack.org/542008 | 03:44 |
| openstackgerrit | Julia Kreger proposed openstack/bifrost master: Update release notes for release https://review.openstack.org/542009 | 03:46 |
| *** dsneddon has quit IRC | 03:49 | |
| *** dsneddon has joined #openstack-ironic | 03:51 | |
| *** dsneddon has quit IRC | 03:56 | |
| openstackgerrit | Merged openstack/bifrost master: Update release notes for release https://review.openstack.org/542009 | 03:58 |
| *** fragatina has quit IRC | 03:58 | |
| *** dsneddon has joined #openstack-ironic | 04:00 | |
| *** fragatina has joined #openstack-ironic | 04:02 | |
| *** dsneddon has quit IRC | 04:05 | |
| *** fragatina has quit IRC | 04:07 | |
| *** hieulq has quit IRC | 04:15 | |
| openstackgerrit | Shivanand Tendulker proposed openstack/ironic master: Add validate_rescue() method to boot interface https://review.openstack.org/540954 | 04:16 |
| *** rwsu has quit IRC | 04:24 | |
| *** rwsu has joined #openstack-ironic | 04:24 | |
| *** dsneddon has joined #openstack-ironic | 04:25 | |
| *** dsneddon has quit IRC | 04:34 | |
| sapd | Hello, How to build ramdisk image which support intel RAID? help me. | 04:55 |
| *** dsneddon has joined #openstack-ironic | 05:04 | |
| *** dsneddon has quit IRC | 05:09 | |
| *** dsneddon has joined #openstack-ironic | 05:12 | |
| *** rpioso is now known as rpioso|afk | 05:16 | |
| *** dsneddon has quit IRC | 05:17 | |
| *** hieulq has joined #openstack-ironic | 05:31 | |
| *** pmannidi has quit IRC | 05:34 | |
| *** lbragstad has quit IRC | 05:37 | |
| *** fragatina has joined #openstack-ironic | 05:42 | |
| *** zhangfei has joined #openstack-ironic | 05:49 | |
| *** threestrands has quit IRC | 05:57 | |
| *** mjturek has quit IRC | 06:01 | |
| *** yangyapeng has quit IRC | 06:10 | |
| *** yangyapeng has joined #openstack-ironic | 06:15 | |
| openstackgerrit | Nguyen Van Trung proposed openstack/ironic master: [WIP] Implement raid configuration for BM via irmc drive (OOB-1) https://review.openstack.org/512979 | 06:18 |
| openstackgerrit | fupingxie proposed openstack/ironic-lib master: Add attemps for the partprobe command https://review.openstack.org/542035 | 06:19 |
| openstackgerrit | Kaifeng Wang proposed openstack/ironic-inspector master: Introducing an api_workers option to inspector https://review.openstack.org/540722 | 06:22 |
| *** dsneddon has joined #openstack-ironic | 06:25 | |
| *** hshiina2 has joined #openstack-ironic | 06:29 | |
| *** hshiina3 has joined #openstack-ironic | 06:30 | |
| *** hshiina has quit IRC | 06:32 | |
| *** hshiina2 has quit IRC | 06:34 | |
| *** hoonetorg has joined #openstack-ironic | 06:44 | |
| openstackgerrit | fupingxie proposed openstack/ironic master: Add more parameter explanation when create a node https://review.openstack.org/531941 | 06:49 |
| *** rbudden has quit IRC | 06:54 | |
| *** absubram has joined #openstack-ironic | 06:54 | |
| *** absubram_ has joined #openstack-ironic | 06:56 | |
| *** marios has joined #openstack-ironic | 06:57 | |
| *** moshele has joined #openstack-ironic | 06:57 | |
| *** absubram has quit IRC | 06:59 | |
| *** absubram_ is now known as absubram | 06:59 | |
| *** mjura has joined #openstack-ironic | 07:02 | |
| *** absubram has quit IRC | 07:04 | |
| *** hoonetorg has quit IRC | 07:06 | |
| *** rbartal has joined #openstack-ironic | 07:14 | |
| *** rcernin has quit IRC | 07:16 | |
| *** mrunge_ is now known as mrunge | 07:16 | |
| openstackgerrit | Mallikarjuna Reddy Kolagatla proposed openstack/ironic master: Updates boot mode on the baremetal as per `boot_mode` https://review.openstack.org/540793 | 07:17 |
| *** AlexeyAbashkin has joined #openstack-ironic | 07:20 | |
| *** azaid_ has joined #openstack-ironic | 07:22 | |
| *** hoonetorg has joined #openstack-ironic | 07:23 | |
| *** jtomasek has joined #openstack-ironic | 07:23 | |
| *** jtomasek has quit IRC | 07:24 | |
| *** jtomasek has joined #openstack-ironic | 07:25 | |
| *** ankkumar has joined #openstack-ironic | 07:28 | |
| ankkumar | Hi All | 07:28 |
| ankkumar | I am facing issues in running tempest testing of ironic(ironic-tempest-plugin). Tox is failing because while booting it is taking private network to create port but after that it is trying to attach with public network port | 07:28 |
| azaid_ | Any core-reviewer here ? | 07:28 |
| ankkumar | for booting it is creating port using private net.. https://github.com/openstack/ironic-tempest-plugin/blob/master/ironic_tempest_plugin/tests/scenario/baremetal_standalone_manager.py#L47 | 07:30 |
| *** pcaruana has joined #openstack-ironic | 07:31 | |
| ankkumar | but it is trying to assign port created using public net. https://github.com/openstack/ironic-tempest-plugin/blob/master/ironic_tempest_plugin/tests/scenario/baremetal_standalone_manager.py#L132 | 07:31 |
| *** e0ne has joined #openstack-ironic | 07:34 | |
| *** vsaienk0 has joined #openstack-ironic | 07:35 | |
| *** charcol has quit IRC | 07:36 | |
| *** AlexeyAbashkin has quit IRC | 07:38 | |
| *** e0ne has quit IRC | 07:39 | |
| *** e0ne has joined #openstack-ironic | 07:40 | |
| *** AlexeyAbashkin has joined #openstack-ironic | 07:43 | |
| *** AlexeyAbashkin has quit IRC | 07:47 | |
| *** hshiina2 has joined #openstack-ironic | 07:48 | |
| *** hshiina3 has quit IRC | 07:52 | |
| *** AlexeyAbashkin has joined #openstack-ironic | 07:52 | |
| *** hshiina2 is now known as hshiina | 07:53 | |
| *** vsaienk0 has quit IRC | 07:53 | |
| *** hshiina2 has joined #openstack-ironic | 07:55 | |
| *** hshiina has quit IRC | 07:58 | |
| *** stendulker has joined #openstack-ironic | 08:04 | |
| openstackgerrit | Merged openstack/ironic-inspector master: Update reno for stable/queens https://review.openstack.org/541750 | 08:04 |
| openstackgerrit | melissaml proposed openstack/ironic-inspector master: Remove redundant "the" from a docstring https://review.openstack.org/541973 | 08:05 |
| *** dtantsur|afk is now known as dtantsur | 08:09 | |
| dtantsur | morning ironic | 08:09 |
| stendulker | Good morning dtantsur :) | 08:11 |
| dtantsur | stendulker: hey, mind reviewing https://review.openstack.org/#/c/541961/ please? | 08:14 |
| patchbot | patch 541961 - ironic - [reno] timeout parameter worked | 08:14 |
| stendulker | dtantsur: ok | 08:14 |
| *** olivierbourdon38 has joined #openstack-ironic | 08:14 | |
| *** robcresswell has quit IRC | 08:18 | |
| *** amarao has joined #openstack-ironic | 08:23 | |
| openstackgerrit | Merged openstack/ironic master: [reno] timeout parameter worked https://review.openstack.org/541961 | 08:25 |
| *** tesseract has joined #openstack-ironic | 08:26 | |
| *** vsaienk0 has joined #openstack-ironic | 08:37 | |
| hjensas | dtantsur: Regarding maintenance mode in https://review.openstack.org/#/c/541678. I responded, don't we have to? Would the node be in one of the other states allowed at that point? | 08:41 |
| patchbot | patch 541678 - tripleo-quickstart-extras - Prep-CI for routed-networks changes | 08:41 |
| openstackgerrit | Merged openstack/ironic master: correct grammar, duplicate the found https://review.openstack.org/541640 | 08:41 |
| dtantsur | hjensas: maintenance mode means literally "the node is broken so much that it needs an operator intervention" | 08:41 |
| dtantsur | if it's not your case, you should not use maintenance mode | 08:41 |
| dtantsur | in your case, you're trying to update AVAILABLE nodes | 08:42 |
| hjensas | dtantsur: ah, so skipp the --provide and set the port option then provide? | 08:42 |
| dtantsur | I think it should be changed to first enroll nodes to MANAGEABLE, then update/introspect, then provide | 08:42 |
| dtantsur | correct | 08:42 |
| *** mgoddard_ has joined #openstack-ironic | 08:46 | |
| *** vsaienk0 has quit IRC | 08:47 | |
| *** robcresswell has joined #openstack-ironic | 08:49 | |
| *** tiendc has quit IRC | 08:52 | |
| *** trungnv has quit IRC | 08:53 | |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/bifrost master: Test OPNFV CI verify job. Please ignore. https://review.openstack.org/367608 | 08:59 |
| *** tiendc has joined #openstack-ironic | 09:02 | |
| openstackgerrit | Hironori Shiina proposed openstack/ironic master: Remove too large configdrive for handling error https://review.openstack.org/542097 | 09:07 |
| *** hw_wutianwei has quit IRC | 09:07 | |
| *** hw_wutianwei has joined #openstack-ironic | 09:08 | |
| *** derekh has joined #openstack-ironic | 09:10 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-tempest-plugin master: Add [baremetal] available_nodes configuration option https://review.openstack.org/541383 | 09:16 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-tempest-plugin master: Rescue mode tempest tests https://review.openstack.org/528699 | 09:16 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-tempest-plugin master: Test cases for iLO drivers in standalone ironic https://review.openstack.org/535652 | 09:16 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-tempest-plugin master: Rescue mode standalone tests https://review.openstack.org/538119 | 09:17 |
| *** vsaienk0 has joined #openstack-ironic | 09:17 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-tempest-plugin master: Replace use of functools.wraps() with six.wraps() https://review.openstack.org/540187 | 09:17 |
| dtantsur | just rebases to pick up the gate fix ^^^ | 09:18 |
| *** azaid_ has quit IRC | 09:25 | |
| openstackgerrit | Merged openstack/ironic-tempest-plugin master: Updated from global requirements https://review.openstack.org/534933 | 09:25 |
| *** vsaienk0 has quit IRC | 09:27 | |
| *** vsaienk0 has joined #openstack-ironic | 09:28 | |
| ankkumar | dtantsur, Hi | 09:29 |
| *** sapd has quit IRC | 09:29 | |
| ankkumar | dtantsur, we are trying to setup HPE CI | 09:29 |
| ankkumar | dtantsur, In our setup we have flat network. we dont have vlan so the problem we are facing is.. it is booting the with private network but trying to associate the port with public network | 09:31 |
| ankkumar | dtantsur, but upstream is using vlan network that we cant use | 09:31 |
| *** hshiina2 is now known as hshiina|afk | 09:32 | |
| ankkumar | dtantsur, is there any solution for us? | 09:32 |
| *** sapd has joined #openstack-ironic | 09:33 | |
| dtantsur | ankkumar: hi, sorry, I don't quite get the question. "upstream" we use both vlans and flat networks, mostly flat networks | 09:37 |
| dtantsur | maybe worth double-checking tempest.conf for correct networks? e.g. compare with one of our jobs? | 09:37 |
| *** vsaienk0 has quit IRC | 09:38 | |
| *** tiendc_ has joined #openstack-ironic | 09:38 | |
| ankkumar | dtantsur, in our case whats happening is, ironic node is creating with private network | 09:39 |
| dtantsur | what do you mean by "node is creating with network" here? | 09:39 |
| ankkumar | dtantsur, and during the tempest run, it will try to create one floating ip and try to associate with private port | 09:39 |
| ankkumar | dtantsur, ironic vif port is neutron private port | 09:40 |
| dtantsur | okay | 09:40 |
| *** tiendc has quit IRC | 09:40 | |
| *** hshiina|afk has quit IRC | 09:41 | |
| ankkumar | dtantsur, so at the end tempest is failing because its not able to associate floating ip with private port | 09:41 |
| dtantsur | ankkumar: check tempest.conf in one of our jobs: http://logs.openstack.org/89/541289/5/gate/ironic-tempest-dsvm-ipa-partition-redfish-tinyipa/d534182/logs/tempest_conf.txt.gz | 09:42 |
| dtantsur | maybe you can find some discrepancies? | 09:42 |
| dtantsur | shared_physical_network, public_network_id, for example | 09:43 |
| *** hjensas has quit IRC | 09:45 | |
| ankkumar | dtantsur, yes even for us also it is same | 09:46 |
| *** hjensas has joined #openstack-ironic | 09:46 | |
| ankkumar | dtantsur, upstream is using both public and private network but we cant use public network | 09:48 |
| ankkumar | dtantsur, it has to be private | 09:48 |
| *** sambetts|afk is now known as sambetts | 09:51 | |
| openstackgerrit | Ilya Etingof proposed openstack/sushy-tools master: Abstract away the virtualization management interface https://review.openstack.org/523321 | 09:53 |
| sambetts | ankkumar: I'm not sure what you mean by we can't use public, in the case of the CI the public network is a fake public network so if the problem is it conflicting with your real networks that shouldn't be a problem | 09:53 |
| yolanda | dtantsur, https://build.opnfv.org/ci/job/xci-verify-ubuntu-deploy-virtual-master/698/console | 09:53 |
| dtantsur | morning sambetts, thanks for jumping in | 09:53 |
| dtantsur | yolanda: inspector failed to start? | 09:53 |
| sambetts | o/ dtantsur | 09:53 |
| dtantsur | where are its logs? | 09:54 |
| ankkumar | sambetts, https://github.com/openstack/ironic-tempest-plugin/blob/master/ironic_tempest_plugin/tests/scenario/baremetal_standalone_manager.py#L339-L340 | 09:54 |
| ankkumar | sambetts, Here in line 339 it calls boot_node | 09:54 |
| ankkumar | and in L340 it calls add_floatingip | 09:55 |
| yolanda | mm, i see that on logs Failed to restart ironic-inspector.service: Unit ironic-inspector.service not found. | 09:55 |
| ankkumar | the boot_node() creates the network with private network | 09:55 |
| sambetts | ankkumar: correct | 09:56 |
| ankkumar | but add_floatingip_to_node() tries to add the IP from pubic network | 09:56 |
| dtantsur | yolanda: ouch :) | 09:56 |
| ankkumar | sambetts, dtantsur https://github.com/openstack/ironic-tempest-plugin/blob/master/ironic_tempest_plugin/tests/scenario/baremetal_standalone_manager.py#L141-L142 | 09:57 |
| yolanda | dtantsur, no... false error, this entry is first on the logs, then it starts inspector | 09:58 |
| yolanda | i guess that error is for some cleanup | 09:58 |
| sambetts | ankkumar: yes that is pretty standard for a deployment of Ironic, but your problem is you don't have a separate public network and private network right? | 09:58 |
| ankkumar | sambetts, yes | 09:59 |
| ankkumar | earlier we were using only private network in the CI(jenkins) | 10:00 |
| ankkumar | and it was working | 10:00 |
| sambetts | this logic hasn't changed for months if not years, was your old CI testing with Ironic? | 10:00 |
| ankkumar | yes | 10:00 |
| *** tiendc_ has quit IRC | 10:02 | |
| sambetts | devstack typically stands up neutron with L3 enabled, with a fake public network where the host running devstack is the gateway router for that network | 10:02 |
| sambetts | perhaps your old CI was doing that without you realising? | 10:02 |
| *** athomas has joined #openstack-ironic | 10:05 | |
| sambetts | ankkumar: I think the other configuration setting that would affect this is "[validation] connect_method='floating' or 'fixed' | 10:07 |
| sambetts | ankkumar: you also want to ensure your running on the baremetal basic ops test not the standalone baremetal basic opts test which is designed for running against a bifrost/standalone ironic deployment not an OpenStack deployment | 10:09 |
| *** etingof has quit IRC | 10:09 | |
| ankkumar | sambetts, Thanks. Yes thats set to floating in our case It might be happening due to that. | 10:10 |
| sambetts | if you change it to "fixed" it will not try to create a floating ip, it'll use the private one instead as long as it can connect to it :) | 10:11 |
| *** etingof has joined #openstack-ironic | 10:11 | |
| *** vsaienk0 has joined #openstack-ironic | 10:12 | |
| *** dsneddon has quit IRC | 10:14 | |
| *** dsneddon has joined #openstack-ironic | 10:16 | |
| *** devananda has quit IRC | 10:16 | |
| *** hw_wutianwei has quit IRC | 10:22 | |
| *** hw_wutianwei has joined #openstack-ironic | 10:22 | |
| *** vsaienk0 has quit IRC | 10:22 | |
| * dtantsur is about to release ironic queens | 10:25 | |
| * sambetts runs and hides | 10:26 | |
| dtantsur | anything non-backportable that we need to (and can) urgently land? | 10:26 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: Add some missed test cases in node object tests https://review.openstack.org/268045 | 10:27 |
| openstackgerrit | Ilya Etingof proposed openstack/sushy-tools master: Abstract away the virtualization management interface https://review.openstack.org/523321 | 10:27 |
| ankkumar | sambetts, sorry even we do have "[validation] connect_method=fixed". i just checked | 10:31 |
| ankkumar | sambetts, and we are trying to do standalone test not the baremetal basic ops | 10:32 |
| sambetts | ankkumar: standalone there are no APIs for floating IP so it shouldn't be trying to use it | 10:34 |
| ankkumar | sambetts, but it is using in standalone https://github.com/openstack/ironic-tempest-plugin/blob/master/ironic_tempest_plugin/tests/scenario/baremetal_standalone_manager.py#L132-L144 | 10:36 |
| openstackgerrit | Shivanand Tendulker proposed openstack/ironic master: Implements rescue interface for 'ilo' hardware type https://review.openstack.org/453012 | 10:38 |
| sambetts | huh... well thats odd... I could have sworn we weren't running neutron in the standalone scenario but apprently we are | 10:38 |
| sambetts | the normal baremetal basic ops takes the connect_method into account | 10:38 |
| sambetts | but it seems like the standalone one doesn't | 10:39 |
| ankkumar | sambetts, yes thats why it is failing | 10:39 |
| ankkumar | sambetts, any suggestion.. what to do? | 10:39 |
| sambetts | seems we need to mimic the logic like https://github.com/openstack/tempest/blob/45e4c45209bc8d1c6664a0a3fc1530183dfd951a/tempest/scenario/manager.py#L634-L661 into our custom standalone test case | 10:41 |
| sambetts | basically if connect_method = flaoting create a flaoting, else just use the node ip | 10:42 |
| *** hw_wutianwei has quit IRC | 10:42 | |
| sambetts | pas-ha: ^ this more your area of expertise | 10:42 |
| ankkumar | sambetts, yes | 10:43 |
| ankkumar | sambetts, shall i raise one bug and patch for that? | 10:43 |
| sambetts | if you can write a patch to the ironic-tempest-plugin that would be great :) | 10:44 |
| openstackgerrit | Shivanand Tendulker proposed openstack/ironic master: Implements rescue interface for 'ilo' hardware type https://review.openstack.org/453012 | 10:44 |
| ankkumar | sambetts, ok will do.. thanks | 10:44 |
| openstackgerrit | Merged openstack/ironic master: Add some missed test cases in node object tests https://review.openstack.org/268045 | 10:49 |
| *** hw_wutianwei has joined #openstack-ironic | 10:50 | |
| *** links has joined #openstack-ironic | 10:51 | |
| *** vsaienk0 has joined #openstack-ironic | 10:55 | |
| *** dsneddon has quit IRC | 11:03 | |
| openstackgerrit | James E. Blair proposed openstack/bifrost stable/pike: Zuul: Remove project name https://review.openstack.org/542131 | 11:03 |
| *** vsaienk0 has quit IRC | 11:05 | |
| *** lucas-afk is now known as lucasagomes | 11:05 | |
| *** dtantsur is now known as dtantsur|bbl | 11:06 | |
| *** dsneddon has joined #openstack-ironic | 11:10 | |
| *** stendulker_ has joined #openstack-ironic | 11:12 | |
| *** dsneddon has quit IRC | 11:13 | |
| *** stendulker has quit IRC | 11:15 | |
| *** olivierbourdon38 has quit IRC | 11:17 | |
| *** vsaienk0 has joined #openstack-ironic | 11:17 | |
| *** stendulker_ has quit IRC | 11:19 | |
| *** dsneddon has joined #openstack-ironic | 11:19 | |
| openstackgerrit | Ilya Etingof proposed openstack/sushy-tools master: Abstract away the virtualization management interface https://review.openstack.org/523321 | 11:23 |
| *** vsaienk0 has quit IRC | 11:27 | |
| openstackgerrit | Sam Betts proposed openstack/ironic master: Ensure we unbind flat network ports and clear BM mac addresses https://review.openstack.org/539981 | 11:30 |
| *** hw_wutianwei has quit IRC | 11:33 | |
| *** hw_wutianwei has joined #openstack-ironic | 11:34 | |
| *** hw_wutianwei has quit IRC | 11:36 | |
| *** links has quit IRC | 11:44 | |
| *** vsaienk0 has joined #openstack-ironic | 11:45 | |
| *** zhangfei has quit IRC | 11:45 | |
| *** hamzy_ is now known as hamzy | 11:50 | |
| *** hw_wutianwei has joined #openstack-ironic | 11:51 | |
| *** vsaienk0 has quit IRC | 11:54 | |
| *** links has joined #openstack-ironic | 11:57 | |
| openstackgerrit | Ilya Etingof proposed openstack/sushy-tools master: Abstract away the virtualization management interface https://review.openstack.org/523321 | 12:00 |
| *** zhangfei has joined #openstack-ironic | 12:02 | |
| openstackgerrit | Nikolay Fedotov proposed openstack/ironic-python-agent master: Try to find partition 3 times https://review.openstack.org/525577 | 12:02 |
| *** mjturek has joined #openstack-ironic | 12:08 | |
| *** makowals has joined #openstack-ironic | 12:12 | |
| makowals | Hello Ironic | 12:12 |
| makowals | On my ironic-api.log I constantly see connection pool full error | 12:13 |
| makowals | ``` | 12:13 |
| makowals | WARNING requests.packages.urllib3.connectionpool [-] Connection pool is full, discarding connection | 12:13 |
| makowals | ``` | 12:13 |
| makowals | Any ideas where can I change the config to allow for more connections? In the config file I did not see anything obvious | 12:14 |
| *** vsaienk0 has joined #openstack-ironic | 12:17 | |
| *** zhangfei has quit IRC | 12:19 | |
| *** links has quit IRC | 12:21 | |
| *** hw_wutianwei has quit IRC | 12:28 | |
| *** hw_wutianwei has joined #openstack-ironic | 12:28 | |
| *** rbudden has joined #openstack-ironic | 12:30 | |
| *** zhangfei has joined #openstack-ironic | 12:32 | |
| *** SerenaFeng has joined #openstack-ironic | 12:34 | |
| makowals | Also this error causes problems with creating instances (it happens creation fails with this as an error message), so it's not just a to-be-ignored warning | 12:37 |
| openstackgerrit | Dao Cong Tien proposed openstack/ironic master: devstack: add support for rescue mode https://review.openstack.org/524118 | 12:38 |
| *** rwsu has quit IRC | 12:41 | |
| *** vsaienk0 has quit IRC | 12:41 | |
| *** dsneddon has quit IRC | 12:42 | |
| openstackgerrit | Dao Cong Tien proposed openstack/ironic master: Update a CI job for rescue mode https://review.openstack.org/528704 | 12:44 |
| *** bfournie has quit IRC | 12:45 | |
| *** bfournie has joined #openstack-ironic | 12:45 | |
| *** ameeda has joined #openstack-ironic | 12:49 | |
| *** bfournie has quit IRC | 12:49 | |
| *** olivierbourdon38 has joined #openstack-ironic | 12:49 | |
| ameeda | Hello, I have 21 nodes . after introspection finish I got 5 servers timed out. I checked ironic-conductor and ironic-inspector logs, that doesn't help me. | 12:51 |
| openstackgerrit | OpenStack Release Bot proposed openstack/ironic-python-agent stable/queens: Update .gitreview for stable/queens https://review.openstack.org/542155 | 12:57 |
| openstackgerrit | OpenStack Release Bot proposed openstack/ironic-python-agent stable/queens: Update UPPER_CONSTRAINTS_FILE for stable/queens https://review.openstack.org/542156 | 12:57 |
| openstackgerrit | OpenStack Release Bot proposed openstack/ironic-python-agent master: Update reno for stable/queens https://review.openstack.org/542157 | 12:57 |
| *** SerenaFeng has quit IRC | 12:57 | |
| *** jaganathan has quit IRC | 12:59 | |
| *** zhangfei has quit IRC | 13:01 | |
| *** links has joined #openstack-ironic | 13:02 | |
| *** jlabarre has joined #openstack-ironic | 13:02 | |
| *** AlexeyAbashkin has quit IRC | 13:03 | |
| *** AlexeyAbashkin has joined #openstack-ironic | 13:04 | |
| *** jaganathan has joined #openstack-ironic | 13:06 | |
| *** dsneddon has joined #openstack-ironic | 13:07 | |
| openstackgerrit | Olivier Bourdon proposed openstack/bifrost master: Fix DHCP informations according to network interface settings https://review.openstack.org/542167 | 13:10 |
| *** vsaienk0 has joined #openstack-ironic | 13:10 | |
| *** dsneddon has quit IRC | 13:13 | |
| *** rbudden has quit IRC | 13:13 | |
| *** zhangfei has joined #openstack-ironic | 13:13 | |
| *** ankkumar has quit IRC | 13:20 | |
| *** dsneddon has joined #openstack-ironic | 13:20 | |
| *** vsaienk0 has quit IRC | 13:20 | |
| openstackgerrit | OpenStack Release Bot proposed openstack/bifrost stable/queens: Update .gitreview for stable/queens https://review.openstack.org/542172 | 13:21 |
| openstackgerrit | OpenStack Release Bot proposed openstack/bifrost stable/queens: Update UPPER_CONSTRAINTS_FILE for stable/queens https://review.openstack.org/542173 | 13:21 |
| openstackgerrit | OpenStack Release Bot proposed openstack/bifrost master: Update reno for stable/queens https://review.openstack.org/542175 | 13:21 |
| *** zhangfei has quit IRC | 13:22 | |
| *** etingof has quit IRC | 13:23 | |
| *** trown|outtypewww is now known as trown | 13:23 | |
| *** bfournie has joined #openstack-ironic | 13:25 | |
| *** zhangfei has joined #openstack-ironic | 13:34 | |
| *** etingof has joined #openstack-ironic | 13:35 | |
| *** d0ugal has quit IRC | 13:36 | |
| *** etingof has quit IRC | 13:42 | |
| *** d0ugal has joined #openstack-ironic | 13:43 | |
| *** rwsu has joined #openstack-ironic | 13:44 | |
| TheJulia | good morning | 13:45 |
| TheJulia | brrrraaaaaiiiinnsssss | 13:45 |
| mgoddard | morning TheJulia, congrats | 13:47 |
| *** rloo has joined #openstack-ironic | 13:47 | |
| TheJulia | Thanks! | 13:47 |
| hwoarang | pas-ha: hello :) we are seeing some ironic-inspector + keystone noauth issues and I suspect http://git.openstack.org/cgit/openstack/ironic-inspector/commit/ironic_inspector?id=918775cb0109209561fd62f1b52a3d9f7cc226c6 may have caused it. Here is the trace http://paste.opensuse.org/34885094 | 13:48 |
| *** absubram has joined #openstack-ironic | 13:49 | |
| *** crushil has quit IRC | 13:52 | |
| TheJulia | hwoarang: that sounds like a misconfiuration... | 13:53 |
| hwoarang | TheJulia: yeah that happens in bifrost but nothing changed in bifrost or in the OPNFV world | 13:53 |
| *** etingof has joined #openstack-ironic | 13:54 | |
| *** dsneddon has quit IRC | 13:54 | |
| *** absubram has quit IRC | 13:54 | |
| *** absubram has joined #openstack-ironic | 13:54 | |
| TheJulia | hmmm | 13:55 |
| hwoarang | also there is this call to the keystone API https://github.com/openstack/ironic-inspector/blob/master/ironic_inspector/common/ironic.py#L87 but maybe that shouldn't happen when noauth is used | 13:55 |
| hwoarang | at least that's where the trace is leading me | 13:55 |
| *** AlexeyAbashkin has quit IRC | 13:56 | |
| TheJulia | ugh, yeah that would do it | 13:56 |
| *** AlexeyAbashkin has joined #openstack-ironic | 13:56 | |
| TheJulia | we need to determine if noauth and act upon it first | 13:56 |
| TheJulia | should be an easy fix if you have the code handy. Should also be easy for you to verify it works for you in that cse | 13:56 |
| hwoarang | TheJulia: here is the inspector config file https://bpaste.net/show/4107e2ecb090 | 13:57 |
| hwoarang | pretty basic | 13:57 |
| openstackgerrit | Ilya Etingof proposed openstack/sushy-tools master: Abstract away the virtualization management interface https://review.openstack.org/523321 | 13:57 |
| TheJulia | ameeda: If you try to re-inspect the five nodes that timed out, do they respond? | 13:57 |
| hwoarang | yeah let me check | 13:57 |
| *** dtantsur|bbl is now known as dtantsur | 13:59 | |
| dtantsur | morning TheJulia | 13:59 |
| dtantsur | TheJulia: should I officially hand you the keys to $everything? :) here they are! (and a biiig bottle of whisky) | 14:00 |
| TheJulia | makowals: you can scale out APIs and conductors, I think that might be what you want to do? Also... make sure you don't have a monitoring system checking to see that the api is alive. I've seen some that do mean things like leave connections open or that are checking every second.... | 14:00 |
| *** Goneri has joined #openstack-ironic | 14:02 | |
| TheJulia | dtantsur: over the next few days, hoping to take today/tomorrow off | 14:03 |
| *** crushil has joined #openstack-ironic | 14:04 | |
| rloo | good morning all ironic'ers, TheJulia, dtantsur, hwoarang | 14:05 |
| rloo | dtantsur: are we done wrt releases? | 14:06 |
| dtantsur | TheJulia: if you hope to take today off, now is about the right time ;) | 14:06 |
| dtantsur | rloo: ironic is pending the release team review | 14:06 |
| rloo | dtantsur: heh, we didn't do much wrt docs. | 14:06 |
| dtantsur | networking-baremetal and n-g-s will be done today | 14:06 |
| rloo | dtantsur, TheJulia; we should backport the vif-detach. what about the glance no-image? | 14:06 |
| dtantsur | rloo: what exactly? any docs can (and should) be backported | 14:06 |
| dtantsur | rloo: you mean, to pike? | 14:07 |
| rloo | dtantsur: yeah, to pike | 14:07 |
| rloo | dtantsur: wrt docs, there weren't many that were ready to land. i think you need to update one related to classic drivers :) | 14:07 |
| dtantsur | rloo: my thought was to get busy with pike after we get rid of queens for real | 14:07 |
| rloo | dtantsur: ok | 14:07 |
| dtantsur | right, I was slacking on docs because they are backportable without any problems | 14:07 |
| dtantsur | will get back to them really soon | 14:07 |
| TheJulia | I think it would be good to backport to pike but brains | 14:07 |
| rloo | dtantsur will be happy when I start asking TheJulia about stuff | 14:07 |
| ameeda | TheJulia: thanks for reply, I am not sure, I guess that to inspect just one node and see tcpdump to see if the node got ip, any other idea ? | 14:07 |
| dtantsur | rloo: SOOOOOOOON :D | 14:08 |
| *** etingof has quit IRC | 14:08 | |
| rloo | dtantsur: :) | 14:08 |
| dtantsur | TheJulia: in case you missed my ping: I've proposed stable releases for essentially everything that contained meaningful changes | 14:08 |
| TheJulia | ameeda: so... inspector is single-threaded with connection queuing... so I could imagine 5 notes suddenly timing out if they are shipping lots of data back or if the node running inspector is not able to process as quickly | 14:09 |
| rloo | OH, I guess we still need to fix multinode grenade... | 14:09 |
| TheJulia | dtantsur: I was looking ofr htat ping because I saw it on my phone, but not my laptop | 14:09 |
| TheJulia | dtantsur: awesome | 14:09 |
| * TheJulia slides whiskey to rloo | 14:09 | |
| rloo | TheJulia: i was thinking we could wait til next week, I'm guessing you'll be official then, to look at multinode grenade :) | 14:10 |
| dtantsur | TheJulia: also if you missed my comment on n-g-s release note: I'd like the next release to be 1.0.0, not 0.6 | 14:10 |
| dtantsur | any objections? | 14:10 |
| rloo | fresh eyes, fresh leader, new era err cycle... | 14:10 |
| dtantsur | and ditto for networking-baremetal <-- hjensas, sambetts | 14:10 |
| TheJulia | dtantsur: none, just the reno prelude neeeds to be fixed | 14:10 |
| dtantsur | TheJulia: I will fix and approve | 14:11 |
| sambetts | 1.0.0 for net-baremetal? | 14:11 |
| TheJulia | dtantsur: none there either | 14:11 |
| TheJulia | ++++ | 14:11 |
| TheJulia | dtantsur: <3 | 14:11 |
| openstackgerrit | Dmitry Tantsur proposed openstack/networking-generic-switch master: Add initial release note https://review.openstack.org/534201 | 14:11 |
| TheJulia | rloo: I _think_ we're going to have to get a conductor up and online, however there is a smoke test that expects things to be fully back online, so tl;dr, we're stuck atm. :( | 14:12 |
| dtantsur | sambetts: yes, for it as well. do you expect breaking changes in Rocky? | 14:12 |
| sambetts | dtantsur: not that I can think of | 14:12 |
| *** zhangfei has quit IRC | 14:12 | |
| rloo | TheJulia: :-( We really need to talk to folks at ptg, about how to correctly test rolling upgrades | 14:12 |
| dtantsur | sambetts: my thinking is that we now recommend to use it, so we better start semvering it for real | 14:13 |
| * rloo knows that TheJulia will make it happen :) | 14:13 | |
| sambetts | dtantsur: yeah, sounds good | 14:13 |
| dtantsur | ironic release is on its way \o/ | 14:13 |
| rloo | wheeeeeee! | 14:14 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/ironic-inspector master: ironic_inspector: common: ironic: Do not query keystone when in noauth https://review.openstack.org/542206 | 14:14 |
| dtantsur | sambetts: mind reviewing https://review.openstack.org/#/c/541697/ then? | 14:14 |
| patchbot | patch 541697 - networking-baremetal - Fix devstack example | 14:14 |
| dtantsur | this seems the only outstanding patch | 14:14 |
| rloo | melissaml sure likes to fix 'the the' occurrences... | 14:14 |
| ameeda | TheJulia: as I see at the logs . after finish introspect of 16's node, it wait around 1 hours before raise the error ironic_inspector.node_cache [-] Introspection for nodes [...] timed out | 14:14 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/ironic-inspector master: ironic_inspector: common: ironic: Do not query keystone when in noauth https://review.openstack.org/542206 | 14:15 |
| ameeda | TheJulia: I think that the nodes can't boot PXE, so I guessed to watch tcpdump. is that will help ? | 14:16 |
| TheJulia | ameeda: Hmm, that is a strong possibility and a good place to start. | 14:16 |
| openstackgerrit | OpenStack Release Bot proposed openstack/ironic stable/queens: Update .gitreview for stable/queens https://review.openstack.org/542207 | 14:16 |
| openstackgerrit | OpenStack Release Bot proposed openstack/ironic stable/queens: Update UPPER_CONSTRAINTS_FILE for stable/queens https://review.openstack.org/542208 | 14:16 |
| openstackgerrit | OpenStack Release Bot proposed openstack/ironic master: Update reno for stable/queens https://review.openstack.org/542209 | 14:16 |
| rloo | yay! | 14:17 |
| ameeda | TheJulia: thanks for help, I will let you know the details :) | 14:18 |
| *** rbudden has joined #openstack-ironic | 14:19 | |
| openstackgerrit | Ruby Loo proposed openstack/bifrost master: Update reno for stable/queens https://review.openstack.org/542175 | 14:20 |
| *** jcoufal has joined #openstack-ironic | 14:22 | |
| dtantsur | rloo, sambetts, TheJulia, oh, the last release thing: we have a few outstanding patches against ironic-tempest-plugin that now started passing the CI (thanks to Sam). let's try to get them in, if possible? | 14:22 |
| *** etingof has joined #openstack-ironic | 14:23 | |
| rloo | dtantsur: sure, you want to get them in before releasing ironic-tempest-plugin? | 14:23 |
| dtantsur | rloo: well, rescue for sure, because we released ironic with rescue | 14:23 |
| dtantsur | anything else I don't care much | 14:23 |
| rloo | dtantsur: ok | 14:24 |
| *** zhangfei has joined #openstack-ironic | 14:24 | |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/ironic-inspector master: ironic_inspector: common: ironic: Do not query keystone when in noauth https://review.openstack.org/542206 | 14:25 |
| openstackgerrit | Ruby Loo proposed openstack/ironic-python-agent master: Update reno for stable/queens https://review.openstack.org/542157 | 14:26 |
| *** rloo has quit IRC | 14:26 | |
| *** lucasagomes is now known as lucas-hungry | 14:26 | |
| *** rloo has joined #openstack-ironic | 14:26 | |
| openstackgerrit | Merged openstack/networking-baremetal master: Fix devstack example https://review.openstack.org/541697 | 14:26 |
| *** zhangfei has quit IRC | 14:26 | |
| *** dsneddon has joined #openstack-ironic | 14:27 | |
| openstackgerrit | Bob Fournier proposed openstack/ironic-inspector master: Only set switch_id in local_link_connection if mac address https://review.openstack.org/542214 | 14:27 |
| dtantsur | rloo: actually, I don't quite like the rescue tests patch | 14:29 |
| rloo | dtantsur: if you don't quite like it, i won't either :) | 14:29 |
| dtantsur | on the other hand, ironic-tempest-plugin is branchless | 14:29 |
| dtantsur | so new tests will always be picked | 14:30 |
| dtantsur | meaning: I can release it now :) | 14:30 |
| rloo | dtantsur: if it is branchless, why do we bother releasing? | 14:30 |
| dtantsur | rloo: for downstream's convenience | 14:31 |
| rloo | dtantsur: OH. :) | 14:31 |
| rloo | dtantsur: well, wouldn't downstream want to test rescue? | 14:31 |
| rloo | dtantsur: how bad is 'not liking it'? is it good enough to merge then modify later, or not worth merging? | 14:31 |
| rloo | dtantsur: i'm good with not testing rescue right now. it isn't all hooked up yet. | 14:32 |
| dtantsur | rloo: quite bad. the test is a part of baremetal_basic_ops and is silently skipped when rescue is not available | 14:32 |
| dtantsur | which is hard to detect. I'd prefer a separate test, to be honest. | 14:32 |
| rloo | dtantsur: in that case, forget it. let's spend some time doing it better. | 14:33 |
| *** dsneddon has quit IRC | 14:33 | |
| sambetts | +1 should be separate test with a "skipif" so we get a nice message in the logs | 14:33 |
| TheJulia | ++ | 14:37 |
| *** jcoufal has quit IRC | 14:37 | |
| *** lbragstad has joined #openstack-ironic | 14:37 | |
| hjensas | dtantsur: I saw a ping, reno prelude? How soon? I am out to another activity with the little one in 5 min. | 14:39 |
| dtantsur | hjensas: only a question if you're fine with the new networking-baremetal release receiving version 1.0.0 | 14:40 |
| dtantsur | as a designation of its (supposed) maturity | 14:40 |
| openstackgerrit | Merged openstack/networking-generic-switch master: Add initial release note https://review.openstack.org/534201 | 14:40 |
| hjensas | dtantsur: I guess if we recommend people using it in docs it should? | 14:40 |
| dtantsur | that's my thinking as well | 14:41 |
| * hjensas goes to update CV ... :D | 14:41 | |
| dtantsur | lol | 14:42 |
| *** etingof has quit IRC | 14:42 | |
| *** hjensas is now known as hjensas|afk | 14:42 | |
| openstackgerrit | Ruby Loo proposed openstack/ironic master: Update reno for stable/queens https://review.openstack.org/542209 | 14:43 |
| TheJulia | hjensas|afk: it just means you get to support it for the rest of time :) | 14:43 |
| dtantsur | TheJulia: don't tell him so quickly :D | 14:44 |
| TheJulia | dtantsur: doh! | 14:45 |
| TheJulia | hjensas|afk: on a serious note, you've done a great job with it! | 14:45 |
| dtantsur | yes, hjensas|afk++ | 14:45 |
| openstackgerrit | Ilya Etingof proposed openstack/sushy-tools master: Abstract away the virtualization management interface https://review.openstack.org/523321 | 14:46 |
| *** jcoufal has joined #openstack-ironic | 14:47 | |
| *** shekar has joined #openstack-ironic | 14:47 | |
| *** shekarvajjula has quit IRC | 14:48 | |
| *** links has quit IRC | 14:50 | |
| openstackgerrit | Ruby Loo proposed openstack/ironic master: Update reno for stable/queens https://review.openstack.org/542209 | 14:51 |
| openstackgerrit | Ruby Loo proposed openstack/ironic master: Update reno for stable/queens https://review.openstack.org/542209 | 14:52 |
| TheJulia | Okay, stepping away to hopefully let this headache go away and do something today that lets my brain unwind. Ping me if needed. | 14:53 |
| *** mjturek has quit IRC | 14:53 | |
| rloo | TheJulia: enjoy! | 14:54 |
| dtantsur | TheJulia: take some rest indeed | 14:54 |
| *** olivierbourdon38 has quit IRC | 14:55 | |
| *** dtantsur is now known as dtantsur|brb | 14:55 | |
| *** etingof has joined #openstack-ironic | 14:56 | |
| *** r-daneel has joined #openstack-ironic | 14:56 | |
| *** SerenaFeng has joined #openstack-ironic | 15:02 | |
| *** devananda has joined #openstack-ironic | 15:02 | |
| openstackgerrit | Merged openstack/bifrost stable/pike: Zuul: Remove project name https://review.openstack.org/542131 | 15:03 |
| *** links has joined #openstack-ironic | 15:03 | |
| rloo | here's the glance bug wrt the postgres job failure. doesn't look like it'll be addressed any time soon. https://bugs.launchpad.net/glance/+bug/1747869 | 15:09 |
| openstack | Launchpad bug 1747869 in Glance "./stack.sh doesn't work with postgres" [Critical,Triaged] - Assigned to Brian Rosmaita (brian-rosmaita) | 15:09 |
| rloo | oh sorry, they are targetting queens rc2 to fix, so next week hopefully | 15:11 |
| *** hamzy has quit IRC | 15:12 | |
| *** mjturek has joined #openstack-ironic | 15:18 | |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/ironic-inspector master: ironic_inspector: common: ironic: Do not query keystone when in noauth https://review.openstack.org/542206 | 15:18 |
| *** mjturek has quit IRC | 15:23 | |
| *** dsneddon has joined #openstack-ironic | 15:24 | |
| *** lucas-hungry is now known as lucasagomes | 15:26 | |
| *** vsaienk0 has joined #openstack-ironic | 15:26 | |
| *** slaweq has joined #openstack-ironic | 15:27 | |
| *** links has quit IRC | 15:28 | |
| *** dsneddon has quit IRC | 15:29 | |
| openstackgerrit | Bob Fournier proposed openstack/ironic-inspector master: Only set switch_id in local_link_connection if it is a mac address https://review.openstack.org/542214 | 15:29 |
| * rloo thinks we have to do a better job with clarifying what was decided cuz I don't know what we decided: http://eavesdrop.openstack.org/meetings/ironic/2017/ironic.2017-12-11-17.00.log.html#l-229 :-( | 15:34 | |
| *** dtantsur|brb is now known as dtantsur | 15:34 | |
| *** vsaienk0 has quit IRC | 15:36 | |
| *** SerenaFeng has quit IRC | 15:38 | |
| *** moshele has quit IRC | 15:38 | |
| *** rpioso|afk is now known as rpioso | 15:42 | |
| rpioso | Good morning, ironicers. | 15:43 |
| rloo | morning rpioso | 15:46 |
| *** hamzy has joined #openstack-ironic | 15:46 | |
| rpioso | rloo: :) | 15:46 |
| *** e0ne has quit IRC | 15:47 | |
| rloo | crushil: did you work on the xclarity stuff? I noticed that we don't have any docs on it: https://docs.openstack.org/ironic/latest/admin/drivers.html | 15:47 |
| bfournie | TheJulia: thanks for review on https://review.openstack.org/#/c/542214/, sambetts: do you think we should remove ironic validation here https://github.com/openstack/ironic/blob/master/ironic/api/controllers/v1/types.py#L313 to allow all values to be accepted for switch_id? | 15:48 |
| patchbot | patch 542214 - ironic-inspector - Only set switch_id in local_link_connection if it ... | 15:48 |
| crushil | rloo, Yes, I did, I am working on a patch that will address the fixes in the driver code and then add docs. I should be able to push that up soon | 15:48 |
| rloo | crushil: oh, so xclarity isn't done yet. I thought it was. | 15:49 |
| crushil | rloo, The code is. There were some nits I need to fix in a follow up patch | 15:49 |
| crushil | And I was going to add docs as well to go with that patch | 15:49 |
| *** hamzy_ has joined #openstack-ironic | 15:50 | |
| rloo | crushil: ok. we released ironic queens today. so we'll probably want to backport the doc | 15:50 |
| *** e0ne has joined #openstack-ironic | 15:51 | |
| crushil | rloo, Ok. Makes sense, thanks | 15:51 |
| rloo | crushil: fwiw, probably better to have separate patches, one to address nits, one for docs. | 15:51 |
| crushil | rloo, Ok. That's what I was shooting for. My comments above were probably confusing | 15:51 |
| TheJulia | bfournie: I really think it is valid to take it all, but largely because the switch in my basement is configurable to what I want it to transmit... and I think it defaults on hostname or something funky | 15:51 |
| rloo | crushil: thx :) | 15:52 |
| crushil | rloo, :) | 15:52 |
| * TheJulia re-engages cloaking device of trying to rest | 15:52 | |
| *** links has joined #openstack-ironic | 15:52 | |
| *** hamzy has quit IRC | 15:52 | |
| *** mjura has quit IRC | 15:53 | |
| *** rbartal has quit IRC | 15:53 | |
| *** SerenaFeng has joined #openstack-ironic | 15:54 | |
| bfournie | thanks just want to make sure ml2 driver isn't relying on mac for switch_id | 15:55 |
| *** slaweq has quit IRC | 15:56 | |
| *** slaweq has joined #openstack-ironic | 15:56 | |
| devananda | bfournie: it seems to also take a data path id today | 15:56 |
| devananda | Though, I haven't gotten it to actually connect to the switch, it doesn't yield any errors | 15:57 |
| bfournie | devananda: yes either mac address or data path id, from the lldp data can only be a mac address (OpenFlow data path id can be set manually) | 15:58 |
| bfournie | devananda: if the LLDP chassis id is an IP address, ifname etc. it will cause an Ironic exception on validation of link_local_connection and fail introspection | 15:59 |
| TheJulia | bfournie: That is truly a question for the ml2 driver... :\ I'm 95% sure n-g-s just ignores it | 15:59 |
| *** dsneddon has joined #openstack-ironic | 15:59 | |
| sambetts | I've responded on the patch, I think our current contract with the ml2 drivers is mac_address or openflow, personally I was campaining from the begining for it to have more relaxed validation because our Ml2 driver completely ignores the switch_id field today because mac address is useless to us | 15:59 |
| devananda | sambetts: what info does your driver need, then? | 16:00 |
| sambetts | we rely on putting the mangement IP address in the not-validated, switch_info field instead | 16:00 |
| devananda | Also, has anyone gotten that working with ovs / openflow, rather than a vendor ml2 driver? | 16:01 |
| *** slaweq has quit IRC | 16:01 | |
| *** vsaienk0 has joined #openstack-ironic | 16:01 | |
| sambetts | devananda: gotten our neutron driver working with OVS/openflow? | 16:01 |
| *** pcaruana has quit IRC | 16:02 | |
| *** dsneddon has quit IRC | 16:03 | |
| devananda | sambetts: nope... Still trying | 16:04 |
| devananda | I mean, yes, that's me question... Because I haven't made it work yet | 16:04 |
| dtantsur | TheJulia, rloo, https://bugs.launchpad.net/ironic/+bug/1748224 | 16:06 |
| openstack | Launchpad bug 1748224 in Ironic "[RFE] Stop special handling of mime extensions in the API" [Wishlist,Confirmed] | 16:06 |
| sambetts | devananda: I havn't got the neutron network driver working with openflow, but I've been working on a side project which uses a special ironic network driver + the standard neutron OVS ml2 driver, to do multi-tenancy where each conductor becomes the TOR and the physical switchs are statically configures and just provide fabric extension to the baremetals | 16:06 |
| rloo | dtantsur: thx, I'm good with approving it. Will comment :) | 16:07 |
| devananda | !!! .... I'm intrigued by this idea. | 16:08 |
| openstack | devananda: Error: "!!" is not a valid command. | 16:08 |
| rloo | good morning stranger...errrr devananda! | 16:08 |
| devananda | sambetts: does this use l2gw or something similar, to create routes between the vlans and tenant networks? | 16:08 |
| TheJulia | o/ devananda | 16:09 |
| devananda | Morning, rloo :) | 16:09 |
| * rloo would love to see devananda at a PTG | 16:09 | |
| devananda | I would love to go, but don't have funding .... | 16:09 |
| *** SerenaFeng has quit IRC | 16:09 | |
| rloo | :-( | 16:10 |
| dtantsur | sigh | 16:10 |
| dtantsur | I was hoping for 4 PTL photo now that Jim is back :D | 16:10 |
| lucasagomes | devananda, :-( | 16:11 |
| rloo | dtantsur: one of these years maybe, yeah, that would have been cool! | 16:11 |
| *** vsaienk0 has quit IRC | 16:11 | |
| devananda | brb | 16:14 |
| sambetts | devananda: basic principle of my idea (although bare in mind this is a ruff first poc) is that each port on physical switches and therefore each NIC on each baremetal is assigned a unique carrier vlan, and then on the conductor side it plumbs vlan sub-interfaces (which indirectly are the baremetal NICs) into OVS exactly the same way nova plumbs a VMs tap interface in | 16:18 |
| dtantsur | sambetts: do you have anything public? I'd certainly be very interested in it. | 16:19 |
| dtantsur | bfournie: FYI ^^^ | 16:19 |
| sambetts | dtantsur: ruff POC is -> https://github.com/Tehsmash/ironic/blob/ironic-ovs/ironic/drivers/modules/network/ovs.py | 16:19 |
| dtantsur | sambetts: thanks! any docs, devstack, ...? | 16:20 |
| rloo | dtantsur: you seem ok with this, i am thinking it is an RFE, I'm good with approving the bug thingy: https://review.openstack.org/#/c/540954/. Or we consider it part of original rescue work? | 16:20 |
| patchbot | patch 540954 - ironic - Add validate_rescue() method to boot interface | 16:20 |
| dtantsur | rloo: I was treating it more as a problem in how rescue interface is composed. but I don't think this change is backportable. | 16:21 |
| sambetts | dtantsur: I've not had much chance to work on it recently but I did get it working in devstack locally with some tweaks to how we plumb in our virtual baremetal nodes | 16:21 |
| sambetts | dtantsur: I'm planning on getting some more time to work on it soon though | 16:21 |
| dtantsur | hmm, it may be easier to try it on tripleo, it has less networking hacks | 16:21 |
| dtantsur | sambetts: if you write something resembling a formal RFE, I can try getting some time to work on it as well | 16:22 |
| rloo | dtantsur: right, i think it makes sense to do. i'll tag it as rfe (cuz it changes the interface API) and approve | 16:22 |
| *** e0ne has quit IRC | 16:22 | |
| sambetts | dtantsur: https://bugs.launchpad.net/ironic/+bug/1735169 | 16:22 |
| openstack | Launchpad bug 1735169 in Ironic "[RFE] Network interface implementation that allows for isolated networking (multitenancy) without TOR programming" [Wishlist,Confirmed] - Assigned to Sam Betts (sambetts) | 16:22 |
| dtantsur | sambetts: thanks, I'll show it to my PM :) | 16:23 |
| sambetts | I was also looking at taking the same principles and trying to make an implementation of multi-tenancy that would work for standalone without neutron | 16:23 |
| sambetts | but need to see how it works out | 16:24 |
| openstackgerrit | Merged openstack/ironic-inspector master: Remove redundant "the" from a docstring https://review.openstack.org/541973 | 16:25 |
| bfournie | sambetts: thanks for update to https://review.openstack.org/#/c/542214/, so it seems for now we should implement it this way, i.e. only set switch_id if mac address | 16:25 |
| patchbot | patch 542214 - ironic-inspector - Only set switch_id in local_link_connection if it ... | 16:25 |
| sambetts | bfournie: I think so, and then I think we need to take a serious look at the current limitations we impose on the llc data | 16:26 |
| bfournie | sambetts: agree, thanks | 16:26 |
| *** absubram has quit IRC | 16:26 | |
| sambetts | bfournie: I think the orignal reason for the limitations was to prevent the ML2 driver implementations from become too different in the information they require | 16:27 |
| sambetts | but I think the ml2 drivers are so different in the information they require it a little restricting | 16:27 |
| dtantsur | TheJulia: are you looking into https://docs.google.com/spreadsheets/d/1J2MRdVQzSyakz9HgTHfwYPe49PaoTypX66eNURsopQY/edit#gid=0 btw? | 16:27 |
| bfournie | sambetts: yep | 16:27 |
| *** tesseract has quit IRC | 16:29 | |
| *** tesseract has joined #openstack-ironic | 16:29 | |
| TheJulia | dtantsur: It is on my radar, trying to give my brain rest today. Failing at that of course. | 16:31 |
| dtantsur | TheJulia: I'm just making sure you've seen it, no hurry at all | 16:31 |
| dtantsur | in the worst case we'll make our photo ourselves :D or make lucasagomes do it for us | 16:32 |
| TheJulia | bfournie: pulled my -1 and commented | 16:32 |
| TheJulia | dtantsur: it could be sillier that way.... ;) | 16:32 |
| * lucasagomes reads | 16:32 | |
| bfournie | thanks TheJulia | 16:33 |
| *** dsneddon has joined #openstack-ironic | 16:33 | |
| lucasagomes | dtantsur, haha sure, but we also need a pic with me included in it :-) | 16:34 |
| lucasagomes | dtantsur, btw, bring pixie with ya | 16:36 |
| dtantsur | lucasagomes: last time Pixie was in TheJulia's caring hands | 16:36 |
| lucasagomes | ah TheJulia bring it then :D | 16:36 |
| *** dsneddon has quit IRC | 16:38 | |
| openstackgerrit | Merged openstack/ironic-tempest-plugin master: Replace use of functools.wraps() with six.wraps() https://review.openstack.org/540187 | 16:38 |
| TheJulia | Pixie boots is ready! https://usercontent.irccloud-cdn.com/file/Kba8iZvr/irccloudcapture957928365.jpg | 16:38 |
| dtantsur | sambetts: the code looks reasonably small and clear btw | 16:39 |
| *** absubram has joined #openstack-ironic | 16:39 | |
| sambetts | dtantsur: yeah its quite clean, it can probably been cleaned up even more if we can pull in the os-vif library which abstracts the ovs commands | 16:41 |
| dtantsur | sambetts: so, time to write a spec? ;) as I said, I can help with it, but I'll need to try it first | 16:41 |
| sambetts | dtantsur: yeah, I started putting on together locally | 16:42 |
| sambetts | one* | 16:42 |
| rloo | dtantsur, TheJulia: my suggestion, Thurs right after lunch. so we don't lose too much time. | 16:42 |
| *** amarao has quit IRC | 16:43 | |
| dtantsur | sambetts: cool, ping me for review | 16:44 |
| sambetts | dtantsur: will do :D | 16:46 |
| *** openstackgerrit has quit IRC | 16:48 | |
| NobodyCam | Good Morning Ironic'ers | 16:49 |
| NobodyCam | and congrats to TheJulia | 16:49 |
| dtantsur | morning NobodyCam | 16:52 |
| NobodyCam | hey hey dtantsur :) good morning :) | 16:53 |
| *** athomas has quit IRC | 16:57 | |
| *** marios has quit IRC | 16:57 | |
| *** dsneddon has joined #openstack-ironic | 17:00 | |
| *** slaweq has joined #openstack-ironic | 17:04 | |
| *** dsneddon has quit IRC | 17:08 | |
| *** slaweq has quit IRC | 17:09 | |
| *** ijw has joined #openstack-ironic | 17:18 | |
| *** derekh has quit IRC | 17:18 | |
| *** AlexeyAbashkin has quit IRC | 17:19 | |
| jroll | morning everyone, y'all talk a lot, I'm declaring scrollback bankruptcy | 17:19 |
| jroll | :P | 17:19 |
| jroll | \o/ for the ironic release | 17:19 |
| NobodyCam | Good morning jroll | 17:19 |
| *** sambetts is now known as sambetts|af | 17:20 | |
| * jroll is in training things for the next 3+ hours, jfyi | 17:20 | |
| *** sambetts|af is now known as sambetts|afk | 17:20 | |
| sambetts|afk | Night all o/ | 17:20 |
| NobodyCam | night sambetts|afk | 17:20 |
| *** AlexeyAbashkin has joined #openstack-ironic | 17:23 | |
| *** openstackgerrit has joined #openstack-ironic | 17:26 | |
| openstackgerrit | OpenStack Release Bot proposed openstack/networking-baremetal stable/queens: Update .gitreview for stable/queens https://review.openstack.org/542327 | 17:26 |
| openstackgerrit | OpenStack Release Bot proposed openstack/networking-baremetal stable/queens: Update UPPER_CONSTRAINTS_FILE for stable/queens https://review.openstack.org/542329 | 17:26 |
| openstackgerrit | OpenStack Release Bot proposed openstack/networking-baremetal master: Update reno for stable/queens https://review.openstack.org/542331 | 17:26 |
| openstackgerrit | OpenStack Release Bot proposed openstack/networking-generic-switch stable/queens: Update .gitreview for stable/queens https://review.openstack.org/542333 | 17:26 |
| openstackgerrit | OpenStack Release Bot proposed openstack/networking-generic-switch stable/queens: Update UPPER_CONSTRAINTS_FILE for stable/queens https://review.openstack.org/542334 | 17:26 |
| openstackgerrit | OpenStack Release Bot proposed openstack/networking-generic-switch master: Update reno for stable/queens https://review.openstack.org/542335 | 17:26 |
| *** AlexeyAbashkin has quit IRC | 17:28 | |
| openstackgerrit | Shivanand Tendulker proposed openstack/ironic-python-agent master: Follow-up patch for documentation of rescue mode https://review.openstack.org/542339 | 17:30 |
| *** AlexeyAbashkin has joined #openstack-ironic | 17:33 | |
| *** trown is now known as trown|lunch | 17:36 | |
| *** moshele has joined #openstack-ironic | 17:36 | |
| *** AlexeyAbashkin has quit IRC | 17:37 | |
| *** moshele has quit IRC | 17:54 | |
| *** baha has joined #openstack-ironic | 17:56 | |
| *** dtantsur is now known as dtantsur|afk | 17:58 | |
| dtantsur|afk | o/ | 17:58 |
| rloo | mgoddard, johnthetubaguy: wrt node traits, the spec mentions validating an instance's traits -- so this isn't quite finished? | 17:59 |
| rloo | bye dtantsur|afk | 17:59 |
| * mgoddard looks at traits spec | 17:59 | |
| *** dsneddon has joined #openstack-ironic | 17:59 | |
| rloo | mgoddard, johnthetubaguy: https://specs.openstack.org/openstack/ironic-specs/specs/approved/node-traits.html | 18:00 |
| rloo | under Proposed change | 18:00 |
| rloo | mgoddard: Ironic needs to validate that instance_info[‘traits’] is consistent with the list of traits currently associated with Ironic node, i | 18:00 |
| johnthetubaguy | ah, that could be right, I had forgot about that | 18:01 |
| johnthetubaguy | to stop races and suprizes | 18:01 |
| rloo | johnthetubaguy: so does that mean we shouldn't use traits for scheduling yet? | 18:01 |
| mgoddard | I think this is a belt and braces thing | 18:02 |
| * rloo not sure what 'belt and braces' means, must be English-eze. | 18:02 | |
| mgoddard | nova *should* only be pushing down instance traits that exist on the node | 18:02 |
| mgoddard | like johnthetubaguy said, there could be races | 18:02 |
| persia | rloo: English-ese for the Armericanese "Belt and Suspenders" | 18:03 |
| rloo | persia: ahhh (and who wears suspenders these daze) | 18:03 |
| johnthetubaguy | yeah, its safe ish without the check | 18:03 |
| rloo | mgoddard: so i wonder if we should push a patch up and backport to queens. | 18:03 |
| johnthetubaguy | its just for when you update traits, it takes a while to sync up | 18:03 |
| rloo | mgoddard: 'we' == you or john :) | 18:03 |
| mgoddard | but if you're removing traits at the same time as deploying instances then you're asking for trouble | 18:04 |
| johnthetubaguy | yeah, its an edge case bug thing | 18:04 |
| jroll | +1 on backporting it | 18:04 |
| rloo | so we should prevent updates of traits on nodes that are in deploy* state. (i've already forgotten how traits was coded) | 18:05 |
| mgoddard | we decided not to limit setting of traits too strictly | 18:05 |
| mgoddard | I did ask about this during spec review but we kept it simple | 18:06 |
| rloo | hmm, if we actually do the validation of traits at this point and it fails, it means the node deploy fails and user/nova won't be too happy? | 18:06 |
| *** e0ne has joined #openstack-ironic | 18:07 | |
| rloo | mgoddard: i think we tend to go with simple when we land a feature, and then we refine it. we're refining it a bit now. maybe... | 18:07 |
| jroll | the reason to validate them at deploy time is primarily for the traits that can change some hardware config, right? | 18:09 |
| mgoddard | I think the logic was - (assuming deploy-time validation of instance traits) at the time of the deploy, the requested instance traits were valid, so no need to prevent changes to node traits | 18:09 |
| jroll | e.g. set UEFI mode if using the flavor that has the UEFI trait | 18:09 |
| jroll | we want to make sure the node is marked as capable of doing that thing | 18:10 |
| jroll | (and I guess also to prevent races in scheduling) | 18:10 |
| mgoddard | and perhaps also for standalone | 18:11 |
| mgoddard | would adding instance trait validation count as an API change? | 18:13 |
| *** moshele has joined #openstack-ironic | 18:13 | |
| jroll | no, not in the sense of API semantics changing | 18:13 |
| jroll | actually not in any sense IMO | 18:13 |
| rloo | mgoddard: depends on where/how we add it. if an explicit API, then yes. if internally in our code, then no. | 18:13 |
| *** slaweq_ has joined #openstack-ironic | 18:13 | |
| jroll | just a change in the way deploys happen. yet another pre-flight check. | 18:14 |
| rloo | mgoddard: i think we'd want to add this internally. as jroll sez. | 18:14 |
| jroll | ++ | 18:14 |
| *** mjturek has joined #openstack-ironic | 18:14 | |
| rloo | i think that's what it sez in the spec; i recall discussing a bit with johnthetubaguy, and i figured we'd figure out the best place to validate once the coding started :) | 18:15 |
| mgoddard | probably needs to be in conductor's update_node() to guarantee consistency | 18:15 |
| johnthetubaguy | mgoddard: so the patch API fails during a deploy? | 18:16 |
| rloo | i actually feel like we shouldn't allow trait updates when a node is 'available' or in 'deploy*' | 18:16 |
| * johnthetubaguy has to run, just as things get interesting :( | 18:17 | |
| mgoddard | yeah, the patch instance_info API call would fail | 18:17 |
| rloo | mgoddard: ++ that would work | 18:18 |
| *** ijw_ has joined #openstack-ironic | 18:18 | |
| *** slaweq_ has quit IRC | 18:19 | |
| mgoddard | rloo: I'll take a look what happens elsewhere | 18:19 |
| rloo | mgoddard: thx | 18:20 |
| rloo | mgoddard: we released ironic today, it isn't a big deal to do another release next week. so this isn't urgent (ie, I don't think it needs to be done today) | 18:20 |
| mgoddard | rloo: ok, good :) | 18:21 |
| openstackgerrit | Merged openstack/ironic stable/queens: Update .gitreview for stable/queens https://review.openstack.org/542207 | 18:21 |
| rloo | mgoddard: yeah, no need to go overboard on this part :) | 18:21 |
| *** ijw has quit IRC | 18:21 | |
| mgoddard | rloo: good catch by the way | 18:22 |
| *** vsaienk0 has joined #openstack-ironic | 18:23 | |
| rloo | mgoddard: i always have to remember (and force myself) to go over specs of features that are 'done' to verify ... | 18:23 |
| jroll | hm, I was thinking more along the lines of ensuring instance_info and traits were compatible at deploy time | 18:24 |
| *** hw_wutianwei has quit IRC | 18:25 | |
| *** hw_wutianwei has joined #openstack-ironic | 18:25 | |
| jroll | e.g. `all(t in node.traits for t in node.instance_info['traits'])` | 18:25 |
| rloo | jroll: isn't that what mgoddard is proposing? oh wait, there are 3? possible places to do this check. | 18:26 |
| jroll | I could live with not updating node.traits when in the deploy* state, but I think we should allow it in the available state. I think. not 100%. | 18:26 |
| *** moshele has quit IRC | 18:26 | |
| jroll | rloo: he is suggesting checking at the time instance_info is updated. I'm suggesting checking it when we begin the deploy. | 18:26 |
| rloo | jroll: 1. when updating the node with instance_info['traits']; 2. when calling node.validate() or node.deploy.validate() (can't recall which), and 3. when node.set-provision-state active | 18:26 |
| jroll | ah yep, forgot about 2 | 18:27 |
| rloo | jroll: so would there be a reason when updating node with instance info, that you don't want/care that the traits match? | 18:27 |
| rloo | jroll: maybe cuz someone has some weird case or is in the middle of updating the node... | 18:28 |
| rloo | jroll: so we probably want to check at step 2 & 3. the code for 3 usually calls 2. | 18:28 |
| jroll | rloo: good question, probably not. probably best to do it early like that. I haven't fully thought it through yet. | 18:28 |
| jroll | rloo: yeah I guess I'd definitely put it in DeployInterface.validate(), whether we do it at step 1 or not. | 18:29 |
| * jroll does not have 100% attention here | 18:30 | |
| rloo | someone could still change the node traits between steps 1 & 2 & 3 :-( | 18:31 |
| jroll | we should probably just ignore that case, honestly | 18:32 |
| jroll | traits should be relatively stable - it's qualitative properties of the hardware | 18:32 |
| *** vsaienk0 has quit IRC | 18:32 | |
| jroll | any sane person will not be removing traits often, and adding traits won't fail the validation | 18:33 |
| *** etingof has quit IRC | 18:33 | |
| rloo | jroll: yeah, it is ok (cuz we can handle with all the validations) if someone changes the traits in between. i'm just not sure how that affects nova scheduling, to have traits change 'in the middle'. but from ironic point of view and provisioning, it'll be fine. | 18:33 |
| jroll | rloo: scheduling is already complete before step 1 | 18:33 |
| rloo | jroll: ah, then we should be good. | 18:34 |
| jroll | rloo: yep. and if someone f's up the traits for a node during a deploy, I vote we just go ahead and let the build fail and reschedule. that'll teach em. | 18:35 |
| rloo | jroll: yup, and i'll quote you on that! :D | 18:35 |
| jroll | gladly! | 18:35 |
| *** tesseract has quit IRC | 18:37 | |
| *** ijw_ has quit IRC | 18:38 | |
| *** etingof has joined #openstack-ironic | 18:39 | |
| *** dsneddon has quit IRC | 18:39 | |
| *** moshele has joined #openstack-ironic | 18:40 | |
| rloo | i'm sorry i am looking at the code. it occurs to me that having DeployInterface.validate() do the trait validation may be tricky, unless we assume that every DeployInterface that cares about traits, will validate it. also, conductor calls deploy.validate before locking node & deploying, so there is a bit of a race thing but i think we're ok with that; we should add jroll's comment ^^ to the code that mgoddard is going to write | 18:42 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/ironic-inspector master: ironic_inspector: common: ironic: Do not query keystone when in noauth https://review.openstack.org/542206 | 18:42 |
| *** links has quit IRC | 18:42 | |
| rloo | mgoddard: ok, i'll leave it for you to deal with. Thanks! :) | 18:42 |
| *** trown|lunch is now known as trown | 18:43 | |
| mgoddard | rloo: good discussion here, thanks. | 18:44 |
| jroll | rloo: I think ironic needs to always validate it or never validate it, it should be above the driver layer | 18:44 |
| jroll | so I guess we should say at DeployInterface.validate() time, rather than in DeployInterface.validate() | 18:44 |
| rloo | jroll: the problem with that, is if you call the node.validate() API, you won't get that trait validation being done :-( | 18:45 |
| openstackgerrit | James E. Blair proposed openstack/ironic-inspector stable/pike: Zuul: Remove project name https://review.openstack.org/542367 | 18:45 |
| mgoddard | jroll: +1, shouldn't be up to the driver | 18:45 |
| *** Sukhdev has joined #openstack-ironic | 18:46 | |
| mgoddard | there's another way to make this consistent. don't allow addition of instance traits that aren't node traits, and don't allow removal of node traits that are instance traits | 18:46 |
| jroll | rloo: unless we have some sort of validate call we add above the driver (which I'm a bit surprised we don't have yet) | 18:46 |
| mgoddard | that way the node is always consistent | 18:47 |
| rloo | jroll: yes, that is what i was thinking, but not sure how easy it would be to add and not break out-of-tree stuff. | 18:47 |
| *** etingof has quit IRC | 18:48 | |
| jroll | rloo: has nothing to do with out of tree drivers | 18:49 |
| rloo | mgoddard: hmm. so you're going with step 1, when updating the node. | 18:49 |
| rloo | jroll: depending on how we implement it, it might. i'd have to look at the code to see what we do. | 18:49 |
| jroll | rloo: I'm thinking (as an example) that we would call task.node.validate() here: https://github.com/openstack/ironic/blob/master/ironic/conductor/manager.py#L822 | 18:50 |
| jroll | the other option is call some _validate_traits() method there | 18:50 |
| jroll | same thing just less portable to other actions | 18:51 |
| * jroll has to run for now, bbl | 18:51 | |
| rloo | jroll: ok, that'd work, we'd have to do that validate call here too: https://github.com/openstack/ironic/blob/master/ironic/conductor/manager.py#L1827 | 18:53 |
| rloo | mgoddard: no clear answer. i am wondering if we make the check in the node.update API call, whether we need to microversion it. | 18:55 |
| *** slaweq has joined #openstack-ironic | 18:56 | |
| rloo | mgoddard: i think you can think about it :) I put a comment (and link to irc) in the bug: https://bugs.launchpad.net/ironic/+bug/1722194 | 19:00 |
| openstack | Launchpad bug 1722194 in Ironic "[RFE] Allow Traits to be set on Nodes " [Wishlist,In progress] - Assigned to Mark Goddard (mgoddard) | 19:00 |
| *** slaweq has quit IRC | 19:01 | |
| *** dsneddon has joined #openstack-ironic | 19:05 | |
| *** hieulq has quit IRC | 19:06 | |
| *** etingof has joined #openstack-ironic | 19:09 | |
| *** lucasagomes is now known as lucas-afk | 19:09 | |
| *** ijw has joined #openstack-ironic | 19:12 | |
| *** ijw has quit IRC | 19:15 | |
| *** vsaienk0 has joined #openstack-ironic | 19:18 | |
| *** ijw has joined #openstack-ironic | 19:20 | |
| *** ijw has quit IRC | 19:25 | |
| *** ElCoyote_ has joined #openstack-ironic | 19:26 | |
| *** vsaienk0 has quit IRC | 19:28 | |
| *** lbragstad has quit IRC | 19:31 | |
| *** lbragstad has joined #openstack-ironic | 19:32 | |
| *** ijw has joined #openstack-ironic | 19:35 | |
| *** e0ne has quit IRC | 19:36 | |
| *** AlexeyAbashkin has joined #openstack-ironic | 19:37 | |
| *** fragatina has quit IRC | 19:40 | |
| *** mgoddard_ has quit IRC | 19:43 | |
| pas-ha | hwoarang: checked the patch, there is a problem indeed, should've only catched it in backward compat handling of config (when you set [ironic]auth_strategy=noauth). But I think the fix should be done it somewhat other way.. | 19:53 |
| *** ElCoyote_ has quit IRC | 19:56 | |
| pas-ha | sambetts|afk: ankkumar: the 'standalone' tests in the ironic-tempest-plugin are not actually designed to be run against standalone ironic, but instead for running w/o nova, basically replicating the order of calls to ironic-api made by nova's ironic virt driver. the aim was to get rid of nova's scheduling woes and uncertainty to fix many things in ironic in parallel. | 19:57 |
| *** ijw has quit IRC | 20:05 | |
| pas-ha | about floating vs fixed ports in those tests - we could probably fix that part to support the mentioned scenario indeed... But OTOH these tests are designed to test different types of drivers as they are hot-swapping those during tests - how applicable is that to the 3rd party CI? shouldn't 3rd party CI just setup/test their own driver? would the HW the 3rd party CI testing even work with generic IPMI | 20:05 |
| pas-ha | hardware type or standard 'agent_ipmitool' or 'pxe_ipmitool' drivers? The more I think about it, the more I lean to these standalone tests should be skipped by the 3rd party CI. | 20:05 |
| * pas-ha was saving a lot of words.. | 20:06 | |
| *** moshele has quit IRC | 20:08 | |
| *** ijw has joined #openstack-ironic | 20:19 | |
| *** moshele has joined #openstack-ironic | 20:19 | |
| *** e0ne has joined #openstack-ironic | 20:23 | |
| *** ijw has quit IRC | 20:24 | |
| *** moshele has quit IRC | 20:25 | |
| *** hw_wutianwei has quit IRC | 20:31 | |
| *** hw_wutianwei has joined #openstack-ironic | 20:33 | |
| *** moshele has joined #openstack-ironic | 20:35 | |
| *** e0ne has quit IRC | 20:50 | |
| *** hamzy_ has quit IRC | 21:00 | |
| *** mgoddard_ has joined #openstack-ironic | 21:00 | |
| *** ricardoas is now known as ricardoas-away | 21:00 | |
| *** moshele has quit IRC | 21:03 | |
| openstackgerrit | Merged openstack/networking-generic-switch stable/queens: Update .gitreview for stable/queens https://review.openstack.org/542333 | 21:07 |
| *** dsneddon has quit IRC | 21:08 | |
| *** dsneddon has joined #openstack-ironic | 21:09 | |
| *** fragatina has joined #openstack-ironic | 21:27 | |
| *** e0ne has joined #openstack-ironic | 21:34 | |
| *** trown is now known as trown|outtypewww | 21:38 | |
| *** Sukhdev has quit IRC | 21:39 | |
| *** mgoddard_ has quit IRC | 21:56 | |
| *** zshi has quit IRC | 22:08 | |
| *** crushil has quit IRC | 22:09 | |
| *** mjturek has quit IRC | 22:10 | |
| *** baha has quit IRC | 22:12 | |
| *** Goneri has quit IRC | 22:20 | |
| *** AlexeyAbashkin has quit IRC | 22:23 | |
| *** mgoddard_ has joined #openstack-ironic | 22:24 | |
| *** e0ne has quit IRC | 22:26 | |
| *** mgoddard_ has quit IRC | 22:30 | |
| *** rcernin has joined #openstack-ironic | 22:31 | |
| *** dsneddon_ has joined #openstack-ironic | 22:36 | |
| *** dsneddon_ has quit IRC | 22:39 | |
| *** dsneddon_ has joined #openstack-ironic | 22:39 | |
| *** Sukhdev has joined #openstack-ironic | 22:40 | |
| *** bfournie has quit IRC | 22:58 | |
| *** Sukhdev has quit IRC | 23:07 | |
| openstackgerrit | Ruby Loo proposed openstack/ironic master: Remove validate_boot_option_for_trusted_boot metric https://review.openstack.org/542477 | 23:08 |
| *** ijw has joined #openstack-ironic | 23:10 | |
| *** zshi has joined #openstack-ironic | 23:31 | |
| *** MattMan has quit IRC | 23:38 | |
| *** MattMan has joined #openstack-ironic | 23:38 | |
| *** jcoufal has quit IRC | 23:39 | |
| openstackgerrit | James E. Blair proposed openstack/ironic stable/ocata: Zuul: Remove project name https://review.openstack.org/542498 | 23:47 |
| *** moshele has joined #openstack-ironic | 23:49 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!
