| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2019-02-21 | |||
| 14:11:25 | openstackgerrit | Rodolfo Alonso Hernandez proposed openstack/os-vif master: Add function "has_table_columns" to OVSDB implementation API https://review.openstack.org/634967 | |
| 14:12:58 | openstackgerrit | Rodolfo Alonso Hernandez proposed openstack/os-vif master: Add native implementation OVSDB API https://review.openstack.org/482226 | |
| 14:24:33 | jackding | takashin: mdbooth: Could you please revisit https://review.openstack.org/#/c/616692/? Thanks | |
| 14:25:14 | takashin | jackdin: Okay. I will check it later. | |
| 14:25:23 | takashin | jackding: Okay. I will check it later. | |
| 14:29:57 | jackding | mriedem: Could you please revisit https://review.openstack.org/#/c/620706/ and https://review.openstack.org/#/c/616692/? Thanks | |
| 14:30:14 | jackding | takashin: thanks | |
| 14:48:37 | openstackgerrit | Merged openstack/os-vif master: Clean up versioned object backlevelling code https://review.openstack.org/630976 | |
| 15:02:04 | mdbooth | jackding: Done. Change looks good. I think there's some missing unit test coverage, though. | |
| 15:02:27 | mdbooth | jackding: Hopefully trivial to add. | |
| 15:03:11 | openstackgerrit | Surya Seetharaman proposed openstack/nova master: [Doc] Best practices for effectively tolerating down cells https://review.openstack.org/638173 | |
| 15:07:58 | mriedem | dansmith: you want to get this rocky backport for that live migration volumes thing https://review.openstack.org/#/c/636895/ | |
| 15:08:24 | openstackgerrit | ya.wang proposed openstack/nova master: Select cpu model from a list of cpu models https://review.openstack.org/637834 | |
| 15:16:55 | openstackgerrit | sean mooney proposed openstack/os-vif master: remove brctl from vif_plug_ovs https://review.openstack.org/636821 | |
| 15:16:56 | openstackgerrit | sean mooney proposed openstack/os-vif master: remove use of brctl from vif_plug_linux_bridge https://review.openstack.org/636822 | |
| 15:42:20 | mriedem | lyarwood: mgoddard: i know this backport has been hanging out awhile but i'm pretty uncomfortable with it https://review.openstack.org/#/c/608573/ | |
| 15:42:30 | mriedem | given the bug report and commit message says "since change in rocky" which was actually a change in stein | |
| 15:43:53 | mriedem | i guess this is an example of the warning http://logs.openstack.org/95/636895/1/check/ironic-tempest-dsvm-ipa-wholedisk-bios-agent_ipmitool-tinyipa/c4e5305/controller/logs/screen-n-cpu.txt.gz#_Feb_15_14_25_05_143091 | |
| 15:44:57 | mgoddard | mriedem: the "since change in rocky" == https://review.openstack.org/#/c/565841/? It merged in August '18 | |
| 15:45:55 | mriedem | included in on that change says master | |
| 15:46:43 | mriedem | https://github.com/openstack/nova/commits/stable/rocky/nova/virt/ironic/driver.py | |
| 15:46:46 | mriedem | not in there | |
| 15:47:03 | mriedem | i remember us talking about merging that late in rocky but decided to defer to early stein | |
| 15:47:11 | mriedem | and it was merged right after the ptg as i recall | |
| 15:47:22 | mriedem | and proceeded to break tripleo | |
| 15:47:37 | mriedem | because they were still using the scheduler filters that needed those inventory values | |
| 15:48:12 | mgoddard | ok, I agree it's not in stable/rocky | |
| 15:48:35 | mriedem | https://wiki.openstack.org/wiki/Nova/Rocky_Release_Schedule so that was merged probably right after rc1 | |
| 15:48:39 | mriedem | when we cut stable/rocky | |
| 15:51:05 | openstackgerrit | Pierre Riteau proposed openstack/nova master: Fix a typo in configuration description https://review.openstack.org/638445 | |
| 15:51:06 | mgoddard | ok, let's abandon it then. the backport was based on the assumption that that change was in rocky | |
| 15:51:07 | mriedem | god this whole transition has been a mess | |
| 15:51:15 | mgoddard | yup | |
| 15:51:18 | mriedem | mgoddard: i left a comment, but if we do have the warnings, | |
| 15:51:24 | mriedem | i think they have to be conditional based on the https://docs.openstack.org/nova/rocky/configuration/config.html#workarounds.report_ironic_standard_resource_class_inventory option in rocky | |
| 15:51:34 | mriedem | because if that's true, the warnings are legit | |
| 15:51:41 | mgoddard | that's another option | |
| 15:51:42 | mriedem | if false, then we can avoid the warnings | |
| 15:52:02 | mriedem | if that option is false we're not reporting standard resoruce class inventory anyway so we don't need to warn about 0 values | |
| 15:52:53 | mgoddard | that seems reasonable. I'll update it to check that flag | |
| 15:53:30 | mriedem | the commit message is going to need some updating | |
| 15:53:33 | mriedem | to sort out the confusion | |
| 15:53:52 | mgoddard | sure | |
| 15:54:15 | mgoddard | thanks for following up | |
| 15:55:26 | mgoddard | actually, hold up | |
| 15:55:44 | mgoddard | "They were not required to be set since the Pike release, but still this | |
| 15:55:47 | mgoddard | warning is emitted. | |
| 15:56:15 | mgoddard | that's why I marked it as backport potential for pike - rocky | |
| 15:56:52 | mgoddard | so since pike, it's valid not to set them, but if you don't then you get a warning | |
| 15:58:11 | mgoddard | it doesn't matter what is reported to placement, this is about the ironic node config | |
| 15:58:38 | mriedem | ok i see https://review.openstack.org/#/c/608573/3/nova/virt/ironic/driver.py@815 | |
| 15:58:56 | mriedem | so if the values are 0 we won't report those resource classes for inventory | |
| 16:00:00 | mriedem | i guess https://docs.openstack.org/ironic/rocky/install/configure-nova-flavors.html doesn't mention anything about 0ing out the node property values, | |
| 16:00:04 | mriedem | is that in the ironic docs somewhere? | |
| 16:01:15 | bauzas | gibi: I'm just starting to sharpen my pen for reviewing network-bw-providers | |
| 16:01:26 | bauzas | gibi: did the implementation diverged from the spec https://specs.openstack.org/openstack/nova-specs/specs/stein/approved/bandwidth-resource-provider.html ? | |
| 16:01:38 | bauzas | gibi: that helps me catching up quickier | |
| 16:02:04 | gibi | bauzas: the name of the resource classes got shorter, and there are extra object model addition that was not covered by the spec | |
| 16:02:16 | mriedem | and the microversion... | |
| 16:02:20 | gibi | bauzas: but I think the basic logic and goal did not changef | |
| 16:02:25 | gibi | mriedem: ++ | |
| 16:02:57 | bauzas | gibi: okay, gtk | |
| 16:03:19 | bauzas | mriedem: a new API microversion ? | |
| 16:03:27 | bauzas | if so, I'll be able to get it | |
| 16:03:42 | gibi | bauzas: yes a new microversion to enable the feature for server create/delete | |
| 16:04:00 | bauzas | gotcha | |
| 16:04:50 | mgoddard | mriedem: let me check the docs | |
| 16:05:17 | mriedem | i guess i just don't know how those values are zeroed out - the node API? something in ironic-inspector? | |
| 16:08:03 | mgoddard | mriedem: the node 'properties' field is just a dict, it defaults to being empty | |
| 16:09:04 | openstackgerrit | Matt Riedemann proposed openstack/nova stable/rocky: Don't emit warning when ironic properties are zero https://review.openstack.org/608573 | |
| 16:09:25 | mriedem | mgoddard: https://review.openstack.org/#/c/608573/3..4//COMMIT_MSG | |
| 16:09:34 | mgoddard | mriedem: I can't find anything that explicitly says not to set those properties, but that is what we test in devstack: https://github.com/openstack/ironic/blob/master/devstack/lib/ironic | |
| 16:09:42 | mgoddard | mriedem: search --property | |
| 16:10:21 | mgoddard | mriedem: nice, thanks for updating | |
| 16:10:38 | mriedem | i only see https://github.com/openstack/ironic/blob/master/devstack/lib/ironic#L2082 but that's something else | |
| 16:10:40 | mriedem | that's the flavor not the node | |
| 16:11:00 | mgoddard | L1996 is node creation | |
| 16:11:32 | openstackgerrit | Merged openstack/python-novaclient master: Make Server.networks use a predictable sort order https://review.openstack.org/638214 | |
| 16:11:33 | openstackgerrit | Merged openstack/python-novaclient master: API microversion 2.69: Handles Down Cells https://review.openstack.org/579563 | |
| 16:11:36 | mgoddard | only property is cpu_arch, and possibly root_device | |
| 16:12:29 | mriedem | ok and we don't get a KeyError because of https://review.openstack.org/#/c/608573/4/nova/virt/ironic/driver.py@245 | |
| 16:12:59 | mgoddard | right | |
| 16:13:05 | mriedem | so we'll warn there too :) | |
| 16:13:25 | mriedem | oh wait | |
| 16:13:29 | mriedem | nevermind we default to - | |
| 16:13:31 | mriedem | *0 | |
| 16:13:41 | mgoddard | yeah | |
| 16:14:26 | mriedem | ok approved | |
| 16:14:37 | mriedem | please re-do the queens backport using the updated stable/rocky change | |
| 16:15:01 | mgoddard | great, thanks. will do | |
| 16:15:12 | mriedem | sorry for the trouble. i had put off sorting this one out for weeks. | |
| 16:16:13 | mgoddard | no problem, I know the kind of task :) | |
| 16:16:39 | tbachman | z;. ¸d fgi`q1e | |
| 16:16:47 | tbachman | ||
| 16:27:26 | melwitt | o/ | |
| 16:28:00 | mriedem | cdent: efried: i'm kind of at a loss as to what to do about this stable/rocky nova backport which is a fix from extracted placement in stein https://review.openstack.org/#/c/608903/ | |
| 16:28:24 | mriedem | as far as i know, nothing is "using" nested providers in rocky - i know nova isn't, but i'm not aware of anything else either | |
| 16:28:33 | mriedem | so i'm leaning to not backport this if we don't need it | |
| 16:28:45 | cdent | mriedem: my inclication would be to wait until somebody asks | |
| 16:28:58 | mriedem | same | |