Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2020-01-20
13:47:57 stephenfin gibi: Care to hit this now that others have looked at it? https://review.opendev.org/#/c/696745/
13:49:09 openstackgerrit Merged openstack/nova master: Set instance CPU policy to 'share' through image property https://review.opendev.org/688603
13:50:10 openstackgerrit Stephen Finucane proposed openstack/nova stable/train: Set instance CPU policy to 'share' through image property https://review.opendev.org/703393
13:51:06 gibi stephenfin: looking
14:08:22 openstackgerrit Kashyap Chamarthy proposed openstack/nova master: libvirt: Add a default VirtIO-RNG device to guests https://review.opendev.org/616603
14:08:33 kashyap stephenfin: When you get a moment --^ does that look better?
14:09:57 kashyap stephenfin: Quick change-log: (a) we should keep the both extra spec properties (see comments); (b) modified the driver function to make (correct) use of the extra spec props; (c) added a rel-note for change in semantics
14:10:05 kashyap stephenfin: Thanks for the earlier review. That was useful
14:16:45 gibi stephenfin: +A, thanks
14:16:49 brinzhang__ gibi: https://review.opendev.org/#/c/701763/2 I leave some comment on
14:16:53 stephenfin gibi: Thanks
14:16:56 stephenfin kashyap: Will do
14:17:04 gibi brinzhang__: ack
14:20:09 gibi brinzhang__: replied
14:24:33 brinzhang__ Yeah, I think "rule:project_reader_api" role is good to me
14:24:42 brinzhang__ gibi
14:25:44 brinzhang__ gmann: are you around?
14:28:35 gmann stephenfin: one question on hacking cap - https://review.opendev.org/#/c/695733/4/test-requirements.txt@5
14:28:44 brinzhang__ gmann: 'rule:project_reader_api' and 'rule:project_member_api' which different from 'rule:system_admin_or_owner' and 'rule:system_or_project_reader'? I am not sure
14:29:14 gmann brinzhang__: did not get completely.
14:30:09 brinzhang__ gmann: which scope?
14:30:11 stephenfin gmann: Makes sense. Will submit a follow-up
14:30:24 gmann brinzhang__: 'system_*' rules are with system level scope and project only rules are project scopped.
14:30:29 gmann stephenfin: +1 thanks
14:30:56 gmann brinzhang__: 'rule:system_admin_or_owner' and 'rule:system_or_project_reader' are [system, project] both scoped.
14:33:18 openstackgerrit Stephen Finucane proposed openstack/nova master: functional: Add '_create_server' helper https://review.opendev.org/697540
14:33:42 brinzhang__ gmann: other words, if we are *system_* role, we can get all projects resource. and if we are *project_* role, we just can get his own project's resource, right?
14:34:46 gmann brinzhang__: right.
14:35:10 brinzhang__ gmann: Thanks :)
14:35:57 gmann in addition to that, new default roles also considered like reader etc. so admin can get member,reader but not vice versa (on both scope case)
14:36:22 brinzhang__ gibi: so filter instances by az, I think use "rule:project_reader_api" is ok, what do you think?
14:37:04 openstackgerrit Stephen Finucane proposed openstack/nova master: requirements: Limit hacking to one minor version https://review.opendev.org/703405
14:37:46 brinzhang__ gmann: yeah, I reviewed your docs of this :P
14:38:28 stephenfin gmann: ^
14:38:53 gibi brinzhang__: based on you discussion with gmann it looks OK to me
14:39:06 gmann brinzhang__: gibi link ?
14:39:11 gmann stephenfin: thanks
14:39:36 gibi gmann: we are reviewing https://review.opendev.org/#/c/701763/ with brinzhang__
14:39:36 brinzhang__ gmann, https://review.opendev.org/#/c/701763/
14:39:54 gmann ok, thanks.
14:45:40 gmann brinzhang__: gibi i left comment in review. we should not need policy there. we already have policy to allow admin filters to access by non-admin.
14:46:01 gmann this case of when operator/user want AZ to be always allowed for non-admin not based on configuration.
14:46:50 openstackgerrit Stephen Finucane proposed openstack/nova master: Remove unused code https://review.opendev.org/703408
14:47:54 gibi gmann: thanks. does os_compute_api:servers:allow_all_filters'
14:48:17 gibi gmann: does os_compute_api:servers:allow_all_filters makes every filtering available for the user?
14:48:26 gmann gibi: yes, but that allow all admin filters.
14:48:54 gibi gmann: so that cannot be used to selectively enable AZ filtering for non-admin
14:49:06 gmann as per discussion with author on IRC, this policy is not they want to use for making AZ for non-admin and wanted AZ to be available by default
14:49:40 gmann gibi: yeah, i mean if we default any filter for non-admin then we do not need policy for that.
14:50:48 gibi gmann: so simply adding availiability_zone to nova.api.openstack.compute.servers.serverscontroller._get_server_search_options would make the AZ filter available to non-admin by default
14:50:53 gmann if we think AZ is for non-admin then we should not add config way to disallow for non-admin.
14:51:02 gmann gibi: yes
14:51:16 gibi gmann: I got it, thanks
14:52:10 gmann but i am little concern to bump microversion for this single fitler. i commented to do a audit for all such filter first. like we did for API clenup BP(doing such cleanup together in single microvesion)
14:52:55 gibi gmann: and I've just agreed with you about that
14:53:00 gibi :)
14:53:02 gibi in the review
14:53:12 gibi gmann: thanks for the review
14:53:19 gmann thanks.
14:54:27 brinzhang__ gmann, gibi: so this spec's context is ok, we donot car about the policy's limit :)
14:55:09 brinzhang__ gibi: I am ok with that ^
14:55:22 gmann yeah, policy is not required for that. we should add another configurable way which is not required.
14:55:58 gibi brinzhang__: I still think this is a valid spec. we don't need a new policy. And we should review the other existing admin-only filters to se other need to be moved to the non-admin set
14:56:41 gmann s/should/should not
14:57:12 brinzhang__ gibi: yeah, like gmann said, *that* should done in a micoroversion, avoid more microversions to do one thing
14:57:26 gibi brinzhang__: exactly
14:59:41 brinzhang__ gibi: I think I would like to do, but it will be start after our holiday~
15:00:55 gibi brinzhang__: sure. holiday recharging is important :) I also expect the author to do the first sweep of the existing filters
15:01:59 brinzhang__ gibi: ^^
15:17:07 openstackgerrit Sylvain Bauza proposed openstack/nova master: Avoid PlacementFixture silently swallowing kwargs https://review.opendev.org/695180
15:17:07 openstackgerrit Sylvain Bauza proposed openstack/nova master: Add a placement audit command https://review.opendev.org/670112
15:21:57 openstackgerrit Kashyap Chamarthy proposed openstack/nova master: libvirt: Add a default VirtIO-RNG device to guests https://review.opendev.org/616603
15:22:22 kashyap sean-k-mooney: ^ I reworded the commit message; and reworded the puropose more clearly.
15:22:42 kashyap sean-k-mooney: (And tweaked the driver.py; hop that looks OK.)
15:25:07 sean-k-mooney i would still delete the backgroud section to be honest. that should have beeen in the patch that chagne to use /dev/urandom by default not this chagne but i wont block on it
15:26:05 sean-k-mooney you could have just referencce the change id of that commit instead of providing the history again
15:30:29 gibi stephenfin: I have one question in https://review.opendev.org/#/c/697153/15/nova/network/neutron.py
15:37:06 stephenfin gibi: Replied (tl;dr: nothing has really changed, but we could probably improve it in a follow-up patch)
15:38:54 gibi stephenfin: ahh _setup_net_dict could raise a similar unhandled exception before. Then your change does not make it worse. thanks for the explanation
15:39:04 stephenfin yup, exactly
15:40:55 gibi stephenfin: do you have an estimate how much work still remains in the nova-net removal?
15:41:11 stephenfin gibi: I think the patches that are up there are it
15:41:43 stephenfin I'm sure I missed some stuff but that should have the bulk of the work (services, API, docs, tests) taken care of
15:41:44 gibi stephenfin: wow soo colse then I gather my brain and try to read through
15:41:59 gibi close
15:51:23 openstackgerrit sean mooney proposed openstack/nova stable/stein: Disable NUMATopologyFilter on rebuild https://review.opendev.org/702973
15:51:23 openstackgerrit sean mooney proposed openstack/nova stable/stein: Block rebuild when NUMA topology changed https://review.opendev.org/702972
15:51:24 openstackgerrit sean mooney proposed openstack/nova stable/stein: FUP for in-place numa rebuild https://review.opendev.org/702974
16:01:59 gibi stephenfin: thank you for your perseverance, it was an awesome job to remove all this logically dead code
16:22:47 KeithMnemonic elod: thank you for the review on https://review.opendev.org/#/c/683008/ any idea who else i can ask to get the "W"?
16:27:27 elod KeithMnemonic: no problem. well, I was a bit hesitant to give a +2+W, but decided to wait for other stable cores (list of stable cores: https://review.opendev.org/#/admin/groups/540,members )
16:42:44 KeithMnemonic melwitt: can you please join the party here https://review.opendev.org/#/c/683008/ ;-) ?
16:43:22 melwitt KeithMnemonic: I can go through it later today. thanks for getting more people to review it, that helps
16:44:01 KeithMnemonic thank you and as i mentioned the customer with the issue confirmed those fixes correct the broken behavior
16:56:57 openstackgerrit Elod Illes proposed openstack/nova stable/stein: Enhance service restart in functional env https://review.opendev.org/703103
17:01:35 kashyap gibi: If you're looking a change of scenery, an "easy win", if I may say so myself :-) -- https://review.opendev.org/#/c/616603/
17:02:01 gibi kashyap: I've saved it for tomorrow morning :)
17:02:26 kashyap gibi: Yep, it's EoD for us in CET anyway...
17:06:04 gibi yepp, I'm about to leave for today
17:31:56 openstackgerrit Ilya Etingof proposed openstack/nova master: Add JSON schema for network_data.json https://review.opendev.org/703133

Earlier   Later