| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2020-09-15 | |||
| 12:22:57 | sean-k-mooney | gibi: interesting | |
| 12:23:08 | sean-k-mooney | should we ignore or raise an error | |
| 12:23:29 | sean-k-mooney | i think the flavor validation code should raise an error no? | |
| 12:23:40 | gibi | I can change that to an error | |
| 12:24:08 | gibi | I choose to ignore as alex_xu said that we will never use that trait in a named group | |
| 12:24:09 | sean-k-mooney | im kind of ok ignoring but when creating or updating a flavor i think the validation code should make it an error | |
| 12:24:15 | sean-k-mooney | so ignore where your currently ignoring | |
| 12:24:16 | gibi | and it made the code a lot simpler | |
| 12:24:39 | sean-k-mooney | well we might | |
| 12:24:53 | sean-k-mooney | but we dont right now | |
| 12:25:39 | sean-k-mooney | so the reason that HW_CPU_HYPERTHREADING was originally not ment ot be set by the virt driver | |
| 12:25:52 | sean-k-mooney | was if we have numa in placment you might want to treat each numa node differently | |
| 12:26:29 | sean-k-mooney | but when stephenfin implemented the pcpus in placment feature we chose to allow the virt driver to report it | |
| 12:26:41 | sean-k-mooney | whihc broke tha tuse case | |
| 12:26:52 | sean-k-mooney | so right now we wont use it in a named group | |
| 12:27:30 | sean-k-mooney | the intent was to allow the isolate policy to work by only whitelisting half the treads on the scond numa node | |
| 12:27:55 | sean-k-mooney | instead of having to partion you cloud into servers with hypertreading enabled and servers with it disabled | |
| 12:28:25 | sean-k-mooney | as i said we decided to drop that usecase | |
| 12:29:59 | gibi | OK. I will try to add a case to the extra_spec validation | |
| 12:30:10 | gibi | to reject this trait in a named group | |
| 13:03:25 | openstackgerrit | Lee Yarwood proposed openstack/nova master: libvirt: Don't age swap files in the image cache https://review.opendev.org/641093 | |
| 13:30:00 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: Reject HW_CPU_HYPERTHREADING trait in named groups https://review.opendev.org/735947 | |
| 13:30:48 | gibi | sean-k-mooney: added an extra_spec validator ^^ | |
| 13:37:14 | lyarwood | does anyone know how/if we are tracking the pypy mirror outage(s)? | |
| 13:38:51 | sean-k-mooney | gibi: cool | |
| 13:39:13 | sean-k-mooney | gibi: i was just wondering should we extend this to all HW_CPU traits? | |
| 13:39:43 | sean-k-mooney | the traits in the image are always merged into the unnamed group right | |
| 13:40:06 | sean-k-mooney | but for a flavor does it ever make sense to allow any HW_CPU_* trait in the named groups? | |
| 13:40:23 | sean-k-mooney | stephenfin: ^ | |
| 13:40:59 | stephenfin | sean-k-mooney: Just left a comment suggesting there might be more traits of this nature to worry about | |
| 13:41:12 | gibi | sean-k-mooney: I did not think about that | |
| 13:41:17 | stephenfin | Personally, I'd have tackled this separately | |
| 13:41:50 | stephenfin | PS2 was good enough, IMO | |
| 13:50:12 | sean-k-mooney | i guess we could condier extendign this in a followup | |
| 13:52:25 | sean-k-mooney | by the way shoudl that bug be marked invalid. i mean it was not ment to work in named groups so its invalid but we are not blocking it so that is a bug | |
| 13:52:46 | sean-k-mooney | so i guess its valid but the title is a little strange | |
| 13:53:23 | sean-k-mooney | oh i see it worked with numbered before | |
| 13:53:48 | sean-k-mooney | ya its vaild then. | |
| 13:54:29 | openstackgerrit | Merged openstack/nova stable/train: Add note and daxio version to the vPMEM document https://review.opendev.org/750222 | |
| 14:22:45 | nightmare_unreal | is tempest and osc-functional-devstack related to user or issue with zuul / infra ? | |
| 14:32:05 | bauzas | gibi: ack, good evening \o | |
| 14:41:50 | openstackgerrit | Stephen Finucane proposed openstack/nova master: functional: Enable real policy fixture by default https://review.opendev.org/745906 | |
| 14:41:51 | openstackgerrit | Stephen Finucane proposed openstack/nova master: functional: Add and use 'GlanceFixture' https://review.opendev.org/745907 | |
| 14:41:51 | openstackgerrit | Stephen Finucane proposed openstack/nova master: tests: Remove '_FakeImageService' https://review.opendev.org/745908 | |
| 14:42:32 | stephenfin | bauzas, lyarwood: Could I ask one of you to take a look at those? ^ Ignore the line count - they're not complex :) | |
| 14:48:44 | sean-k-mooney | stephenfin: will the glance figure remove the need to use the fake image service stuff | |
| 14:48:52 | stephenfin | yup | |
| 14:49:06 | sean-k-mooney | cool there are like 3 differnt ways to stub glance currently | |
| 14:49:18 | sean-k-mooney | having one fixture for it would be nice | |
| 14:49:21 | stephenfin | indeed; that's pretty much the entire point of that fixture :) | |
| 14:50:14 | sean-k-mooney | ya in the rare case i have to care about the image i always have to first check which why is its stubbed which causes a context switch from the actual problem i was trying to solve | |
| 14:58:01 | bauzas | stephenfin: ack, /me clicks | |
| 15:04:08 | stephenfin | bauzas: ta | |
| 15:04:55 | bauzas | stephenfin: that's an easy one to review but I need to verify you haven't forgotten any | |
| 15:05:49 | stephenfin | Wise. fwiw, I simply used "ag -Q 'image.fake'" and "ag 'image import fake'" to find the imports | |
| 15:06:04 | stephenfin | (ag = the silver searcher, btw. Great tool) | |
| 15:21:06 | openstackgerrit | Stephen Finucane proposed openstack/nova master: Handle libvirt reporting incorrect 4k page quantities https://review.opendev.org/631038 | |
| 15:21:07 | openstackgerrit | Stephen Finucane proposed openstack/nova master: Make overcommit check for pinned instance pagesize aware https://review.opendev.org/631053 | |
| 15:22:22 | openstackgerrit | Lee Yarwood proposed openstack/nova master: DNM libvirt: Support extending locally attached rbd volumes https://review.opendev.org/752069 | |
| 15:23:34 | gmann | brinzhang_: i will check that today | |
| 15:39:09 | gmann | gibi: bauzas brinzhang_ do we need to fix this in older microversion or ask users to use new microversion 2.64 or higher which already has fix - https://bugs.launchpad.net/nova/+bug/1894966 | |
| 15:39:09 | openstack | Launchpad bug 1894966 in OpenStack Compute (nova) "Create servergroup failed with unexpected error" [Low,Confirmed] | |
| 15:40:04 | gmann | if we decide to fix in older microversion then we can only fix the error code from 500->400. I feel we should ask to use new microversion. | |
| 15:42:42 | openstackgerrit | Federico Ressi proposed openstack/nova master: Requires default python3-devel package on CentOS https://review.opendev.org/752073 | |
| 16:29:38 | ganso | Hi Nova folks! If anyone could please take one last look at https://review.opendev.org/#/c/748533/ it already has +2 and +1 (thanks stephenfin and sean-k-mooney!) | |
| 16:34:28 | sean-k-mooney | oh yes :) gibi i think is done for the day. bauzas maybe you can take a look its libvirt/networking related but its relivitly simple to review | |
| 16:34:56 | openstackgerrit | Federico Ressi proposed openstack/nova master: Log stack trace when fails launching a defined domain https://review.opendev.org/750652 | |
| 18:31:30 | openstackgerrit | Lee Yarwood proposed openstack/nova master: compute: Skip cinder_encryption_key_id check when booting from volume https://review.opendev.org/752090 | |
| #openstack-nova - 2020-09-16 | |||
| 00:24:30 | brinzhang0 | gmann: agree with you, allowing to request array([]) of the policy is itself an error, improve the error code from 500 to 400, it's good for me | |
| 00:33:49 | brinzhang0 | gmann, gibi, bauzas: Looks like it's a small change, and it's happened in microversion 2.1~2.63, | |
| 00:35:04 | gmann | brinzhang0: as it is already fixed in 2.64, my suggestion is that there is no much benefits to improve error code for older microversion. | |
| 00:36:22 | brinzhang0 | gmann: yeah, but improve it's error code seems friendly to end users | |
| 00:37:29 | brinzhang0 | but as you said, in the privous release, add new microversion is no much benefits | |
| 00:38:02 | gmann | it does not need microversion bump as it is error code improvement to existing error code only. | |
| 00:38:04 | brinzhang0 | how about add a NOTE or WARNING in the request api docs? | |
| 00:38:52 | gmann | yeah, that can be good. saying not to pass 'policies' as empty array. | |
| 00:39:02 | brinzhang0 | yes | |
| 00:40:00 | brinzhang0 | it's better than to improve the error from 500 to 400, after all it's in old release | |
| 00:41:40 | gmann | if you can push the doc change, i am good on that. | |
| 00:42:20 | brinzhang0 | gmann: it's my pleasure, later will do.. | |
| 00:42:35 | gmann | brinzhang0: thanks | |
| 00:42:42 | brinzhang0 | np | |
| 01:33:20 | openstackgerrit | Brin Zhang proposed openstack/nova master: Add warning to os-server-groups POST api https://review.opendev.org/752133 | |
| 01:34:17 | brinzhang0 | gmann: done by https://review.opendev.org/#/c/752133/ | |
| 06:32:32 | openstackgerrit | wu.chunyang proposed openstack/nova master: remove python warnning from tox https://review.opendev.org/752190 | |
| 06:50:13 | gibi | gmann: regarding https://bugs.launchpad.net/nova/+bug/1894966 I would only fix the error code to 500 -> 400 by enhancing the request validation | |
| 06:50:13 | openstack | Launchpad bug 1894966 in OpenStack Compute (nova) "Create servergroup failed with unexpected error" [Low,In progress] - Assigned to Brin Zhang (zhangbailin) | |
| 08:07:16 | brinzhang_ | gibi, gmann: I think that fix or the docs need to backport to the privous release, right? | |
| 08:07:31 | gibi | brinzhang_: yes that is good to be backported | |
| 08:07:44 | gibi | it is broken since the begining of time I guess | |
| 08:07:54 | brinzhang_ | yes | |
| 08:08:03 | brinzhang_ | so that we should target the bug to the stable branch | |
| 08:08:36 | gibi | yes please | |
| 08:08:46 | brinzhang_ | from v2.1, it's wrong | |
| 08:10:18 | brinzhang_ | gibi: bug 1894966 was backport to ocata, I think it's enough, please check | |
| 08:10:18 | openstack | bug 1894966 in OpenStack Compute (nova) "Create servergroup failed with unexpected error" [Low,In progress] https://launchpad.net/bugs/1894966 - Assigned to Brin Zhang (zhangbailin) | |
| 08:25:01 | gibi | brinzhang_: sure, it is OK. older than ocata branches are already marked end of life and deleted | |
| 08:25:27 | brinzhang_ | gibi: ack | |
| 08:28:50 | nightmare_unreal | I am writing a testcase for live_migration_abort feature but I keep getting this <module 'novaclient.v2.server_migrations' from '.tox/py38/lib/python3.8/site-packages/novaclient/v2/server_migrations.py'> does not have the attribute 'live_migration_abort' | |
| 08:28:55 | nightmare_unreal | any suggestions | |