| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2020-07-16 | |||
| 09:44:49 | stephenfin | which uses API version 2 (.0) by default | |
| 09:45:03 | stephenfin | any reason those later tests shouldn't be using v2.1 + latest microversion too? | |
| 09:45:23 | stephenfin | you need the latest microversion so you can request a host https://github.com/openstack/nova/blob/master/nova/tests/functional/libvirt/test_vgpu.py#L330 | |
| 09:45:30 | stephenfin | well, a reasonably new one | |
| 09:46:25 | bauzas | stephenfin: yeah, we need a specific microversion for at least one test AFAICU | |
| 09:47:27 | bauzas | stephenfin: at least for the resize tests : https://github.com/openstack/nova/commit/d2e0afc1f72db5cb56ed987e2873798fc1e89545 | |
| 09:49:01 | stephenfin | bauzas: okay, so I'm going to move this https://github.com/openstack/nova/blob/master/nova/tests/functional/libvirt/test_vgpu.py#L134-L137 | |
| 09:49:14 | stephenfin | to the base class https://github.com/openstack/nova/blob/master/nova/tests/functional/libvirt/test_vgpu.py#L42 | |
| 09:49:36 | stephenfin | doesn't sound like you wanted to use the legact API v2 for the VGPUMultipleTypesTests tests | |
| 09:50:40 | bauzas | stephenfin: I just wanted to provide some tests for vGPU by using the v2 API, but then when I wanted to provide the resize ones, it needed the microversion | |
| 09:50:58 | bauzas | stephenfin: so, no worries if you want to use the latest API for all of them, it should work | |
| 09:51:09 | stephenfin | okay, sweet. Thanks | |
| 09:52:34 | bauzas | the main issue I had with the tests was the fact that I needed to move some methods to the InstanceHelperMixin one | |
| 09:52:59 | stephenfin | You'll probably be interested in https://review.opendev.org/#/c/741283/1 so | |
| 09:53:12 | bauzas | stephenfin: hence the https://github.com/openstack/nova/blob/d2e0afc1f72db5cb56ed987e2873798fc1e89545/nova/tests/functional/integrated_helpers.py#L538 | |
| 09:53:45 | stephenfin | bauzas: You're *definitely* interested in the above so | |
| 09:55:41 | bauzas | stephenfin: :) | |
| 11:06:15 | openstackgerrit | Merged openstack/nova master: notifications: add the pcpuset info to instance numa cell payload https://review.opendev.org/728480 | |
| 11:06:23 | openstackgerrit | Merged openstack/nova master: objects: Introduce the 'CPUAllocationPolicy.MIXED' enum https://review.opendev.org/716267 | |
| 11:18:57 | openstackgerrit | Merged openstack/nova master: hardware: refactor the code for CPU policy sanity check https://review.opendev.org/740925 | |
| 11:21:06 | lyarwood | stephenfin / gibi ; https://review.opendev.org/#/c/699291/ - Would you mind taking a look at this fix from mriedem that I've been trying to push over the line for a while? | |
| 11:22:07 | stephenfin | sure | |
| 11:24:47 | lyarwood | thanks, I'll get back to your series later today btw | |
| 11:26:07 | gibi | lyarwood: I will check it | |
| 11:56:20 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: Regenerate device metadata during interface detach https://review.opendev.org/741436 | |
| 12:10:23 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: [WIP] Support SRIOV interface attach and detach https://review.opendev.org/740995 | |
| 12:10:50 | stephenfin | gibi, sean-k-mooney: Do either of you know why we do this in the resize operation instead of accessing instance.image_meta? https://github.com/openstack/nova/blob/master/nova/compute/api.py#L3886-L3887 | |
| 12:11:15 | stephenfin | I suspect it's something to do with filtering out undesirable image metadata properties but I'm not sure | |
| 12:11:50 | sean-k-mooney | let me check | |
| 12:12:23 | sean-k-mooney | stephenfin: we are using the cached metadta | |
| 12:12:50 | sean-k-mooney | we store the image metadta in the system metadata table | |
| 12:13:46 | sean-k-mooney | if we looked up the image by its uuid and got the metadata it could have change or the image could have been deleted | |
| 12:15:33 | sean-k-mooney | stephenfin: we only have the image uuid in the instance https://github.com/openstack/nova/blob/master/nova/objects/instance.py#L127 | |
| 12:15:41 | sean-k-mooney | thre is no image_meta | |
| 12:16:00 | sean-k-mooney | but if you want to add a property that just calls that function i dont have any issue with that | |
| 12:17:32 | stephenfin | Hmm, I think they may be equivalent. 'Instance.image_meta' calls 'ImageMeta.from_instance' https://github.com/openstack/nova/blob/master/nova/objects/instance.py#L250-L251 | |
| 12:18:00 | sean-k-mooney | oh we do have a property https://github.com/openstack/nova/blob/master/nova/objects/instance.py#L127 | |
| 12:18:00 | stephenfin | which calls 'get_image_from_system_metadata' https://github.com/openstack/nova/blob/master/nova/objects/image_meta.py#L127 | |
| 12:18:26 | stephenfin | hmm, those sound equivalent so. Probably legacy stuff | |
| 12:18:26 | sean-k-mooney | yes https://github.com/openstack/nova/blob/057891ac247fadfb9637d3e834f8b1ef3e39bc27/nova/objects/image_meta.py#L126-L128 | |
| 12:18:38 | gibi | seem equivalent, yes | |
| 12:18:48 | stephenfin | So my series is gaining yet another tech debt reduction patch \o/ | |
| 12:18:51 | stephenfin | every. damn. time. | |
| 12:19:18 | sean-k-mooney | you could pretend you did not see it :P | |
| 12:19:29 | sean-k-mooney | also this seams backwards to me | |
| 12:20:00 | sean-k-mooney | i would have had instance.image_meta just call utils.get_image_from_system_metadata(sysmeta) | |
| 12:20:24 | gibi | stephenfin: you are a hero removing those depts | |
| 12:20:28 | gibi | debts | |
| 12:20:36 | sean-k-mooney | and had image_meta.from_instance(instance) return instance.image_meta | |
| 12:21:20 | stephenfin | fair point | |
| 12:21:23 | stephenfin | much of a muchness | |
| 12:21:38 | stephenfin | worth noting the 'ImageMeta.from_instance' also handles multiple different ways of storing system_metadata in the instance object | |
| 12:21:41 | sean-k-mooney | there are extra db queries curently | |
| 12:21:43 | stephenfin | if indeed instance is an object | |
| 12:22:18 | stephenfin | I've no idea if that call to 'utils.instance_sys_meta' is necessary, vs. just using 'instance.system_metadata' | |
| 12:22:45 | sean-k-mooney | it should not bey system_metadata is always loaded by default | |
| 12:23:14 | sean-k-mooney | https://github.com/openstack/nova/blob/master/nova/objects/instance.py#L70-L71 | |
| 12:23:15 | stephenfin | An it's a 'DictOfNullableStringsField' type field | |
| 12:23:23 | stephenfin | so it couldn't be anything *except* a dict | |
| 12:23:51 | stephenfin | so this check seems like garbage https://github.com/openstack/nova/blob/master/nova/utils.py#L547-L549 | |
| 12:23:57 | sean-k-mooney | ya it might be empty but it will be a dict | |
| 12:24:24 | stephenfin | yup, undefined or a dict. Nothing else since it's not nullable | |
| 12:24:51 | sean-k-mooney | well i dont know it might be there for some reason in the past | |
| 12:25:00 | stephenfin | pre-o.vo, most likely | |
| 12:25:33 | sean-k-mooney | https://github.com/openstack/nova/commit/96164cecc6b7bf447d06896f723dab416e63ee06 | |
| 12:25:53 | openstackgerrit | Merged openstack/nova stable/rocky: Reject boot request for unsupported images https://review.opendev.org/708608 | |
| 12:26:46 | sean-k-mooney | so ya it was part of the ovo migration | |
| 12:27:06 | sean-k-mooney | its proably not needed 7 years later. | |
| 13:35:57 | artom | gibi, how did you notice https://bugs.launchpad.net/nova/+bug/1887799 ? | |
| 13:35:57 | openstack | Launchpad bug 1887799 in OpenStack Compute (nova) "interface metadata is not regenerated after interface detach" [Low,In progress] - Assigned to Balazs Gibizer (balazs-gibizer) | |
| 13:36:04 | artom | Code inspection, or "real life"? | |
| 13:36:14 | artom | Because... we *do* update device metadata on restart | |
| 13:36:17 | artom | Just no in the driver | |
| 13:36:21 | artom | Lemme comment on the patch | |
| 13:36:44 | gibi | artom: working on the SRIOV interface attach feature and during detach I see that there is metadata left in the database | |
| 13:36:52 | gibi | then I inspected the code | |
| 13:37:34 | artom | gibi, something else is going on then: https://opendev.org/openstack/nova/src/branch/master/nova/network/neutron.py#L1755 | |
| 13:37:35 | sean-k-mooney | gibi: your actully workign on that | |
| 13:38:14 | sean-k-mooney | gibi: do you have any patches up | |
| 13:38:17 | gibi | sean-k-mooney: I've spent two day on it and I have a WIP patch, now writin a small spec | |
| 13:38:27 | sean-k-mooney | gibi: cool | |
| 13:38:34 | gibi | sean-k-mooney: https://review.opendev.org/#/c/740995/ | |
| 13:38:42 | gibi | sean-k-mooney: it works in devstack already | |
| 13:39:03 | artom | gibi, shot in the dark, but maybe your bug is related to https://review.opendev.org/#/c/670593/ ? | |
| 13:39:09 | sean-k-mooney | the code is more or less already there in regards to actully doing the attach. all that was missing was the claim which wew can do form the api using the new rpc calls added for live migration | |
| 13:39:42 | gibi | artom: thanks for the pointers I wil dig | |
| 13:40:03 | gibi | sean-k-mooney: yes, I needed to put together the pieces | |
| 13:40:12 | gibi | but most of the pieces exist | |
| 13:40:35 | sean-k-mooney | yep we say that when we filed the bug a few year ago but it was too close to code freeze to get it in | |
| 13:40:49 | sean-k-mooney | it was deemed to need a spec and we were past m2 | |
| 13:41:17 | sean-k-mooney | so the curent api check was added but we never went and fixed it. that was pre sriov migration so it was more work then | |
| 13:42:32 | jsuchome | gibi: thanks for your +2, what/who do we need for starting the workflow? | |
| 13:43:10 | sean-k-mooney | gibi: ill try an test out your change today or tomorow im currently working on the fucntial test for my pci device lookup patches but ill test yours when i test that on real hardware | |
| 14:07:45 | gibi | jsuchome: you need one more +2 . But I also suggest that pull this back to the original series so that the reviewers of the original series will see it and review it | |
| 14:08:11 | gibi | sean-k-mooney: thanks. any test is useful | |
| 14:08:28 | gibi | sean-k-mooney: I now have access with a physical machine with SRIOV card | |
| 14:13:51 | jsuchome | lyarwood: hi, could you please take a look at https://review.opendev.org/#/c/738738/ ? that's a new prerequisite to direct-rbd-download to make things more readable | |
| 14:14:31 | lyarwood | jsuchome: ack queued, might not get to it today however | |
| 14:14:46 | jsuchome | great, thanks! | |
| 14:34:40 | huaqiang | hi, I have a question, do we support specify such flavor and image metadata conbination：vcpus=4, hw:cpu_realtime=yes, hw_cpu_policy=dedicated , hw_cpu_realtime_mask=^0 | |