| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2020-04-07 | |||
| 08:05:13 | lyarwood | stephenfin: asking as I've added them in https://review.opendev.org/#/c/701430/ assuming we needed to do this regardless, if we don't I can rm -rf this from that change. | |
| 08:05:34 | bauzas | gibi: looking | |
| 08:06:02 | bauzas | gibi: FWIW, I can upload a DNM patch to see how to create pGPUs | |
| 08:06:43 | brinzhang_ | gibi: and I think there is an issue inline, but I dont have time to search how to fix it, so I add TODO in test case. The issue is filter by config_drive=False, cannt get the servers, because if we request config=False, that will be changed to None, that will be ignored | |
| 08:07:17 | gibi | brinzhang_: ack, I have to look at that client patch today. thanks for taking it over | |
| 08:07:34 | gibi | brinzhang_: I will try to see if I can fix that while you are asleep. | |
| 08:08:39 | brinzhang_ | gibi: thanks | |
| 08:09:55 | brinzhang_ | gibi: I was added TODO in test_shell, and test_servers file, if you are looking into, you can see it, while you fix that you can remove that TODO and complete the tes cases | |
| 08:11:13 | luyao | gibi: Hi do you have time to help review vpmem live migration? https://review.opendev.org/#/q/status:open+project:openstack/nova+branch:master+topic:bp/support-live-migration-with-virtual-persistent-memory. I think stephenfin is still and will be busy on mixed cpu patches. | |
| 08:12:02 | gibi | brinzhang_: thanks | |
| 08:12:20 | brinzhang_ | gibi:np | |
| 08:12:29 | openstackgerrit | Sylvain Bauza proposed openstack/nova master: DNM: Functional test with pGPUs https://review.opendev.org/717975 | |
| 08:12:41 | gibi | luyao: I'm knee deep in the vgpu types series, then I will jump to the novaclient patch above. So I don't think I will get to vpmem today. sorry | |
| 08:13:11 | bauzas | gibi: so | |
| 08:13:17 | bauzas | gibi: thanks for your comment | |
| 08:13:33 | bauzas | gibi: fwiw, I'm working on https://review.opendev.org/717975 | |
| 08:13:49 | gibi | bauzas: ack about the func test | |
| 08:13:52 | gibi | will look | |
| 08:13:56 | bauzas | gibi: see https://review.opendev.org/#/c/717975/1/nova/virt/libvirt/driver.py | |
| 08:14:07 | bauzas | maybe I should add it to make sure it works | |
| 08:14:20 | bauzas | (without waiting for the driver.init_host()) | |
| 08:15:16 | luyao | gibi: OK, it does not matter | |
| 08:16:35 | gibi | bauzas: I have to make a step back. now that we are cleare about the initialization order of the compute service. What will potentially break? and then how the driver change in 717975 prevents that? (sorry I don't have the full context yet) | |
| 08:17:14 | bauzas | gibi, so about your concern | |
| 08:17:39 | bauzas | say we call recreate_assigned_mediated_devices() before driver.init_host() | |
| 08:18:11 | bauzas | gibi: then we'll get the existing mdevs from libvirt XML guests | |
| 08:18:19 | bauzas | guest XMLs | |
| 08:18:21 | bauzas | * | |
| 08:18:58 | bauzas | gibi: so the 'parent' argument we pass down to _create_new_mediated_device is got from the libvirt XML, not from placement | |
| 08:20:01 | bauzas | gibi: my only concern without having https://review.opendev.org/#/c/717975/1/nova/virt/libvirt/driver.py means that if _recreate() is called before init_host() then the self.supported_vgpu_types value would be None | |
| 08:20:24 | bauzas | gibi: that's why I I'll provide a new revision to add the property | |
| 08:20:39 | gibi | OK. so the parent will not be None just the vgpu_types | |
| 08:21:04 | bauzas | gibi: tbc, we don't call Placement within _recreate() | |
| 08:21:13 | bauzas | gibi: it's just for recreating mdevs in sysfs | |
| 08:21:25 | bauzas | so it's okay to not have a reshape | |
| 08:21:45 | bauzas | gibi: but you actually found a related problem, so thanks ! | |
| 08:22:14 | gibi | bauzas:OK, thanks. I continue reading the patches... | |
| 08:22:39 | bauzas | gibi: I'll provide a new revision by the next 5 mins | |
| 08:23:04 | bauzas | gibi: just by adding the property and then just modifying the tests to no longer call init_host() | |
| 08:23:11 | bauzas | it will simplify the testrs | |
| 08:23:34 | bauzas | (and I'll clarify the comment to explain that we don't call placement so we're all fine) | |
| 08:24:46 | gibi | thanks | |
| 08:34:14 | jkulik | gibi, you've given your +2 on https://review.opendev.org/#/c/713863/ and after some discussion the -1 is gone, too. Is there anything left to do? Do we wait for other reviewers? | |
| 08:35:37 | gibi | jkulik: I hope melwitt will have time to get back to it as in general we need two +2 votes to merge a patch. | |
| 08:36:08 | jkulik | Ah, didn't know that. Thank you. | |
| 08:36:11 | openstackgerrit | xulei proposed openstack/nova master: Add unit test for nova-api _check_metadata_properties_quota https://review.opendev.org/717981 | |
| 08:38:10 | gibi | jkulik: this week is feature freeze so most of the reviews are busy finishing up feature reviews so it might take time to get the another +2 on your bug. I suggest to ping melwitt early next week | |
| 08:42:54 | jkulik | Sure, thanks. Just wanted to know what's the holdup and if I can do anything. | |
| 08:44:28 | openstackgerrit | Stephen Finucane proposed openstack/nova master: tox: Integrate mypy https://review.opendev.org/676208 | |
| 08:44:29 | openstackgerrit | Stephen Finucane proposed openstack/nova master: hardware: Update and correct typing information https://review.opendev.org/714694 | |
| 08:44:29 | openstackgerrit | Stephen Finucane proposed openstack/nova master: libvirt: Add typing information https://review.opendev.org/714695 | |
| 08:44:30 | openstackgerrit | Stephen Finucane proposed openstack/nova master: tests: Split instance NUMA object tests https://review.opendev.org/714696 | |
| 08:44:30 | openstackgerrit | Stephen Finucane proposed openstack/nova master: objects: Replace 'cpu_pinning_requested' helper https://review.opendev.org/714697 | |
| 08:44:31 | openstackgerrit | Stephen Finucane proposed openstack/nova master: hardware: Don't consider overhead CPUs for unpinned instances https://review.opendev.org/714698 | |
| 08:44:31 | openstackgerrit | Stephen Finucane proposed openstack/nova master: hardware: Remove handling of pre-Train compute nodes https://review.opendev.org/714699 | |
| 08:44:32 | openstackgerrit | Stephen Finucane proposed openstack/nova master: hardware: Add validation for 'cpu_realtime_mask' https://review.opendev.org/468203 | |
| 08:44:32 | openstackgerrit | Stephen Finucane proposed openstack/nova master: hardware: Tweak the 'cpu_realtime_mask' handling slightly https://review.opendev.org/461456 | |
| 08:44:33 | openstackgerrit | Stephen Finucane proposed openstack/nova master: hardware: Rework 'get_realtime_constraint' https://review.opendev.org/714700 | |
| 08:44:33 | openstackgerrit | Stephen Finucane proposed openstack/nova master: hardware: Invert order of NUMA topology generation https://review.opendev.org/714701 | |
| 08:44:34 | openstackgerrit | Stephen Finucane proposed openstack/nova master: hardware: Remove '_numa_fit_instance_cell_with_pinning' https://review.opendev.org/714703 | |
| 08:44:34 | openstackgerrit | Stephen Finucane proposed openstack/nova master: Introduce 'pcpuset' field for InstanceNUMACell https://review.opendev.org/714658 | |
| 08:51:45 | openstackgerrit | Sylvain Bauza proposed openstack/nova master: Provide the parent pGPU when creating a new vGPU https://review.opendev.org/715489 | |
| 08:51:46 | openstackgerrit | Sylvain Bauza proposed openstack/nova master: Support different vGPU types per pGPU https://review.opendev.org/715490 | |
| 08:51:46 | openstackgerrit | Sylvain Bauza proposed openstack/nova master: DNM: Functional test with pGPUs https://review.opendev.org/717975 | |
| 08:51:53 | bauzas | gibi: updated based on your remarks ^ | |
| 08:52:41 | gibi | bauzas: ack | |
| 08:52:59 | bauzas | gibi: had to change a few unittests, you helped me to fix a few details | |
| 08:53:07 | gibi | cool | |
| 08:53:38 | bauzas | TIL init_host() is not really called when initing the host :p | |
| 08:56:34 | gibi | I alway have to look up the code to see what is the init order | |
| 09:13:44 | openstackgerrit | Stephen Finucane proposed openstack/nova master: Drop concept of '?validation' parameter https://review.opendev.org/717789 | |
| 09:13:44 | openstackgerrit | Stephen Finucane proposed openstack/nova master: api: Add microversion for extra spec validation https://review.opendev.org/708436 | |
| 09:13:45 | openstackgerrit | Stephen Finucane proposed openstack/nova master: docs: Add documentation for flavor extra specs https://review.opendev.org/710037 | |
| 09:34:01 | openstackgerrit | Guo Jingyu proposed openstack/nova master: WIP: nova supports password encrypted VNC https://review.opendev.org/622336 | |
| 09:38:05 | nightmare_unreal | hello can someone review this : https://review.opendev.org/#/c/715395/ . Not sure why Zuul verification is failing | |
| 09:38:06 | nightmare_unreal | thanks | |
| 09:42:58 | gibi | nightmare_unreal: the zuul failure seems to be unrelated. | |
| 09:43:53 | nightmare_unreal | Yeah, I was not sure why it is failing. It had merge conflict. I resolved it and since then it's -1 from zuul | |
| 09:44:09 | nightmare_unreal | I did checked before submitting, ( pre-commit, pep8 and others) | |
| 09:58:14 | nightmare_unreal | thanks for the review | |
| 09:59:23 | huaqiang | stephenfin: can you have a look for this ? url:https://review.opendev.org/#/c/714658/4/nova/objects/instance_numa.py@87 | |
| 09:59:37 | huaqiang | and do you agree with my reply for your comments? | |
| 10:01:13 | stephenfin | huaqiang: I need to dig into it, to be honest. I think you should just include a summary of what you said as a note for now. We can bring it up again if needed | |
| 10:02:04 | huaqiang | sure, I'll provide a note. | |
| 10:02:51 | lyarwood | stephenfin: not sure if you saw my ping re https://review.opendev.org/#/c/708436/ earlier? | |
| 10:02:56 | stephenfin | I did not | |
| 10:03:01 | lyarwood | stephenfin: https://review.opendev.org/#/c/708436/17 - Stupid question, so this doesn't need doc/api_samples and nova/tests/functional/api_sample_tests/api_samples/ updates as you're not changing request/responses in this microversion right? | |
| 10:03:06 | lyarwood | stephenfin: asking as I've added them in https://review.opendev.org/#/c/701430/ assuming we needed to do this regardless, if we don't I can rm -rf this from that change. | |
| 10:03:29 | lyarwood | stephenfin: just about to rebase on your change and take the next microversion btw | |
| 10:03:36 | stephenfin | lyarwood: I'd assumed not, but tbh I'm not certain. Maybe gibi would know? | |
| 10:03:56 | stephenfin | I'm not sure why you'd need them if you're not changing the request or response | |
| 10:05:25 | lyarwood | stephenfin: yeah right, I can't recall why I added these originally tbh, I assume something was failing but that clearly isn't the case in your change | |
| 10:05:28 | gibi | stephenfin, lyarwood: did query param removed too? | |
| 10:05:49 | stephenfin | gibi: In my change? Yeah | |
| 10:06:15 | gibi | I guess I mixing changes | |
| 10:06:28 | openstackgerrit | jayaditya gupta proposed openstack/nova master: Support for --overwrite flag for nova-manage placement heal_allocations command Closes-Bug:#1868997 https://review.opendev.org/715395 | |
| 10:06:36 | lyarwood | yeah that isn't part of the microversion change | |
| 10:06:49 | lyarwood | nothing changes in the request or response for that change | |
| 10:07:14 | lyarwood | I'm just going to rm -rf my samples changes from mine while I rebase | |
| 10:07:28 | gibi | so neither lyarwood nor stephenfin's change introduce any request or response or query string change | |