| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2020-06-26 | |||
| 16:40:08 | stephenfin | Placement is mandatory from OSP 13 or so | |
| 16:40:10 | stephenfin | Sorry, Newton | |
| 16:48:34 | openstackgerrit | Merged openstack/os-traits master: Fix hacking min version to 3.0.1 https://review.opendev.org/727559 | |
| 17:40:36 | openstackgerrit | Merged openstack/nova master: Update interop repo link https://review.opendev.org/735427 | |
| 19:40:45 | openstackgerrit | Merged openstack/nova master: Delete resource provider in tree by top-down traversable order https://review.opendev.org/719163 | |
| 19:52:23 | openstackgerrit | Elod Illes proposed openstack/nova master: Fix cherry-pick check for merge patch https://review.opendev.org/738271 | |
| 20:02:22 | elod | melwitt dansmith gmann : ^^^^ I've tested it locally, but will do some testing on my DNM patch in zuul as well | |
| 20:02:51 | dansmith | elod: ah, makes sense | |
| 20:02:56 | melwitt | elod: awesome, thanks for tackling it | |
| 20:03:02 | dansmith | I dunno why we're on a merge and sometimes not, but probably good to account for that | |
| 20:05:12 | melwitt | it's something about how zuul applies the patch in CI | |
| 20:06:05 | elod | if patch's parent is not the latest patch on stable/xxxxx then zuul rebases it, to see if the patch works with the latest state | |
| 20:06:23 | elod | yes, as melwitt says :) | |
| 20:13:36 | dansmith | ah, zuul not git or gerrit? | |
| 20:20:10 | elod | i think so | |
| 20:26:55 | openstackgerrit | Elod Illes proposed openstack/nova stable/stein: DNM: debug check-cherry-picks.sh on gate https://review.opendev.org/738189 | |
| 20:50:56 | openstackgerrit | Merged openstack/nova stable/train: Remove stale nested backport from InstancePCIRequests https://review.opendev.org/725932 | |
| 21:01:15 | openstackgerrit | Elod Illes proposed openstack/nova stable/stein: DNM: debug check-cherry-picks.sh on gate https://review.opendev.org/738189 | |
| 23:29:22 | openstackgerrit | Merged openstack/nova stable/ussuri: Guard against missing image cache directory https://review.opendev.org/737990 | |
| #openstack-nova - 2020-06-27 | |||
| 00:44:30 | openstackgerrit | sean mooney proposed openstack/nova master: add special function to ProviderTree and _Provier https://review.opendev.org/737594 | |
| 01:47:31 | openstackgerrit | sean mooney proposed openstack/nova master: add special function to ProviderTree and _Provier https://review.opendev.org/737594 | |
| 02:43:22 | openstackgerrit | Brin Zhang proposed openstack/nova master: Define and catch InstanceAgentNotRunning for set-password failure https://review.opendev.org/661466 | |
| 05:31:11 | openstackgerrit | Merged openstack/nova master: Fix aggregate placement sync issue https://review.opendev.org/726201 | |
| 08:57:33 | openstackgerrit | Merged openstack/nova master: hardware: Rework 'get_realtime_constraint' https://review.opendev.org/714700 | |
| #openstack-nova - 2020-06-28 | |||
| 05:03:43 | openstackgerrit | Hao Sun proposed openstack/nova master: Add checks for volume status when rebuilding https://review.opendev.org/734734 | |
| 07:21:21 | openstackgerrit | Hao Sun proposed openstack/nova master: Add checks for volume status when rebuilding https://review.opendev.org/734734 | |
| 07:32:23 | openstackgerrit | Hao Sun proposed openstack/nova master: Add checks for volume status when rebuilding https://review.opendev.org/734734 | |
| 09:08:29 | brinzhang_ | gibi: hello | |
| 09:10:03 | brinzhang_ | gibi: Do we add a judgment condition to control "continue" when instance.state=BUILDING, otherwise continue execution? https://opendev.org/openstack/nova/src/branch/master/nova/compute/resource_tracker.py#L1499-L1500 | |
| 09:11:53 | brinzhang_ | s/otherwise continue execution/ otherwise continue do below code logical | |
| 09:38:32 | brinzhang_ | gibi: I am sorry you are in weekend ^^ | |
| #openstack-nova - 2020-06-29 | |||
| 07:53:53 | bauzas | good morning Nova | |
| 08:37:36 | gibi | bauzas: good morning | |
| 08:37:52 | bauzas | gibi: any spec you'd like me to review ? | |
| 08:38:08 | gibi | bauzas: let me check the list | |
| 08:41:31 | gibi | bauzas: I only see a spec update from dansmith https://review.opendev.org/#/c/728778/ | |
| 08:41:54 | gibi | bauzas: sorry mixed up the link | |
| 08:42:06 | gibi | bauzas: this is a spec update https://review.opendev.org/#/c/735192 | |
| 08:42:30 | gibi | bauzas: https://review.opendev.org/#/c/728778/ is a re-propose | |
| 08:43:08 | bauzas | gibi: okay, I'll look | |
| 08:44:33 | gibi | thanks | |
| 08:49:24 | gibi | brinzhang_: did you hit the archiving case described in the comment https://opendev.org/openstack/nova/src/branch/master/nova/compute/resource_tracker.py#L1499-L1500 ? | |
| 09:03:58 | brinzhang_ | gibi: yes | |
| 09:04:00 | brinzhang_ | gibi: we meet the case that the RequestSpec data of instance isnot exist, but the allocations data still exist, maybe it is migration operation causes this scence, but i am not sure. | |
| 09:09:06 | openstackgerrit | Merged openstack/nova-specs master: Support volume local cache https://review.opendev.org/728778 | |
| 09:13:15 | gibi | brinzhang_: do you have a reproduction? if yes then please file a bug | |
| 09:15:03 | brinzhang_ | gibi: I am sorry, I cannot provide exact steps :( | |
| 09:15:26 | brinzhang_ | gibi: Can you think of the anomalies that will occur here? | |
| 09:15:58 | brinzhang_ | whatever you can think | |
| 09:17:34 | openstackgerrit | Merged openstack/nova-specs master: Update the testing section of the rbd-glance-multistore spec https://review.opendev.org/735192 | |
| 09:19:28 | gibi | brinzhang_: so by the fact that not even the RequestSpec of the instance exists in the db you concluded that this instance was deleted + archived. | |
| 09:20:12 | gibi | brinzhang_: if that the case then there is no way to check instance.state == BUILDING | |
| 09:22:25 | gibi | brinzhang_: also you cannot simply check the existence of the RequestSpec from the compute as RequestSpec is in the api database | |
| 09:27:52 | brinzhang_ | gibi: I discussed it with my colleagues. In the following situation, the resource cannot be cleaned up: | |
| 09:27:59 | brinzhang_ | Steps to reproduce: | |
| 09:28:07 | brinzhang_ | 3. Timed task update_available_resource() fails to synchronize resource occupancy to placement | |
| 09:28:07 | brinzhang_ | 2. Delete instance A, the placement API is abnormal and cannot synchronize resource information. RequestSPEC will store instance A resource allocation information | |
| 09:28:07 | brinzhang_ | 1. Failed to create instance A, instance A exist in cell0 | |
| 09:30:27 | gibi | brinzhang_: i think what you need is "nova-manage placement audit | |
| 09:30:28 | gibi | " | |
| 09:31:03 | gibi | brinzhang_: and later on https://specs.openstack.org/openstack/nova-specs/specs/victoria/approved/nova-audit.html | |
| 09:31:10 | gibi | when it is implemented | |
| 09:34:17 | brinzhang_ | gibi: later I will review this spec | |
| 09:36:10 | brinzhang_ | gibi: in Lin1487, we just get the instance from nova cell1 db, but cannot get from cell0(store the error instances) https://opendev.org/openstack/nova/src/branch/master/nova/compute/resource_tracker.py#L1487 | |
| 09:37:13 | gibi | I don't think you can talk to the cell0 db from a compute host | |
| 09:37:44 | brinzhang_ | I know | |
| 09:37:50 | gibi | but placement audit runs on the top therefore it can inspect every cell db | |
| 09:38:38 | brinzhang_ | if so, I think it can resolve my question | |
| 11:02:19 | openstackgerrit | Merged openstack/nova stable/stein: Make greande jobs n-v for EM and oldest stable https://review.opendev.org/737332 | |
| 11:05:13 | openstackgerrit | Johannes Kulik proposed openstack/nova master: vmware: Handle exception in destroy with attached volumes https://review.opendev.org/738416 | |
| 12:00:48 | openstackgerrit | Wenping Song proposed openstack/nova master: Add instance project_id for cyborg arq https://review.opendev.org/738428 | |
| 12:37:02 | openstackgerrit | renminmin proposed openstack/nova master: SRIOV: Fix instance with macvtap port live migration failed https://review.opendev.org/738432 | |
| 13:06:34 | openstackgerrit | Stephen Finucane proposed openstack/nova master: Update keypairs in saving an instance object https://review.opendev.org/683043 | |
| 13:13:58 | stephenfin | gibi: I've reproposed the use-pcpu-and-vcpu-in-one-instance BP work for a runway. What was approved has merged (thanks for the reviews!), and now there are four patches at the top good to go and huaquiang is reworking the remainder at the moment | |
| 13:14:15 | stephenfin | Will ping alex_xu when he's around and ask for reviews | |
| 13:14:19 | gibi | stephenfin: cool | |
| 13:14:27 | gibi | stephenfin: I will move it to the last slot then | |
| 13:14:36 | stephenfin | \o/ | |
| 13:15:54 | stephenfin | Also, I took a look through Gerrit on Friday trying to find more stuff to add. No luck. melwitt and johnthetubaguy both have stuff that could conceivably go in (nova-audit and oslo.limit integration, respectively) but they both have open comments :( | |
| 13:16:37 | johnthetubaguy | snowed under at the moment sadly | |
| 13:16:42 | stephenfin | johnthetubaguy: I'll review if/when you do | |
| 13:16:50 | stephenfin | The provider.yaml file work looks stalled as well :( | |
| 13:17:06 | gibi | stephenfin: thanks for looking. | |
| 13:17:13 | johnthetubaguy | stephenfin: thanks, I will try get to that ASAP | |
| 13:17:16 | gibi | yeah I noticed that provider.yaml is stopped | |
| 13:17:26 | gibi | stephenfin: what is the resolution of https://review.opendev.org/#/c/737026/2/nova/virt/hardware.py@1708 ? | |
| 13:17:35 | johnthetubaguy | I wanted to reboot that, but right now don't have bandwidth | |
| 13:17:52 | johnthetubaguy | (on the OpenDev call into) | |
| 13:19:13 | stephenfin | I mean, I could take it over but I don't know who'll be available to review. I've already got the vTPM stuff and mixed CPU types work against me too | |
| 13:19:36 | stephenfin | gibi: Oh, I've a patch for that somewhere. Let me try find it | |
| 13:19:59 | stephenfin | gibi: nvm, pushed it. https://review.opendev.org/#/c/737026/3/nova/virt/libvirt/driver.py | |
| 13:20:48 | gibi | stephenfin: thanks, will look | |
| 13:23:04 | stephenfin | gibi: Sweet. Last thing, it's not ready but I've started work on adding type hints for objects. It's kind of ugly and will probably need a mypy plugin to make it not ugly, but it should fix the gap you were seeing with types not being checked for stuff from nova.objects https://review.opendev.org/#/c/738020/2 | |
| 13:23:25 | stephenfin | To be clear, obviously not ready for review yet | |
| 13:24:42 | gibi | stephenfin: interesting. so basically that mypy plugin could look at the ovo field declaration directly and generate the type hints on the fly for ovos | |
| 13:25:52 | stephenfin | gibi: Yup. It seems that's what they do for Django, which has similar magic behavior for things like models https://github.com/typeddjango/django-stubs/tree/master/mypy_django_plugin | |
| 13:26:04 | stephenfin | I just need to figure out how mypy works internally first :( | |
| 13:27:21 | gibi | just opened the code of that plugin and got scared by the internals :) | |
| 13:27:51 | gibi | this feel heavy lifting for me right now | |
| 13:28:29 | stephenfin | Yeah, it's pretty low priority. If I can get the ugly but effective thing working for now, I'll do that | |