| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2026-07-15 | |||
| 10:25:08 | gibi | not right now | |
| 10:25:12 | gibi | I will ping you later | |
| 10:25:36 | sean-k-mooney | ok cool my idea is pretty simple and i tought something placement already idd but apprently not | |
| 10:26:40 | sean-k-mooney | effectivly we shoudl be flattenting both the provider tree and aagreating the counts of all free resouce classe in the tree and flatteing the request and droping any root provider where the recquest exceed any RC in the flatttened tree | |
| 10:27:05 | sean-k-mooney | that tells use that regardless of the toplogy that root provider cant supprot the request | |
| 10:27:26 | sean-k-mooney | if that flanned check passes then and only then do we process it in python | |
| 10:28:49 | sean-k-mooney | i think that can be expressed as a select root_uuid, rc, capcity group by root_uuid or similar to get the flattend trees | |
| 10:29:28 | sean-k-mooney | we can trivially flaten the request in python to constuct a where clause | |
| 10:30:34 | sean-k-mooney | and basiclly seleect out the root_uuids form that where the flattend free capsity is greater then then the requested ammount | |
| 10:31:18 | sean-k-mooney | obviously its more complicated then that as we need to compute the free capsty not just the total and genericly we need to account for the allocation ratio as well | |
| 10:31:35 | sean-k-mooney | so conceptually simple but expressign that in placment could be tricky | |
| 10:32:54 | sean-k-mooney | if we can implemnt that optimistation we will never do the expnesive combinitorial decent if it cant fit | |
| 10:38:10 | sean-k-mooney | when i type it out it feels more complicated the it is in my head :) | |
| 12:15:14 | opendevreview | Kamil Sambor proposed openstack/nova master: libvirt: retry get_guest() on InstanceNotFound during post-copy VIF reattach https://review.opendev.org/c/openstack/nova/+/997380 | |
| 12:15:52 | sean-k-mooney | gibi: this shoudl not require a spec but if i was to have ai formualt my idea into spec format it woudl looks somethign like https://paste.opendev.org/show/bVKIYH6Om0svZbMwlCbg/ :) | |
| 14:10:33 | gibi | sean-k-mooney: I feel that this flatten and do a pre-flight check make sense. It solve the current problem. There is probably more problems. I.e. when this pre-flight check passes but the normal expansion still finds, in the long way, that the host is not having enough resources. | |
| 14:10:54 | gibi | this is a wack-a-mole as we knew since the first perf bug was triaged :) | |
| 14:11:05 | sean-k-mooney | https://github.com/SeanMooney/tachyon/blob/master/reports/almost-fitting-root-bug-report.md | |
| 14:11:32 | sean-k-mooney | im not saying my ai generate placment replacemnt is a valid solution | |
| 14:11:52 | sean-k-mooney | but i quickly tested the prefilterign there and it seam to resolve the currnt issue | |
| 14:13:11 | gibi | sean-k-mooney: if you would like to take the bug (once I file it upstream) I'm happy to pass it on :) | |
| 14:13:32 | sean-k-mooney | gibi: i proably have to many other things on my plate | |
| 14:13:50 | sean-k-mooney | if you have time to do it properly in placment then that might be more efficent | |
| 14:14:05 | sean-k-mooney | i just left ai run on the idea to validate the concept | |
| 14:14:16 | sean-k-mooney | im interested in this because cyborg makes this problem worse | |
| 14:14:31 | sean-k-mooney | becaus it report every VF as its onw rp today | |
| 14:15:13 | sean-k-mooney | that on my todo list to fix next cycle but i am generally conferedn/interested in the placment performance | |
| 14:16:16 | gibi | yeah. right now I have no time to read the 1000 lines proposal. If this bug gets into our next sprint then I will load context on the shape of the placement code and then it would make sense to me to read you proposal to see if that fits in to the code | |
| 14:16:21 | gibi | if/how | |
| 14:38:10 | opendevreview | Kamil Sambor proposed openstack/nova master: Replace StaticallyDelayingWrapper with futurist DelayingExecutor https://review.opendev.org/c/openstack/nova/+/997410 | |
| 14:47:01 | gibi | fyi, nova-alt-configuration is green again, so the gate is unblocked | |
| 15:28:10 | gibi | nova upstream bug triage call https://meet.google.com/zjr-rxus-hzj | |
| 15:28:26 | sean-k-mooney | https://meetings.opendev.org/#Nova_Team_Meeting is there but we shoudl proably add that meetign sepreatly | |
| 15:28:29 | gibi | (I expect it to be short as Uggla is on PTO) | |
| 17:07:04 | opendevreview | Balazs Gibizer proposed openstack/placement master: Reproduce bug 2160721 https://review.opendev.org/c/openstack/placement/+/997362 | |
| 18:30:57 | opendevreview | Merged openstack/nova master: Remove unused method for image v1 API https://review.opendev.org/c/openstack/nova/+/995532 | |
| 18:33:36 | opendevreview | Merged openstack/nova master: Remove unused UpdateGlanceImage https://review.opendev.org/c/openstack/nova/+/995726 | |
| 19:31:43 | opendevreview | Merged openstack/nova master: Accept hw_firmware_stateless=True for BIOS firmware type https://review.opendev.org/c/openstack/nova/+/995835 | |
| 20:42:10 | opendevreview | Merged openstack/nova stable/2025.1: Make QEMU_IMG_LIMITS process limits configurable https://review.opendev.org/c/openstack/nova/+/997029 | |
| 21:08:33 | opendevreview | Merged openstack/os-vif master: ovs: Preserve versioned field compatibility https://review.opendev.org/c/openstack/os-vif/+/997297 | |
| 21:08:35 | opendevreview | Merged openstack/os-vif master: ovs: Pass ``multi_queue`` inside ``ifr`` dict for tuntap devices https://review.opendev.org/c/openstack/os-vif/+/995933 | |
| 21:34:36 | opendevreview | Merged openstack/nova master: Add vTPM startup validation reproducer https://review.opendev.org/c/openstack/nova/+/990551 | |
| 22:22:28 | opendevreview | Merged openstack/nova master: Delegate startup instance processing to drivers https://review.opendev.org/c/openstack/nova/+/990552 | |
| #openstack-nova - 2026-07-16 | |||
| 02:54:47 | opendevreview | Takashi Kajinami proposed openstack/nova master: compute: Validate SEV/SEV-ES support at start up https://review.opendev.org/c/openstack/nova/+/994342 | |
| 02:54:48 | opendevreview | Takashi Kajinami proposed openstack/nova master: libvirt: Ignore SEV-ES when SEV-SNP is detected https://review.opendev.org/c/openstack/nova/+/994343 | |
| 02:55:26 | opendevreview | Takashi Kajinami proposed openstack/nova master: libvirt: Detect AMD SEV-SNP support https://review.opendev.org/c/openstack/nova/+/994764 | |
| 02:55:37 | opendevreview | Takashi Kajinami proposed openstack/nova master: libvirt: Support launchSecurity element for SEV-SNP https://review.opendev.org/c/openstack/nova/+/994920 | |
| 02:56:19 | opendevreview | Takashi Kajinami proposed openstack/nova master: libvirt: Omit optional fields for SEV/SEV-ES https://review.opendev.org/c/openstack/nova/+/994921 | |
| 02:56:27 | opendevreview | Takashi Kajinami proposed openstack/nova master: libvirt: Drop sev capability check in launch instance flow https://review.opendev.org/c/openstack/nova/+/994929 | |
| 02:56:35 | opendevreview | Takashi Kajinami proposed openstack/nova master: AMD SEV: omit iommu='on' for virtio devices https://review.opendev.org/c/openstack/nova/+/909635 | |
| 03:00:55 | opendevreview | Takashi Kajinami proposed openstack/nova master: libvirt: Launch instances with SEV-SNP memory encryption https://review.opendev.org/c/openstack/nova/+/994930 | |
| 03:02:20 | opendevreview | Takashi Kajinami proposed openstack/nova master: compute: Validate SEV/SEV-ES support at start up https://review.opendev.org/c/openstack/nova/+/994342 | |
| 03:02:25 | opendevreview | Takashi Kajinami proposed openstack/nova master: libvirt: Ignore SEV-ES when SEV-SNP is detected https://review.opendev.org/c/openstack/nova/+/994343 | |
| 03:03:04 | opendevreview | Takashi Kajinami proposed openstack/nova master: libvirt: Detect AMD SEV-SNP support https://review.opendev.org/c/openstack/nova/+/994764 | |
| 03:03:23 | opendevreview | Takashi Kajinami proposed openstack/nova master: libvirt: Support launchSecurity element for SEV-SNP https://review.opendev.org/c/openstack/nova/+/994920 | |
| 03:03:33 | opendevreview | Takashi Kajinami proposed openstack/nova master: libvirt: Omit optional fields for SEV/SEV-ES https://review.opendev.org/c/openstack/nova/+/994921 | |
| 03:03:47 | opendevreview | Takashi Kajinami proposed openstack/nova master: libvirt: Drop sev capability check in launch instance flow https://review.opendev.org/c/openstack/nova/+/994929 | |
| 03:04:02 | opendevreview | Takashi Kajinami proposed openstack/nova master: AMD SEV: omit iommu='on' for virtio devices https://review.opendev.org/c/openstack/nova/+/909635 | |
| 03:05:17 | opendevreview | Takashi Kajinami proposed openstack/nova master: libvirt: Launch instances with SEV-SNP memory encryption https://review.opendev.org/c/openstack/nova/+/994930 | |
| 03:05:21 | opendevreview | Takashi Kajinami proposed openstack/nova master: Update documentations for AMD SEV-SNP support https://review.opendev.org/c/openstack/nova/+/995090 | |
| 03:05:28 | opendevreview | Takashi Kajinami proposed openstack/nova master: Remove [libvirt] num_memory_encrypted_guests https://review.opendev.org/c/openstack/nova/+/995120 | |
| 03:05:35 | opendevreview | Takashi Kajinami proposed openstack/nova master: Use domain capabilities to detect SEV-SNP support https://review.opendev.org/c/openstack/nova/+/996316 | |
| 03:44:24 | opendevreview | Takashi Kajinami proposed openstack/nova master: Remove [DEFAULT] vcpu_pin_set https://review.opendev.org/c/openstack/nova/+/996016 | |
| 08:27:18 | opendevreview | Takashi Kajinami proposed openstack/nova master: Remove [workarounds] disable_libvirt_livesnapshot https://review.opendev.org/c/openstack/nova/+/997492 | |
| 08:27:28 | opendevreview | Merged openstack/nova stable/2025.1: Fix mutating global config in websocket proxy https://review.opendev.org/c/openstack/nova/+/995958 | |
| 09:38:11 | gibi | fyi tox cover is timing out a lot recently in test_schemas I filed https://bugs.launchpad.net/nova/+bug/2160901 | |
| 09:48:24 | sean-k-mooney | gibi: i know there was some slownees with those in the past | |
| 09:48:33 | sean-k-mooney | but i dont remember exactly why | |
| 09:48:43 | sean-k-mooney | i just recall mentioning it to stephen | |
| 09:53:33 | sean-k-mooney | gibi:lol we were looking at the same review | |
| 10:06:19 | tkajinam | :-) | |
| 10:15:46 | opendevreview | Balazs Gibizer proposed openstack/nova master: [cover]Disable slow test_schemas test case https://review.opendev.org/c/openstack/nova/+/997503 | |
| 10:16:18 | gibi | I found no better way given our current cirumstances (py314+threading mode fails) | |
| 10:17:40 | sean-k-mooney | gibi: i think that is fien for now | |
| 10:17:52 | sean-k-mooney | gibi: i think the answer might be to split it into multipel tests | |
| 10:20:23 | sean-k-mooney | currently its looping over all our api contoler and then looping over all the route and then testing every schema in one masive test | |
| 10:20:27 | sean-k-mooney | https://github.com/openstack/nova/blob/master/nova/tests/unit/api/openstack/compute/test_schemas.py#L69-L123 | |
| 10:20:56 | sean-k-mooney | gibi: the fix i think is to use subtests or ddt to split up the test per contoler or route | |
| 10:21:12 | sean-k-mooney | so that each schema baseiclly get mapped to its own test | |
| 10:21:52 | sean-k-mooney | that way we wont hit the per test timeout | |
| 10:26:50 | sean-k-mooney | tkajinam: speaking of masive tests that should be broken up you sev functional test have lots of repatead code that should be in shared helpers and the test shoudl be broken up in to small tests | |
| 10:26:58 | sean-k-mooney | tkajinam: but im not going to block on that for now | |
| 10:27:10 | sean-k-mooney | but we shoudl clean that up in a dedicated followup | |
| 10:29:58 | gibi | sean-k-mooney: yeah that is also an option | |
| 10:30:38 | sean-k-mooney | im also fine with not runing that in coverage | |
| 10:31:17 | sean-k-mooney | in this particalar case im not sure that the coverage report will matirally be any better/worse by runing that one test | |
| 10:31:34 | sean-k-mooney | i just think it might be nice to do that break anyway for the unit tests | |
| 10:31:45 | sean-k-mooney | its extra slow in cover because of the profiler overhead | |
| 10:31:59 | sean-k-mooney | but its still the slowest or one of the slowest unit test today | |
| 10:47:31 | sean-k-mooney | tkajinam: i didnt know that sev-snp actully needed sateless firmware? | |
| 10:48:57 | sean-k-mooney | tkajinam: is that a qemu limiation or amd one | |
| 10:51:24 | sean-k-mooney | ok ya that called ou tin the spec i just missed it https://specs.openstack.org/openstack/nova-specs/specs/2026.2/approved/amd-sev-snp-libvirt-support.html#:~:text=Also%20amd%2Dsev%2Dsnp%20requires%20stateless%20firmware%2C%20so%20explicit%20hw%5Ffirmware%5Fstateless%3DTrue%20image%20property%20setting%20is%20also%20required | |
| 10:59:50 | tkajinam | sean-k-mooney, that's the limitation enforced by qemu, and is also needed so that libvirt can pick up the appropriate ovmf file | |
| 11:00:50 | sean-k-mooney | ack | |
| 11:00:53 | tkajinam | sean-k-mooney, yeah I agree the current tests are quite redundant. I'll try refactoring these later | |
| 11:01:13 | sean-k-mooney | im basiclly +2w on most of your serise now ill add +w to the bottom patch when im done | |
| 11:01:36 | tkajinam | thanks ! | |
| 11:01:51 | sean-k-mooney | im on the last two clean up patchs so we shoudl get this into the ci later today | |
| 11:02:05 | tkajinam | :-D | |
| 11:02:35 | sean-k-mooney | i have left some comments along the way but nothing was worth holding or respinign the patches for | |