| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2025-05-13 | |||
| 10:37:09 | sean-k-mooney | we didnt entirly forget we added the help text as part of the config option but ya the improvment looks reasonable | |
| 10:39:49 | sean-k-mooney | pas-ha[m]: ritght it basically create enouch VF for the smallest vgpu type it can support | |
| 10:40:23 | sean-k-mooney | wich i get why they do that but its also not a great ux | |
| 10:45:25 | hertrste | Hi. I would like to ask to get some further feedback on https://review.opendev.org/c/openstack/nova-specs/+/945549. Tank you! | |
| 10:52:31 | gokhan | pas-ha[m], yes it is correct. L40S created 32 vf. I select nvidia-1153 as l40s-12q and nova/placement created 4 resource providers beacuse max instances are 4 in l40s-12. If I change from l40s-12q to l40s-1q, nova/placement creates 32 new resource providers. It maps every pci slot to one vm. It seem it is working as expected. | |
| 10:53:10 | gokhan | https://paste.openstack.org/show/bdL9DHdT9pnIbGEFrbEO/ | |
| 10:54:49 | gokhan | but sometimes when I try to change types, nova is deleting previous resource providers and not create new ones | |
| 10:58:51 | gokhan | pas-ha[m], host vgpu mode is SR-IOV https://paste.openstack.org/show/bg5WeH9FWvpWhenTJA70/ | |
| 11:00:03 | gokhan | I wonder can we create different vgpu types per VFs? | |
| 11:00:07 | pas-ha[m] | gokhan: the problem is that very rarely you can really create that many resources, which leads to overreporting in placement. Nova does not know that of those 32 resource providers that each can handle 1 vgpu resource, only say 12 can really be occupied (because you set Nova to create that vgpu type that can fit only 12), mdev initially still reports that every of those 32 pci devs can hold 1 vgpu of your type... for that recent nova | |
| 11:00:08 | pas-ha[m] | versions have 'max instances' option in the [mdev_...] section, to limit the amount of resource providers created. | |
| 11:00:08 | pas-ha[m] | https://docs.nvidia.com/vgpu/17.0/grid-vgpu-user-guide/index.html#virtual-gpu-types-grid-reference - here is a comprehensive list of all the vgpu types _and their combinations_ (for the heterogeneous vGPU mode which your card supports, and which I guess is enabled when you put the card into the SR-IOV mode). | |
| 11:03:06 | pas-ha[m] | again, check mdevctl types, it will show you 'available instances' for each mdev type on each PCI device - and mind that this number will change (everywhere, on all the PCI VFs) as you create new mdevs. | |
| 11:05:03 | opendevreview | Stephen Finucane proposed openstack/nova master: api: Adjust validation helpers for a single-method future https://review.opendev.org/c/openstack/nova/+/936365 | |
| 11:05:03 | opendevreview | Stephen Finucane proposed openstack/nova master: api: Stop using wsgi.Controller.api_version to switch between API versions https://review.opendev.org/c/openstack/nova/+/936366 | |
| 11:05:04 | opendevreview | Stephen Finucane proposed openstack/nova master: api: Only check minimum API version https://review.opendev.org/c/openstack/nova/+/940325 | |
| 11:05:04 | opendevreview | Stephen Finucane proposed openstack/nova master: api: Add new, simpler api_version decorator https://review.opendev.org/c/openstack/nova/+/936367 | |
| 11:05:05 | opendevreview | Stephen Finucane proposed openstack/nova master: api: Only run format checks on strings https://review.opendev.org/c/openstack/nova/+/936368 | |
| 11:05:05 | opendevreview | Stephen Finucane proposed openstack/nova master: api: Simplify parameter types https://review.opendev.org/c/openstack/nova/+/936369 | |
| 11:05:06 | opendevreview | Stephen Finucane proposed openstack/nova master: tests: Ensure all APIs have a response body schema https://review.opendev.org/c/openstack/nova/+/924598 | |
| 11:05:07 | opendevreview | Stephen Finucane proposed openstack/nova master: doc: Add missing API samples https://review.opendev.org/c/openstack/nova/+/937012 | |
| 11:05:07 | opendevreview | Stephen Finucane proposed openstack/nova master: api: Add response body schemas for hosts APIs https://review.opendev.org/c/openstack/nova/+/937047 | |
| 11:05:09 | opendevreview | Stephen Finucane proposed openstack/nova master: api: Add response body schemas for instance actions https://review.opendev.org/c/openstack/nova/+/937048 | |
| 11:05:09 | opendevreview | Stephen Finucane proposed openstack/nova master: api: Add response body schemas for hypervisors APIs (1/3) https://review.opendev.org/c/openstack/nova/+/937245 | |
| 11:05:11 | opendevreview | Stephen Finucane proposed openstack/nova master: api: Add response body schemas for hypervisors APIs (2/3) https://review.opendev.org/c/openstack/nova/+/937246 | |
| 11:05:11 | opendevreview | Stephen Finucane proposed openstack/nova master: api: Add response body schemas for hypervisors APIs (3/3) https://review.opendev.org/c/openstack/nova/+/937247 | |
| 11:05:13 | opendevreview | Stephen Finucane proposed openstack/nova master: api: Add response body schemas for server IPs APIs https://review.opendev.org/c/openstack/nova/+/937608 | |
| 11:05:13 | opendevreview | Stephen Finucane proposed openstack/nova master: api: Add response body schemas for remote consoles https://review.opendev.org/c/openstack/nova/+/940114 | |
| 11:05:15 | opendevreview | Stephen Finucane proposed openstack/nova master: api: Add response body schemas for keypairs APIs https://review.opendev.org/c/openstack/nova/+/940115 | |
| 11:05:15 | opendevreview | Stephen Finucane proposed openstack/nova master: api: Add response body schemas for image metadata APIs https://review.opendev.org/c/openstack/nova/+/940299 | |
| 11:05:17 | opendevreview | Stephen Finucane proposed openstack/nova master: api: Add response body schemas for server topology API https://review.opendev.org/c/openstack/nova/+/940300 | |
| 11:05:17 | opendevreview | Stephen Finucane proposed openstack/nova master: api: Add response body schemas for server diagnostics API https://review.opendev.org/c/openstack/nova/+/940506 | |
| 11:05:19 | opendevreview | Stephen Finucane proposed openstack/nova master: api: Add response body schemas for server external events API https://review.opendev.org/c/openstack/nova/+/940508 | |
| 11:05:19 | opendevreview | Stephen Finucane proposed openstack/nova master: api: Add response body schemas for server password APIs https://review.opendev.org/c/openstack/nova/+/945736 | |
| 11:12:12 | gokhan | thanks for great explanation pas-ha[m] I got it :) | |
| 11:52:39 | opendevreview | Stephen Finucane proposed openstack/nova stable/2024.1: add pyproject.toml to support pip 23.1 https://review.opendev.org/c/openstack/nova/+/949617 | |
| 12:07:33 | gibi | bauzas: is it just me or nova-next is failing with mdev stuff https://zuul.opendev.org/t/openstack/builds?job_name=nova-next&skip=0 | |
| 12:07:54 | gibi | 3237 Async compile_mdev_samples:45784]: finished compile_mdev_samples with result 1 in 37 seconds | |
| 12:08:40 | gibi | https://zuul.opendev.org/t/openstack/build/b8c32faa528e4bddbfb027f43f1f184c/log/job-output.txt#21834 | |
| 12:08:58 | sean-k-mooney | gibi: perhaps ocationally but it seam sto be passign for the most part unless it just started today? | |
| 12:10:47 | gibi | filed bug https://bugs.launchpad.net/nova/+bug/2110545 | |
| 12:10:54 | gibi | it started today | |
| 12:11:30 | sean-k-mooney | https://zuul.opendev.org/t/openstack/build/f28f307aa79646f6a9a54965ae3592e6/log/controller/logs/devstacklog.txt#19105 | |
| 12:11:46 | sean-k-mooney | so it looks like the kernel we downloaded is gettign a 404 | |
| 12:12:18 | gibi | yeah it seems | |
| 12:12:45 | sean-k-mooney | we may need to check how we are selectign that package | |
| 12:12:54 | sean-k-mooney | this is either an infra issue where the package should be avioable | |
| 12:13:10 | sean-k-mooney | or perhaps we hard coded it instead fo gettign the one for the currently installed kernel | |
| 12:13:42 | sean-k-mooney | https://github.com/openstack/nova/blob/master/devstack/lib/mdev_samples#L40 | |
| 12:13:58 | sean-k-mooney | sorry https://github.com/openstack/nova/blob/master/devstack/lib/mdev_samples#L30-L32 | |
| 12:14:28 | sean-k-mooney | or perhaps https://github.com/openstack/nova/blob/master/devstack/lib/mdev_samples#L36 is what actully failing | |
| 12:14:42 | sean-k-mooney | in any case we are using uname to get the runnign kernel | |
| 12:15:04 | sean-k-mooney | so this is proably an issue with our mirros or an issue with the ubuntu repos | |
| 12:15:41 | sean-k-mooney | we could temproally disabele the compile and reenable it in bauzas's serise where he is actully usign the compiled module | |
| 12:16:07 | bauzas | mmmm, this is just for the moment compiling the kernel indeed | |
| 12:16:11 | sean-k-mooney | right now we are building the mdev support but not actully useing it to test with mdevs becuase that patch is not merged yet | |
| 12:16:18 | bauzas | yeah exacylu | |
| 12:17:20 | sean-k-mooney | ok so the tar is gone form the ubuntu archive... | |
| 12:18:04 | gibi | I vote for disabling this compliation and then search for a permanent solution | |
| 12:18:12 | sean-k-mooney | +1 | |
| 12:18:26 | bauzas | fine by me | |
| 12:19:21 | bauzas | I need to look how to get the kernel examples then | |
| 12:19:23 | sean-k-mooney | it looks like the drop all the kernle tars from that archive | |
| 12:19:24 | gibi | I can push a patch... | |
| 12:19:28 | sean-k-mooney | so they might have moved them | |
| 12:19:32 | bauzas | gibi: please do | |
| 12:19:34 | gibi | OK | |
| 12:19:34 | gibi | on it | |
| 12:20:03 | sean-k-mooney | their is nolonger any "linux_*" packages | |
| 12:20:32 | sean-k-mooney | actuly there are no tars there anymroe so maybe that location now only has the debs and the source is at a new location | |
| 12:20:40 | sean-k-mooney | so we might need to update the repos we nebale | |
| 12:20:41 | bauzas | that's weird | |
| 12:21:08 | sean-k-mooney | we are enableing the soruce repos https://github.com/openstack/nova/blob/master/devstack/lib/mdev_samples#L21-L23 | |
| 12:21:38 | sean-k-mooney | so that shoudl be all that is required but perhaps we can reach out to some of the cannocal folks to understand what to do | |
| 12:22:35 | sean-k-mooney | perhaps we should be using https://launchpad.net/ubuntu/noble/+package/linux-source | |
| 12:22:45 | sean-k-mooney | instead of linux-image-unsigned-$kver | |
| 12:24:02 | sean-k-mooney | hum https://canonical-kteam-docs.readthedocs-hosted.com/en/public/how-to/build-kernel.html#get-kernel-source-for-version-installed-on-build-machine at least suggest we are doing it properly | |
| 12:24:46 | opendevreview | Balazs Gibizer proposed openstack/nova master: Disable NOVA_COMPILE_MDEV_SAMPLES to unblock CI https://review.opendev.org/c/openstack/nova/+/949622 | |
| 12:24:54 | gibi | fix ^^ | |
| 13:00:13 | jlejeune | ||
| 13:15:24 | opendevreview | Elod Illes proposed openstack/nova master: [tool] Fix backport validator for non-SLURP https://review.opendev.org/c/openstack/nova/+/949628 | |
| 13:25:18 | opendevreview | Antony Messerli proposed openstack/nova master: docs: Corrects a typo on a placement audit command https://review.opendev.org/c/openstack/nova/+/922850 | |
| 14:23:34 | opendevreview | Takashi Natsume proposed openstack/nova master: Update contributor guide for 2025.2 Flamingo https://review.opendev.org/c/openstack/nova/+/944603 | |
| 14:35:43 | dansmith | gibi: argh, that's frustrating | |
| 14:36:17 | dansmith | if the kernel sources for the kernel we're running are ever not present, that should mean our image is (way) too old | |
| 14:36:40 | dansmith | I wonder if they moved/hid them for the same reason RHEL did a while back.. that would really suck | |
| 14:45:41 | gibi | dansmith: I did not dig but I think Sean did more | |
| 14:46:03 | opendevreview | Konrad Gube proposed openstack/nova master: Use Cinder's os-extend_volume_completion volume action. https://review.opendev.org/c/openstack/nova/+/873560 | |
| 15:10:24 | Uggla | Nova meeting in ~50mn | |
| 15:56:53 | Uggla | Meeting in 5mn | |
| 16:01:07 | Uggla | #startmeeting nova | |
| 16:01:07 | opendevmeet | Meeting started Tue May 13 16:01:07 2025 UTC and is due to finish in 60 minutes. The chair is Uggla. Information about MeetBot at http://wiki.debian.org/MeetBot. | |
| 16:01:07 | opendevmeet | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | |
| 16:01:07 | opendevmeet | The meeting name has been set to 'nova' | |
| 16:01:14 | elodilles | o/ | |
| 16:01:18 | fwiesel | o/ | |
| 16:01:25 | Uggla | Hello everyone | |
| 16:01:30 | masahito | o/ | |
| 16:01:36 | gibi | o/ | |
| 16:01:39 | Uggla | awaiting a moment for people to join. | |