| lajoskatona | Uggla: Hi, our downstream team recently run into this (or to a very similar issue): https://redhat.atlassian.net/browse/RHEL-48801 (thanks gibi for pointing to it) | 11:59 |
|---|---|---|
| lajoskatona | Uggla: I am not a nova->libvirt->qemu chain specialist but with some reading I started to believe that there is a fix: https://gitlab.com/qemu-project/qemu/-/commit/3214bec13d8d4c40f707d21d8350d04e4123ae97 in qemu | 12:00 |
| lajoskatona | at least Qemu folks did something they believe is related to the paused state issues (fix for assertions during paused migrations; rework of late-block-activate logic (#2395, #686)) | 12:01 |
| lajoskatona | It was released in Qemu v10.0.0, but currently Nova runs between 6.2.0 and 8.2.2 | 12:02 |
| lajoskatona | Uggla: after the long story: wehat do you think about opening a Nova bug, "fix it" with a doc mentioning that it is a qemu bug, and when qemu 10.0.0 is in the nova support matrix, remove the warning from the doc? | 12:04 |
| lajoskatona | I checked this support page: https://docs.openstack.org/nova/latest/reference/libvirt-distro-support-matrix.html | 12:04 |
| *** polyedre_ is now known as polyedre | 12:13 | |
| opendevreview | Lucas HENRY proposed openstack/nova master: Shut down executors after the RPC servers drain https://review.opendev.org/c/openstack/nova/+/1009164 | 12:24 |
| na-te | hi, could someone take a look at https://review.opendev.org/c/openstack/nova/+/1009444 (libvirt rescue fix), CI is green | 12:29 |
| na-te | * hi, could someone take a look at https://review.opendev.org/c/openstack/nova/+/1009444 (libvirt rescue fix), CI is green, thanks in advance ! | 12:36 |
| *** ykarel__ is now known as ykarel | 13:08 | |
| Uggla | Hi lajoskatona, reading back but that sounds good. | 13:13 |
| lajoskatona | Uggla: so shall I open a bug to track it in Nova launchpad? as I see with ubuntu26 (https://review.opendev.org/c/openstack/governance/+/996925/1/reference/runtimes/2027.1.rst ) we will have qemu 10+ for Indri, but I also have to test it myselv to see if the find in gitlog is really fixing the issue with pause | 13:27 |
| Uggla | lajoskatona, yes please. | 13:36 |
| lajoskatona | Uggla: ack | 13:36 |
| opendevreview | sean mooney proposed openstack/os-vif stable/2026.1: ovs: Pass ``multi_queue`` inside ``ifr`` dict for tuntap devices https://review.opendev.org/c/openstack/os-vif/+/1009778 | 14:16 |
| dansmith | lajoskatona: do you know for sure the fix won't be backported by distros? | 14:19 |
| dansmith | it seems weird to me for us to maintain a bug pointer like that, unless it's a fundamental pitfall | 14:19 |
| lajoskatona | dansmith: good question, I am not sure, for example in my test ubuntu24.04 it seems it is not backported. | 15:25 |
| dansmith | point being we could mirror all the potential bugs in our docs, but it seems like a maintenance issue for us and I'm not sure why this one would warrant it necessarily | 15:26 |
| lajoskatona | dansmith: I wait anyway as on the above mentioned devstack I was not able to reproduce, so I have to check how our downstream team run into it and understand better the bugzilla also | 15:27 |
| dansmith | we previously say things like "multifd with postcopy is not supported in <10.0" but that's sort of a thing we are telling them not to turn on in nova until it's supported.. a bug like this could be different across distros | 15:27 |
| sean-k-mooney | i guess they have left but the other thing to note is that nova does not have a max libvir/qemu or a supproted range | 17:32 |
| sean-k-mooney | we only have a min and next min | 17:32 |
| sean-k-mooney | where next min is just telling distro and oepratroat what we wil move too for plannign reaons | 17:32 |
| sean-k-mooney | so qemu 10 is supproted by nova today | 17:33 |
| sean-k-mooney | and it was since it was relesaed and shiped in a supproted distro | 17:33 |
| sean-k-mooney | lajoskatona ^ if you read scroll back later vai the logs. basiclly the idea that "Nova supprot between 6.2.0 and 8.2.2" is invliad | 17:36 |
| sean-k-mooney | so tl;dr i was invovled in digagnoing that oring issue and advicsed agaisn opening a nova bug for this or mentioning it in our docs as this was a qemu issue nto a nova one and i dont think we shoudl include it in our docs | 17:43 |
| eduardomorais[m] | Hey everyone, I'm having an issue with my OpenStack. I'm using kolla-ansible 2024.1, and when I try to create a VM through the Horizon dashboard, it gets stuck in an infinite 'spawning' state. However, when I create it via the CLI, it works perfectly fine. Has anyone seen anything like this before? | 17:43 |
| sean-k-mooney | are they both created as the saem user with the same parmaters. i have had issue where for exmple the default in horizon diveraged form nova i..e using BFV by defualt or simialr | 17:46 |
| sean-k-mooney | in general no i the request is the saem there shoudl be 0 delta as nova does not actully know if its coming form horizon or not | 17:47 |
| eduardomorais[m] | same user, flavor, network and image | 17:47 |
| sean-k-mooney | and both are using boot form image using nova provisioned storage? | 17:48 |
| sean-k-mooney | as i said there is a config option to make booting form a cinder volume created form a glance image the default in horizon and i think that is the default at least in devstack when cinder is deployed | 17:49 |
| eduardomorais[m] | I created both of them using ephemeral disks. When I try to create them through the dashboard using a volume, Cinder times out; however, it is working for everything else and there are no error logs, which makes me believe it is not the problem. | 17:51 |
| sean-k-mooney | os that the only casie i have had that happen thace comes to mind. horizong cna have some other defult liek cofnig driver or inlcuding a ssh key by default but neither shoudl cause wahat you describe | 17:53 |
| sean-k-mooney | eduardomorais[m]: i woudl propably set the nova api/service into debug mode | 17:53 |
| sean-k-mooney | and then create a vm with horizon then find the request id and see if you can compare that to a cli request | 17:53 |
| eduardomorais[m] | ok, i will try that, thx | 17:54 |
| sean-k-mooney | if its getting stuck in spawning its proably gettin gstuck in the comptue agent somewhere but again that only makes sense to me if the api request were somehow differnt | 17:54 |
| eduardomorais[m] | ok | 17:55 |
| opendevreview | Merged openstack/placement master: Replace license classifier https://review.opendev.org/c/openstack/placement/+/1004422 | 19:01 |
| opendevreview | Stephen Finucane proposed openstack/nova master: Replace use of datetime.timezone.UTC https://review.opendev.org/c/openstack/nova/+/989515 | 22:37 |
| opendevreview | Stephen Finucane proposed openstack/nova master: trivial: Fix outstanding flake8 issues https://review.opendev.org/c/openstack/nova/+/999714 | 22:37 |
| opendevreview | Stephen Finucane proposed openstack/nova master: Add ruff-check https://review.opendev.org/c/openstack/nova/+/974441 | 22:37 |
| opendevreview | Stephen Finucane proposed openstack/nova master: Enable G (flake8-logging-format) rules https://review.opendev.org/c/openstack/nova/+/998557 | 22:37 |
| opendevreview | Stephen Finucane proposed openstack/nova master: tests: Fix DB migration URL test for SQLAlchemy 2.1+ https://review.opendev.org/c/openstack/nova/+/1009877 | 22:37 |
| opendevreview | Merged openstack/nova master: Add reproducer for bug #2164869 https://review.opendev.org/c/openstack/nova/+/1006717 | 22:52 |
Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!