Skip to content

qa/distros: add ubuntu 24.04 as supported distro/container host#63912

Closed
cbodley wants to merge 8 commits intoceph:mainfrom
cbodley:wip-qa-distros-t
Closed

qa/distros: add ubuntu 24.04 as supported distro/container host#63912
cbodley wants to merge 8 commits intoceph:mainfrom
cbodley:wip-qa-distros-t

Conversation

@cbodley
Copy link
Contributor

@cbodley cbodley commented Jun 12, 2025

once ubuntu 24 packages are made available in shaman (depends on https://tracker.ceph.com/issues/66914), we can start testing them in teuthology

  • replace any subsuite distro folders that only contain ubuntu_latest,centos_latest with a symlink to supported-random-distro$ so they'll pick up the changes there
  • bump the ubuntu_latest.yaml symlink to 24.04
  • add 22.04 back as supported so users of supported-random-distro$ will test both ubuntu versions
  • add ubuntu 24.04 as a supported container host for cephadm suites
Show available Jenkins commands

@github-actions github-actions bot added cephfs Ceph File System tests labels Jun 12, 2025
@cbodley
Copy link
Contributor Author

cbodley commented Jun 18, 2025

we got a successful build of ubuntu noble packages in https://shaman.ceph.com/builds/ceph/wip-kefu-pr-64001/ so i tried scheduling this suite-branch against it, but all of the jobs that selected ubuntu_latest (now 24.04) failed with:

Packages for os_type 'ubuntu', flavor default and ceph hash '04324fa1ac1916b7a3db7928b26d4660c099d35c' not found

@ljflores
Copy link
Member

ljflores commented Jul 10, 2025

I scheduled a rados suite against this build: https://shaman.ceph.com/builds/ceph/main/94b8a156e3c59813b5354f484da97122b49be851/

./teuthology/virtualenv/bin/teuthology-suite -v -m smithi -c main -S 94b8a156e3c59813b5354f484da97122b49be851 --suite-repo https://github.com/cbodley/ceph.git --suite-branch wip-qa-distros-t -s rados --subset 111/120000 -p 150

https://pulpito.ceph.com/lflores-2025-07-10_16:37:06-rados-main-distro-default-smithi/

@cbodley
Copy link
Contributor Author

cbodley commented Jul 11, 2025

example from https://qa-proxy.ceph.com/teuthology/lflores-2025-07-10_16:37:06-rados-main-distro-default-smithi/8380254/teuthology.log:

2025-07-11T08:45:52.041 INFO:teuthology.orchestra.run.smithi090.stdout:The following packages have unmet dependencies:
2025-07-11T08:45:52.153 INFO:teuthology.orchestra.run.smithi090.stdout: ceph-mgr : Depends: python3-distutils but it is not installable
2025-07-11T08:45:52.153 INFO:teuthology.orchestra.run.smithi090.stdout:            Recommends: ceph-mgr-dashboard but it is not going to be installed
2025-07-11T08:45:52.153 INFO:teuthology.orchestra.run.smithi090.stdout:            Recommends: ceph-mgr-diskprediction-local but it is not going to be installed
2025-07-11T08:45:52.153 INFO:teuthology.orchestra.run.smithi090.stdout:            Recommends: ceph-mgr-k8sevents but it is not going to be installed
2025-07-11T08:45:52.153 INFO:teuthology.orchestra.run.smithi090.stdout:            Recommends: ceph-mgr-cephadm but it is not going to be installed

this was resolved with #64386 for https://tracker.ceph.com/issues/72020. but package installation still fails due to https://tracker.ceph.com/issues/72083

@adk3798
Copy link
Contributor

adk3798 commented Sep 10, 2025

Did this orch run just for noble https://pulpito.ceph.com/adking-2025-09-09_01:28:09-orch:cephadm-wip-adk2-testing-2025-09-08-1425-distro-default-smithi/. It looks like the nvme-loop task will have to be updated to work with noble at the least. For whatever reason a bunch of tests timed out reimaging machines as well. Didn't bother with reruns since I know we need the nvme-loop changes

@cbodley
Copy link
Contributor Author

cbodley commented Sep 10, 2025

It looks like the nvme-loop task will have to be updated to work with noble at the least

thanks @adk3798 - does this deserve a tracker issue? is it clear who's responsible?

@adk3798
Copy link
Contributor

adk3798 commented Sep 10, 2025

It looks like the nvme-loop task will have to be updated to work with noble at the least

thanks @adk3798 - does this deserve a tracker issue? is it clear who's responsible?

Guillaume and I have fixed that up in the past. I guess we might as well have a tracker, especially if we plan to backport this to tentacle

@adk3798
Copy link
Contributor

adk3798 commented Sep 16, 2025

It looks like the nvme-loop task will have to be updated to work with noble at the least

thanks @adk3798 - does this deserve a tracker issue? is it clear who's responsible?

I've added a commit to your branch that I think should fix the nvme_loop task for ubuntu noble. If you'd prefer it be its own separate PR I can create one. Either way, will do another test run

Copy link
Contributor

@adk3798 adk3798 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

with the new commit to fix nvme_loop, orch run on noble looks good https://pulpito.ceph.com/adking-2025-09-17_13:14:17-orch:cephadm-wip-adk2-testing-2025-09-16-1635-distro-default-smithi/, just one known issue failure and a bunch of tests that failed to re-image machines.

@cbodley did you also see imaging issues in any of your runs? I'm approving since it's not an orch related thing, but I've now seen a number of those in both runs I've done. I could also just be unlucky.

@cbodley
Copy link
Contributor Author

cbodley commented Sep 17, 2025

@cbodley did you also see imaging issues in any of your runs?

@adk3798 i do remember seeing some failures like that, but it's been a while since i tried schedule against ubuntu 24

cbodley and others added 8 commits October 17, 2025 15:55
Signed-off-by: Casey Bodley <cbodley@redhat.com>
Signed-off-by: Casey Bodley <cbodley@redhat.com>
Signed-off-by: Casey Bodley <cbodley@redhat.com>
Signed-off-by: Casey Bodley <cbodley@redhat.com>
Signed-off-by: Casey Bodley <cbodley@redhat.com>
and add ubuntu_22.04.yaml back to distros/supported and
distros/supported-random-distro$

Signed-off-by: Casey Bodley <cbodley@redhat.com>
Signed-off-by: Casey Bodley <cbodley@redhat.com>
CEPH-BUILD-JOB: ceph-dev-pipeline
DISTROS: noble jammy centos9
ARCHS: x86_64
FLAVORS: default
Compared to older distros, this one complains if
you include `-q hostnqn` in the nvme connect command,
saying "Failed to write to /dev/nvme-fabrics: Invalid argument".
Removing that argument gets passed that error and
doesn't seem to have any downsides

Signed-off-by: Adam King <adking@redhat.com>
@cbodley
Copy link
Contributor Author

cbodley commented Oct 17, 2025

pushed a rebase

i also prepared a branch, based on this one, that goes on to add yamls for rocky 10 in https://github.com/cbodley/ceph/commits/wip-qa-distros-t-rocky. but we're not producing rocky 10 builds yet to test

@cbodley
Copy link
Contributor Author

cbodley commented Oct 24, 2025

i also prepared a branch, based on this one, that goes on to add yamls for rocky 10 in https://github.com/cbodley/ceph/commits/wip-qa-distros-t-rocky. but we're not producing rocky 10 builds yet to test

raised a draft pr #66055 with just the rocky bits (not based on this pr)

@batrick
Copy link
Member

batrick commented Mar 5, 2026

Moving to #67683.

Thanks for starting this @cbodley

@batrick batrick closed this Mar 5, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants