Conversation
|
/test |
learnitall
left a comment
There was a problem hiding this comment.
Docs changes look good, thanks!
tklauser
left a comment
There was a problem hiding this comment.
The "Build commits" workflow failure looks legitimate. Probably the commits need to be reordered/squashed?
Otherwise LGTM, thanks!
94b3d66 to
90d6e04
Compare
|
CI should be fixed now. There was a breaking change in the K8s code generation dependencies that was tricky to nail down. See first commit msg for more details. |
e29632f to
3d465ec
Compare
|
/test |
3d465ec to
94f11e9
Compare
|
/test |
|
Force pushed to fix the related CI failures + rebase with main branch. All CI jobs seem green now, marking this ready for review if you don't mind. |
To support the upgrade to Kubernetes v1.31, controller-tools needs to be bumped. See kubernetes-sigs/controller-tools#1017 for more background. This commit partially bumps additionally dependencies for Kubernetes v1.31 as controller-tools v0.16.1 relies on the former. Therefore, this commit adjusts Cilium code to the new APIs from the v1.31 dependencies. A subsequent commit will add the remaining support for v1.31. Additionally, add the `groupName` directive to core API group slim files. Due to the change in kubernetes/code-generator@3d2e34f, this causes the `make generate-k8s-api` to generate code with the wrong API path for objects under core/v1. This causes all watchers for core/v1 objects in Cilium to break. Signed-off-by: Chris Tarazi <chris@isovalent.com>
The kind node images have moved to a new location, so update the instructions accordingly. Signed-off-by: Chris Tarazi <chris@isovalent.com>
Bump the remaining dependencies of Kubernetes from the previous commit and update CI to the this Kubernetes version. Remove v1.27 from the testing matrix. Signed-off-by: Chris Tarazi <chris@isovalent.com>
94f11e9 to
0a52987
Compare
|
/test |
|
Given that this touches CI - should it have been tested from a cilium/cilium branch, instead of a forked repo? |
Probably yes. It seems as Trying to fix it over here: #34773 |
Hm, that's something I didn't think about. I'll update the docs to explicitly mention this detail so we can avoid this situation on the next K8s bump. Edit: #34791 |
Uh oh!
There was an error while loading. Please reload this page.