-
Notifications
You must be signed in to change notification settings - Fork 1.3k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
🐛 fix fail-swap-on=false flag not being part of kind images anymore #8767
🐛 fix fail-swap-on=false flag not being part of kind images anymore #8767
Conversation
/test help |
@chrischdi: The specified target(s) for
The following commands are available to trigger optional jobs:
Use
In response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
/test pull-cluster-api-e2e-workload-upgrade-1-27-latest-main |
Ah damn, test-infra issue ongoing 🤦 |
/lgtm Assuming CI is green once Prow is not overloaded anymore |
LGTM label has been added. Git tree hash: f699b1c54d14e8372e72941c4fe6f2b3af6257a6
|
test/e2e/data/infrastructure-docker/main/clusterclass-quick-start-runtimesdk.yaml
Outdated
Show resolved
Hide resolved
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Should this be added in the template for the machinePools as well? Seems like those would fail at 1.27+
Otherwise looks good to me.
c69395a
to
a651aed
Compare
a651aed
to
faf1cb6
Compare
/test pull-cluster-api-e2e-workload-upgrade-1-27-latest-main |
/test pull-cluster-api-e2e-full-main |
@chrischdi: The
The following commands are available to trigger optional jobs:
Use
In response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
/test pull-cluster-api-e2e-informing-main |
Q: We technically have the same issue for v0.4-v1.3 right? We didn't fix it yet as those those tests are not run locally that frequently (?) and it only fails if someone has a Machine where swap is enabled? Would it make sense to just set it on all versions? If possible I would like to avoid running into this issue again. |
We don't have it in ci, because there is no swap enabled there. For <= v1.3, it depends on the used versions of the kind images (dunno if there will be a released kind node image < v1.27 which faces the same issue, also v1.27 is not supported in v1.3). |
The change in kind was only done to images for Kubernetes >= 1.27 right? They didn't change the older Kubenetes versions? (just because they are always pushing all versions: https://github.com/kubernetes-sigs/kind/releases/tag/v0.19.0) |
This needs to get researched (but looks like that). |
Would be good to know before we merge ideally |
Just talked to Killian, what do you think about just setting the swap setting everywhere? Shouldn't hurt and we don't have to do more research. |
faf1cb6
to
639e201
Compare
/test help |
@chrischdi: The specified target(s) for
The following commands are available to trigger optional jobs:
Use
In response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
/test pull-cluster-api-apidiff-main |
/test pull-cluster-api-e2e-mink8s-main |
@chrischdi: The following tests failed, say
Full PR test history. Your PR dashboard. Please help us cut down on flakes by linking to an open issue when you hit one in your PR. Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. I understand the commands that are listed here. |
@killianmuldoon looks like we have an issue in mink8s. I think we have to clone k/k for this job as well so we can build the node image if necessary |
Or maybe we should better pin to a 1.24 version we have a kind image for. I think we don't have to use latest 1.24 and it makes the job quicker |
The way to do this is to set I think it might be better to let this job build instead of trying to constantly keep track of which version we should be using in the job. |
But because this is specifically a minimum lower bound I'd also be happy to pin it to something like 1.24.2 and just never bump it without reason until we officially bump our mininimum supported K8s version |
Talked to Killian. We would pin to v1.24.13. Overall it shouldn't matter too much, but with the highest available patch version we have probably better compatibility with the kind version we use and Kubernetes has more bugfixes :). There is not a big benefit to build 1.24 latest as the goal is to roughly test against "min". It would also just take more time and resources to always build. |
Thx! /lgtm |
LGTM label has been added. Git tree hash: 8fc4ba71427a55f530303262b4602e475ce6f97c
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: sbueringer The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
@sbueringer : something worth cherry-picking? I think so! |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@sbueringer : something worth cherry-picking? I think so!
Probably needs to be done manually - let's see 😄
/cherry-pick release-1.4 |
@killianmuldoon: #8767 failed to apply on top of branch "release-1.4":
In response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
Fine for me! |
/area provider/infrastructure-docker |
/retitle 🐛 fix fail-swap-on=false flag not being part of kind images anymore |
What this PR does / why we need it:
This PR adds a json patch for KubeadmControlPlanes and MachineDeployments, to set the
fail-swap-on
flag for the kubelet tofalse
in our ClusterClasses for the release, main and v1.4 tests.This prevents that the kubelet refuses to start for newer kind images which don't specify the flag anymore (because kind sets the configuration parameter in KubeletConfiguration.
Which issue(s) this PR fixes (optional, in
fixes #<issue number>(, fixes #<issue_number>, ...)
format, will close the issue(s) when PR gets merged):Fixes #8766