Skip to content
This repository has been archived by the owner on Jul 30, 2021. It is now read-only.

pkg/asset: Add critical pod annotations to control-plane #435

Merged
merged 1 commit into from
Apr 14, 2017

Conversation

derekparker
Copy link
Contributor

For more information: https://kubernetes.io/docs/concepts/cluster-administration/guaranteed-scheduling-critical-addon-pods/.

Since the hack setups (and I believe all cluster setups) only use a single master we could run the rescheduler, however for this PR I am not including it as we have concerns over the rescheduler not being leader elected, and potential issues with multi-master setups. See kubernetes-retired/contrib#2551.

@k8s-ci-robot k8s-ci-robot added the cncf-cla: yes Indicates the PR's author has signed the CNCF CLA. label Apr 14, 2017
@@ -145,6 +145,7 @@ spec:
component: kube-apiserver
annotations:
checkpointer.alpha.coreos.com/checkpoint: "true"
scheduler.alpha.kubernetes.io/critical-pod: ''
Copy link
Contributor

Choose a reason for hiding this comment

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

Given this is a daemonset, should this also get the toleration? or should it not have the annotation at all? The taint is currently only scheduler side (not kubelet enforced) -- daemonsets don't use scheduler -- so my guess is that this would be ignored (unless daemon controller respects the taint?)

Copy link
Contributor Author

Choose a reason for hiding this comment

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

The annotation also exists on kube-proxy but not the toleration, so I added it here for consistency. However, daemonsets do respect taints / tolerations so I don't think it hurts to add the toleration to the daemonsets as well. Thoughts?

Copy link
Contributor

Choose a reason for hiding this comment

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

Ah right. The annotation is basically so it can check if it isn't being scheduled (and would evict other work).

Yeah, if the daemon controller respects it (and maybe later kubelet enforces taints) might not be bad to have the toleration already set.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Added.

Copy link
Contributor

@aaronlevy aaronlevy left a comment

Choose a reason for hiding this comment

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

lgtm

@derekparker derekparker merged commit 1ad279e into kubernetes-retired:master Apr 14, 2017
@derekparker derekparker deleted the critical-pods branch April 14, 2017 22:14
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
cncf-cla: yes Indicates the PR's author has signed the CNCF CLA.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants