Support syncing a set of labels from MachineDeployment/MachineSet/Machine to Nodes

[sidharthsurana]

This issue is to add the capability of defining user defined labels, annotations & taints to be created automatically on the created nodes.

for e.g.: The following MachineSet when realized on any of the provider, should result in the 2 user defined labels latency-sensitive: yes and service-class: gold to be created on the nodes created as part of this machineset.

apiVersion: cluster.k8s.io/v1alpha1
kind: MachineSet
metadata:
  name: gold-workers
spec:
  replicas: 3
  selector:
    matchLabels:
      node-type: worker-node
  template:
    metadata:
      labels:
        node-type: worker-node
    spec:
      metadata:
        labels:
          latency-sensitive: yes
          service-class: gold
      providerConfig:
      ...
      ...

The same mechanics should be extended for annotations as well as taints.

[dgoodwin]

This would be nice, we had to write a little controller ourselves to get those applied.

[derekwaynecarr]

the list of labels, taints, and annotations need to be additive (not authoritative)

[vikaschoudhary16]

@dgoodwin

This would be nice, we had to write a little controller ourselves to get those applied.

Wondering what is the issue in extending existing machine controller in cluster-api to perform this task as well in the reconcile loop?

[roberthbailey]

/assign @hardikdr

[enxebre]

Related to #493 and #658. Also for reference kubernetes/kubernetes#73097

[ncdc]

@hardikdr are you currently working on this?

@sidharthsurana do you believe this is a hard requirement for v1alpha1, or could it slip to v1alpha2?

[sidharthsurana]

@ncdc I don't this is a hard requirement from the cluster-api point of view for the v1alpha1 milestone. Given that individual provider implementations can implement these independently as well. I opened this Issue so that if we can figure out a provider agnostic way of implementing this behavior.

[detiber]

/milestone Next

[hardikdr]

@ncdc Yes I have started looking into it, I will soon be providing a proposal-doc for the same.

[jichenjc]

#881
this one can also be covered ?

[gyliu513]

@hardikdr What is the progress of this? If you are not working, can I take this over?

[gyliu513]

@qiujian16 @xunpan @clyang82

[vincepri]

/area api

[hardikdr]

@gyliu513 I had prepared and presented the proposal doc. There was not any conclusion though. Doc has to be converted into the defined template as well.
I am more than happy to collaborate with you, to push it forward.

[gyliu513]

Thanks @hardikdr , the doc looks good ;-)

@vincepri @detiber can you please post some comments to the design doc as well? I think this is a very important project for cluster life cycle management.

[detiber]

The linked doc proposes full management of Labels, Annotations, and Taints for Machines. It might be better to initial target initial state (Labels and Taints are already supported in this manner if using kubeadm for bootstrapping).

[Karthik-K-N]

Hi team, awaiting to use this feature, Even we wanted to have this capability where we can bring up a cluster with predefined labels on the nodes with the label values are applied based on some internal computation / vm or cloud specific details.

@Arvinderpal If needed i would like to contribute to it. Please let me know. Thank you

[k8s-triage-robot]

The Kubernetes project currently lacks enough contributors to adequately respond to all issues and PRs.

This bot triages issues and PRs according to the following rules:

• After 90d of inactivity, lifecycle/stale is applied
• After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
• After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

• Mark this issue or PR as fresh with /remove-lifecycle stale
• Mark this issue or PR as rotten with /lifecycle rotten
• Close this issue or PR with /close
• Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale

[sbueringer]

/remove-lifecycle stale
/lifecycle active

[fabriziopandini]

/triage accepted

[deepakpunjabi]

@hardikdr was this merged into CAPI?

[hardikdr]

The proposal has been merged: https://github.com/kubernetes-sigs/cluster-api/blob/main/docs/proposals/20220927-label-sync-between-machine-and-nodes.md

[sbueringer]

/reopen

given that the issue is only resolved after the implementation merges

[k8s-ci-robot]

@sbueringer: Reopened this issue.

In response to this:

/reopen

given that the issue is only resolved after the implementation merges

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.

[fabriziopandini]

given that the issue is only resolved after the implementation merges

What about linking the issues where we are tracking implementation work breakdown and closing this one, otherwise we are tracking this in two places

[sbueringer]

I wasn't aware of the umbrella issue at that time :)

Fine for me to close. Looks like the umbrella issue links directly to the PR instead of to this issue. But we can either just have it as a sub-task in the umbrella issue or create a new issue if we really need one.

/close
as it's covered by #7731 now

[k8s-ci-robot]

@sbueringer: Closing this issue.

In response to this:

I wasn't aware of the umbrella issue at that time :)

Fine for me to close. Looks like the umbrella issue links directly to the PR instead of to this issue. But we can either just have it as a sub-task in the umbrella issue or create a new one if necessary.

/close
as it's covered by #7731 now

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.