tracking issue for HA to GA

[neolit123]

long term we need to enumerate what tasks remain for HA support to be considered GA by the kubeadm project.

some ideas:

• more robustness for etcd e.g. using learner mode:
  RFE: use learner mode for joining etcd members #1793
• instance specific component config?
  Allow usage of KubeletConfiguration next to JoinConfiguration on kubeadm join #1682
• ?

---

tracking issue at k/e:
kubernetes/enhancements#357

tracking issue for HA to Beta:
#1567

Beta KEPS:
https://github.com/kubernetes/enhancements/blob/master/keps/sig-cluster-lifecycle/kubeadm/0015-kubeadm-join-control-plane.md

https://github.com/kubernetes/enhancements/blob/master/keps/sig-cluster-lifecycle/kubeadm/20190122-Certificates-copy-for-kubeadm-join--control-plane.md

[fabriziopandini]

IMO kubeadm HA

• should be agnostic to lb solutions (integrated lb should not be a GA criteria)
• should not be linked to the node specific/component config discussions (this should not be a graduation criteria)

Instead I think that GA criteria should be related to robustness and flexibility. WRT to this on top of my mind the etcd learner feature can be used both for improving robustness and as a sync mechanism for improving reliable parallel joins

[BenTheElder]

/subscribe 🙃

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[neolit123]

/remove-lifecycle stale

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[neolit123]

/remove-lifecycle stale

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[neolit123]

/remove-lifecycle stale

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-contributor-experience at kubernetes/community.
/lifecycle stale

[neolit123]

/remove-lifecycle stale

[fabriziopandini]

@neolit123 wdyt about graduating this feature?

As far as I'm concerned we have two point still on the table

• Use etcd learner mode
• Parallel join for control plane nodes

Both of them are nice to have, but at the same time they requires investigation and I don't see use getting on them in the near future. So I'm starting to wonder if we have to go GA with the current implementation, which is now stable since a couple of versions, and reconsider the points above eventually in future

[neolit123]

Parallel join for control plane nodes

i guess this is already supported within the limits of our hardcoded timeouts / retries on join.

Both of them are nice to have, but at the same time they requires investigation and I don't see use getting on them in the near future. So I'm starting to wonder if we have to go GA with the current implementation, which is now stable since a couple of versions, and reconsider the points above eventually in future

i see what you mean, but maybe we can ask contributors to investigate learners mode and someone could create a prototype before going GA without it:
#1793

another item in the OP here is instance specific config:
#1682
i don't remember why we added it and it has a "?". outside of the KubeletConfiguration topic, current workaround for configuring a CP node differently is with patches.
your operator proposal had some ideas about instance specific config per CP node.

[fejta-bot]

Stale issues rot after 30d of inactivity.
Mark the issue as fresh with /remove-lifecycle rotten.
Rotten issues close after an additional 30d of inactivity.

If this issue is safe to close now please do so with /close.

Send feedback to sig-contributor-experience at kubernetes/community.
/lifecycle rotten

[neolit123]

/remove-lifecycle rotten

[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

[k8s-triage-robot]

The Kubernetes project currently lacks enough active 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 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 rotten

[neolit123]

/lifecycle frozen