Windows Support

[rhockenbury]

User Story

As an operator, I would like to manage windows server worker nodes with the cluster api.

Detailed Description

Since I just went through setting up windows nodes to work with the cluster api (CAPV), I'm documenting areas of improvements / issues for better supporting windows nodes.

• Support for meeting netbios requirements for hostnames on windows server - Windows Support: NetBIOS and Active Directory LDAP SAMAccountName restrictions on Hostname #2217
• Ensure that additional credentials (admin kubeconfig, domain join password) for windows server nodes can be set securely through CABPK - [WIP] ✨Add support for defining KubeadmConfig files by reference (ConfigMap & Secret) #1860
• Official release of cloudbase-init with additional directives required for CABPK - https://github.com/cloudbase/cloudbase-init (all required features have been merged to master) - add the vmware guestinfo service cloudbase/cloudbase-init#20 (comment)
• Publication of packer and ansible provisioning scripts for building base images - Windows machine image image-builder#30

Anything else you would like to add:

Windows worker nodes are not officially support by the cluster api, but it is up for discussion about where it belongs in the roadmap - https://github.com/kubernetes-sigs/cluster-api/pull/2148/files#diff-767f66541aad47089dd5ded720dede6bR32.

/kind feature

[ksubrmnn]

Hi @rhockenbury I'm currently working on this, and I've already created a POC here. I will be working on the design for this in the next month regarding the bootstrap provider and will put up the design and PR soon.

[ksubrmnn]

/assign @ksubrmnn

[rhockenbury]

That sounds great. Did you have any thoughts about #2217 ?

[michmike]

cc: @benmoss

[ksubrmnn]

That sounds great. Did you have any thoughts about #2217 ?

Yes, that's an issue that I have in my POC. I currently just work around it via hardcoding (not great). I think we'd have to open this to a broader community discussion to agree upon a standard. Perhaps this should be discussed in the overall CAPI meeting.

Other than that, I haven't found any other significant differences beyond the bootstrap script (I've created my own that uses the KubeCluster scripts in sig-windows tools).

In my implementation we are using custom script extensions (as is custom for Azure). Haven't had a chance to read up on @benmoss 's work.

@JocelynBerrendonner ill also be working on this from MSFT.

[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

[vincepri]

/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

[detiber]

/lifecycle frozen