Support for Windows nodes in 1.14

[olegchorny]

Feature Request

Feature

Support of Windows nodes

Tradeoffs

Windows nodes are GA starting from 1.14 and it would be interesting to know if typhoon going to support them.

Cons: Huge effort required
Pros: Anyone who wants to leverage full .NET framework in containers will be blessed to run typhoon.

[dghubble]

If Microsoft folks stumble across this, I'd much rather their efforts go toward improving the maturity of Azure as a cloud provider. Its still considered Alpha with known shortcomings summarized here.

For Windows nodes as a feature, I'm not interested in spending time on that. Lots of more useful and interesting big areas for 2019 in my opinion.

[olegchorny]

@dghubble got it, thanks.

[dghubble]

Yep, sorry about that. Since its just workers, a motivated person could have their own module that's just for joining workers to a Typhoon cluster. Similar to the Azure/AWS/GCP worker pools, except instead of declarative provisioning at boot with Container Linux Configs, you'd need to use whatever Windows offers.

[remoe]

I've tried to run a Windows Server 2019 node in a Typhoon setup. Typhoon (terraform-render-bootkube) creates DaemonSets which are deployed on all nodes. But flanneld and kube-proxy linux container should not be deployed on Windows machines:

node label: kubernetes.io/os=windows

because this binaries should run on a windows environment.

Is this something for #429 ?

Name:               win-6hglsrhbdgi
Roles:              <none>
Labels:             beta.kubernetes.io/arch=amd64
                    beta.kubernetes.io/os=windows
                    kubernetes.io/arch=amd64
                    kubernetes.io/hostname=win-6hglsrhbdgi
                    kubernetes.io/os=windows
Annotations:        flannel.alpha.coreos.com/backend-data: {"VNI":4096,"VtepMAC":""}
                    flannel.alpha.coreos.com/backend-type: vxlan
                    flannel.alpha.coreos.com/kube-subnet-manager: true
                    flannel.alpha.coreos.com/public-ip: xxx.xxx.xx.xx
                    node.alpha.kubernetes.io/ttl: 0
                    volumes.kubernetes.io/controller-managed-attach-detach: true
CreationTimestamp:  Tue, 25 Jun 2019 23:05:44 +0200
Taints:             <none>
Unschedulable:      false
Conditions:
  Type             Status  LastHeartbeatTime                 LastTransitionTime                Reason                       Message
  ----             ------  -----------------                 ------------------                ------                       -------
  MemoryPressure   False   Wed, 26 Jun 2019 18:11:49 +0200   Wed, 26 Jun 2019 18:11:49 +0200   KubeletHasSufficientMemory   kubelet has sufficient memory available
  DiskPressure     False   Wed, 26 Jun 2019 18:11:49 +0200   Wed, 26 Jun 2019 18:11:49 +0200   KubeletHasNoDiskPressure     kubelet has no disk pressure
  PIDPressure      False   Wed, 26 Jun 2019 18:11:49 +0200   Wed, 26 Jun 2019 18:11:49 +0200   KubeletHasSufficientPID      kubelet has sufficient PID available
  Ready            True    Wed, 26 Jun 2019 18:11:49 +0200   Wed, 26 Jun 2019 18:11:49 +0200   KubeletReady                 kubelet is posting ready status
Addresses:
  InternalIP:  xxx.xxx.xx.xx
  Hostname:    win-6hglsrhbdgi
Capacity:
 cpu:                1
 ephemeral-storage:  41378812Ki
 memory:             4193844Ki
 pods:               110
Allocatable:
 cpu:                1
 ephemeral-storage:  38134713077
 memory:             4091444Ki
 pods:               110
System Info:
 Machine ID:                 WIN-6HGLSRHBDGI
 System UUID:
 Boot ID:
 Kernel Version:             10.0.17763.557
 OS Image:                   Windows Server 2019 Standard
 Operating System:           windows
 Architecture:               amd64
 Container Runtime Version:  docker://18.9.6
 Kubelet Version:            v1.15.0
 Kube-Proxy Version:         v1.15.0
PodCIDR:                     10.2.4.0/24
Non-terminated Pods:         (3 in total)
  Namespace                  Name                   CPU Requests  CPU Limits  Memory Requests  Memory Limits  AGE
  ---------                  ----                   ------------  ----------  ---------------  -------------  ---
  kube-system                flannel-5slxl          100m (10%)    0 (0%)      0 (0%)           0 (0%)         19h
  kube-system                kube-proxy-r8h4k       0 (0%)        0 (0%)      0 (0%)           0 (0%)         19h
  monitoring                 node-exporter-pn6pm    100m (10%)    200m (20%)  50Mi (1%)        100Mi (2%)     19h
Allocated resources:
  (Total limits may be over 100 percent, i.e., overcommitted.)
  Resource           Requests    Limits
  --------           --------    ------
  cpu                200m (20%)  200m (20%)
  memory             50Mi (1%)   100Mi (2%)
  ephemeral-storage  0 (0%)      0 (0%)
Events:
  Type    Reason                   Age                From                         Message
  ----    ------                   ----               ----                         -------
  Normal  Starting                 43m                kube-proxy, win-6hglsrhbdgi  Starting kube-proxy.
  Normal  Starting                 24m                kubelet, win-6hglsrhbdgi     Starting kubelet.
  Normal  NodeHasSufficientMemory  24m (x3 over 24m)  kubelet, win-6hglsrhbdgi     Node win-6hglsrhbdgi status is now: NodeHasSufficientPID
  Normal  NodeReady                24m (x2 over 24m)  kubelet, win-6hglsrhbdgi     Node win-6hglsrhbdgi status is now: NodeReady
  Normal  Starting                 84s                kubelet, win-6hglsrhbdgi     Starting kubelet.
  Normal  NodeHasSufficientPID     33s (x3 over 84s)  kubelet, win-6hglsrhbdgi     Node win-6hglsrhbdgi status is now: NodeHasSufficientPID
  Normal  NodeReady                33s (x2 over 83s)  kubelet, win-6hglsrhbdgi     Node win-6hglsrhbdgi status is now: NodeReady```